Skip to content
Tutorials Center
My learning

Topics

Lesson 23 of 38 Β· Colour, Contrast and Media

Captions, Transcripts and Audio Description

Make video and audio accessible: WebVTT captions, transcripts, audio description, accessible player controls and autoplay rules, using the native video element.

Video and audio carry information through sound and picture. Deaf and hard-of-hearing people cannot use the sound; blind people cannot use the picture; people in noisy offices, on trains or in quiet libraries simply have the sound off. Alternatives for time-based media are among the highest-impact accessibility work you can do, and they help search and translation too.

What WCAG asks for

Content Level A Level AA Level AAA
Audio only (podcast) Transcript (1.2.1)
Video only (silent animation) Transcript or audio track (1.2.1)
Prerecorded video with sound Captions (1.2.2); audio description or transcript (1.2.3) Audio description (1.2.5) Sign language (1.2.6)
Live video with sound Captions (1.2.4)

In short: captions for speech and meaningful sound, an audio description for important visuals, and a transcript for everything.

Captions vs subtitles

  • Captions include speech and non-speech information: "[door slams]", "[tense music]", speaker names. They are for people who cannot hear.
  • Subtitles translate or transcribe dialogue for people who can hear but do not understand the language.

Auto-generated captions are a start, not a finish. Review them for names, jargon and punctuation; accuracy below about 99% fails users in practice.

Adding captions with the video element

The HTML way is a <track> element pointing at a WebVTT file.

HTML
<video controls poster="/samples/media/fractal-poster.jpg" width="480">
  <source src="/samples/media/fractal.mp4" type="video/mp4">
  <source src="/samples/media/fractal.webm" type="video/webm">
  <track kind="captions" src="/captions/fractal.en.vtt" srclang="en" label="English" default>
  <track kind="captions" src="/captions/fractal.es.vtt" srclang="es" label="EspaΓ±ol">
  <p>Your browser cannot play this video. <a href="/samples/media/fractal.mp4">Download it</a>.</p>
</video>

A WebVTT file looks like this:

TEXT
WEBVTT

00:00:00.000 --> 00:00:03.500
[Soft electronic music]

00:00:03.500 --> 00:00:07.000
NARRATOR: Every zoom reveals a smaller copy of the whole.

Rules for good captions: one or two lines, about 32 to 42 characters per line, on screen long enough to read (roughly 1 second per 3 words plus a bit), in sync with speech, and without covering on-screen text.

The next demo builds a tiny caption track in JavaScript so it runs without any file, and shows the cues as they become active. Press play on the mock timeline.

Caption cues in sync

Audio description

Audio description (AD) is a narrated track that describes important visual information (actions, scene changes, on-screen text) that is not clear from the dialogue. Strategies, from cheapest to richest:

  1. Write the video so it does not need AD: say what is on screen in the narration ("Click the blue Save button in the top right").
  2. A text alternative: a transcript that includes visual descriptions (satisfies level A).
  3. A second video or audio track with narration mixed in (needed for AA when the visuals carry meaning not in the dialogue).

<track kind="descriptions"> exists but browsers do not read it aloud by default, so most sites publish a separate described version and link to it near the player.

Transcripts

A transcript is plain text on the page (or a link to it) with speaker names, spoken words, and important sounds and visuals. Put it beside the media or in a <details> element. For audio-only content it is the alternative.

HTML
<details>
  <summary>Read the transcript</summary>
  <p><strong>Maya:</strong> Welcome back. Today we test with a keyboard.</p>
  <p><em>[She unplugs her mouse.]</em></p>
</details>

Player and autoplay rules

  • Use native controls or make sure a custom player is fully keyboard operable, labeled, with captions toggle (see keyboard operability).
  • Do not autoplay sound. If audio plays more than 3 seconds, give a way to pause, stop or lower the volume independent of system volume (1.4.2).
  • Muted autoplay of background video needs a visible pause control if it lasts over 5 seconds (2.2.2).
  • Do not trap focus in an embedded player.
  • Give the <video> or iframe a title that says what it is.

Workflow tips

Write a script first; captions and transcript fall out of it. Ask for caption files from the start of a production. Use a style guide for speaker names and sound cues. Re-check the captions after every edit of the video.

Check your understanding

Quick check Β· 4 questions

of 4 answered You scored / 4 Β· perfect!

  1. 1. What is the difference between captions and subtitles?

    Captions describe sound as well as speech. Subtitles usually translate or transcribe dialogue only.
  2. 2. Which element adds a WebVTT caption file to a video?

    The track element, with kind, src, srclang and label, attaches captions to audio or video.
  3. 3. A podcast has no video. What is the minimum text alternative at level A?

    1.2.1 requires a transcript for prerecorded audio-only content.
  4. 4. Which is best practice for autoplay?

    Sudden sound disrupts screen reader users, and long-running motion needs a way to pause (1.4.2, 2.2.2).

Summary

  • Provide captions, a transcript and, where visuals matter, audio description.
  • Attach captions with <track kind="captions"> and WebVTT files, and review auto captions.
  • Prefer narration that describes what is on screen.
  • Avoid autoplay with sound; give pause controls and accessible players.

Next: Motion, Animation and Flashing