one suggestion, create a subtitle system that is similar to the transcription in relation to the cuts, for example, the video transcription is separated when we make the cuts in the audio, so the subtitles could follow this idea too, the timecode follows the cuts of the audio on timeline