{
  "schema_version": "1.0.0",
  "model_id": "001",
  "status": "complete",
  "reviewed_at": "2026-10-25T15:24:28.296914+00:00",
  "reviewer": "root",
  "review_scope": "Completed current short reference pack. Machine voice review is bound to the exact stored clip; future scenes require their own QA. No human listening claim.",
  "voice": {
    "status": "accepted",
    "qa_status": "machine_verified_short_proof",
    "output_sha256": "f79524ccf9cb746ba92831ec7a038499f375c98b37a5f021c0c6c7fa3d5f0c58",
    "source_audio_sha256": "9e69e5d0a238287cc8e8361debffa1719f5c67c38936cf5e6657a7a05be0f13f",
    "native_video_duration_seconds": 7.466667,
    "source_audio_duration_seconds": 7.275,
    "approved_scope": "exact_stored_short_voice_proof",
    "human_listening_performed": false,
    "checks": {
      "visual_identity": "passed_sampled_frames",
      "waveform_match": "passed",
      "lip_sync": "passed_full_valid_window_machine_review",
      "spoken_content": "passed_contraction_normalized_asr",
      "human_listening": "not_performed_not_required_for_machine_scope",
      "source_audio_preserved": true
    },
    "waveform": {
      "waveform_correlation": 0.9995828234500848,
      "candidate_offset_seconds": 0.0,
      "source_audio_duration_seconds": 7.275,
      "candidate_audio_duration_seconds": 7.291083333333333,
      "overlap_seconds": 7.275,
      "metric_scope": "Automatic waveform alignment; not an absolute perceptual or voice-identity guarantee."
    },
    "syncnet": {
      "source_video_sha256": "f79524ccf9cb746ba92831ec7a038499f375c98b37a5f021c0c6c7fa3d5f0c58",
      "source_commit": "907c0b579c2e2d83f0eae1b2ac9e720cde4e5623",
      "model_url": "https://www.robots.ox.ac.uk/~vgg/software/lipsync/data/syncnet_v2.model",
      "model_sha256": "961e8696f888fce4f3f3a6c3d5b3267cf5b343100b238e79b2659bff2c605442",
      "source_files_sha256": {
        "SyncNetModel.py": "41b85aed0006546cb2d6dc9180c5bc3dac737ab80eff643022e397a349c30c22",
        "SyncNetInstance.py": "bd722bf844700d6a923dec2f10d7cacd65c0c6b22d09aa5d44d2509660e900d6",
        "run_pipeline.py": "3654a3cf22c74cb3071fca6fb97ca9fb7dd1a5943fe864169ff8c0b57eb80acf"
      },
      "runtime": {
        "torch": "2.7.1+cpu",
        "opencv": "4.11.0",
        "numpy": "1.26.4",
        "python_speech_features": "0.6",
        "scipy": "1.13.1"
      },
      "crop": {
        "frames_25fps": 187,
        "face_detections": 187,
        "missing_detections": 0,
        "crop_dimensions": [
          224,
          224
        ],
        "crop_scale": 0.4,
        "crop_track_sha256": "cadc4b97517201f07700c62d4865ebb85b4c2a3632c3d707757babc1aa1accde",
        "face_detector": "OpenCV HaarCascade frontalface_default, largest central track, median13 smoothing",
        "face_detector_xml_sha256": "0f7d4527844eb514d4a4948e822da90fbb16a34a0bbbbc6adc6498747a5aafb0"
      },
      "actual": {
        "offset_frames": -1,
        "offset_seconds": -0.04,
        "confidence": 7.5897111892700195,
        "minimum_distance": 7.0868988037109375,
        "analysis_feature_windows": 177,
        "window_hop_seconds": 0.04,
        "network_context_seconds": 0.2,
        "coverage_seconds_from_first_to_last_window_end": 7.24,
        "mean_distance_by_shift": [
          15.062797546386719,
          15.604902267456055,
          15.83265209197998,
          15.382081031799316,
          14.831375122070312,
          14.741670608520508,
          15.05629825592041,
          15.467184066772461,
          15.30886173248291,
          14.265058517456055,
          13.526315689086914,
          13.520204544067383,
          14.195324897766113,
          14.676609992980957,
          13.49779987335205,
          10.31911563873291,
          7.0868988037109375,
          8.489754676818848,
          12.299443244934082,
          14.65534496307373,
          15.04289436340332,
          14.452119827270508,
          14.091731071472168,
          14.309476852416992,
          14.661645889282227,
          14.73039722442627,
          14.591573715209961,
          14.85916805267334,
          15.496158599853516,
          15.980709075927734,
          16.142486572265625
        ],
        "per_second_windows": [
          {
            "start_seconds": 0.0,
            "feature_windows": 25,
            "offset_frames": -1,
            "offset_seconds": -0.04,
