video-assemble
Synthesize the final video narration cut: lay the narration audio onto the source video, duck the original sound according to the narration windows, generate SRT / ASS subtitles with optional burn-in, and finally perform loudness normalization. Used as the final assembly stage. Inputs are the source video, tts_meta.json, and narration positions; outputs are the recap cut and subtitles. Trigger words: video compositing, mixing, subtitles, subtitle burn-in, assemble…