w0rldw3aver360-motion

6 curated 15-second 360° equirectangular video clips sourced from NASA public-domain media — a redistributable, legally-clean dataset for Gaussian splatting, dynamic scene reconstruction, and 360-video model training.

Why this exists

Hugging Face hosts several large "360 video" collections, but none carry a usable license (YouTube rips / gated robotics data). This dataset fills that gap with verified public-domain U.S. Government footage, quality-filtered per frame.

Contents

Field	Value
Clips	6 × 15.0s MP4 (h264, yuv420p, faststart, no audio)
Sources	2 verified true-equirect NASA videos
Projection	equirectangular 2:1 (validated programmatically + visually)
Resolution	1920×960 (Stennis) · 1752×876 (Perseverance, zenith-compensated crop)
Captions	clips/<name>.txt — w0rldw3aver360 trigger style, matching names
Manifest	clips_manifest.csv — every clip: source, NASA ID/URL, offset, license
QC frames	qc_frames/*.jpg — mid-clip stills for fast visual audit

Scenes

• Stennis B-2 Test Stand ×4 windows (62s/155s/285s/415s of a 519s source) — industrial rocket test facility, 1920×960 native equirect, 30fps
• Perseverance first 360° of Mars (Jezero Crater) ×2 windows (13s/15s) — letterboxed source auto-cropped to true 2:1 with zenith compensation

Source audit (what was rejected and why)

Every NASA "360" candidate was frame-audited before inclusion:

Source	Verdict
Stennis B-2 Test Stand	✅ native 2:1 equirect, clean
Perseverance first 360	✅ true equirect (letterboxed) — cropped to 2:1
Curiosity Gediz Vallis 8K (both)	❌ partial cylindrical Navcam mosaics, jagged unimaged zones + burned-in annotation text
Hurricane Maria 360	❌ flat 3D data-visualization render, no spherical content

Provenance

All footage: NASA (images.nasa.gov), U.S. Government work, public domain (17 U.S.C. § 105). Curated & captioned by TheMindExpansionNetwork / Sonic Forage, 2026-08.

Recommended use

• Gaussian splatting from monocular 360 video (rotate-pan camera path = parallax baseline)
• Dynamic-scene NeRF/splat benchmarks needing permissive video
• 360-video caption→video model training (trigger: w0rldw3aver360)