Subtitle converter

SRT is the plain one that everything accepts and is the safe default. WebVTT is what browsers want for HTML video, and is close enough to SRT that the conversion is nearly mechanical. ASS and SSA carry positioning, fonts and colours, which is why they are used for anything typeset rather than merely captioned.

Converting from a styled format to a plain one keeps the words and the timings and drops the styling, because there is nowhere for it to go. That is worth knowing before you convert a carefully positioned sign translation into SRT.

A subtitle file is only useful if it stays in step. The formats express time slightly differently — commas against full stops, hours sometimes omitted — and getting that translation wrong shifts every line in the file. Timings are converted exactly rather than rounded, so a track that was in sync stays in sync.

SRT came from a DVD ripping tool in the late 1990s and is the simplest thing that works: a number, a time range, the text, a blank line. VTT is its web descendant, standardised for HTML video, with styling and positioning added. SUB and SBV came from DVD and YouTube workflows respectively, and ASS came from anime fansubbing, where captions had to sit around on-screen Japanese text.

That history explains the capability gap. SRT is universal and can express almost nothing beyond the words and when they appear; ASS can position, colour, animate and typeset. Converting downward from ASS to SRT necessarily discards all of it, which the pages say rather than leaving you to find out.

Subtitle files carry no declaration of their character encoding. UTF-8 is the sane default and is what almost everything writes now, but older files are frequently in a regional code page — Windows-1252 for Western European, Windows-1251 for Cyrillic — and nothing in the file says so.

The symptom is distinctive: the timings are perfect and every accented character is a question mark or a pair of nonsense symbols. Nothing is damaged. The file needs to be read as the encoding it actually is and written back out as UTF-8, which is what converting it here does.

SUB files in the MicroDVD style store positions as frame numbers rather than as times, which means they only make sense against a video at a specific frame rate. Convert one assuming 25 frames per second when the video is 23.976 and the subtitles start correct and end minutes adrift.

The drift is progressive rather than sudden, which is how it is told apart from a fixed offset. A constant offset means the subtitles start at the wrong moment and stay parallel; drift means the frame rate assumption was wrong. Time-based formats such as SRT and VTT do not have this problem at all, which is a good reason to convert away from frame-based ones.

SRT entries are supposed to be sequentially numbered and non-overlapping. Many players tolerate gaps in the numbering, entries out of order and two captions on screen at once; some refuse the file outright, and browsers reading VTT are among the stricter readers.

If a subtitle file works in VLC and fails in a browser or on a television, this is nearly always why rather than the format being unsupported. Conversion renumbers and reorders as a side effect, which is why converting a file to its own format sometimes fixes it.