{
  "schema": "adfactory.style-pace-observations/v1",
  "created_utc": "2026-09-10T02:44:20.174146+00:00",
  "purpose": "Evidence and production directions for same new dynamic/lipsync iteration. No generated-film acceptance claim.",
  "sources": [
    {
      "path": "/workspaces/UGS/AdFactoryNarrativeFormat/data/references_liked/resilia-aged-garlic-family-cholesterol-story.mp4",
      "sha256": "708cb005be27b73d43fa248bd8f003489098a48a43fcc59d6ce31bde841f7ff6"
    },
    {
      "path": "delivery/new-film/film-480p.mp4",
      "sha256": "12d63a16ead46b5cff0e23098767cf73b00c5ee8d4472198905fcc19f1a289c1"
    }
  ],
  "speech_samples": [
    {
      "excerpt": "source-open",
      "source_start_s": 0,
      "excerpt_duration_s": 30,
      "media": {
        "path": "analysis/dynamic-reference/source-open.mp4",
        "sha256": "c81a8ca3e299be3f291cb1cb1ac984d4822b3d0d5958702e1d48b00d0103b330"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-source-open/analysis.md",
        "sha256": "8f693db4bc33ecc2094642f6886eaa662e6603ae195d5f84738caa5317020576"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 51,
      "tokens_per_elapsed_minute": 102.0,
      "model_story_turn_count": 4,
      "model_story_turns_per_excerpt_minute": 8.0,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention."
    },
    {
      "excerpt": "source-search",
      "source_start_s": 30,
      "excerpt_duration_s": 30,
      "media": {
        "path": "analysis/dynamic-reference/source-search.mp4",
        "sha256": "b9c435de8d379328295f21fd2ad3ff5d9090664a709d8c7fce591e7f7fbab89a"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-source-search/analysis.md",
        "sha256": "b14af35ceaa5df97058f0e5038962e54fb437c7635772664961a073847749fd5"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 18,
      "tokens_per_elapsed_minute": 36.0,
      "model_story_turn_count": 5,
      "model_story_turns_per_excerpt_minute": 10.0,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention."
    },
    {
      "excerpt": "source-teaching",
      "source_start_s": 77,
      "excerpt_duration_s": 28,
      "media": {
        "path": "analysis/dynamic-reference/source-teaching.mp4",
        "sha256": "cec678b9edf81404b3b59e4e5134b87f9d1893fe985f31e81bad9adba2190930"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-source-teaching/analysis.md",
        "sha256": "31ee86e8cc7e5109c2461a3cc0fe9373807e2cbe794ac2319cae71e9a30e65e4"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 92,
      "tokens_per_elapsed_minute": 197.14,
      "model_story_turn_count": 2,
      "model_story_turns_per_excerpt_minute": 4.29,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention."
    },
    {
      "excerpt": "source-objections",
      "source_start_s": 160,
      "excerpt_duration_s": 30,
      "media": {
        "path": "analysis/dynamic-reference/source-objections.mp4",
        "sha256": "94bb7c1a6de3cbfb40fa9a833fcb1efd4c2b14ea7bf828dfb0a2685b2409b4de"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-source-objections/analysis.md",
        "sha256": "230fe588b45566f2e7574ea1c9764b4a2fdd2a60e04e2831aad6c39cef22e127"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 95,
      "tokens_per_elapsed_minute": 190.0,
      "model_story_turn_count": 3,
      "model_story_turns_per_excerpt_minute": 6.0,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention."
    },
    {
      "excerpt": "source-proof",
      "source_start_s": 241,
      "excerpt_duration_s": 43,
      "media": {
        "path": "analysis/dynamic-reference/source-proof.mp4",
        "sha256": "e6e9048e591032888e5c0f8f178df76d69c6148b836ff07afd97db5ecb2eb3b6"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-source-proof/analysis.md",
        "sha256": "ee9235c34bef597ba8381b374700b3b711110a0953799487b8496aa11fc7e3a6"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 64,
      "tokens_per_elapsed_minute": 89.3,
      "model_story_turn_count": 5,
      "model_story_turns_per_excerpt_minute": 6.98,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention."
    },
    {
      "excerpt": "source-close",
      "source_start_s": 284,
      "excerpt_duration_s": 41.56,
      "media": {
        "path": "analysis/dynamic-reference/source-close.mp4",
        "sha256": "043da3074bd58d842b2f04277953b016bad101a77cd76128d7fd656809a9b16f"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-source-close/analysis.md",
        "sha256": "6e3178d01abad88ce3e7c5493000c2f45d1450828d27df5e1b77a1ccf6b8ff86"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 120,
      "tokens_per_elapsed_minute": 173.24,
      "model_story_turn_count": 3,
      "model_story_turns_per_excerpt_minute": 4.33,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention."
