diff --git a/bench_sampled/selected_cases.json b/bench_sampled/selected_cases.json new file mode 100644 index 0000000000000000000000000000000000000000..52c63cc6f41d965743d0078e98832dc6222bf8b1 --- /dev/null +++ b/bench_sampled/selected_cases.json @@ -0,0 +1,124 @@ +{ + "seed": 42, + "total": 100, + "by_task": { + "at_rest": [ + "at_rest/0", + "at_rest/2", + "at_rest/3", + "at_rest/4", + "at_rest/7", + "at_rest/8", + "at_rest_newbatch/3", + "at_rest_newbatch/10", + "at_rest_newbatch/12", + "at_rest_newbatch/13", + "at_rest_newbatch/14" + ], + "circular": [ + "circular/0", + "circular/2", + "circular/6", + "circular/7", + "circular_newbatch/2", + "circular_newbatch/5", + "circular_newbatch/6", + "circular_newbatch/8", + "circular_newbatch/11", + "circular_newbatch/12", + "circular_newbatch/14" + ], + "freefall": [ + "freefall/0", + "freefall/4", + "freefall/5", + "freefall/6", + "freefall/8", + "freefall_newbatch/0", + "freefall_newbatch/2", + "freefall_newbatch/3", + "freefall_newbatch/4", + "freefall_newbatch/8", + "freefall_newbatch/12" + ], + "inclined_plane": [ + "inclined_plane/1", + "inclined_plane/2", + "inclined_plane/3", + "inclined_plane/7", + "inclined_plane/8", + "inclined_plane_newbatch/1", + "inclined_plane_newbatch/2", + "inclined_plane_newbatch/3", + "inclined_plane_newbatch/8", + "inclined_plane_newbatch/9", + "inclined_plane_newbatch/13" + ], + "inelastic_collision": [ + "inelastic_collision/1", + "inelastic_collision/2", + "inelastic_collision/3", + "inelastic_collision/6", + "inelastic_collision/9", + "inelastic_collision_newbatch/0", + "inelastic_collision_newbatch/1", + "inelastic_collision_newbatch/2", + "inelastic_collision_newbatch/7", + "inelastic_collision_newbatch/11", + "inelastic_collision_newbatch/14" + ], + "perfectly_elastic_collision": [ + "perfectly_elastic_collision/2", + "perfectly_elastic_collision/3", + "perfectly_elastic_collision/5", + "perfectly_elastic_collision/7", + "perfectly_elastic_collision/8", + "perfectly_elastic_collision/9", + "perfectly_elastic_collision_newbatch/0", + "perfectly_elastic_collision_newbatch/1", + "perfectly_elastic_collision_newbatch/3", + "perfectly_elastic_collision_newbatch/4", + "perfectly_elastic_collision_newbatch/13", + "perfectly_elastic_collision_newbatch/14" + ], + "perfectly_inelastic_collision": [ + "perfectly_inelastic_collision/2", + "perfectly_inelastic_collision/5", + "perfectly_inelastic_collision/6", + "perfectly_inelastic_collision/7", + "perfectly_inelastic_collision/8", + "perfectly_inelastic_collision_newbatch/1", + "perfectly_inelastic_collision_newbatch/6", + "perfectly_inelastic_collision_newbatch/7", + "perfectly_inelastic_collision_newbatch/8", + "perfectly_inelastic_collision_newbatch/11", + "perfectly_inelastic_collision_newbatch/14" + ], + "projectile": [ + "projectile/1", + "projectile/7", + "projectile/8", + "projectile/10", + "projectile/12", + "projectile_newbatch/1", + "projectile_newbatch/4", + "projectile_newbatch/5", + "projectile_newbatch/6", + "projectile_newbatch/9", + "projectile_newbatch/11" + ], + "uniform_linear": [ + "uniform_linear/2", + "uniform_linear/6", + "uniform_linear/8", + "uniform_linear_newbatch/0", + "uniform_linear_newbatch/2", + "uniform_linear_newbatch/4", + "uniform_linear_newbatch/5", + "uniform_linear_newbatch/8", + "uniform_linear_newbatch/10", + "uniform_linear_newbatch/13", + "uniform_linear_newbatch/14" + ] + } +} \ No newline at end of file diff --git a/bench_sampled/unified.json b/bench_sampled/unified.json new file mode 100644 index 0000000000000000000000000000000000000000..cfb880a92b4fe17f24676599f6f57ce2198f189a --- /dev/null +++ b/bench_sampled/unified.json @@ -0,0 +1,12152 @@ +[ + { + "case": "at_rest/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΣF = m·Δv / (Δt·cosθ)\n B) ΣF_ext = 0\n C) T = 2π·√(m/k)\n D) F_drag = (1/2)·C_d·ρ·A·v²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) ΣF_ext = 0\n C) G = m·g\n D) τ_net = Σ(F_i·r_i·sinθ_i) · e^(-μ·t)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) G = m·g\n B) ΣF = m·Δv / (Δt·cosθ)\n C) ΣF_ext = 0\n D) U = -G·m_1·m_2 / r\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest/4", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) τ_net = Σ(F_i·r_i·sinθ_i) · e^(-μ·t)\n B) v = f·λ\n C) ΣF_ext = 0\n D) G = m·g\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΣF_ext = 0\n B) τ_net = Σ(F_i·r_i·sinθ_i) · e^(-μ·t)\n C) T = 2π·√(m/k)\n D) G = m·g\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n B) I = V / R\n C) ΣF = m·Δv / (Δt·cosθ)\n D) ΣM_ext = 0\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_eq = m·g / (1 + μ_s·tanθ)²\n B) v = f·λ\n C) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n D) ΣF_ext = 0\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΣF = m·Δv / (Δt·cosθ)\n B) ΣF_ext = 0\n C) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n D) U = -G·m_1·m_2 / r\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) G = m·g\n B) ΔS = Q_rev / T\n C) ΣM_ext = 0\n D) ΣM_ext = F·r·cos²θ / (1 + sinθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_drag = (1/2)·C_d·ρ·A·v²\n B) T = 2π·√(m/k)\n C) ΣF_ext = 0\n D) ΣM_ext = F·r·cos²θ / (1 + sinθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΣM_ext = F·r·cos²θ / (1 + sinθ)\n B) ΔS = Q_rev / T\n C) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n D) ΣM_ext = 0\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "circular/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) a_c = v²·cosθ / (R·sinθ)\n B) PV = nRT\n C) F = (3/4)·E*·√(R·d³)\n D) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "circular/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) C = Q / V\n B) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n C) e = (v_2f - v_1f) / (v_1i - v_2i)\n D) F_N = m·g·(cosθ - 2) + m·v²/R\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "circular/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) PV = nRT\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) F_N = m·g·(cosθ - 2) + m·v²/R\n D) F_N - m·g·cosθ = m·v²/R\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n B) F = (3/4)·E*·√(R·d³)\n C) F_N = m·g·(cosθ - 2) + m·v²/R\n D) v = f·λ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_N - m·g·cosθ = m·v²/R\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) ΔS = Q_rev / T\n D) ω = √(g·cosθ / R) · (1 + sinθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/5", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ω = √(g·cosθ / R) · (1 + sinθ)\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) Q = m·c·ΔT\n D) F_N - m·g·cosθ = m·v²/R\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "circular_newbatch/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) E_total = m·g·R·(3/2 - cosθ)²\n B) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n C) λ = h/(m·v)\n D) F_drag =(1/2)·C_d·ρ·A·v²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "circular_newbatch/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v² = g·R·(2 - 3·cosθ) / (1 + sinθ)\n B) f_k = μ_k·F_N\n C) ΔE = -13.6·(1/n_f² - 1/n_i²) eV\n D) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "circular_newbatch/11", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_drag =(1/2)·C_d·ρ·A·v²\n B) E_total = m·g·R·(3/2 - cosθ)²\n C) ΔS = Q_rev / T\n D) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular_newbatch/12", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_N - m·g·cosθ = m·v²/R\n B) E_total = m·g·R·(3/2 - cosθ)²\n C) e = (v_2f - v_1f) / (v_1i - v_2i)\n D) B = μ_0·n·I\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "circular_newbatch/14", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔS = Q_rev / T\n B) F_N - m·g·cosθ = m·v²/R\n C) f_s ≤ μ_s·F_N\n D) a_c = v²·cosθ / (R·sinθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "freefall/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) y(t) = h + v_0·t + (1/2)·g·t²\n C) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n D) P + (1/2)·ρ·v² + ρ·g·h = constant\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall/4", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) Q = m·c·ΔT\n B) I = (1/10)·m·d²\n C) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall/5", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v = v_0·cos(g·t / v_0)\n B) ΔS = Q_rev / T\n C) I = (1/10)·m·d²\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_final = -e·v_impact\n B) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n C) y(t) = h + v_0·t + (1/2)·g·t²\n D) n_1·sinθ_1 = n_2·sinθ_2\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n B) v = v_0 + g·t\n C) Q = m·c·ΔT\n D) P + (1/2)·ρ·v² + ρ·g·h = constant\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "freefall/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "freefall_newbatch/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_impact = √(g·h) · (1 + μ·t)\n B) v = v_0 + g·t\n C) v_final = -e·v_impact\n D) T = 2π·√(m/k)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "freefall_newbatch/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y(t) = h + v_0·t + (1/2)·g·t²\n B) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n C) Q = m·c·ΔT\n D) I = (1/10)·m·d²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "freefall_newbatch/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y(t) = h + v_0·t + (1/2)·g·t²\n B) F·Δt = m·v_f - m·v_i\n C) τ = r × F\n D) y(t) = h·(1 - e^(-g·t²/2))\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "freefall_newbatch/4", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y(t) = h·(1 - e^(-g·t²/2))\n B) v = f·λ\n C) v_final = -e·v_impact\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall_newbatch/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_impact = √(g·h) · (1 + μ·t)\n B) F = k·q_1·q_2 / r²\n C) I = (1/10)·m·d²\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "freefall_newbatch/12", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔS = Q_rev / T\n B) y(t) = h + v_0·t + (1/2)·g·t²\n C) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n D) F·Δt = m·v_f - m·v_i\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inclined_plane/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v² = v_0² + 2·a_slope·s\n B) ΔS = Q_rev / T\n C) E_p = m·g·(H - h)\n D) s = v_0·t·cosθ + (1/2)·g·sinθ·t²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) N = m·g·(cosθ - μ_k·sinθ)²\n B) Q = m·c·ΔT\n C) E_p = m·g·(H - h)\n D) v² = v_0² + 2·a_slope·s\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) U = -G·m_1·m_2 / r\n B) a_slope = g·(sinθ - μ_k·cosθ)\n C) E_p = m·g·(H - h)\n D) F_net = m·g·(sin²θ - μ_k·cos²θ) / sinθ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inclined_plane/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) T = 2π·√(L/g)\n B) v² = v_0² + 2·a_slope·s\n C) h_cm·sinθ > (a/2)·cosθ\n D) N = m·g·(cosθ - μ_k·sinθ)²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) f_k = μ_k·m·g\n B) v² = v_0² + g·sinθ·s·(1 - μ_k·tanθ)\n C) v² = v_0² + 2·a_slope·s\n D) F = k·q_1·q_2 / r²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane_newbatch/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) s = v_0·t·cosθ + (1/2)·g·sinθ·t²\n B) f_k = μ_k·m·g\n C) T = 2π·√(L/g)\n D) v² = v_0² + 2·a_slope·s\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane_newbatch/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) s = v_0·t·cosθ + (1/2)·g·sinθ·t²\n B) B·dl = μ_0·I\n C) a_slope = g·(sinθ - μ_k·cosθ)\n D) E_p = m·g·(H - h)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane_newbatch/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) Q = m·c·ΔT\n B) a_slope = g·(sinθ - μ_k·cosθ)\n C) v² = v_0² + g·sinθ·s·(1 - μ_k·tanθ)\n D) h_cm·sinθ > (a/2)·cosθ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inclined_plane_newbatch/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) N = m·g·(cosθ - μ_k·sinθ)²\n B) v² = v_0² + 2·a_slope·s\n C) U = -G·m_1·m_2 / r\n D) h_cm·sinθ > (a/2)·cosθ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane_newbatch/9", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) E = (1/2)·L·I²\n B) v_t = g·d²·(ρ_p - ρ_f) / (18·η)\n C) a_slope = g·(sinθ - μ_k·cosθ)\n D) F_net = m·g·(sin²θ - μ_k·cos²θ) / sinθ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane_newbatch/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane_newbatch/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inclined_plane_newbatch/13", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_t = g·d²·(ρ_p - ρ_f) / (18·η)\n B) ΔU = q·ΔV\n C) a_slope = g·(sinθ + μ_k·cosθ) / (1 + sinθ)\n D) a_slope = g·(sinθ - μ_k·cosθ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inclined_plane_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inclined_plane_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) U = -G·m_1·m_2 / r\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) ΔE_loss = (1/2)·e²·(m_1·m_2)/(m_1+m_2)·(v_1i + v_2i)²\n D) W_f = ∫ μ_k·m·g·ds = ΔE_k\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n B) ΔE = (1/2)·m_1·m_2·e²·(v_1i - v_2i)² / (m_1 + m_2)\n C) F = k·q_1·q_2 / r²\n D) e = (v_2f - v_1f) / (v_1i - v_2i)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "inelastic_collision/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_1f = (m_1 - e·m_2)·v_1i / (m_1 + m_2) + v_2i·cosθ\n B) σ = F/a² = E·ε\n C) E = h·f\n D) e = (v_2f - v_1f) / (v_1i - v_2i)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) v_cm = (m_1·v_1i² + m_2·v_2i²) / (m_1·v_1i + m_2·v_2i)\n C) σ = F/a² = E·ε\n D) T = 2π·√(m/k)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision/9", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) σ = F/a² = E·ε\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) J = m_1·m_2·(v_1i - v_2i)·(1+e) / (m_1 + m_2) · sinθ\n D) v = f·λ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 4.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 4.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) e = 1 - ΔE / (m_1·v_1i² + m_2·v_2i²)\n B) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n C) E_rot = (1/2)·(1/6·m·a²)·ω²\n D) PV = nRT\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) E_rot = (1/2)·(1/6·m·a²)·ω²\n C) B = μ_0·I / (2π·r)\n D) ΔE_loss = (1/2)·e²·(m_1·m_2)/(m_1+m_2)·(v_1i + v_2i)²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F = k·q_1·q_2 / r²\n B) e = 1 - ΔE / (m_1·v_1i² + m_2·v_2i²)\n C) W_f = ∫ μ_k·m·g·ds = ΔE_k\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) λ = h/(m·v)\n B) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n C) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n D) J = m_1·m_2·(v_1i - v_2i)·(1+e) / (m_1 + m_2) · sinθ\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v = f·λ\n B) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n C) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n D) e = 1 - ΔE / (m_1·v_1i² + m_2·v_2i²)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n B) v = f·λ\n C) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n D) ΔE_loss = (1/2)·e²·(m_1·m_2)/(m_1+m_2)·(v_1i + v_2i)²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n B) B = μ_0·I / (2π·r)\n C) ΔE = μ_k·m·g·d\n D) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) v_1f = v_1i · e^(-m_2/m_1) + v_2i·(1 - e^(-m_2/m_1))\n C) v_2f - v_1f = e·(v_1i - v_2i)\n D) L = I·ω\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔE = μ_k·m·g·d\n B) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n C) λ = h/(m·v)\n D) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ∫ F(t) dt = Δ(mv)\n B) v_1f = v_1i·(m_1² - m_2²) / (m_1 + m_2)²\n C) λ = h/(m·v)\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) I = I_cm + m·(a/2)²\n B) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n C) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n D) F = k·q_1·q_2 / r²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F = k·q_1·q_2 / r²\n B) ΔE = μ_k·m·g·d\n C) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 4.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 4.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) ∫ F(t) dt = Δ(mv)\n C) v_1f = v_1i · e^(-m_2/m_1) + v_2i·(1 - e^(-m_2/m_1))\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔS = Q_rev / T\n B) v_2f - v_1f = e·(v_1i - v_2i)\n C) ΔE = (1/2)·m_1·(v_1f - v_1i)²·(m_1/m_2 + 1)\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) I = I_cm + m·(a/2)²\n B) F = k·q_1·q_2 / r²\n C) E_k = (1/2)·(m_1·v_1i + m_2·v_2i)² / (m_1 + m_2)\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) Q = m·c·ΔT\n C) I = I_cm + m·(a/2)²\n D) v_1f = v_1i·(m_1² - m_2²) / (m_1 + m_2)²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) E_k = (1/2)·(m_1·v_1i + m_2·v_2i)² / (m_1 + m_2)\n B) v_2f - v_1f = e·(v_1i - v_2i)\n C) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n D) Q = m·c·ΔT\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) T = 2π·√(m/k)\n B) v_cm = (m_1·v_1i² + m_2·v_2i²) / (2·(m_1·v_1i + m_2·v_2i))\n C) ΔE = μ_k·m·g·d\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·(m_1 + m_2)·v_f²\n B) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n C) ΔE = (1/2)·(m_1·m_2)·(v_1i + v_2i)² / (m_1 + m_2)²\n D) PV = nRT\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) I = V / R\n B) m = ρ·a³\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) ΔE_loss = m_1·m_2·(v_1i - v_2i) / (2·(m_1 + m_2))\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) ΔS = Q_rev / T\n B) ΔE_loss = m_1·m_2·(v_1i - v_2i) / (2·(m_1 + m_2))\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) v_2f - v_1f = e·(v_1i - v_2i)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) W_friction = μ_k·(m_1 + m_2)·g·d\n B) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n C) E_k_f = (m_1²·v_1i² + m_2²·v_2i²) / (2·(m_1 + m_2))\n D) ΔS = Q_rev / T\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n B) ΔE_loss = m_1·m_2·(v_1i - v_2i) / (2·(m_1 + m_2))\n C) F_b = ρ_f·g·V\n D) m = ρ·a³\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) PV = nRT\n B) v_f = (m_1·v_1i + m_2·v_2i)² / ((m_1 + m_2)·(v_1i + v_2i))\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) m = ρ·a³\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_f² = (m_1·v_1i + m_2·v_2i)² / (m_1² + m_2²)\n B) W_friction = μ_k·(m_1 + m_2)·g·d\n C) E_n = -13.6/n² eV\n D) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_f² = (m_1·v_1i + m_2·v_2i)² / (m_1² + m_2²)\n B) m = ρ·a³\n C) n_1·sinθ_1 = n_2·sinθ_2\n D) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) W_friction = μ_k·(m_1 + m_2)·g·d\n B) v_f² = (m_1·v_1i + m_2·v_2i)² / (m_1² + m_2²)\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) PV = nRT\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 18, + "time_point": 9.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 10.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 19, + "time_point": 10.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) B = μ_0·I / (2π·r)\n B) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n C) v_f = √(m_1/m_2)·v_1i + √(m_2/m_1)·v_2i\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·(m_1 + m_2)·v_f²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v_f = (m_1·v_1i + m_2·v_2i)² / ((m_1 + m_2)·(v_1i + v_2i))\n B) I = V / R\n C) m = ρ·a³\n D) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "projectile/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) h_next = e²·h_current\n B) B = μ_0·I / (2π·r)\n C) y = h·cos(g·x / v_0²) + x·tanθ\n D) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile/7", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n B) ΔS = Q_rev / T\n C) x(t) = v_0·t·cosθ · (1 - g·t/v_0)\n D) h_next = e²·h_current\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile/7", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n B) a_n = v² / ρ\n C) t_flight = 2·v_0·sin²θ / g\n D) L = I·ω\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile/10", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) a_n = v² / ρ\n B) v_x(t) = v_0·cosθ · e^(-g·t/v_0)\n C) T = 2π·√(L/g)\n D) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile/12", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y = h·cos(g·x / v_0²) + x·tanθ\n B) L = I·ω\n C) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n D) m·g·h + (1/2)·m·v_0² = constant\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 0.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile/12", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/1", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n B) I = V / R\n C) x(t) = v_0·t·cosθ · (1 - g·t/v_0)\n D) m·g·h + (1/2)·m·v_0² = constant\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/1", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile_newbatch/4", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) h_next = e²·h_current\n B) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n C) PV^γ = constant\n D) x(t) = v_0·t·cosθ · (1 - g·t/v_0)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "projectile_newbatch/5", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m·g·h + (1/2)·m·v_0² = constant\n B) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n C) y_max = v_0²·sin²θ / (2·g·cosθ)\n D) B = μ_0·I / (2π·r)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "projectile_newbatch/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) h_next = e²·h_current\n B) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n C) v_x(t) = v_0·cosθ · e^(-g·t/v_0)\n D) U = (1/2)·k·x²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.75 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 3.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "projectile_newbatch/9", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) m·g·h + (1/2)·m·v_0² = constant\n B) y = h·cos(g·x / v_0²) + x·tanθ\n C) ΔU = q·ΔV\n D) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile_newbatch/9", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/11", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) h_next = e²·h_current\n B) U = (1/2)·k·x²\n C) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n D) t_flight = 2·v_0·sin²θ / g\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "projectile_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "projectile_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "projectile_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "projectile_newbatch/11", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) P = ρ·g·h\n B) x(t) = x_0 + v_0·t\n C) F_D = (1/2)·C_d·ρ·a²·v²\n D) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "uniform_linear/6", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) T = 2π·√(L/g)\n B) p = m·v_0 · ln(1 + t/τ)\n C) F_push ≤ μ_s·m·g\n D) x(t) = x_0 + v_0·t\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 12, + "time_point": 6.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 13, + "time_point": 7.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 7.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 14, + "time_point": 7.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 15, + "time_point": 8.0 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 8.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 16, + "time_point": 8.5 + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 9.