{
  "schema": 1,
  "language": "en",
  "lang_code": "a",
  "dialect": "us",
  "voice": "am_liam",
  "speed": 0.88,
  "sample_rate": 24000,
  "total_duration": 39.16,
  "media_sha256": "fad41971fb80d2061ef303f6eec92faef95472ff5d4a0427761e6eb96e1bc9e3",
  "media_bytes": 247272,
  "render_fingerprint": "ccec58c2068a53b2743d74d578a67b8f67375a8680193950c8f3027b08a32d61",
  "render_inputs": {
    "renderer": "render_all_voices-v4",
    "lesson": "39_classifier_evaluation",
    "language": "en",
    "lang_code": "a",
    "dialect": "us",
    "voice": "am_liam",
    "speed": 0.88,
    "scenes": [
      {
        "narration": "This tutorial covers Classifier Evaluation. Audit held-out predictions, nested cross-validation, calibration, leakage, and per-plate performance.",
        "speech_text": [
          "This tutorial covers [Classifier](/klˈæsəfIəɹ/) Evaluation.",
          "Audit held-out predictions, nested cross-validation, calibration, leakage, and per-plate performance."
        ],
        "authored_hold": 0.45,
        "resolved_hold": 0.45,
        "speed_multiplier": 1.0,
        "effective_speed": 0.88
      },
      {
        "narration": "Choose predictions with ground truth, split metadata, plate identifiers, probabilities, and model-run records.",
        "speech_text": [
          "Choose predictions with ground truth, split metadata, plate identifiers, probabilities, and model-run records."
        ],
        "authored_hold": 0.45,
        "resolved_hold": 0.45,
        "speed_multiplier": 1.0,
        "effective_speed": 0.88
      },
      {
        "narration": "In the main configuration area, configure class order, thresholds, calibration bins, grouping, bootstrap uncertainty, and leakage checks.",
        "speech_text": [
          "In the main configuration area, configure class order, thresholds, calibration bins, grouping, bootstrap uncertainty, and leakage checks."
        ],
        "authored_hold": 0.45,
        "resolved_hold": 0.45,
        "speed_multiplier": 1.0,
        "effective_speed": 0.88
      },
      {
        "narration": "Evaluate held-out and per-plate behavior before using a classifier for biological conclusions.",
        "speech_text": [
          "Evaluate held-out and per-plate behavior before using a [classifier](/klˈæsəfIəɹ/) for biological conclusions."
        ],
        "authored_hold": 0.45,
        "resolved_hold": 0.45,
        "speed_multiplier": 1.0,
        "effective_speed": 0.88
      },
      {
        "narration": "When the workflow finishes, spaCR produces discrimination, calibration, confusion, leakage, and robustness reports.",
        "speech_text": [
          "When the workflow finishes, spacer produces discrimination, calibration, confusion, leakage, and robustness reports."
        ],
        "authored_hold": 0.65,
        "resolved_hold": 0.52,
        "speed_multiplier": 1.0,
        "effective_speed": 0.88
      }
    ],
    "assembly": {
      "version": "2026-08-11-natural-sentences-v1",
      "sample_rate": 24000,
      "lead_silence_seconds": 0.1,
      "tail_silence_seconds": 0.18,
      "sentence_pause_seconds": 0.12,
      "scene_hold_min_seconds": 0.32,
      "scene_hold_max_seconds": 0.52,
      "default_scene_hold_seconds": 0.42,
      "vad_frame_seconds": 0.005,
      "vad_peak_ratio": 0.002,
      "vad_absolute_floor": 2e-05
    },
    "cadence_profile": {
      "version": "2026-08-11-af-heart-scene-cadence-v6",
      "sha256": "4d728d67e73eea7e6203878879992f9050ae4a2677d717efdd68545104eb7812",
      "bytes": 23008,
      "enabled": false,
      "targets": {
        "slow_wpm": 125.0,
        "fast_wpm": 145.0,
        "calibration_wpm": 150.0
      },
      "safe_multiplier_bounds": {
        "minimum": 0.68,
        "maximum": 1.18
      }
    },
