WCAG 1.2.5: Audio Description (Prerecorded)
Captions make video work without sound; audio description makes it work without sight. This criterion requires that prerecorded video content carries audio description of the important visual information — the actions, on-screen text, and scene changes that a blind viewer would otherwise miss. If your video’s meaning lives in the pictures, someone has to say it out loud.
The success criterion, in full
Audio description is provided for all prerecorded video content in synchronized media.
Synchronized media is video presented together with audio. Audio description is narration added to the soundtrack that describes important visual details that cannot be understood from the main soundtrack alone. One crucial relief valve from the Understanding document: if the existing audio already conveys everything the video shows, no additional description is needed.
Who this helps
The primary audience is people who are blind or have low vision. They hear the dialogue and the music, but everything the camera is doing is invisible: the wordless product demo, the chart the presenter gestures at, the name and title that appear under a speaker, the plot beat delivered by a glance. Audio description restores that channel by narrating it — traditionally during the natural pauses in dialogue.
Description helps other groups too: people with cognitive disabilities who benefit from key visual events being named and reinforced verbally, people listening to a video like a podcast while driving or walking, and anyone whose attention is on another task. Like captions, audio description is an accessibility feature with a large mainstream dividend.
What the requirement covers
For each prerecorded video with a soundtrack that your site publishes, ask: does a listener who cannot see the screen receive all the information a viewer does? If not, audio description must fill the gap. Conforming approaches include:
- A described version of the video — the description narration is mixed into the soundtrack, and the version is offered alongside (or instead of) the original.
- A selectable alternate audio track in the player that carries the program audio plus description, so users flip to it like a language track.
- Authoring the primary soundtrack to be self-describing — the presenter narrates all meaningful visual content as it happens (integrated description), which for screencasts and tutorials is often the cheapest and best option.
Where 1.2.5 sits among its neighbors
1.2.3 Audio Description or Media Alternative — A
At Level A you may choose: audio description or a full descriptive transcript. At AA, 1.2.5 requires the audio description itself — meeting 1.2.5 automatically meets 1.2.3.
1.2.7 Extended Audio Description — AAA
When dialogue leaves no pauses long enough to describe the visuals, extended description pauses the video to make room. That is a AAA enhancement — 1.2.5 itself expects description fitted into the existing soundtrack.
Pass and fail examples
- Pass — Described version offered next to the original. A product launch video is published in two versions: the original, and one with professional audio description mixed in, linked directly beside it as 'Watch with audio description'.
- Pass — Selectable description audio track. A training video's player offers an alternate audio track — 'English (Audio Description)' — that adds narration of the on-screen demonstrations during pauses in the instructor's speech.
- Pass — Self-describing screencast. A tutorial presenter narrates every action and reads all on-screen text aloud as they perform it. The audio track already conveys all the visual information, so no separate description is needed.
- Fail — Demo video with silent visual steps. A how-to video shows a settings panel while the presenter says 'just click here, then here, and you're done'. A blind viewer has no idea what 'here' is — the key visual information is never spoken.
- Fail — Transcript instead of description at AA. A marketing film full of visual storytelling ships with captions and a descriptive transcript but no audio description. That satisfies Level A (1.2.3) — it does not satisfy 1.2.5 at Level AA.
Implementation examples
Offer the described version alongside the original
The simplest robust pattern: publish two renditions and let the user pick. Make the described option a first-class, clearly labeled link — not a footnote.
<!-- ✗ Failing: one version, visual steps never spoken aloud -->
<video controls src="/media/product-demo.mp4"></video>
<!-- ✓ Passing: described rendition offered as an equal choice -->
<video controls src="/media/product-demo.mp4"></video>
<p>
<a href="/media/product-demo-described.mp4">
Watch this demo with audio description
</a>
</p>Or switch audio tracks in the player
Streaming setups can carry description as an alternate audio rendition, exactly like a dubbed language, so one video element serves both audiences.
# HLS multivariant playlist with a described audio rendition
#EXTM3U
#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="aud",NAME="English",\
LANGUAGE="en",DEFAULT=YES,URI="audio_en.m3u8"
#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="aud",NAME="English (Audio Description)",\
LANGUAGE="en",CHARACTERISTICS="public.accessibility.describes-video",\
AUTOSELECT=YES,URI="audio_en_ad.m3u8"
#EXT-X-STREAM-INF:BANDWIDTH=2500000,RESOLUTION=1280x720,AUDIO="aud"
video_720p.m3u8A descriptions text track is a supplement, not the solution
HTML defines kind="descriptions" for timed text meant to be spoken by assistive technology, but browser support is inconsistent — ship it as an extra, never as your only mechanism.
<video controls>
<source src="/media/product-demo-described.mp4" type="video/mp4" />
<track kind="captions" src="/media/demo.en.vtt" srclang="en"
label="English captions" default />
<!-- Supplemental text descriptions; support varies across browsers -->
<track kind="descriptions" src="/media/demo.desc.en.vtt" srclang="en"
label="English descriptions" />
</video>Common failures
- Publishing demo and how-to videos where crucial steps happen silently on screen — 'click here, then here' with nothing named aloud.
