What are video captions: captions explained
Video captions are timed text shown on top of a video and synchronised to the audio, carrying the spoken words plus any meaningful non speech sound. They let a viewer follow the content with the sound off, or with no hearing at all.
The word people reach for interchangeably is subtitles, and the two are not quite the same thing.
What does captioning mean? Closed, open and subtitles
- Closed captions are carried as a separate text track alongside the video and can be switched on and off by the viewer. On the web this is usually a WebVTT file, a plain text format that lists each line of text with a start and end timecode.
- Open captions are burned into the picture during editing. They cannot be turned off, which makes them reliable on players that ignore caption tracks and useless for anyone who wants them gone.
- Subtitles assume the viewer can hear and are usually a translation of the dialogue. Captions assume the viewer cannot hear, so they also identify who is speaking and note sound that carries meaning.
That last distinction matters more than it sounds. In a video about a knocking suspension, the line describing the knock is the content. Subtitles would drop it.
Why captions matter when the sound is off
Most video sent to a phone is opened somewhere the sound is inconvenient. An office, a waiting room, a shop floor, a bus, a house with someone asleep in it. The viewer either finds headphones, unmutes and hopes, or watches without audio.
Without captions, a muted video of a person talking conveys nothing at all. With captions it still works. That is the whole case, and it is a bigger driver of watch time than most people assume, because the alternative is not a quieter viewing, it is no viewing.
There is a second effect that is less obvious. Text on screen holds attention through the first few seconds of a video in a way a silent talking head does not, which shows up in video engagement as fewer people leaving before the message lands.
Accessibility obligations vary by jurisdiction
Captions are also an access requirement, and the specifics depend on where you operate. Two reference points do most of the work.
The W3C's Web Content Accessibility Guidelines set Success Criterion 1.2.2, Captions (Prerecorded), at Level A, the most basic conformance level: captions are required for all prerecorded audio content in synchronised media. Level A is not the ambitious tier. It is the floor.
In the European Union, Directive (EU) 2019/882, the European Accessibility Act, has applied since 28 June 2025 and covers services including e-commerce and access to audiovisual media, with captioning and subtitles named among the accessibility requirements. Member states transposed it into national law, so the exact obligations differ country by country.
Elsewhere the picture is different again, and public sector bodies, regulated industries and consumer facing businesses are not held to the same standard in every market. Check what applies to you rather than assuming a single global rule, because there is not one.
Captions and transcripts: how they differ
They are a pair, and they solve different halves of the same problem.
- Captions are timed and sit on the video. They work during playback, in sync, for someone who is watching.
- A video transcript is untimed and sits beside the video. It is the full text as a document, readable without playing anything, searchable, and copyable.
Producing both from the same transcription pass costs very little more than producing one, since the timed file and the flat document come out of the same source text.
Video captions explained: a worked example
An advisor records a 45 second clip pointing at a worn brake pad and explaining the measurement. The captions carry the spoken explanation and the figure, so the customer, sitting in an open plan office with the phone on mute, reads what the pad measures and why it matters.
Without captions that customer sees a person gesturing at a dark under car shot and understands nothing. The video was delivered, opened and wasted.
What good captions look like
- Accurate on the words that carry money. Part names, measurements, prices and dates are exactly where automatic captioning fails, and exactly where an error is expensive. Read those lines back.
- Two lines at a time, at most. Longer blocks cover the picture and outrun the reader.
- Positioned clear of the subject. Captions sitting over the damage you are pointing at defeat the purpose, which is a real risk in a vertical video where the safe area is narrow.
- Readable against the picture. White text on a bright bonnet is invisible. Use a background plate or an outline.
- Punctuated properly. Punctuation is what makes a caption scan at speed instead of reading as a wall of words.