    },
    {
      "excerpt": "previous-open",
      "source_start_s": 0,
      "excerpt_duration_s": 32.44,
      "media": {
        "path": "analysis/dynamic-reference/previous-open.mp4",
        "sha256": "52377941b1c4438e78de3f7fd48795e703b48ea43c5f2d35ef039ed5287415c9"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-previous-open/analysis.md",
        "sha256": "cc97de3d457b4d0813fc4dc7ef750a41844b3bfc74bc5912ddcf5296b77d643d"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 56,
      "tokens_per_elapsed_minute": 103.58,
      "model_story_turn_count": 2,
      "model_story_turns_per_excerpt_minute": 3.7,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention.",
      "accepted_word_alignment_count": 56,
      "accepted_words_per_elapsed_minute": 103.58,
      "accepted_alignment": {
        "path": "videos/gopure-fresh/assets/audio/words.json",
        "sha256": "81f25a73a9b6b2ac8c60c2482527bb99cf414f3782eea6106ef7945631b925f1"
      }
    },
    {
      "excerpt": "previous-explain",
      "source_start_s": 109.36,
      "excerpt_duration_s": 48.32,
      "media": {
        "path": "analysis/dynamic-reference/previous-explain.mp4",
        "sha256": "6443de76a24d60d26d189c758f7fbbdc3ad626a6ba350626fcfd55c7792e95b5"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-previous-explain/analysis.md",
        "sha256": "f1ad7cfdb2fa6145cde472b458511b13e669e01c56d19bf17f847466a943f080"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 82,
      "tokens_per_elapsed_minute": 101.82,
      "model_story_turn_count": 5,
      "model_story_turns_per_excerpt_minute": 6.21,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention.",
      "accepted_word_alignment_count": 82,
      "accepted_words_per_elapsed_minute": 101.82,
      "accepted_alignment": {
        "path": "videos/gopure-fresh/assets/audio/words.json",
        "sha256": "81f25a73a9b6b2ac8c60c2482527bb99cf414f3782eea6106ef7945631b925f1"
      }
    },
    {
      "excerpt": "previous-proof",
      "source_start_s": 205.88,
      "excerpt_duration_s": 49.76,
      "media": {
        "path": "analysis/dynamic-reference/previous-proof.mp4",
        "sha256": "1c6d0c0fec5c2fc6b3de80184c52db8122d6f9878e6f760f9783836e6e642316"
      },
      "analysis": {
        "path": "analysis/evidence/dynamic-previous-proof/analysis.md",
        "sha256": "6e8d4f57955d7932a668a15b9a979c5cd6041354b24732b83356a98cdc60272a"
      },
      "count_method": "Unicode orthographic tokens; internal apostrophes and hyphens kept; numeral token counts as one, not expanded spoken number words. Native model-assisted transcript, approximate segment timings.",
      "transcript_tokens": 78,
      "tokens_per_elapsed_minute": 94.05,
      "model_story_turn_count": 3,
      "model_story_turns_per_excerpt_minute": 3.62,
      "turn_count_status": "Analyst/model segmentation of meaningful knowledge, agency or story-state changes; not physical measurement or retention.",
      "accepted_word_alignment_count": 78,
      "accepted_words_per_elapsed_minute": 94.05,
      "accepted_alignment": {
        "path": "videos/gopure-fresh/assets/audio/words.json",
        "sha256": "81f25a73a9b6b2ac8c60c2482527bb99cf414f3782eea6106ef7945631b925f1"
      }
    }
  ],
  "previous_whole_voice": {
    "accepted_words": 605,
    "audio_duration_s": 367.44,
    "words_per_elapsed_minute": 98.79,
    "audio": {
      "path": "videos/gopure-fresh/assets/audio/narration.wav",
      "sha256": "9ad37eacd310bdf8ae2ec4842e4b48ba68c899d965eec3f457e908cd5c0bd957"
    }
  },
  "picture_measurements": {
    "path": "analysis/dynamic-reference/picture-measurements.json",
    "sha256": "5308bb642a48339ce2347825dee6c8756218341624f947bd44bbff100068a251"
  },
  "style_frames": {
    "path": "analysis/dynamic-reference/style-frames/manifest.json",
    "sha256": "a0fb195e1e2caa97e57fd0674982f6690ae867ff3320315eb69bf0ca1ab2354f"
  },
  "assistant_observations": [
    {
      "status": "observed",
      "reviewer": "Codex assistant direct frame-triplet inspection",
      "source": "main reference",
      "interval_s": [
        4.1,
        7.8
      ],
      "observation": "Doctor reverse retains stable ceiling/cabinet alignment and scale; face and paper act inside a locked frame.",
      "evidence": {
        "path": "analysis/dynamic-reference/source-camera-triplets.jpg",
        "sha256": "218a6c0521d809a866edcd9b3bd03a6f5f83a3b8351c3a4438e34c0d36804c33"
      }
    },
    {
      "status": "observed",
      "reviewer": "Codex assistant direct frame-triplet inspection",
      "source": "main reference",
      "interval_s": [
        11.8,
        13.2
      ],
      "observation": "Fixed hand insert shows partner hand arriving, resting, gripping; physical state advances without camera push.",
      "evidence": {
        "path": "analysis/dynamic-reference/source-camera-triplets.jpg",
        "sha256": "218a6c0521d809a866edcd9b3bd03a6f5f83a3b8351c3a4438e34c0d36804c33"
      }
    },
    {
      "status": "observed",
      "reviewer": "Codex assistant direct frame-triplet inspection",
      "source": "main reference",
      "interval_s": [
        46,
        48
      ],