    "mastering": {
      "filters": [
        "loudnorm=I=-16:TP=-3:LRA=11",
        "loudnorm=I=-16:TP=-5:LRA=11,alimiter=limit=0.7:attack=5:release=50:level=disabled",
        "loudnorm=I=-16:TP=-5:LRA=11,alimiter=limit=0.5:attack=5:release=50:level=disabled"
      ],
      "codec": "aac",
      "bitrate": "48k",
      "channels": 1,
      "maximum_decoded_true_peak_dbfs": -1.0
    },
    "pronunciation": {
      "role": "pronunciation_profile",
      "version": "2026-08-11-en-natural-v6",
      "sha256": "1fbaa615cfa904094bdef5cba963539f0e8bec5927192a76b0d503e90c4bb83e",
      "bytes": 13263
    },
    "synthesis_runtime": {
      "model": {
        "repository": "hexgrad/Kokoro-82M",
        "snapshot": "f3ff3571791e39611d31c381e3a41a3af07b4987",
        "config": {
          "role": "kokoro_config",
          "name": "14a726edd3718279eac426630879ff743955b16a",
          "bytes": 2351,
          "sha256": "5abb01e2403b072bf03d04fde160443e209d7a0dad49a423be15196b9b43c17f"
        },
        "weights": {
          "role": "kokoro_weights",
          "name": "496dba118d1a58f5f3db2efc88dbdc216e0483fc89fe6e47ee1f2c53f18ad1e4",
          "bytes": 327212226,
          "sha256": "496dba118d1a58f5f3db2efc88dbdc216e0483fc89fe6e47ee1f2c53f18ad1e4"
        }
      },
      "voice_pack": {
        "role": "kokoro_voice_pack",
        "name": "c82550757ddb31308b97f30040dda8c2d609a9e2de6135848d0a948368138518",
        "bytes": 523420,
        "sha256": "c82550757ddb31308b97f30040dda8c2d609a9e2de6135848d0a948368138518",
        "tensor_sha256": "52403be32fd047c6a44517cb0bcd6b134f2a18baa73e70ef41651e0eab921ade",
        "loaded_tensor_sha256": "52403be32fd047c6a44517cb0bcd6b134f2a18baa73e70ef41651e0eab921ade"
      },
      "software": {
        "kokoro": "0.9.4",
        "misaki": "0.9.4",
        "torch": "2.6.0",
        "soundfile": "0.14.0",
        "python": "3.12.4",
        "implementation": "cpython"
      },
      "inference": {
        "device": "cuda"
      },
      "mastering_tool": {
        "version": "ffmpeg version 6.1.1-3ubuntu5+esm10 Copyright (c) 2000-2023 the FFmpeg developers",
        "version_output_sha256": "92dfb2468e6f5f06919b5433c0f14459ac19b81b1bc2d9fdceb83c6ad42e37b5",
        "executable": {
          "role": "ffmpeg_executable",
          "name": "ffmpeg",
          "bytes": 342488,
          "sha256": "ad63ef3f2f18fa92e89f0ec2bde6941e81f83724c8d265414cb659c28077392d"
        }
      }
    }
  },
  "scenes": [
    {
      "scene": 1,
      "speech_start": 0.0,
      "speech_end": 9.23,
      "scene_end": 9.68,
      "duration": 9.68,
      "text": "This tutorial covers Classifier Evaluation. Audit held-out predictions, nested cross-validation, calibration, leakage, and per-plate performance.",
      "speech_text": "This tutorial covers [Classifier](/klˈæsəfIəɹ/) Evaluation. Audit held-out predictions, nested cross-validation, calibration, leakage, and per-plate performance.",
      "phonemes": "ðˌɪs tutˈɔɹiəl kˈʌvəɹz klˈæsəfIəɹ əvˌæljʊˈAʃən. ˈɔdət hˈɛldˌWt pɹidˈɪkʃənz, nˈɛstᵻd kɹˌɔsvˌælɪdˈAʃən, kˌæləbɹˈAʃən, lˈikɪʤ, ænd pɜɹplˈAt pəɹfˈɔɹməns.",
      "authored_hold_after": 0.45,
      "hold_after": 0.45,
      "base_speed": 0.88,
      "speed_multiplier": 1.0,
      "effective_speed": 0.88,
      "sentences": [
        {
          "sentence": 1,
          "speech_start": 0.0,
          "speech_end": 3.26,
          "audible_start": 0.1,
          "audible_end": 3.08,
          "duration": 3.26,
          "gap_from_previous": null,
          "trimmed_lead": 0.15,
          "trimmed_tail": 0.39,
          "text": "This tutorial covers Classifier Evaluation.",
          "speech_text": "This tutorial covers [Classifier](/klˈæsəfIəɹ/) Evaluation.",
          "phonemes": "ðˌɪs tutˈɔɹiəl kˈʌvəɹz klˈæsəfIəɹ əvˌæljʊˈAʃən."
