The criteria at a glance
| Criterion | What it requires |
|---|---|
| 1.2.1 Audio-only and Video-only (Prerecorded) | Audio-only or video-only recordings need a text alternative, such as a transcript. |
| 1.2.2 Captions (Prerecorded) | Recorded video with sound needs captions. |
| 1.2.3 Audio Description or Media Alternative (Prerecorded) | Audio description or a full text alternative for the video. |
| 1.2.4 Captions (Live) | Live video with sound needs live captions. |
| 1.2.5 Audio Description (Prerecorded) | Recorded video needs audio description (Level AA). |
| 1.4.2 Audio Control | Sound that plays on its own for more than three seconds must be pausable or have its own volume control. |
Captions are more than subtitles
Subtitles usually translate dialogue for people who can hear but do not speak the language. Captions are for people who cannot hear: they include speaker names, relevant sounds such as “doorbell rings” and music that carries meaning. WCAG asks for captions.
Auto-generated captions are a starting point, but check them before publishing; names, technical terms and punctuation often go wrong. Make sure they have enough contrast and do not cover important parts of the picture.
Live streams (1.2.4 Captions (Live)) need captions in real time, usually from a professional captioner.
Audio description and text alternatives
Audio description narrates, in natural pauses, what is happening on screen that matters for understanding: actions, on-screen text, scene changes. It gives blind viewers what they cannot see.
At Level A (1.2.3 Audio Description or Media Alternative (Prerecorded)) a full text alternative describing sound and picture may be used instead. At Level AA (1.2.5 Audio Description (Prerecorded)) audio description is required. If everything important is already said aloud, as in a talk without slides, little or no extra description may be needed. Planning for it in the script is the cheapest route.
The technical side: WebVTT and the track element
On the web, captions are usually stored as a WebVTT file (.vtt) and added to a <video> with a <track> element. Use kind="captions" for captions, kind="descriptions" for text descriptions, srclang for the language and label for the name shown in the player menu.
Because they are text, captions can be turned off, resized and indexed by search engines. Captions burned into the picture can do none of that.
The player must be usable
- Every control can be reached and used by keyboard (2.1.1 Keyboard), with visible focus (2.4.7 Focus Visible).
- Buttons have an accessible name such as “Play” or “Captions on” (4.1.2 Name, Role, Value).
- Controls have enough contrast against the video (1.4.11 Non-text Contrast) and are big enough to tap (2.5.8 Target Size (Minimum)).
- Captions can be switched on directly in the player.
No sound on page load
Sound that starts when a page loads drowns out a screen reader. If audio plays automatically for more than three seconds, there must be a way to pause or stop it, or to control its volume separately from the system volume (1.4.2 Audio Control). Better still: never start sound automatically. Silent background videos that move for more than five seconds also need a way to pause them (2.2.2 Pause, Stop, Hide).
Frequently asked questions
Are a platform’s auto-captions enough?
Only if they are accurate. Review and correct them before publishing; errors in names and technical terms are common.
Does a video without speech need captions?
Not if there is no meaningful sound. A video-only recording still needs a text or audio alternative under 1.2.1.
What about podcasts?
For audio-only recordings, 1.2.1 asks for a text alternative, usually a transcript.
Sources
How does your website do?
The free check tests your home page in two minutes. The audit decides all 55 criteria, by a person.