      "observation": "House/street framing remains fixed as child walks across lower frame. Walking is subject movement, not tracking. Native search shot times are approximate and not the edit clock.",
      "evidence": {
        "path": "analysis/dynamic-reference/source-camera-triplets.jpg",
        "sha256": "218a6c0521d809a866edcd9b3bd03a6f5f83a3b8351c3a4438e34c0d36804c33"
      }
    },
    {
      "status": "observed",
      "reviewer": "Codex assistant direct frame-triplet inspection",
      "source": "main reference",
      "interval_s": [
        80,
        84.5
      ],
      "observation": "Helper close-up retains background cabinets/practical alignment while eyes, mouth and hands change.",
      "evidence": {
        "path": "analysis/dynamic-reference/source-camera-triplets.jpg",
        "sha256": "218a6c0521d809a866edcd9b3bd03a6f5f83a3b8351c3a4438e34c0d36804c33"
      }
    },
    {
      "status": "observed",
      "reviewer": "Codex assistant direct frame-triplet inspection",
      "source": "main reference",
      "interval_s": [
        171.7,
        176
      ],
      "observation": "Group master keeps stable kitchen/window geometry, speaker gestures and listeners respond.",
      "evidence": {
        "path": "analysis/dynamic-reference/source-camera-triplets.jpg",
        "sha256": "218a6c0521d809a866edcd9b3bd03a6f5f83a3b8351c3a4438e34c0d36804c33"
      }
    },
    {
      "status": "observed",
      "reviewer": "Codex assistant direct frame-triplet inspection",
      "source": "previous generated clips",
      "shot_ids": [
        "F01a",
        "F01b",
        "F06a",
        "F11a"
      ],
      "observation": "Repeated inward changes of framing: two-shot to daughter close; same setup to mother close; seated search to phone/head; patch test to forearm. Background scale/crop changes, disproving generic native static labels.",
      "evidence": {
        "path": "analysis/dynamic-reference/previous-camera-triplets.jpg",
        "sha256": "b950157af39c5af906139d282e394b53b0efcbee5dc5e57062d6db85523cee4a"
      }
    },
    {
      "status": "observed",
      "reviewer": "Codex assistant direct entry-frame inspection",
      "source": "previous generated clips",
      "observation": "Paired F01,F02,F04,F05,F06 and many later entries repeat composition with small pose changes. Eight F14–F17 proof entries hold frontal portrait grammar. Unique pixel files or changed expressions do not mean a new starting composition.",
      "evidence": [
        {
          "path": "analysis/dynamic-reference/previous-entries-1.jpg",
          "sha256": "8294034ef101f730f9c459dddef60832216989d2dc1c465e4c19bcd4c3ff5bb0"
        },
        {
          "path": "analysis/dynamic-reference/previous-entries-2.jpg",
          "sha256": "4849a6a8035ad7bd3c165f67585145c8f0a8d9564dccdc5bbc571f3727806c65"
        },
        {
          "path": "analysis/dynamic-reference/previous-entries-3.jpg",
          "sha256": "e74a010ee2eec720873c439abe210e798f69f13389eeb1bb156b90a780331cfa"
        }
      ]
    }
  ],
  "production_targets": {
    "status": "intended creative choices, not measured results",
    "speech_words_per_minute": [
      165,
      185
    ],
    "typical_cut_s": [
      1.5,
      3.5
    ],
    "longer_spoken_sentence_hold_s": [
      4,
      6
    ],
    "opening_goal": "Question/conflict, specific concern and character choice within first 10–15 seconds, using real exchange and one reaction/prop insertion.",
    "starting_compositions": "Each new commissioned shot has a distinct scene composition and new board; references condition identity. Judge actual first-frame pairs, not only hashes. Proof comparison should be a single concise continuous comparison with four distinct states, not paired recycled clips.",
    "camera": "Default locked speaker/reverse/listener coverage; choose occasional lateral follow, motivated pan/reveal, low/top-down action angle. Every move has a story reason; no repeated push-in recipe.",
    "dialogue": "Present-tense characters with their own voices/emotions; Seedance lip-sync validation checks actual mouths against accepted spoken turns. Offscreen speech may continue over distinct reaction/prop coverage.",
    "style": "Dimensional source drama: directional face light, true shadow depth, foreground shoulders/objects, conversational axes, restrained coherent material palette; local life remains appealing and specific."
  },
  "uncertainties": [
    "Source transcripts are model-assisted and not independently human-certified; source clocks within excerpts are approximate.",
    "English and Ukrainian WPM differ with tokenization and morphology; targets must pass Ukrainian listening, not a speed number alone.",
    "Candidate-cut detector is not whole-film ground truth; native shot lists also contain timing omissions.",
    "No optical-flow ground-truth camera analysis or motion-frequency census was performed; explicitly listed triplets support bounded observations.",
    "Story-turn counts depend on semantic segmentation and cannot prove measured attention or dopamine.",
    "No live audience retention/conversion experiment was run."
  ]
}