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 17, + "time_point": 9.0 + }, + { + "case": "uniform_linear/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) v = f·λ\n B) x(t) = x_0 + v_0·t\n C) x = v_0²·t / (2·x_0)\n D) F_push ≤ μ_s·m·g\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/0", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) x(t) = x_0 + v_0·t\n B) F_push ≤ μ_s·m·g\n C) PV^γ = constant\n D) v(t) = v_0 · e^(-t/τ_0)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear_newbatch/2", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) Q = m·c·ΔT\n B) x(t) = x_0·(1 + v_0·t/x_0)^(1/2)\n C) x(t) = x_0 + v_0·t\n D) F_ext - μ_k·m·g = m·a\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear_newbatch/4", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) x = v_0²·t / (2·x_0)\n B) F_push ≤ μ_s·m·g\n C) x(t) = x_0 + v_0·t\n D) B·dl = μ_0·I\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/4", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) x(t) = x_0 + v_0·t\n C) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n D) F_push ≤ μ_s·m·g\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "uniform_linear_newbatch/8", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) x(t) = x_0 + v_0·t\n B) F_ext - μ_k·m·g = m·a\n C) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n D) Q = m·c·ΔT\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/8", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) F_b = ρ_f·g·V\n B) x(t) = x_0 + v_0·t\n C) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n D) -μ_k·m·g · Δx = (1/2)·m·v² - (1/2)·m·v_0²\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 7, + "time_point": 4.0 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 4.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 8, + "time_point": 4.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 9, + "time_point": 5.0 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 5.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 10, + "time_point": 5.5 + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 6.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 11, + "time_point": 6.0 + }, + { + "case": "uniform_linear_newbatch/13", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) τ = I·α\n B) x(t) = x_0·(1 + v_0·t/x_0)^(1/2)\n C) x(t) = x_0 + v_0·t\n D) F_push ≤ μ_s·m·g\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/13", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "perception_text", + "prompt": "Given this physics infographic, generate a new image that shows ONLY the human-drawn annotations on a pure white background. Keep all text labels, coordinate axes, and velocity arrows exactly as they appear — same position, content, color, font, and size. Text must be crisp and readable. Remove everything else: the background, ground, sky, and all physical objects. Do not add any extra elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "perception_graphic", + "prompt": "Show only the subject physical objects from this image on a pure white background. Remove all annotations (text labels, coordinate axes, velocity arrows), the ground surface, sky, and supporting structures (ramps, tracks, rails, platforms). Keep only the objects that have labeled physical properties. Do not transform, resize, or replace any object — a cube must stay a cube, a sphere must stay a sphere. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows a scene with annotated physical parameters. Below are 4 candidate formulas. Exactly ONE is the correct governing equation for the motion shown:\n A) x(t) = x_0 + v_0·t\n B) P = ρ·g·h\n C) F_D = (1/2)·C_d·ρ·a²·v²\n D) a_uniform = v_0 / (x_0·t)\n\nGenerate an image with a pure white background containing THREE lines of large, clearly readable, centered text:\n Line 1: The correct option letter (A/B/C/D).\n Line 2: The correct formula exactly as written above.\n Line 3: The formula with annotated variables replaced by their numeric values from the image, keeping unannotated symbols unchanged.\n\nWhen substituting: if multiple objects exist, pick exactly one consistent set of values.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate an image showing the scene at that instant, where the original annotations (text labels, coordinate axes, arrows) have been cleanly removed while the background, ground, sky, and all scene lighting remain completely unchanged from the input. Show each object at its correct position at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, overlay a borderless orange arrow starting from the object's center pointing in its instantaneous velocity direction, with a label \"v = X.XX\" (speed value, no unit). If an object is at rest, show \"v = 0\" with no arrow. Keep the same viewpoint as the input image — no zoom, no rotation, no crop.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 0.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 0, + "time_point": 0.5 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 1, + "time_point": 1.0 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 1.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 2, + "time_point": 1.5 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 3, + "time_point": 2.0 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 2.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 4, + "time_point": 2.5 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.0 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 5, + "time_point": 3.0 + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "This image is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate what the scene looks like at t = 3.5 seconds after motion begins, following these physical parameters accurately. Remove all annotations (text, axes, arrows). Objects are rigid bodies — same shape, size, color, texture throughout. Static camera, fixed background and ground.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 1, + "size": "1280x720", + "prompt_idx": 6, + "time_point": 3.5 + } +] \ No newline at end of file diff --git a/bench_sampled/videogen.json b/bench_sampled/videogen.json new file mode 100644 index 0000000000000000000000000000000000000000..af3b28fb8f36d147337d94f78698258cf9526a92 --- /dev/null +++ b/bench_sampled/videogen.json @@ -0,0 +1,4002 @@ +[ + { + "case": "at_rest/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΣF = m·Δv / (Δt·cosθ)\n B) ΣF_ext = 0\n C) T = 2π·√(m/k)\n D) F_drag = (1/2)·C_d·ρ·A·v²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) ΣF_ext = 0\n C) G = m·g\n D) τ_net = Σ(F_i·r_i·sinθ_i) · e^(-μ·t)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) G = m·g\n B) ΣF = m·Δv / (Δt·cosθ)\n C) ΣF_ext = 0\n D) U = -G·m_1·m_2 / r\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) τ_net = Σ(F_i·r_i·sinθ_i) · e^(-μ·t)\n B) v = f·λ\n C) ΣF_ext = 0\n D) G = m·g\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/4", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΣF_ext = 0\n B) τ_net = Σ(F_i·r_i·sinθ_i) · e^(-μ·t)\n C) T = 2π·√(m/k)\n D) G = m·g\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n B) I = V / R\n C) ΣF = m·Δv / (Δt·cosθ)\n D) ΣM_ext = 0\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_eq = m·g / (1 + μ_s·tanθ)²\n B) v = f·λ\n C) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n D) ΣF_ext = 0\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΣF = m·Δv / (Δt·cosθ)\n B) ΣF_ext = 0\n C) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n D) U = -G·m_1·m_2 / r\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/10", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) G = m·g\n B) ΔS = Q_rev / T\n C) ΣM_ext = 0\n D) ΣM_ext = F·r·cos²θ / (1 + sinθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/12", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_drag = (1/2)·C_d·ρ·A·v²\n B) T = 2π·√(m/k)\n C) ΣF_ext = 0\n D) ΣM_ext = F·r·cos²θ / (1 + sinθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/13", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΣM_ext = F·r·cos²θ / (1 + sinθ)\n B) ΔS = Q_rev / T\n C) W = ΔE_k = (1/2)·m·v_t² - (1/2)·m·v_0²\n D) ΣM_ext = 0\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "at_rest_newbatch/14", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/at_rest_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) a_c = v²·cosθ / (R·sinθ)\n B) PV = nRT\n C) F = (3/4)·E*·√(R·d³)\n D) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) C = Q / V\n B) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n C) e = (v_2f - v_1f) / (v_1i - v_2i)\n D) F_N = m·g·(cosθ - 2) + m·v²/R\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) PV = nRT\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) F_N = m·g·(cosθ - 2) + m·v²/R\n D) F_N - m·g·cosθ = m·v²/R\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n B) F = (3/4)·E*·√(R·d³)\n C) F_N = m·g·(cosθ - 2) + m·v²/R\n D) v = f·λ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_N - m·g·cosθ = m·v²/R\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) ΔS = Q_rev / T\n D) ω = √(g·cosθ / R) · (1 + sinθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ω = √(g·cosθ / R) · (1 + sinθ)\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) Q = m·c·ΔT\n D) F_N - m·g·cosθ = m·v²/R\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/5", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) E_total = m·g·R·(3/2 - cosθ)²\n B) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n C) λ = h/(m·v)\n D) F_drag =(1/2)·C_d·ρ·A·v²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v² = g·R·(2 - 3·cosθ) / (1 + sinθ)\n B) f_k = μ_k·F_N\n C) ΔE = -13.6·(1/n_f² - 1/n_i²) eV\n D) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_drag =(1/2)·C_d·ρ·A·v²\n B) E_total = m·g·R·(3/2 - cosθ)²\n C) ΔS = Q_rev / T\n D) (1/2)·m·v_0² + m·g·R·(1-cosθ_start) = (1/2)·m·v² + m·g·R·(1-cosθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/11", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_N - m·g·cosθ = m·v²/R\n B) E_total = m·g·R·(3/2 - cosθ)²\n C) e = (v_2f - v_1f) / (v_1i - v_2i)\n D) B = μ_0·n·I\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/12", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔS = Q_rev / T\n B) F_N - m·g·cosθ = m·v²/R\n C) f_s ≤ μ_s·F_N\n D) a_c = v²·cosθ / (R·sinθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "circular_newbatch/14", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/circular_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) y(t) = h + v_0·t + (1/2)·g·t²\n C) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n D) P + (1/2)·ρ·v² + ρ·g·h = constant\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) Q = m·c·ΔT\n B) I = (1/10)·m·d²\n C) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/4", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v = v_0·cos(g·t / v_0)\n B) ΔS = Q_rev / T\n C) I = (1/10)·m·d²\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/5", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_final = -e·v_impact\n B) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n C) y(t) = h + v_0·t + (1/2)·g·t²\n D) n_1·sinθ_1 = n_2·sinθ_2\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n B) v = v_0 + g·t\n C) Q = m·c·ΔT\n D) P + (1/2)·ρ·v² + ρ·g·h = constant\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_impact = √(g·h) · (1 + μ·t)\n B) v = v_0 + g·t\n C) v_final = -e·v_impact\n D) T = 2π·√(m/k)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y(t) = h + v_0·t + (1/2)·g·t²\n B) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n C) Q = m·c·ΔT\n D) I = (1/10)·m·d²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y(t) = h + v_0·t + (1/2)·g·t²\n B) F·Δt = m·v_f - m·v_i\n C) τ = r × F\n D) y(t) = h·(1 - e^(-g·t²/2))\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y(t) = h·(1 - e^(-g·t²/2))\n B) v = f·λ\n C) v_final = -e·v_impact\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/4", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_impact = √(g·h) · (1 + μ·t)\n B) F = k·q_1·q_2 / r²\n C) I = (1/10)·m·d²\n D) y(t) = h + v_0·t + (1/2)·g·t²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔS = Q_rev / T\n B) y(t) = h + v_0·t + (1/2)·g·t²\n C) F_fall = m·g·(1 + v/c) / √(1 - v²/c²)\n D) F·Δt = m·v_f - m·v_i\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "freefall_newbatch/12", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/freefall_newbatch/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v² = v_0² + 2·a_slope·s\n B) ΔS = Q_rev / T\n C) E_p = m·g·(H - h)\n D) s = v_0·t·cosθ + (1/2)·g·sinθ·t²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) N = m·g·(cosθ - μ_k·sinθ)²\n B) Q = m·c·ΔT\n C) E_p = m·g·(H - h)\n D) v² = v_0² + 2·a_slope·s\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) U = -G·m_1·m_2 / r\n B) a_slope = g·(sinθ - μ_k·cosθ)\n C) E_p = m·g·(H - h)\n D) F_net = m·g·(sin²θ - μ_k·cos²θ) / sinθ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) T = 2π·√(L/g)\n B) v² = v_0² + 2·a_slope·s\n C) h_cm·sinθ > (a/2)·cosθ\n D) N = m·g·(cosθ - μ_k·sinθ)²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) f_k = μ_k·m·g\n B) v² = v_0² + g·sinθ·s·(1 - μ_k·tanθ)\n C) v² = v_0² + 2·a_slope·s\n D) F = k·q_1·q_2 / r²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) s = v_0·t·cosθ + (1/2)·g·sinθ·t²\n B) f_k = μ_k·m·g\n C) T = 2π·√(L/g)\n D) v² = v_0² + 2·a_slope·s\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) s = v_0·t·cosθ + (1/2)·g·sinθ·t²\n B) B·dl = μ_0·I\n C) a_slope = g·(sinθ - μ_k·cosθ)\n D) E_p = m·g·(H - h)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 0.