        },
        {
          "sentence": 2,
          "speech_start": 3.38,
          "speech_end": 9.23,
          "audible_start": 3.48,
          "audible_end": 9.05,
          "duration": 5.85,
          "gap_from_previous": 0.4,
          "trimmed_lead": 0.18,
          "trimmed_tail": 0.42,
          "text": "Audit held-out predictions, nested cross-validation, calibration, leakage, and per-plate performance.",
          "speech_text": "Audit held-out predictions, nested cross-validation, calibration, leakage, and per-plate performance.",
          "phonemes": "ˈɔdət hˈɛldˌWt pɹidˈɪkʃənz, nˈɛstᵻd kɹˌɔsvˌælɪdˈAʃən, kˌæləbɹˈAʃən, lˈikɪʤ, ænd pɜɹplˈAt pəɹfˈɔɹməns."
        }
      ]
    },
    {
      "scene": 2,
      "speech_start": 9.68,
      "speech_end": 16.265,
      "scene_end": 16.715,
      "duration": 7.035,
      "text": "Choose predictions with ground truth, split metadata, plate identifiers, probabilities, and model-run records.",
      "speech_text": "Choose predictions with ground truth, split metadata, plate identifiers, probabilities, and model-run records.",
      "phonemes": "ʧˈuz pɹidˈɪkʃənz wɪð ɡɹˈWnd tɹˈuθ, splˈɪt mˈɛTədˌATə, plˈAt IdˈɛntəfˌIəɹz, pɹˌɑbəbˈɪləTiz, ænd mˈɑdᵊlɹˌʌn ɹˈɛkəɹdz.",
      "authored_hold_after": 0.45,
      "hold_after": 0.45,
      "base_speed": 0.88,
      "speed_multiplier": 1.0,
      "effective_speed": 0.88,
      "sentences": [
        {
          "sentence": 1,
          "speech_start": 9.68,
          "speech_end": 16.265,
          "audible_start": 9.78,
          "audible_end": 16.085,
          "duration": 6.585,
          "gap_from_previous": null,
          "trimmed_lead": 0.16,
          "trimmed_tail": 0.405,
          "text": "Choose predictions with ground truth, split metadata, plate identifiers, probabilities, and model-run records.",
          "speech_text": "Choose predictions with ground truth, split metadata, plate identifiers, probabilities, and model-run records.",
          "phonemes": "ʧˈuz pɹidˈɪkʃənz wɪð ɡɹˈWnd tɹˈuθ, splˈɪt mˈɛTədˌATə, plˈAt IdˈɛntəfˌIəɹz, pɹˌɑbəbˈɪləTiz, ænd mˈɑdᵊlɹˌʌn ɹˈɛkəɹdz."
        }
      ]
    },
    {
      "scene": 3,
      "speech_start": 16.715,
      "speech_end": 24.77,
      "scene_end": 25.22,
      "duration": 8.504999999999999,
      "text": "In the main configuration area, configure class order, thresholds, calibration bins, grouping, bootstrap uncertainty, and leakage checks.",
      "speech_text": "In the main configuration area, configure class order, thresholds, calibration bins, grouping, bootstrap uncertainty, and leakage checks.",
      "phonemes": "ɪn ðə mˈAn kənfˌɪɡjəɹˈAʃən ˈɛɹiə, kənfˈɪɡjəɹ klˈæs ˈɔɹdəɹ, θɹˈɛʃhˌOldz, kˌæləbɹˈAʃən bˈɪnz, ɡɹˈupɪŋ, bˈutstɹˌæp ˌʌnsˈɜɹtᵊnti, ænd lˈikɪʤ ʧˈɛks.",
      "authored_hold_after": 0.45,
      "hold_after": 0.45,
      "base_speed": 0.88,
      "speed_multiplier": 1.0,
      "effective_speed": 0.88,
      "sentences": [
        {
          "sentence": 1,
          "speech_start": 16.715,
          "speech_end": 24.77,
          "audible_start": 16.815,
          "audible_end": 24.59,
          "duration": 8.055,
          "gap_from_previous": null,
          "trimmed_lead": 0.185,
          "trimmed_tail": 0.41,
          "text": "In the main configuration area, configure class order, thresholds, calibration bins, grouping, bootstrap uncertainty, and leakage checks.",
          "speech_text": "In the main configuration area, configure class order, thresholds, calibration bins, grouping, bootstrap uncertainty, and leakage checks.",
          "phonemes": "ɪn ðə mˈAn kənfˌɪɡjəɹˈAʃən ˈɛɹiə, kənfˈɪɡjəɹ klˈæs ˈɔɹdəɹ, θɹˈɛʃhˌOldz, kˌæləbɹˈAʃən bˈɪnz, ɡɹˈupɪŋ, bˈutstɹˌæp ˌʌnsˈɜɹtᵊnti, ænd lˈikɪʤ ʧˈɛks."
