A reviewer asks you to hold a tutorial shot long enough for the result to appear. You move the next narration line later, and its caption remains where it was. This small script flags caption starts that have drifted from their corresponding spoken line. Use times from the latest sequence or export.
from dataclasses import dataclass from decimal import Decimal @dataclass class Cue: label: str voice_start: Decimal caption_start: Decimal cues = [ Cue("setup", Decimal("8.45"), Decimal("8.45")), Cue("result", Decimal("10.60"), Decimal("10.15")), Cue("next", Decimal("12.20"), Decimal("12.20")), ] tolerance = Decimal("0.05") for cue in cues: offset = cue.caption_start - cue.voice_start if abs(offset) <= tolerance: continue direction = "late" if offset > 0 else "early" print(f"{cue.label}: caption starts {abs(offset):.2f}s {direction}") Run it with Python 3.10 or later. It prints:
result: caption starts 0.45s early To adapt it to a new revision, enter the times for the changed line and one line on either side. Check the affected cue in your editor and run it again with the new time. Finally, play the exported video through the change. Make sure the action is visible and the source audio sounds continuous.
AI assistance was used to draft and edit this tutorial. The example was run locally.