            "confidence": 8.713030815124512,
            "min_distance": 5.473116874694824
          },
          {
            "start_seconds": 1.0,
            "feature_windows": 25,
            "offset_frames": -2,
            "offset_seconds": -0.08,
            "confidence": 8.128011703491211,
            "min_distance": 6.88123893737793
          },
          {
            "start_seconds": 2.0,
            "feature_windows": 25,
            "offset_frames": -1,
            "offset_seconds": -0.04,
            "confidence": 9.580582618713379,
            "min_distance": 6.420191764831543
          },
          {
            "start_seconds": 3.0,
            "feature_windows": 25,
            "offset_frames": -1,
            "offset_seconds": -0.04,
            "confidence": 8.234185218811035,
            "min_distance": 9.030373573303223
          },
          {
            "start_seconds": 4.0,
            "feature_windows": 25,
            "offset_frames": -1,
            "offset_seconds": -0.04,
            "confidence": 7.4024248123168945,
            "min_distance": 7.38034725189209
          },
          {
            "start_seconds": 5.0,
            "feature_windows": 25,
            "offset_frames": -1,
            "offset_seconds": -0.04,
            "confidence": 6.590740203857422,
            "min_distance": 7.559130668640137
          },
          {
            "start_seconds": 6.0,
            "feature_windows": 25,
            "offset_frames": -1,
            "offset_seconds": -0.04,
            "confidence": 5.37504243850708,
            "min_distance": 6.324090480804443
          },
          {
            "start_seconds": 7.0,
            "feature_windows": 2,
            "offset_frames": 0,
            "offset_seconds": 0.0,
            "confidence": 6.778265953063965,
            "min_distance": 3.332066535949707
          }
        ]
      },
      "project_gates": {
        "official_example_matches_expected_offset_3": true,
        "known_delayed_audio_recovered": true,
        "known_advanced_audio_recovered": true,
        "absolute_actual_offset_within_80ms": true,
        "actual_confidence_above_3": true,
        "reversed_audio_fails_sync_gate": true
      },
      "machine_sync_review_passed": true,
      "scope": "Machine audiovisual synchronization evaluation over every valid 5-frame sliding window. No human listening claim; no broad future-video guarantee. Edge context shorter than 0.2 seconds is excluded by official network.",
      "source_report_sha256": "0728217d02a6c40604e732ea8696f7ddd1fe848b2fb527f06fa3bca79e6b6202",
      "official_example": {
        "offset_frames": 3,
        "confidence": 10.08146858215332,
        "analysis_feature_windows": 312
      },
      "shift_controls": {
        "audio_delayed_480ms": {
          "known_audio_shift_seconds": 0.48,
          "offset_frames": -13,
          "offset_seconds": -0.52,
          "confidence": 7.878454208374023
        },
        "audio_advanced_480ms": {
          "known_audio_shift_seconds": -0.48,
          "offset_frames": 11,
          "offset_seconds": 0.44,
          "confidence": 7.462197780609131
        }
      },
      "mismatched_audio_negative_control": {
        "audio_transform": "full spoken audio reversed in time",
        "audio_sha256": "87527719ad889345e916bbd2b98cd8217bd472ccfe495132a5e5460fddb61d71",
        "offset_frames": -15,
        "offset_seconds": -0.6,
        "confidence": 3.0503578186035156,
        "incorrectly_passes_project_sync_gate": false
      },
      "shift_control_acceptance_policy": "Known +/-480ms shifts must move estimated offset by the corresponding +/-12 frames relative to actual clip estimate, with at most one-frame tolerance.",
      "runtime_version_source": "Imported module attributes; SciPy version corrected from stale distribution metadata and verified in a clean pinned dependency target.",
      "clean_pinned_runtime_replay": {
        "passed": true,
        "actual_metrics_identical": true,
        "all_6_project_gates_passed": true,
        "scipy": "1.13.1",
        "opencv": "4.11.0",
        "numpy": "1.26.4",
        "torch": "2.7.1+cpu",
        "python_speech_features": "0.6",
        "transformers": "5.20.2",
        "huggingface_hub": "2.3.0",
        "wall_seconds": 25.634737014770508
      }
    },
    "asr": {
