Two kinds of captions
Burned-in captions are drawn into the video frames themselves. They always appear, they cannot be turned off, and they survive any platform. They also cannot be searched, translated, or read by a screen reader, and they are stuck at whatever size you chose.
Sidecar captions are a separate timed text file — SRT or VTT — carried alongside the video. They are selectable, translatable, indexable by search engines, and accessible. Not every upload path accepts them.
The pragmatic answer for social video is usually both: a sidecar file where the destination supports it, burned-in where it does not.
Writing captions that read well
Keep lines short — roughly six or seven words — and never more than two lines on screen at once. A caption block should sit long enough to read comfortably, which for a short line is around a second and a half.
Caption what is said, not what you meant to say. Cleaning up filler is fine; rewriting the sentence is not, because viewers who need captions are then getting different content from everyone else.
Position matters. Most platforms overlay their own interface on the lower part of the frame, so captions sitting at the very bottom get covered by the caption box, the username, or the buttons.
- 1
Transcribe first
Get the words down before worrying about timing. Automatic transcription is a reasonable starting point but always needs a pass for names, jargon and homophones.
- 2
Time in blocks, not words
Split at natural clause boundaries so each block is a readable unit. Splitting mid-phrase makes viewers work.
- 3
Watch it muted, end to end
This is the step people skip and the only one that reliably catches captions that lag, overlap, or sit under the interface.
Translation
Machine translation of captions is good enough to be worth doing and not good enough to be left unchecked, particularly with idiom, humour and names. If a language matters to your audience, have someone who speaks it read the file once.
Automatic caption generation and translation inside Frenz are planned rather than available today.
Common questions
- SRT or VTT?
- VTT is the web standard and supports styling and positioning; SRT is simpler and accepted almost everywhere. If you have a choice, VTT.
- Do captions help reach?
- They help retention, which most ranking systems respond to — a muted viewer who can follow along keeps watching. They also make the words available to search engines when carried as a sidecar file.