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) Q = m·c·ΔT\n B) a_slope = g·(sinθ - μ_k·cosθ)\n C) v² = v_0² + g·sinθ·s·(1 - μ_k·tanθ)\n D) h_cm·sinθ > (a/2)·cosθ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) N = m·g·(cosθ - μ_k·sinθ)²\n B) v² = v_0² + 2·a_slope·s\n C) U = -G·m_1·m_2 / r\n D) h_cm·sinθ > (a/2)·cosθ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) E = (1/2)·L·I²\n B) v_t = g·d²·(ρ_p - ρ_f) / (18·η)\n C) a_slope = g·(sinθ - μ_k·cosθ)\n D) F_net = m·g·(sin²θ - μ_k·cos²θ) / sinθ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/9", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_t = g·d²·(ρ_p - ρ_f) / (18·η)\n B) ΔU = q·ΔV\n C) a_slope = g·(sinθ + μ_k·cosθ) / (1 + sinθ)\n D) a_slope = g·(sinθ - μ_k·cosθ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inclined_plane_newbatch/13", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inclined_plane_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) U = -G·m_1·m_2 / r\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) ΔE_loss = (1/2)·e²·(m_1·m_2)/(m_1+m_2)·(v_1i + v_2i)²\n D) W_f = ∫ μ_k·m·g·ds = ΔE_k\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n B) ΔE = (1/2)·m_1·m_2·e²·(v_1i - v_2i)² / (m_1 + m_2)\n C) F = k·q_1·q_2 / r²\n D) e = (v_2f - v_1f) / (v_1i - v_2i)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_1f = (m_1 - e·m_2)·v_1i / (m_1 + m_2) + v_2i·cosθ\n B) σ = F/a² = E·ε\n C) E = h·f\n D) e = (v_2f - v_1f) / (v_1i - v_2i)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) v_cm = (m_1·v_1i² + m_2·v_2i²) / (m_1·v_1i + m_2·v_2i)\n C) σ = F/a² = E·ε\n D) T = 2π·√(m/k)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) σ = F/a² = E·ε\n B) e = (v_2f - v_1f) / (v_1i - v_2i)\n C) J = m_1·m_2·(v_1i - v_2i)·(1+e) / (m_1 + m_2) · sinθ\n D) v = f·λ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 4.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 4.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision/9", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) e = 1 - ΔE / (m_1·v_1i² + m_2·v_2i²)\n B) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n C) E_rot = (1/2)·(1/6·m·a²)·ω²\n D) PV = nRT\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 6 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) E_rot = (1/2)·(1/6·m·a²)·ω²\n C) B = μ_0·I / (2π·r)\n D) ΔE_loss = (1/2)·e²·(m_1·m_2)/(m_1+m_2)·(v_1i + v_2i)²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F = k·q_1·q_2 / r²\n B) e = 1 - ΔE / (m_1·v_1i² + m_2·v_2i²)\n C) W_f = ∫ μ_k·m·g·ds = ΔE_k\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) λ = h/(m·v)\n B) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n C) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n D) J = m_1·m_2·(v_1i - v_2i)·(1+e) / (m_1 + m_2) · sinθ\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v = f·λ\n B) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n C) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n D) e = 1 - ΔE / (m_1·v_1i² + m_2·v_2i²)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔE = (1/2)·[(m_1·m_2) / (m_1 + m_2)]·(v_1i - v_2i)²\n B) v = f·λ\n C) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n D) ΔE_loss = (1/2)·e²·(m_1·m_2)/(m_1+m_2)·(v_1i + v_2i)²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 6 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n B) B = μ_0·I / (2π·r)\n C) ΔE = μ_k·m·g·d\n D) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) v_1f = v_1i · e^(-m_2/m_1) + v_2i·(1 - e^(-m_2/m_1))\n C) v_2f - v_1f = e·(v_1i - v_2i)\n D) L = I·ω\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔE = μ_k·m·g·d\n B) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n C) λ = h/(m·v)\n D) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/5", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ∫ F(t) dt = Δ(mv)\n B) v_1f = v_1i·(m_1² - m_2²) / (m_1 + m_2)²\n C) λ = h/(m·v)\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 6 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) I = I_cm + m·(a/2)²\n B) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n C) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n D) F = k·q_1·q_2 / r²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 9 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F = k·q_1·q_2 / r²\n B) ΔE = μ_k·m·g·d\n C) v_2f = 2·m_1·v_1i·cosθ / (m_1 + m_2)\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 4.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 4.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision/9", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) ∫ F(t) dt = Δ(mv)\n C) v_1f = v_1i · e^(-m_2/m_1) + v_2i·(1 - e^(-m_2/m_1))\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔS = Q_rev / T\n B) v_2f - v_1f = e·(v_1i - v_2i)\n C) ΔE = (1/2)·m_1·(v_1f - v_1i)²·(m_1/m_2 + 1)\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) I = I_cm + m·(a/2)²\n B) F = k·q_1·q_2 / r²\n C) E_k = (1/2)·(m_1·v_1i + m_2·v_2i)² / (m_1 + m_2)\n D) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/3", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision_newbatch/3/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m_1·v_1i + m_2·v_2i = m_1·v_1f + m_2·v_2f\n B) Q = m·c·ΔT\n C) I = I_cm + m·(a/2)²\n D) v_1f = v_1i·(m_1² - m_2²) / (m_1 + m_2)²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/4", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) E_k = (1/2)·(m_1·v_1i + m_2·v_2i)² / (m_1 + m_2)\n B) v_2f - v_1f = e·(v_1i - v_2i)\n C) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n D) Q = m·c·ΔT\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/13", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) T = 2π·√(m/k)\n B) v_cm = (m_1·v_1i² + m_2·v_2i²) / (2·(m_1·v_1i + m_2·v_2i))\n C) ΔE = μ_k·m·g·d\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·m_1·v_1f² + (1/2)·m_2·v_2f²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_elastic_collision_newbatch/14", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_elastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·(m_1 + m_2)·v_f²\n B) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n C) ΔE = (1/2)·(m_1·m_2)·(v_1i + v_2i)² / (m_1 + m_2)²\n D) PV = nRT\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) I = V / R\n B) m = ρ·a³\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) ΔE_loss = m_1·m_2·(v_1i - v_2i) / (2·(m_1 + m_2))\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/5", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 6 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) ΔS = Q_rev / T\n B) ΔE_loss = m_1·m_2·(v_1i - v_2i) / (2·(m_1 + m_2))\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) v_2f - v_1f = e·(v_1i - v_2i)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) W_friction = μ_k·(m_1 + m_2)·g·d\n B) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n C) E_k_f = (m_1²·v_1i² + m_2²·v_2i²) / (2·(m_1 + m_2))\n D) ΔS = Q_rev / T\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n B) ΔE_loss = m_1·m_2·(v_1i - v_2i) / (2·(m_1 + m_2))\n C) F_b = ρ_f·g·V\n D) m = ρ·a³\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) PV = nRT\n B) v_f = (m_1·v_1i + m_2·v_2i)² / ((m_1 + m_2)·(v_1i + v_2i))\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) m = ρ·a³\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_f² = (m_1·v_1i + m_2·v_2i)² / (m_1² + m_2²)\n B) W_friction = μ_k·(m_1 + m_2)·g·d\n C) E_n = -13.6/n² eV\n D) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_f² = (m_1·v_1i + m_2·v_2i)² / (m_1² + m_2²)\n B) m = ρ·a³\n C) n_1·sinθ_1 = n_2·sinθ_2\n D) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) W_friction = μ_k·(m_1 + m_2)·g·d\n B) v_f² = (m_1·v_1i + m_2·v_2i)² / (m_1² + m_2²)\n C) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n D) PV = nRT\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 10 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) B = μ_0·I / (2π·r)\n B) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n C) v_f = √(m_1/m_2)·v_1i + √(m_2/m_1)·v_2i\n D) (1/2)·m_1·v_1i² + (1/2)·m_2·v_2i² = (1/2)·(m_1 + m_2)·v_f²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/11", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 7 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v_f = (m_1·v_1i + m_2·v_2i)² / ((m_1 + m_2)·(v_1i + v_2i))\n B) I = V / R\n C) m = ρ·a³\n D) m_1·v_1i + m_2·v_2i = (m_1+m_2)·v_f\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "perfectly_inelastic_collision_newbatch/14", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/perfectly_inelastic_collision_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) h_next = e²·h_current\n B) B = μ_0·I / (2π·r)\n C) y = h·cos(g·x / v_0²) + x·tanθ\n D) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n B) ΔS = Q_rev / T\n C) x(t) = v_0·t·cosθ · (1 - g·t/v_0)\n D) h_next = e²·h_current\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/7", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile/7/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n B) a_n = v² / ρ\n C) t_flight = 2·v_0·sin²θ / g\n D) L = I·ω\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) a_n = v² / ρ\n B) v_x(t) = v_0·cosθ · e^(-g·t/v_0)\n C) T = 2π·√(L/g)\n D) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/10", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y = h·cos(g·x / v_0²) + x·tanθ\n B) L = I·ω\n C) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n D) m·g·h + (1/2)·m·v_0² = constant\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 0.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 0.