SRT vs VTT: which caption format do you need?
Two formats, decades apart, both still in daily use. The choice is simpler than it looks.
· 5 min read
If you have ever exported captions you have met SRT and VTT, and probably picked one at random. They do largely the same job, and for most uses either will work — but the differences matter in a few specific cases.
What they have in common
Both are plain text files that pair a block of words with a start and end time. Both are human-readable and editable in any text editor, which is genuinely useful — if a caption is off by a second you can fix it without special software.
A caption file is not a transcript. It is a transcript cut into timed chunks sized for reading on screen, which is why the same content works badly as a document.
SRT: older, simpler, accepted almost everywhere
SubRip Text dates from DVD-ripping software and is about as simple as a format can be: a number, a time range, the text, a blank line. It has no official styling and no formal specification, which turns out to be a strength — near-universal support, and nothing to get wrong.
Timestamps use a comma before the milliseconds, which is the usual thing people trip over when hand-editing.
VTT: built for the web, with more capability
WebVTT was designed for HTML5 video and is the format browsers natively understand through the track element. It supports positioning, basic styling, metadata and cue identifiers, and it uses a period before milliseconds rather than a comma.
If you are putting video on your own website with a native player, VTT is the natural choice — it is what the platform expects.
Which to choose
A short decision rule that covers almost every case:
- Uploading to YouTube, Vimeo or a social platform — SRT, and do not overthink it
- Embedding video on your own site with an HTML5 player — VTT
- Handing captions to an editor or a client — SRT, for the widest compatibility
- Needing caption positioning or styling — VTT, since SRT cannot express it
- Archiving alongside a transcript — keep the plain text as well; a caption file is a poor document
A note on line length
Whichever format you use, the readability problem is the same: captions that are too long or change too fast are worse than none. The widely used guidance is around 32 to 42 characters per line, at most two lines at a time, held on screen long enough to read comfortably.
Automatically generated captions often ignore this, which is why a pass by a human still improves them noticeably.