Convert and fix subtitle files.

What each control does

Shift adds the same number of seconds to every cue. Use it when subtitles are consistently early or late from the first line to the last, which usually means the video was trimmed at the start after the subtitles were made.

Stretch multiplies every timestamp by a factor. Use it when subtitles start in sync and drift steadily, which is a frame-rate mismatch: a 25 fps timing played at 23.976 fps needs a factor of 0.959, and the reverse needs 1.0427. The drift article explains how to tell the two problems apart.

Re-wrap rebuilds cues with the same rules the transcriber uses for its exports: at most 42 characters a line, two lines, between one and six seconds on screen, an 80 ms gap between cues. Turn it off to keep the original cue boundaries and only convert or shift.

Plain text input has no timing, so it converts to text only; the timing fields are ignored for it. Plain text output drops timing and keeps one paragraph per run of speech.

Formats

SRT and VTT are read with cue numbers, VTT headers, NOTE blocks, positioning settings and inline tags removed. Cues whose timestamps cannot be read are skipped and counted in the summary. Which format you need is covered in SRT vs VTT vs plain text.