ROGUELIKE — EXPLAINER VIDEO PRODUCTION RECORD
Winning the fight, losing your options

OUTPUT
318.400 s · 1920 × 1080 native · 30 fps · 9,552 frames · no upscaling

FILE
MP4 · H.264 High · yuv420p · AAC 48000 Hz, 2 channels · 17,082,397 bytes (16.29 MiB)

ELAPSED
1816.951 s from 2026-10-27T15:05:12.601466+00:00 to 2026-10-27T15:35:29.552800+00:00; asset/narration preparation through final render verification, including authoring and review; excludes earlier research and later packaging.

RENDER PROCESSING
One completed render: 262.176 s process wall time; internal pipeline 260.633 s. Capture 239.896 s, encoding 17.729 s, assembly 0.466 s. Not pure GPU time.

ANIMATION
Authored SVG/GSAP 3.15.0 · HyperFrames 0.8.200 · local screenshot capture · 2 software browser workers · no generative image/video model or GPU encoder

NARRATION
Kokoro v1.0 82M ONNX · bm_george · en-gb · speed 1.0 · CPUExecutionProvider, 4 threads · kokoro-onnx 0.4.7 / ONNX Runtime 1.30.0 · original 24 kHz mono PCM16 WAV; quantization unverified; no adapters; seed not exposed

VOICE GENERATION
30 completed paragraph jobs · 0 failed · 0 retried · 39.132 s total wall time · 38.410 s summed inference-call time

SCORE AND EFFECTS
Original deterministic DSP · seed 27102026 · 34 cues · 48 kHz stereo stems · 9.000 s successful generation/mix preparation; one earlier normalization attempt failed and was corrected; its elapsed time was not recorded

AUDIO MEASUREMENTS
Finished mix: -15.02 LUFS integrated, -1.54 dBTP true peak, 2.10 LU range. Voice target −18 LUFS; score target −36 LUFS. Audio measured, not listened to.

SETTINGS
Delivery quality; 30 fps; software capture; no upscaling or frame interpolation. Diffusion steps, sampler, scheduler, guidance and denoise are not applicable. Motion has no random seed; procedural audio seed is recorded above.

JOBS AND RESOURCES
Video: 1 completed, 0 failed, 0 retried. Voice: 30 completed, 0 failed, 0 retried. Audio preparation: 1 failed attempt, 1 successful retry. No ComfyUI jobs. GPU not used for this render/voice route. Peak VRAM and energy not recorded.

VERIFICATION
Zero layout/runtime/contrast/lint errors; full video and audio decode passed; nine final decoded frames visually inspected. Out-of-order seek at 185 s was pixel-identical. Short deliberate hold warnings retained. Optional sentence caption timing is approximate.

SCENES — OUTPUT DURATION / VOICE GENERATION TIME
Render processing is not independently attributed by scene; the film is one continuous render job.
1. Progress can hide danger: 37.936 s output / 4.565 s voice generation
2. Retreat while it is easy: 37.448 s output / 4.440 s voice generation
3. What spare capacity buys: 40.629 s output / 5.042 s voice generation
4. Find the assumption that can break: 42.885 s output / 5.211 s voice generation
5. Choose a trigger before the pressure: 43.312 s output / 5.162 s voice generation
6. Use a plan you can actually run: 38.835 s output / 4.603 s voice generation
7. Decide what success means: 37.960 s output / 4.558 s voice generation
8. Keep uncertainty in the picture: 39.387 s output / 4.830 s voice generation

RENDER PHASES
compile: 0.232 s
browser_probe: 0.077 s
video_extract: 0.001 s
audio_process: 2.183 s
file_server: 0.004 s
capture_disk: 239.896 s
encode: 17.729 s
assemble: 0.466 s

VOICE MODEL SHA-256
7d5df8ecf7d4b1878015a32686053fd0eebe2bc377234608764cc0ef3636a6c5

VIDEO SHA-256
22f5003ee5e7656c5c27dc93fe3340b1a46767ccd4d4eb00c19e751d4e1ea725

EDITABLE LOCAL PROJECT
/home/albert/Documents/ChatGPT/Codex/videos/roguelike-options
