Problem
Caption timing shifts after frame-rate conversion because subtitle timestamps were not conformed with picture.
Solution
Root Cause / Diagnostic:
Subtitle formats (SRT, VTT) store timing in absolute hours, minutes, seconds, and milliseconds (HH:MM:SS,mmm), whereas video timelines operate on discrete frame intervals. When a sequence is converted across disparate frame rates (e.g., 29.97 fps drop-frame conformed to 23.976 fps), the video picture undergoes a 0.1% speed change or frame re-sampling. If subtitle timestamps are not conformed alongside the video, subtitles progressively desynchronize, appearing seconds ahead of or behind spoken dialogue by the end of the video.
Actionable Fix:
1. Conform the subtitle file inside the NLE or subtitle editor using a frame-rate retiming tool, explicitly setting source frame rate (29.97) and destination frame rate (23.976).
2. Snap subtitle start times to exact dialogue waveform attack transients on the conformed video timeline.
3. Audit dialogue cues at the 50% and 95% marks of the video timeline to verify that captions appear within 2 frames of spoken audio.
Pro Tip:
When authoring subtitles for global syndication, maintain master subtitle files in frame-indexed formats (such as EBU-STL or SMPTE-TT) so timing scales mathematically with video frame rate transforms.
Subtitle formats (SRT, VTT) store timing in absolute hours, minutes, seconds, and milliseconds (HH:MM:SS,mmm), whereas video timelines operate on discrete frame intervals. When a sequence is converted across disparate frame rates (e.g., 29.97 fps drop-frame conformed to 23.976 fps), the video picture undergoes a 0.1% speed change or frame re-sampling. If subtitle timestamps are not conformed alongside the video, subtitles progressively desynchronize, appearing seconds ahead of or behind spoken dialogue by the end of the video.
Actionable Fix:
1. Conform the subtitle file inside the NLE or subtitle editor using a frame-rate retiming tool, explicitly setting source frame rate (29.97) and destination frame rate (23.976).
2. Snap subtitle start times to exact dialogue waveform attack transients on the conformed video timeline.
3. Audit dialogue cues at the 50% and 95% marks of the video timeline to verify that captions appear within 2 frames of spoken audio.
Pro Tip:
When authoring subtitles for global syndication, maintain master subtitle files in frame-indexed formats (such as EBU-STL or SMPTE-TT) so timing scales mathematically with video frame rate transforms.