SRT / VTT conversion and sync: diagnose before moving the timeline

Bad sync may not come from the subtitle file. A different intro edit, frame-rate conversion, removed scenes, or player behavior can change the timeline. Record the video duration and version, copy the original subtitles, and save every conversion or timing correction as a new file.

Garbled text usually points to character encoding. Prefer UTF-8. If the original displays correctly in one player, keep an untouched copy so encoding problems are not introduced before timing diagnosis.

Find a line with a clear boundary near the beginning and record subtitle time versus audio time. Repeat near the end. If both are 1.5 seconds early or both 1.5 seconds late, the problem is a fixed offset; add the same value to every timecode.

If the start is off by 0.3 seconds but the end by 4 seconds, error grows over time. A fixed offset can repair only one point. Use two anchors for linear correction, or return to the correct video version and frame-rate source.

SRT usually separates milliseconds with a comma, as in 00:01:02,500. WebVTT uses a period and begins with a WEBVTT header. Cue settings, tags, and comments also differ. Renaming .srt to .vtt does not produce a valid file.

After conversion, inspect the first cue, one random middle cue, and the last. Check sequence, start before end, target timecode syntax, blank lines, and text grouping. If the source uses style tags or positioning cues, confirm that the target platform supports them.

Sample fast dialogue, long lines, speaker changes, and the ending. Captions should not linger long after audio or disappear before they can be read. Automatic line breaks and transcript extraction can lose tone, speaker, and visual context; they do not replace editorial review.

Before release, search for phone numbers, emails, accounts, internal project names, and unauthorized names. Redact a text copy, then generate a new delivery file. Do not edit only the web preview and forget to download the result.