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile/12", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile/12/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n B) I = V / R\n C) x(t) = v_0·t·cosθ · (1 - g·t/v_0)\n D) m·g·h + (1/2)·m·v_0² = constant\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/1", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile_newbatch/1/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) h_next = e²·h_current\n B) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n C) PV^γ = constant\n D) x(t) = v_0·t·cosθ · (1 - g·t/v_0)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/4", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m·g·h + (1/2)·m·v_0² = constant\n B) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n C) y_max = v_0²·sin²θ / (2·g·cosθ)\n D) B = μ_0·I / (2π·r)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.5 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.5 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/5", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) h_next = e²·h_current\n B) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n C) v_x(t) = v_0·cosθ · e^(-g·t/v_0)\n D) U = (1/2)·k·x²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 3.75 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 3.75 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile_newbatch/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) m·g·h + (1/2)·m·v_0² = constant\n B) y = h·cos(g·x / v_0²) + x·tanθ\n C) ΔU = q·ΔV\n D) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/9", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 1 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile_newbatch/9/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) h_next = e²·h_current\n B) U = (1/2)·k·x²\n C) y = h + x·tanθ - g·x² / (2·v_0²·cos²θ)\n D) t_flight = 2·v_0·sin²θ / g\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 1.25 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 1.25 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "projectile_newbatch/11", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/projectile_newbatch/11/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) P = ρ·g·h\n B) x(t) = x_0 + v_0·t\n C) F_D = (1/2)·C_d·ρ·a²·v²\n D) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 4 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) T = 2π·√(L/g)\n B) p = m·v_0 · ln(1 + t/τ)\n C) F_push ≤ μ_s·m·g\n D) x(t) = x_0 + v_0·t\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/6", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 9 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear/6/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) v = f·λ\n B) x(t) = x_0 + v_0·t\n C) x = v_0²·t / (2·x_0)\n D) F_push ≤ μ_s·m·g\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) x(t) = x_0 + v_0·t\n B) F_push ≤ μ_s·m·g\n C) PV^γ = constant\n D) v(t) = v_0 · e^(-t/τ_0)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/0", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/0/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) Q = m·c·ΔT\n B) x(t) = x_0·(1 + v_0·t/x_0)^(1/2)\n C) x(t) = x_0 + v_0·t\n D) F_ext - μ_k·m·g = m·a\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/2", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/2/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) x = v_0²·t / (2·x_0)\n B) F_push ≤ μ_s·m·g\n C) x(t) = x_0 + v_0·t\n D) B·dl = μ_0·I\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/4", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/4/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) n_1·sinθ_1 = n_2·sinθ_2\n B) x(t) = x_0 + v_0·t\n C) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n D) F_push ≤ μ_s·m·g\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/5", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/5/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) x(t) = x_0 + v_0·t\n B) F_ext - μ_k·m·g = m·a\n C) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n D) Q = m·c·ΔT\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/8", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/8/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) F_b = ρ_f·g·V\n B) x(t) = x_0 + v_0·t\n C) x(t) = x_0 + v_0·t + (v_0²/c²)·t²\n D) -μ_k·m·g · Δx = (1/2)·m·v² - (1/2)·m·v_0²\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/10", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 6 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/10/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) τ = I·α\n B) x(t) = x_0·(1 + v_0·t/x_0)^(1/2)\n C) x(t) = x_0 + v_0·t\n D) F_push ≤ μ_s·m·g\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/13", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 2 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/13/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "perception_text", + "prompt": "Generate a video where the background, ground, sky, and all physical objects gradually fade to pure white, while all human-drawn annotations (text labels, coordinate axes, velocity arrows) remain completely unchanged. Static camera, no zoom, no rotation. The final frame should show only the original annotations on white — crisp, sharp, and readable. Do not add any extra annotations or elements not in the original.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "perception_graphic", + "prompt": "Generate a video where the background, ground, sky, and all annotations (text labels, coordinate axes, velocity arrows) gradually fade to pure white, while the subject physical objects remain still and unchanged. Static camera, no zoom, no rotation. The final frame should show only the objects with labeled physical properties on white — remove supporting structures (ramps, tracks, rails, platforms). Do not transform, resize, or replace any object. Each object must retain its exact position, size, shape, color, and texture.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "comprehension_text", + "prompt": "This physics infographic shows annotated physical parameters describing a specific type of motion. Below are 4 candidate formulas. Exactly ONE is correct:\n A) x(t) = x_0 + v_0·t\n B) P = ρ·g·h\n C) F_D = (1/2)·C_d·ρ·a²·v²\n D) a_uniform = v_0 / (x_0·t)\n\nGenerate a video ending on a white background with THREE lines of large, clearly readable, centered text:\n Line 1: the correct option letter (A/B/C/D).\n Line 2: the correct formula exactly as written above.\n Line 3: the formula with annotated values substituted in.\n\nWhen substituting: if multiple objects exist, pick one consistent set.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "comprehension_graphic", + "prompt": "Predict the physical state at t = 2.0 s. Generate a video where the original annotations (text labels, coordinate axes, arrows) gradually fade out and cleanly disappear, while the background, ground, sky, and all scene lighting remain completely unchanged throughout. Objects move smoothly to their correct positions at t = 2.0 s — each object must retain its exact texture, material, size, and shape. For each moving object, draw a borderless orange arrow starting from the object's center pointing in its velocity direction, with a label \"v = X.XX\" (speed value, no unit). If at rest, show \"v = 0\" with no arrow. Static camera, no zoom, no rotation.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + }, + { + "case": "uniform_linear_newbatch/14", + "track": "generation", + "prompt": "The first frame is an annotated physics scene. All annotations follow SI units: g is gravitational acceleration, e is restitution (0-1), v0 is initial velocity (arrows show direction), mu is friction coefficient. Generate 3.5 seconds of physically accurate motion strictly following these parameters. Every object AND the ground surface are perfectly rigid — nothing breaks, cracks, deforms, or creates craters. Each object maintains its exact material, color, and texture throughout. All annotations must cleanly disappear after the first frame. Static camera, fixed background. Smooth, continuous motion.", + "image": "firstframe/uniform_linear_newbatch/14/rgb_0000.png", + "duration": 8, + "size": "1280x720" + } +] \ No newline at end of file diff --git a/firstframe_sampled/at_rest/0/rgb_0000.png b/firstframe_sampled/at_rest/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..171d8cbc0d0f3d937c0232970253147f7664feb8 --- /dev/null +++ b/firstframe_sampled/at_rest/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f71c83170890748d841c9c8f094b9998355707079d33fc2545d15913281a6bbc +size 1714598 diff --git a/firstframe_sampled/at_rest/2/rgb_0000.png b/firstframe_sampled/at_rest/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..002aeb689aed3efcff27e2d91726cfacd96a731d --- /dev/null +++ b/firstframe_sampled/at_rest/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d51d9353040678cc0917edbbbbe6fd9932887e988ecf7ed86881bbc265c85405 +size 1756582 diff --git a/firstframe_sampled/at_rest/3/rgb_0000.png b/firstframe_sampled/at_rest/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..7c031d99131defe3b1ae426cfa2f010ed8638a5e --- /dev/null +++ b/firstframe_sampled/at_rest/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ae863cdbf97ed2eac5e162b2f850ead95e0b7a1a62d2ea5563e6f01aad4c9590 +size 1819869 diff --git a/firstframe_sampled/at_rest/4/rgb_0000.png b/firstframe_sampled/at_rest/4/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..26d2ec44b80fd6ed8e1e16eb5bcb9559b6947842 --- /dev/null +++ b/firstframe_sampled/at_rest/4/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c7ab6dd4bf9a03a6634b2eccd2d12ff5dc42c6350186226c2a556edb587c8b0e +size 1803654 diff --git a/firstframe_sampled/at_rest/7/rgb_0000.png b/firstframe_sampled/at_rest/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..ab9fcb2e790a4e4b9ed90d99559021ef96ae555b --- /dev/null +++ b/firstframe_sampled/at_rest/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:95aadcb5f33db3b20125ca060385bcd1a411836ce7880605990ce245cc86073c +size 1827613 diff --git a/firstframe_sampled/at_rest/8/rgb_0000.png b/firstframe_sampled/at_rest/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..f1e0cffeff439cf86f38234987b903f76ebeb9d6 --- /dev/null +++ b/firstframe_sampled/at_rest/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e83def7de3b905c82ad83796d90cc6174497b4557b1677018694a9b0581d0a36 +size 1672281 diff --git a/firstframe_sampled/at_rest_newbatch/10/rgb_0000.png b/firstframe_sampled/at_rest_newbatch/10/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..6348b94f6af0ee92f46945a88e305dc81371ad86 --- /dev/null +++ b/firstframe_sampled/at_rest_newbatch/10/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74c2a1012cbe9f252da49e829a683784ef9c87f1225d998112e6f7a8efc780dc +size 1602350 diff --git a/firstframe_sampled/at_rest_newbatch/12/rgb_0000.png b/firstframe_sampled/at_rest_newbatch/12/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..38cd6567b3f6587fb58eb4c745ee1aafdee44ea7 --- /dev/null +++ b/firstframe_sampled/at_rest_newbatch/12/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a145c3ee1688fb475d074a71e0a1b630f1bfd8107b0d7041e27851f35b54a4d2 +size 1617171 diff --git a/firstframe_sampled/at_rest_newbatch/13/rgb_0000.png b/firstframe_sampled/at_rest_newbatch/13/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..d0fd3a5b91b230d3f6f081dcf3671abc906a1b3b --- /dev/null +++ b/firstframe_sampled/at_rest_newbatch/13/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e98f0741b619a69aaf20fc3e5091ac5c6d2754913502e1731af01537f9224793 +size 1036376 diff --git a/firstframe_sampled/at_rest_newbatch/14/rgb_0000.png b/firstframe_sampled/at_rest_newbatch/14/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..898a1c1d587b8cace8cdff987d3a2e6f61d76b83 --- /dev/null +++ b/firstframe_sampled/at_rest_newbatch/14/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a7031805053e7dd3f2fbcafd5f90efd74b85c09e9727875d995aae7170d39a76 +size 1310313 diff --git a/firstframe_sampled/at_rest_newbatch/3/rgb_0000.png