        }
      ]
    },
    {
      "scene": 4,
      "speech_start": 25.22,
      "speech_end": 31.099999999999998,
      "scene_end": 31.549999999999997,
      "duration": 6.329999999999998,
      "text": "Evaluate held-out and per-plate behavior before using a classifier for biological conclusions.",
      "speech_text": "Evaluate held-out and per-plate behavior before using a [classifier](/klˈæsəfIəɹ/) for biological conclusions.",
      "phonemes": "əvˈæljʊˌAt hˈɛldˌWt ænd pɜɹplˈAt bəhˈAvjəɹ bəfˈɔɹ jˈuzɪŋ ɐ klˈæsəfIəɹ fɔɹ bˌIəlˈɑʤəkᵊl kənklˈuʒənz.",
      "authored_hold_after": 0.45,
      "hold_after": 0.45,
      "base_speed": 0.88,
      "speed_multiplier": 1.0,
      "effective_speed": 0.88,
      "sentences": [
        {
          "sentence": 1,
          "speech_start": 25.22,
          "speech_end": 31.099999999999998,
          "audible_start": 25.32,
          "audible_end": 30.919999999999998,
          "duration": 5.88,
          "gap_from_previous": null,
          "trimmed_lead": 0.175,
          "trimmed_tail": 0.445,
          "text": "Evaluate held-out and per-plate behavior before using a classifier for biological conclusions.",
          "speech_text": "Evaluate held-out and per-plate behavior before using a [classifier](/klˈæsəfIəɹ/) for biological conclusions.",
          "phonemes": "əvˈæljʊˌAt hˈɛldˌWt ænd pɜɹplˈAt bəhˈAvjəɹ bəfˈɔɹ jˈuzɪŋ ɐ klˈæsəfIəɹ fɔɹ bˌIəlˈɑʤəkᵊl kənklˈuʒənz."
        }
      ]
    },
    {
      "scene": 5,
      "speech_start": 31.549999999999997,
      "speech_end": 38.64,
      "scene_end": 39.160000000000004,
      "duration": 7.6100000000000065,
      "text": "When the workflow finishes, spaCR produces discrimination, calibration, confusion, leakage, and robustness reports.",
      "speech_text": "When the workflow finishes, spacer produces discrimination, calibration, confusion, leakage, and robustness reports.",
      "phonemes": "wˌɛn ðə wˈɜɹkflˌO fˈɪnəʃᵻz, spˈAsəɹ pɹədˈusᵻz dəskɹˌɪmənˈAʃən, kˌæləbɹˈAʃən, kənfjˈuʒᵊn, lˈikɪʤ, ænd ɹObˈʌstnəs ɹəpˈɔɹts.",
      "authored_hold_after": 0.65,
      "hold_after": 0.52,
      "base_speed": 0.88,
      "speed_multiplier": 1.0,
      "effective_speed": 0.88,
      "sentences": [
        {
          "sentence": 1,
          "speech_start": 31.549999999999997,
          "speech_end": 38.64,
          "audible_start": 31.65,
          "audible_end": 38.459999999999994,
          "duration": 7.09,
          "gap_from_previous": null,
          "trimmed_lead": 0.165,
          "trimmed_tail": 0.42,
          "text": "When the workflow finishes, spaCR produces discrimination, calibration, confusion, leakage, and robustness reports.",
          "speech_text": "When the workflow finishes, spacer produces discrimination, calibration, confusion, leakage, and robustness reports.",
          "phonemes": "wˌɛn ðə wˈɜɹkflˌO fˈɪnəʃᵻz, spˈAsəɹ pɹədˈusᵻz dəskɹˌɪmənˈAʃən, kˌæləbɹˈAʃən, kənfjˈuʒᵊn, lˈikɪʤ, ænd ɹObˈʌstnəs ɹəpˈɔɹts."
        }
      ]
    }
  ]
}