- Relying on a transcript alone: sufficient for 1.2.3 at Level A, not for 1.2.5 at Level AA.
- Assuming captions cover it — captions serve people who cannot hear; they do nothing for people who cannot see the video.
- On-screen text (speaker names, statistics, URLs, title cards) that is never read out in any audio.
- Buying professional description for the flagship brand film but ignoring the hundreds of tutorial and support videos where users actually need it.
- Hiding the described version behind an unlabeled icon or a help-center article instead of offering it next to the original.
- Shipping only a <track kind='descriptions'> file and assuming browsers will speak it — support is too unreliable to carry conformance alone.
- Scripting videos with wall-to-wall dialogue that leaves no pauses for description, when a small script change would have made room (or made the narration self-describing).
How to test for 1.2.5
- 1
Listen with your eyes closed
The definitive manual test. Play the video without looking at the screen and note every moment you lose information — an unexplained action, an unread name or number, a silent demonstration. Each note is a gap audio description must cover.
- 2
Inventory prerecorded video across the site
Marketing pages, help centers, onboarding flows, course content, embedded YouTube/Vimeo players. For each video with a soundtrack, record whether description exists, in what form, and where the user finds it.
- 3
Judge whether the existing audio self-describes
For each gap-free video from step 1, you're done — the audio track already conveys the visual information and no separate description is needed. Document that judgment; it is your conformance evidence.
- 4
Verify the description mechanism actually works
Play the described version or switch to the description audio track in each supported browser. Confirm the control is labeled, keyboard-accessible, and that the described audio really contains the narration.
- 5
Review description quality
Does the narration identify speakers and on-screen text, describe key actions, and fit the pauses without talking over dialogue? Compare against the eyes-closed notes from step 1 — every noted gap should now be filled.
No automated tool can tell whether a soundtrack conveys the visuals — this criterion is human judgment plus production process. Track it alongside the other media criteria in the WCAG 2.2 checklist.
Frequently asked questions
What does WCAG 1.2.5 Audio Description (Prerecorded) require?
It requires that audio description is provided for all prerecorded video content in synchronized media. Audio description is narration added to (or interleaved with) the soundtrack that describes important visual details which cannot be understood from the main soundtrack alone — actions, characters, scene changes, on-screen text, charts being pointed at. It is a Level AA success criterion in WCAG 2.0, 2.1, and 2.2, and it applies to video that has an accompanying audio track (synchronized media).
If my video's narration already says everything on screen, do I still need a separate audio description?
No. The Understanding document is explicit: if all of the information in the video track is already provided in the audio track, no audio description is necessary. A screencast whose presenter reads out every step and every piece of on-screen text as they go, or a talking-head interview with no meaningful visuals beyond the speaker, effectively self-describes. The test is whether someone listening with their eyes closed misses any information a sighted viewer gets.
How is 1.2.5 different from 1.2.3 Audio Description or Media Alternative?
1.2.3 (Level A) gives you a choice for prerecorded synchronized media: provide either an audio description or a full media alternative (a descriptive transcript covering all audio and visual information). 1.2.5 (Level AA) removes that choice: audio description itself is required — a transcript alone no longer suffices at AA. If you conform to 1.2.5, you automatically satisfy 1.2.3 as well. At AAA, 1.2.7 Extended Audio Description goes further for videos whose soundtrack leaves no pauses long enough for adequate description.
What should good audio description actually include?
Whatever a listener needs that the existing soundtrack doesn't give them: who is on screen and what they are doing, scene and location changes, facial expressions and gestures that carry meaning, on-screen text such as titles, captions, speaker names, and data callouts, and visual demonstrations ('she removes the battery cover'). Descriptions are traditionally timed to fit the natural pauses in dialogue, which is why concise, prioritized wording is a professional skill — describe what matters for understanding, not every visual detail.
Can I satisfy 1.2.5 with a separate 'described version' of the video?
Yes. Common approaches all conform: (1) a second version of the video with description mixed into the soundtrack, offered alongside the original ('Watch with audio description'); (2) a selectable alternate audio track in the player that includes description; or (3) where pauses allow, a single soundtrack authored so the narration itself covers all visual information. A text-based descriptions track (<track kind="descriptions">) that assistive technology reads aloud is a helpful supplement, but browser and screen reader support is too inconsistent to be your only mechanism.
Does 1.2.5 apply to decorative or ambient background video?
The criterion targets video content in synchronized media that conveys information. A purely decorative looping background — abstract shapes, ambient office footage behind a hero heading, with no audio track and no information content — is not synchronized media conveying content, and there is nothing to describe. Be honest in the assessment, though: if the 'ambient' video actually shows your product in use or communicates anything users are meant to take away, that information needs an accessible equivalent.
Related Success Criteria
Provide alternatives for prerecorded audio-only and video-only content.
Captions are provided for all prerecorded audio content in synchronized media.
Audio description or full text alternative is provided for prerecorded video content.
Captions are provided for all live audio content in synchronized media.
Sign language interpretation is provided for all prerecorded audio content.