b/firstframe_sampled/at_rest_newbatch/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..cc2bb4e6a68a52cf4a5cc51464d225cb1f4465e2 --- /dev/null +++ b/firstframe_sampled/at_rest_newbatch/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cad6eeedb21eeef8cb6883acf3a1af971ca2c946719f513539522288437ac49b +size 1484196 diff --git a/firstframe_sampled/circular/0/rgb_0000.png b/firstframe_sampled/circular/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..cdd1f4eaa30b6c634b788d080c0495e0cf98abae --- /dev/null +++ b/firstframe_sampled/circular/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:475c8a3af9a31bd520a90e3485b463467fbcf65e964ad558caa2fe12560772ae +size 965556 diff --git a/firstframe_sampled/circular/2/rgb_0000.png b/firstframe_sampled/circular/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..c8821a6e1d6ba4e011aba534f74361ece13284fc --- /dev/null +++ b/firstframe_sampled/circular/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e06bab341f7e0f02e87449c961fafc263225a503b109cf9278a91012c98615a7 +size 1206189 diff --git a/firstframe_sampled/circular/6/rgb_0000.png b/firstframe_sampled/circular/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..e77e1a1b52c07dad0bb88bcbf616d6ee8ece3d36 --- /dev/null +++ b/firstframe_sampled/circular/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d848656a05b963ac733afdb2afba73ac0bfae7f426f04fcaa7568ed26eebd47e +size 721225 diff --git a/firstframe_sampled/circular/7/rgb_0000.png b/firstframe_sampled/circular/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..7f7c454565439e24aa40715be11499fc99538ab2 --- /dev/null +++ b/firstframe_sampled/circular/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83a5537143c4d5a69d5cb3a6980c1d9c4a07d19760ae78552e2ff77abc7415bb +size 626557 diff --git a/firstframe_sampled/circular_newbatch/11/rgb_0000.png b/firstframe_sampled/circular_newbatch/11/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..042023b42b8f94d691e3038205b3242e2e842001 --- /dev/null +++ b/firstframe_sampled/circular_newbatch/11/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6ea2cfc00de6b28502760474d1ec96e2a54d1b9c0e2b15f6de7cfdb28c67585 +size 540122 diff --git a/firstframe_sampled/circular_newbatch/12/rgb_0000.png b/firstframe_sampled/circular_newbatch/12/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..1bb7eb2f16c7055924b3e76c1d25536cf898dd94 --- /dev/null +++ b/firstframe_sampled/circular_newbatch/12/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b2ae0d4684aa0c04edb042b651b45a5b13217895b28e5c96a9d66dba3652cbc4 +size 1552970 diff --git a/firstframe_sampled/circular_newbatch/14/rgb_0000.png b/firstframe_sampled/circular_newbatch/14/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..a7dd98c1ae5b5d440442add6dbe626930ad727e4 --- /dev/null +++ b/firstframe_sampled/circular_newbatch/14/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3948af855264fae563fa11d7346a255fbb3267fd30385eac9d23ce533b1f096 +size 593329 diff --git a/firstframe_sampled/circular_newbatch/2/rgb_0000.png b/firstframe_sampled/circular_newbatch/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..cd7e564bfe600453f29b33ab662c484eb18e9ec6 --- /dev/null +++ b/firstframe_sampled/circular_newbatch/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9d3ad39bfe3d1507ae7046b4a16597fe5533802a5278bc5e95b5bfa8bd9fae26 +size 591819 diff --git a/firstframe_sampled/circular_newbatch/5/rgb_0000.png b/firstframe_sampled/circular_newbatch/5/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..10184cf100fcc258f492a064b94c6ec545927668 --- /dev/null +++ b/firstframe_sampled/circular_newbatch/5/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ce451e0118b7b5c63066261f8c6692fa23e8519ae3b934e8d0ebd0c5a8df32be +size 550171 diff --git a/firstframe_sampled/circular_newbatch/6/rgb_0000.png b/firstframe_sampled/circular_newbatch/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..a9c494c742326a0e3c18d8eec854cd120398842c --- /dev/null +++ b/firstframe_sampled/circular_newbatch/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b561da6e0dc3a22c8c1c132fda1144e9eb5b4b5128b29440006f112952390544 +size 1806797 diff --git a/firstframe_sampled/circular_newbatch/8/rgb_0000.png b/firstframe_sampled/circular_newbatch/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..7dbdb1d21d7d2d056b6d0f4ae7b70ed71f08379c --- /dev/null +++ b/firstframe_sampled/circular_newbatch/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dde47f8cf565227791418e43536885e9cb0afade29de620e6418d696df8e9a3f +size 1734790 diff --git a/firstframe_sampled/freefall/0/rgb_0000.png b/firstframe_sampled/freefall/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..e98149a2c5dca462971fca129be7f5c28ecdba57 --- /dev/null +++ b/firstframe_sampled/freefall/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a1dcef25456f9c8ddc6a8801210a9222c4f2337248f2388b9be0bad65c164bc +size 1116033 diff --git a/firstframe_sampled/freefall/4/rgb_0000.png b/firstframe_sampled/freefall/4/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..2cb3fda46baaf6bfbc8ed659622e7b1ba96bfd49 --- /dev/null +++ b/firstframe_sampled/freefall/4/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87afd4e692c5474b9189de1f0a593b878e6caa626bb3ac29075023c509cf1e79 +size 1428964 diff --git a/firstframe_sampled/freefall/5/rgb_0000.png b/firstframe_sampled/freefall/5/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..966f685df0dd9b4b3a0ed58bc3bd2c1b9ae1c2b3 --- /dev/null +++ b/firstframe_sampled/freefall/5/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab650f46255170e3df48ecc2e2c474c705ecc10c53ab9e7ebb42b5b4f4c33a0b +size 1429087 diff --git a/firstframe_sampled/freefall/6/rgb_0000.png b/firstframe_sampled/freefall/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..a76c76a99efcc20777fac577f0a646ae1d0df64b --- /dev/null +++ b/firstframe_sampled/freefall/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ed08e4a0fc93ea7b03d3bfb0bf6bc9436df88fe49c0ab674e9beb2436ba6616 +size 1433212 diff --git a/firstframe_sampled/freefall/8/rgb_0000.png b/firstframe_sampled/freefall/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..8683981e98582eb077022c4426a56db49af784c8 --- /dev/null +++ b/firstframe_sampled/freefall/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b1dc984e9e45bf7f4f065362bfd6c78754f894b42946f9514eac1b762e8a7d04 +size 1451704 diff --git a/firstframe_sampled/freefall_newbatch/0/rgb_0000.png b/firstframe_sampled/freefall_newbatch/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..a1a2a7f5df60b5b9f157107e3706d9efae34bc36 --- /dev/null +++ b/firstframe_sampled/freefall_newbatch/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a5c8972c5e0b28116ae0829fd03ed62cf7ad225c7c6982fbc995b9fa63cc738 +size 1566439 diff --git a/firstframe_sampled/freefall_newbatch/12/rgb_0000.png b/firstframe_sampled/freefall_newbatch/12/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..9d12e89b82a845119098c6b71505e7b06361f1b2 --- /dev/null +++ b/firstframe_sampled/freefall_newbatch/12/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d2567207815bd38be3fcf99fe2e758a577b076e45676f17d6ec70aa44ff5786e +size 1838307 diff --git a/firstframe_sampled/freefall_newbatch/2/rgb_0000.png b/firstframe_sampled/freefall_newbatch/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..86fcf2483998cc96849de0882c99deaa2e4bbcec --- /dev/null +++ b/firstframe_sampled/freefall_newbatch/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ffc6ada6d4545bbe2ba0154bb70d19fecef12f1fe5769fd7d2a92f326fb2111 +size 1685327 diff --git a/firstframe_sampled/freefall_newbatch/3/rgb_0000.png b/firstframe_sampled/freefall_newbatch/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..35e556601c9f47fef5d88c3c9c5261e72235ce0e --- /dev/null +++ b/firstframe_sampled/freefall_newbatch/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c963aae7429729eb09de91156ff9d16176fbee3b0a3122e83eb1dcca28c4dfa +size 1445287 diff --git a/firstframe_sampled/freefall_newbatch/4/rgb_0000.png b/firstframe_sampled/freefall_newbatch/4/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..d846bc7e698fb6f71ebf0abc6fd87d096dd215cc --- /dev/null +++ b/firstframe_sampled/freefall_newbatch/4/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e19d1476eec255fca6e044caf6c2462eda134058f26539926e3cb7262e7ee3e5 +size 1307877 diff --git a/firstframe_sampled/freefall_newbatch/8/rgb_0000.png b/firstframe_sampled/freefall_newbatch/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..6e9e3016b6df20603f898c686f7fc034458d72df --- /dev/null +++ b/firstframe_sampled/freefall_newbatch/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e139e721d22a69ab88de9f78be9bac5a45b67590a983c23dd39a7d85a5e5e86f +size 1374094 diff --git a/firstframe_sampled/inclined_plane/1/rgb_0000.png b/firstframe_sampled/inclined_plane/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..acacefad0944428f63f853eed5a5430357921dbe --- /dev/null +++ b/firstframe_sampled/inclined_plane/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e14916833f9dda971373861d3ecc964d40fac85179036e97743e48d203a472a +size 1038963 diff --git a/firstframe_sampled/inclined_plane/2/rgb_0000.png b/firstframe_sampled/inclined_plane/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..84e8c10a1345441d538ba9b6daa83cb0a08869e2 --- /dev/null +++ b/firstframe_sampled/inclined_plane/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bab808ccbf9ba000fae8c8ef7fbe36f549d774c5013b8361f7c53b1beee58868 +size 930761 diff --git a/firstframe_sampled/inclined_plane/3/rgb_0000.png b/firstframe_sampled/inclined_plane/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..35f1216378b9786dc990ed6e18fea3d61365a0f0 --- /dev/null +++ b/firstframe_sampled/inclined_plane/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c00e67fb0676f5424b041b9f0fa5ddbd2b889fe2bc31e223b71d25ad7af7d08b +size 1093199 diff --git a/firstframe_sampled/inclined_plane/7/rgb_0000.png b/firstframe_sampled/inclined_plane/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..3e2c050d31d79fb3cf9cc3738882ff5b191599d2 --- /dev/null +++ b/firstframe_sampled/inclined_plane/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd179dbbb3ab1e84a00ee931572bf090f32f6a89176ca6b4ba51c4690e4a7804 +size 1055730 diff --git a/firstframe_sampled/inclined_plane/8/rgb_0000.png b/firstframe_sampled/inclined_plane/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..7ec8c688a76430c59688d2c9564d91dd05b31332 --- /dev/null +++ b/firstframe_sampled/inclined_plane/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ee81308d6dd83daa7e7fb0237721e16e3f8ac62013fecafdacb776a44a56483 +size 1031432 diff --git a/firstframe_sampled/inclined_plane_newbatch/1/rgb_0000.png b/firstframe_sampled/inclined_plane_newbatch/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..6352f6ac4fe826ff4f9f718b1315c9897a8c18c6 --- /dev/null +++ b/firstframe_sampled/inclined_plane_newbatch/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18b48cd4fd8e4c49c024ba3bf6cf9917c389baf5ea2872837fa82c486206bbea +size 1128079 diff --git a/firstframe_sampled/inclined_plane_newbatch/13/rgb_0000.png b/firstframe_sampled/inclined_plane_newbatch/13/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..1994caf094e0077d3cd592ca8f9f4a88487a3b79 --- /dev/null +++ b/firstframe_sampled/inclined_plane_newbatch/13/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:43200cb94b9ffc7467ef053156a96b0c094fdaae000d040e8654b7beea57779e +size 1042628 diff --git a/firstframe_sampled/inclined_plane_newbatch/2/rgb_0000.png b/firstframe_sampled/inclined_plane_newbatch/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..0e57afdf846b42856a73be82e0a8ee165c481f96 --- /dev/null +++ b/firstframe_sampled/inclined_plane_newbatch/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f9a469b3eead2e3df85266543a95441abfc4c356e607d5a33722382c1e3b044 +size 1137060 diff --git a/firstframe_sampled/inclined_plane_newbatch/3/rgb_0000.png b/firstframe_sampled/inclined_plane_newbatch/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..ca1eeabba943bf3443119941a70b1e3493c107ac --- /dev/null +++ b/firstframe_sampled/inclined_plane_newbatch/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:98330fd4a0c326a9b219a3d687c451d5da70185bb840c66484086a1c7b4210b8 +size 1097063 diff --git a/firstframe_sampled/inclined_plane_newbatch/8/rgb_0000.png b/firstframe_sampled/inclined_plane_newbatch/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..9efd3a0123fb6bce44570276b74e07e067fc0f3f --- /dev/null +++ b/firstframe_sampled/inclined_plane_newbatch/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06aae28de9579a57aecbf65d648f6d46ea586faa46dd1959954680cbf20365f9 +size 1120199 diff --git a/firstframe_sampled/inclined_plane_newbatch/9/rgb_0000.png b/firstframe_sampled/inclined_plane_newbatch/9/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..88e1a9ecbfe55a372b4c81052909715de0d76950 --- /dev/null +++ b/firstframe_sampled/inclined_plane_newbatch/9/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:374084c87e548c483cf6310d37abedf604191c9276e1cbee5f654af861126899 +size 923646 diff --git a/firstframe_sampled/inelastic_collision/1/rgb_0000.png b/firstframe_sampled/inelastic_collision/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..f17d5baba2338c1e1fbbf93152a40d7df884f111 --- /dev/null +++ b/firstframe_sampled/inelastic_collision/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31afaebf8675497b85eb45ee150c9582db5b21519193801989d6d12677194add +size 602242 diff --git a/firstframe_sampled/inelastic_collision/2/rgb_0000.png b/firstframe_sampled/inelastic_collision/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..99e7c48f56128463f261348e572d621cf2cf6f33 --- /dev/null +++ b/firstframe_sampled/inelastic_collision/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4e554909e56f67d0ac203703814c257454bce0f2038ec7506296149998faf395 +size 622217 diff --git a/firstframe_sampled/inelastic_collision/3/rgb_0000.png b/firstframe_sampled/inelastic_collision/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..52e80ba47fbd5b62c1a25e1eba56d892a209c1d7 --- /dev/null +++ b/firstframe_sampled/inelastic_collision/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fdf9b1d77c663a22aaadbe4f0bbfcd1ee787db0b6e1f0c6e224751ca16b1d253 +size 623222 diff --git a/firstframe_sampled/inelastic_collision/6/rgb_0000.png b/firstframe_sampled/inelastic_collision/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..6d264999210acf46f5dfbb65a8aebcd7bfa54c73 --- /dev/null +++ b/firstframe_sampled/inelastic_collision/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eca1b120e10d585a22ecfd680125fc0a1f87d4a3d78a0d3d894c43e2ce2e7aec +size 578011 diff --git a/firstframe_sampled/inelastic_collision/9/rgb_0000.png b/firstframe_sampled/inelastic_collision/9/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..4a2248f294aab4533fcd44c7ee72d7575e848479 --- /dev/null +++ b/firstframe_sampled/inelastic_collision/9/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a4b96890be94de1bce147c31bbff931c25df9bb2e17d7a7fafddc1a712ec23e +size 776076 diff --git a/firstframe_sampled/inelastic_collision_newbatch/0/rgb_0000.png b/firstframe_sampled/inelastic_collision_newbatch/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..43829c26b8542d68cd7a03a0dc64370960b8e198 --- /dev/null +++ b/firstframe_sampled/inelastic_collision_newbatch/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:348a9389161e03e523b5d32b57112aaa9e30c08ddbbe4976576a6c33382aa63f +size 510595 diff --git a/firstframe_sampled/inelastic_collision_newbatch/1/rgb_0000.png b/firstframe_sampled/inelastic_collision_newbatch/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..824d1886677615a0f7248d87931ecb3f7ee8ede9 --- /dev/null +++ b/firstframe_sampled/inelastic_collision_newbatch/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3aff61bc3eb0d6fa449c97e250d731dd7b082fd0f53f31a39b1085da64b0391b +size 520647 diff --git a/firstframe_sampled/inelastic_collision_newbatch/11/rgb_0000.png b/firstframe_sampled/inelastic_collision_newbatch/11/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..ba39dbaef50c2cdbd1e3bf19ddff3dbe12dc31c0 --- /dev/null +++ b/firstframe_sampled/inelastic_collision_newbatch/11/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cc514ba78090d38391bf9c8bbc6257345f2e870e28e699a5236e644cbf26ceed +size 455952 diff --git a/firstframe_sampled/inelastic_collision_newbatch/14/rgb_0000.png b/firstframe_sampled/inelastic_collision_newbatch/14/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..2a06e44a089de1bccc1c43297ff7fcfd5f995d9d --- /dev/null +++ b/firstframe_sampled/inelastic_collision_newbatch/14/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2fcaa9465bdcbedd0257df3d58fcfd8d9d930fb71b5c64ae62413a80e28c040e +size 552289 diff --git a/firstframe_sampled/inelastic_collision_newbatch/2/rgb_0000.png b/firstframe_sampled/inelastic_collision_newbatch/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..761a8b856ced14dc61873e1e0bb91f0a65d90bbc --- /dev/null +++ b/firstframe_sampled/inelastic_collision_newbatch/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1204cec700041ed3b50961f8d50a01ead9e7f6dfd9df756b91a2a2dad1b16cae +size 484751 diff --git a/firstframe_sampled/inelastic_collision_newbatch/7/rgb_0000.png b/firstframe_sampled/inelastic_collision_newbatch/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..90bc737488d4bbcb7b25044750a8cbeab7bf6816 --- /dev/null +++ b/firstframe_sampled/inelastic_collision_newbatch/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2f0bbb2e6621df6848f9c17e07ddb42ac479863f01c045493bede38144bcd39 +size 400911 diff --git a/firstframe_sampled/perfectly_elastic_collision/2/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..808a39b42d70dbe705c280f9a5955ff052a1bf60 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:847345850ac3170cb6748ac17c021a4705441f4506ca07a560c1257ffeb3a4f5 +size 582941 diff --git a/firstframe_sampled/perfectly_elastic_collision/3/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..42004711d48870b9ab73c8d885648ec29eacff42 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c94142ae442530e51beedae84d35464a0da21c59572b1a8f016326138c88083 +size 562057 diff --git a/firstframe_sampled/perfectly_elastic_collision/5/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision/5/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..73701bfc9537e06c148a84b8405bfd3ed4ead523 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision/5/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e38850b870dee98f99112287e3f585f394bd2d23d9cecbfe46328741c73dc57d +size 577433 diff --git a/firstframe_sampled/perfectly_elastic_collision/7/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..fd80dd7da8b6982ad9d5667467ba00524b5aab93 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55c23648debb28b915a13f92c97d966a363ce7ce1c1524c5a64f28a481863754 +size 739445 diff --git a/firstframe_sampled/perfectly_elastic_collision/8/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..6bf81c3b739daf781e58895cf028cb9555322a01 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb6aa56110a267ce4979de63ecef8211be626d9fdb9e660466c3101b3d7f231c +size 817238 diff --git a/firstframe_sampled/perfectly_elastic_collision/9/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision/9/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..965a40a6484c50922ad1587e77f03417159059bf --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision/9/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a4fc8cf4e07acaffa0a4be45a694a3bdc0b7d3121544efe9fb9759b053a031ad +size 588994 diff --git a/firstframe_sampled/perfectly_elastic_collision_newbatch/0/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision_newbatch/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..997b970f52dd772ca981b46ab7e6cb6f86824756 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision_newbatch/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9a4d9355e8a215707baf078d4445b7b26ffa49e4b0c84a0ae91975937a333cb7 +size 582863 diff --git a/firstframe_sampled/perfectly_elastic_collision_newbatch/1/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision_newbatch/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..08d0016abe9e60583921b0a1c270dd0dc74640e8 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision_newbatch/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d47adc6f2c54712abb82ed1dfccc8580f6a68b38da918dfd54f2d0bd0d75b766 +size 572305 diff --git a/firstframe_sampled/perfectly_elastic_collision_newbatch/13/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision_newbatch/13/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..c4f934e3e00acbd80602f3127c588183cba41a4f --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision_newbatch/13/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fbc735bc768b6e352140bb17f4243c045abf7405f365049813f9ed539fbb5e0d +size 594058 diff --git a/firstframe_sampled/perfectly_elastic_collision_newbatch/14/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision_newbatch/14/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..8c7e8564f6636374b7b53465c36bb678081ff4cc --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision_newbatch/14/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:43ad0e5d7d32079c84c7821d2d1166ab7250ccd5383da1297006536de5901317 +size 583263 diff --git a/firstframe_sampled/perfectly_elastic_collision_newbatch/3/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision_newbatch/3/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..36c809ec10fdb32b4381bbaf955b9a799dd5b033 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision_newbatch/3/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d87a188681ed3b2645c1366976bc68d214f2bea023228391ebbec4cd832ea317 +size 585662 diff --git a/firstframe_sampled/perfectly_elastic_collision_newbatch/4/rgb_0000.png b/firstframe_sampled/perfectly_elastic_collision_newbatch/4/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..58fe8d01bae28ac3f3731c429781d2f6a90b8002 --- /dev/null +++ b/firstframe_sampled/perfectly_elastic_collision_newbatch/4/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:699a98f22bb5b02c2fb49b935ca2e34d2fa58b9bf9ede0149c3e9db68f532a0a +size 597062 diff --git a/firstframe_sampled/perfectly_inelastic_collision/2/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..f0281643fba6dc950f527525827653c9adf7abb3 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:75db8218114badff55b396d49e4c4ce3ba0342f95cb483754ec982769aaba2ae +size 640456 diff --git a/firstframe_sampled/perfectly_inelastic_collision/5/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision/5/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..b55565937a4d03cdb2acaac34bc354074eadd5e5 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision/5/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:84999104c980ee5ba70e2b6265153e001403fb755f9e0b35339447f919bf151f +size 588737 diff --git a/firstframe_sampled/perfectly_inelastic_collision/6/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..ef6cb233598696ace9597eb28080029e6446b808 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d67bf4b64adca59638bf8f3279128e23cf64b2cb2f5e1b6e0073c9fb200b9a7 +size 774474 diff --git a/firstframe_sampled/perfectly_inelastic_collision/7/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..39c0f0234e62187e166b9e60c3ce8af00b783747 