      "asr_model": "openai/whisper-tiny.en",
      "revision": "87c7102498dcde7456f24cfd30239ca606ed9063",
      "model_weights_sha256": "db59695928ded6043adaef491a53ef4e12da9611184d77c53baa691a60b958ad",
      "decoding": "greedy, do_sample=False; no prompt or forced transcript provided",
      "runtime": {
        "torch": "2.7.1+cpu",
        "transformers": "5.20.2"
      },
      "expected_text": "Hello, I am Kayla. This short voice reference helps keep my speech clear, natural, and consistent across every video.",
      "source_report_sha256": "1e5fb7e422726bd2ff074522d5fb997923c51837bb938d822b66c97b85fb0803",
      "master_transcript": "Hello, I'm Kayla. This short voice reference helps keep my speech clear natural and consistent across every video.",
      "returned_transcript": "Hello, I'm Kayla. This short voice reference helps keep my speech clear natural and consistent across every video.",
      "master_audio_sha256": "9e69e5d0a238287cc8e8361debffa1719f5c67c38936cf5e6657a7a05be0f13f",
      "returned_audio_pcm_sha256": "d012930bd27efd9e712e1d78f1cc15122baa67b073875a78533042fc5e04b3be",
      "normalization": "Lowercase; expand I'm to I am; ignore punctuation; compare resulting words exactly.",
      "normalized_reference_words": 19,
      "contraction_normalized_word_error_rate": 0.0,
      "semantic_content_match": true,
      "master_vs_returned_transcripts_equal": true,
      "human_listening_performed": false,
      "external_api_spend_usd_micros": 0,
      "local_compute_cost_usd_micros": null
    }
  },
  "turn_reference": {
    "status": "accepted",
    "approved_scope": "front_to_profile_reference_cut_without_final_hold_requirement",
    "output_sha256": "208ba94ac19114bf0bd8075d28c44139f4dfb8633f2398f41a006e746c036d11",
    "source_sha256": "9c2282bce841cb9c035db6e43ae189a453be3ebaf0ec55f813319d514b57a0f9",
    "source_path": "assets/kayla-motion-turn-r2-v004.mp4",
    "path": "assets/kayla-motion-turn-reference-v004.mp4",
    "source_interval": {
      "start_frame": 0,
      "end_frame_exclusive": 85,
      "last_frame": 84,
      "start_seconds": 0,
      "end_seconds_exclusive": 3.5416666666666665
    },
    "duration_seconds": 3.541667,
    "frame_count": 85,
    "same_resolution_encode": {
      "codec": "libx264",
      "preset": "medium",
      "crf": 18,
      "pixel_format": "yuv420p",
      "ssim_all": 0.995883,
      "exact_command": [
        "ffmpeg",
        "-v",
        "error",
        "-nostdin",
        "-y",
        "-i",
        "kayla-motion-turn-r2-v004.mp4",
        "-map",
        "0:v:0",
        "-an",
        "-vf",
        "trim=end_frame=85,setpts=PTS-STARTPTS",
        "-frames:v",
        "85",
        "-c:v",
        "libx264",
        "-preset",
        "medium",
        "-crf",
        "18",
        "-pix_fmt",
        "yuv420p",
        "-movflags",
        "+faststart",
        "kayla-motion-turn-reference-v004.mp4"
      ],
      "ffmpeg_version": "ffmpeg version 6.1.1-3ubuntu5 Copyright (c) 2000-2023 the FFmpeg developers"
    },
    "root_review": "Sampled sequence and every final five face frames preserve identity, bare feet and full side profile with eyes open.",
    "original_full_5s_brief_approved": false,
    "original_two_render_attempts_preserved": true,
    "freeze_frames_added": 0
  },
  "costs": {
    "provider_generation_requests": 6,
    "observed_provider_credits_consumed": 266,
    "listed_rate_equivalent_usd_micros": 1330000,
    "external_reserved_usd_micros": 1650000,
    "new_generation_requests": 0,
    "new_external_api_spend_usd_micros": 0,
    "actual_cash_cost_usd_micros": null,
    "local_machine_review_compute_usd_micros": null,
    "local_finishing_compute_usd_micros": null
  },
  "completion": {
    "current_short_reference_pack_complete": true,
    "current_pending_requirements": [],
    "current_motion_baseline_seconds": 18.6667,
    "voice_proof_seconds": 7.466667,
    "all_scoped_functions_retired": true
  },
  "evidence_reports": {
    "syncnet-full-review.json": {
      "sha256": "0728217d02a6c40604e732ea8696f7ddd1fe848b2fb527f06fa3bca79e6b6202"
    },
    "asr-full-review.json": {
      "sha256": "1e5fb7e422726bd2ff074522d5fb997923c51837bb938d822b66c97b85fb0803"
    },
    "machine-review-summary.json": {
      "sha256": "8e51857f1ea69939997863c389e4f2fb852253a513802bc235477aaaf25d26e4"
    }
  }
}