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0ab7c321a0f58130453b00f339e72cd657c1b37ab5e9fd052e30ffdbab48c77b +size 792275 diff --git a/firstframe_sampled/perfectly_inelastic_collision/8/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..e390d5829c0d95777fd07e9966348e44e2764315 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b69096c413caf525457fe6623beae13c057100dc931bb010888696abf7401e0 +size 722817 diff --git a/firstframe_sampled/perfectly_inelastic_collision_newbatch/1/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision_newbatch/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..cfbe56d7b0612c6e3235335406278f6b4fd11eaa --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision_newbatch/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:578858a0c9dd5afd011fab3f570ef3e6d43fd0226927ce618cdfb16d47ce69c1 +size 604341 diff --git a/firstframe_sampled/perfectly_inelastic_collision_newbatch/11/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision_newbatch/11/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..87003c17a049bdd1325590aa18f64dc9f6f29469 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision_newbatch/11/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8df478e88191b86f6ffefd27f16d03ae0c21c4ddead326ee06d7ef4dfbb4b08b +size 599686 diff --git a/firstframe_sampled/perfectly_inelastic_collision_newbatch/14/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision_newbatch/14/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..16f0c4d6d6e24c11db46212a0d0342f8232237a8 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision_newbatch/14/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:33d03049dc3ee191cc3cc2eb28c9d2dc4fb1358b9c932eeb97e7c4c05d14c664 +size 599060 diff --git a/firstframe_sampled/perfectly_inelastic_collision_newbatch/6/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision_newbatch/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..258a97ae25cb8a7b701d8b20923812e58ec3cb84 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision_newbatch/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:97e1ae96544b545269f028fac18133e6b736f37da64fd015ddbe7d082a4988c2 +size 577270 diff --git a/firstframe_sampled/perfectly_inelastic_collision_newbatch/7/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision_newbatch/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..7642bbfcdb0bfaddc2c94f2fa1f57a958e4a37c6 --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision_newbatch/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c704a42793a1e0a487b8e85578a7e0c3f0c640e5b42e538d0db8454f5a0050d +size 597365 diff --git a/firstframe_sampled/perfectly_inelastic_collision_newbatch/8/rgb_0000.png b/firstframe_sampled/perfectly_inelastic_collision_newbatch/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..cf17371efde1cae98d814f3204417e6b1d8b760d --- /dev/null +++ b/firstframe_sampled/perfectly_inelastic_collision_newbatch/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1496ca564f6971bc6d63d62e00dc143421a22e821027dfd946a6169d9a35c96f +size 603045 diff --git a/firstframe_sampled/projectile/1/rgb_0000.png b/firstframe_sampled/projectile/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..d5c3af880eea5d1ff724ad75e6c4a40b78829fa9 --- /dev/null +++ b/firstframe_sampled/projectile/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:56ff2147ab4badc4abafb2d5c004525ac113f2053aabba3936e6514f774f83f8 +size 1115298 diff --git a/firstframe_sampled/projectile/10/rgb_0000.png b/firstframe_sampled/projectile/10/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..424a279e6d75168ce210db36950d638443a8d2fd --- /dev/null +++ b/firstframe_sampled/projectile/10/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec808e5002608d4c8e9a7b5a2b9ba7d22d1bdb84029d71aa8ea93484580d5e4f +size 1303072 diff --git a/firstframe_sampled/projectile/12/rgb_0000.png b/firstframe_sampled/projectile/12/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..77876d58a8cdf45e6731c5e0870f0e81a85279aa --- /dev/null +++ b/firstframe_sampled/projectile/12/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff1c0cc2b4ff2a04274bf71b07e1a1e965f37ea79c9db7fece74d29a39ab3e4c +size 1347044 diff --git a/firstframe_sampled/projectile/7/rgb_0000.png b/firstframe_sampled/projectile/7/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..9c2a658f210d9fbb4c59554086771a7881844852 --- /dev/null +++ b/firstframe_sampled/projectile/7/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:560a84c81e015fc38f42687ce6bd21d77002e133bffd43291399c088fe9c6f2d +size 1440153 diff --git a/firstframe_sampled/projectile/8/rgb_0000.png b/firstframe_sampled/projectile/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..63816c097ba9d46c0a2696a7834e746eb0e1aa7d --- /dev/null +++ b/firstframe_sampled/projectile/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a482a0db2aa1b68f91070d6199d9d4a3a96bbab5d84068c095380e9c0a7a6d2 +size 1563998 diff --git a/firstframe_sampled/projectile_newbatch/1/rgb_0000.png b/firstframe_sampled/projectile_newbatch/1/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..45c24d5f132581a86066baff175c16490e072249 --- /dev/null +++ b/firstframe_sampled/projectile_newbatch/1/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db1b06b60af5e0e9c4e9b51792a9f3190dda3f085ce686774236e3e2fe712d98 +size 1170076 diff --git a/firstframe_sampled/projectile_newbatch/11/rgb_0000.png b/firstframe_sampled/projectile_newbatch/11/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..2878aa43161e812bb82b0c8a12350c97021e48ae --- /dev/null +++ b/firstframe_sampled/projectile_newbatch/11/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0bd814d53d9ef52543a5cfe7d4609aed4fcace910e3ca04f8215e91f66e731e +size 821463 diff --git a/firstframe_sampled/projectile_newbatch/4/rgb_0000.png b/firstframe_sampled/projectile_newbatch/4/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..ed25cc048e05ff0186dbce39c612e2d18322ce97 --- /dev/null +++ b/firstframe_sampled/projectile_newbatch/4/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d77e36fba010f798b0fcc29b5c1247880f7d30c838dd5ecaeb941785bdd8abc5 +size 834101 diff --git a/firstframe_sampled/projectile_newbatch/5/rgb_0000.png b/firstframe_sampled/projectile_newbatch/5/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..bd24ef7958e45800a5c06db02fd3c982d72301f7 --- /dev/null +++ b/firstframe_sampled/projectile_newbatch/5/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:14871b58cfd91c5c8718abf7a1199eebae7c97ee7179af41eee91b60645b2b7a +size 1248981 diff --git a/firstframe_sampled/projectile_newbatch/6/rgb_0000.png b/firstframe_sampled/projectile_newbatch/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..72b63539066a606af48601e2e2fbf671a08b4ee6 --- /dev/null +++ b/firstframe_sampled/projectile_newbatch/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c43daf5dca3162ff60bf177e9deee1c0200e8ab6e7475ad5be159e2fe7980fc4 +size 1315579 diff --git a/firstframe_sampled/projectile_newbatch/9/rgb_0000.png b/firstframe_sampled/projectile_newbatch/9/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..f143d84a060d32412fcff9cf50f7ff2f43060ef0 --- /dev/null +++ b/firstframe_sampled/projectile_newbatch/9/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ba7dd20b147ee0daa9a3035aa9d3722f2997e3412c03d0f17e3597c73c9f266b +size 824359 diff --git a/firstframe_sampled/uniform_linear/2/rgb_0000.png b/firstframe_sampled/uniform_linear/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..769ac4ecb34c859f82978d1adcfdcc82b3e16c48 --- /dev/null +++ b/firstframe_sampled/uniform_linear/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:75d68dd225e6c47ec911c60836aed8ab1c62c0e5309279d45da81cbb81f11b21 +size 798224 diff --git a/firstframe_sampled/uniform_linear/6/rgb_0000.png b/firstframe_sampled/uniform_linear/6/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..c5a8084202c6878510a1fa026369ab49a473369c --- /dev/null +++ b/firstframe_sampled/uniform_linear/6/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a64ecf4f752f6629add1af5791cc06cb88cd09080688992d7f9fd4a5622a5f48 +size 928547 diff --git a/firstframe_sampled/uniform_linear/8/rgb_0000.png b/firstframe_sampled/uniform_linear/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..f535ffca31177ef991872809050637be212c13fb --- /dev/null +++ b/firstframe_sampled/uniform_linear/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0e9fea0d02effaef1881c90e3ac56e5c301dcf4b160ec132b23e1a5282b3ab80 +size 932377 diff --git a/firstframe_sampled/uniform_linear_newbatch/0/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/0/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..18b36e942fa4cf3fee07c62a6c2fd9674e9a9986 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/0/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:982925e9705722bbcca5f598a8cd4089f2dac6ccfcfcdf883a3a28754b0d13b1 +size 618543 diff --git a/firstframe_sampled/uniform_linear_newbatch/10/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/10/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..418e29905294d5ed43495e65ef2af31740d1ac18 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/10/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1efc28847cce02388ecafea46e5a208fe650c7b69ca80f123c1cc940d0b3e855 +size 619059 diff --git a/firstframe_sampled/uniform_linear_newbatch/13/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/13/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..de5f01b3623305c8c2eef18be8190e2bdd478b16 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/13/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:27ea4b8b503c39b914f1d2a23ce51bc9e1ff08ee7cf655de1528da14b10548fa +size 685824 diff --git a/firstframe_sampled/uniform_linear_newbatch/14/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/14/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..03fdd0e284b93c0196aec3047805c7c9650c8e06 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/14/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34862535060bf08137680a2d905fa8d142529f61066b233761a98976d6be75c4 +size 683849 diff --git a/firstframe_sampled/uniform_linear_newbatch/2/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/2/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..2f42570589c154874c28c377b8766bf072b530a1 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/2/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f25618c6ecf94d4a0e02aa12f12114785288d617c4e5dbe15a1be9a18827dd3f +size 621099 diff --git a/firstframe_sampled/uniform_linear_newbatch/4/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/4/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..80f454cd61e30953f4bcf9547eafaeca0742a123 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/4/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3154c5e8dd8a34047425f3500dedb2aed5fe8eeacc88b6ce66ce7da8f4ff7cef +size 619463 diff --git a/firstframe_sampled/uniform_linear_newbatch/5/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/5/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..4f40abb203a92eb1441b483a96ed6f25745f69bf --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/5/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0eec7352c65451f78c2b0630cbb6e0c48ec5a01c46fcda11cfccd9720266b121 +size 647329 diff --git a/firstframe_sampled/uniform_linear_newbatch/8/rgb_0000.png b/firstframe_sampled/uniform_linear_newbatch/8/rgb_0000.png new file mode 100644 index 0000000000000000000000000000000000000000..642486c0f5379a67a2d9d4dadd2a9bfd8302fcd3 --- /dev/null +++ b/firstframe_sampled/uniform_linear_newbatch/8/rgb_0000.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2f64173d94c5f3cb19fc020d14291acccdd0cb1402127fb59b44a985a1a2f2f +size 689101