1.2.3 Audio Description or Media Alternative (Prerecorded)
- Level A
- Perceivable
- Since WCAG 2.0
Summary
This criterion protects people who are blind or have low vision when they encounter prerecorded video with sound. The dialogue and soundtrack reach them, but anything conveyed only visually (on-screen text, charts, demonstrations, who entered the room, what the presenter is pointing at) does not. WCAG 1.2.3 requires the author to bridge that gap in one of two ways: an audio description track that narrates the important visual content, or a full text alternative for time-based media (in practice, a descriptive transcript that covers both the dialogue and the visual information).
The auditor's mental model at Level A is a choice, not a mandate for audio description. Either route satisfies 1.2.3. At Level AA, SC 1.2.5 Audio Description (Prerecorded) removes the transcript option and requires audio description outright, so in an AA audit you assess the two together but record them separately. The second key judgement is whether the video contains important visual-only information at all: a talking-head clip where the speaker says everything that matters may already be fully conveyed by its own soundtrack, and that is not a fail. Finally, note the exception: media that is explicitly a media alternative for text, and clearly labelled as such, is out of scope.
Official wording
An alternative for time-based media or audio description of the prerecorded video content is provided for synchronized media, except when the media is a media alternative for text and is clearly labeled as such.
EN 301 549 mapping
- Web pages
- Clause 9.1.2.3
- Software and native apps
- Clause 11.1.2.3
- Closed functionality (kiosks, terminals)
Closed systems must provide speech output for video information (clause 5.1.3.7).
Clause 9.1.2.3 applies this criterion to web pages unchanged and clause 11.1.2.3 applies it to software that supports assistive technologies. On closed systems the requirement becomes speech output for video information under clause 5.1.3.7.
Clause references are to EN 301 549 V3.2.1 (2021-03), the harmonised European standard. Descriptions are our own summary, not the text of the standard.
In practice
Web
Play every prerecorded video with an audio track on the page and ask first: does anything important appear only visually? Watch with your eyes closed or the screen turned away and note what you lose: on-screen captions of names and titles, product demonstrations, slides, text overlays, visual gags, scene changes that matter to comprehension. If nothing important is visual-only, record a pass with that reasoning as evidence.
Where visual-only content exists, look for either satisfying mechanism. For audio description, check the player's audio track or settings menu for a described version, or a separately linked "described video". HTML video can carry an alternate audio track or a second described rendition; some sites link the described version alongside the original. Play it and confirm the narration actually covers the important visual content, not just a duplicate soundtrack. For the transcript route, look near the video for a transcript link or expandable panel, then read it critically: a dialogue-only transcript is a captions file in disguise and does not satisfy this criterion. A conforming descriptive transcript includes the visual information (speaker identification where it matters, descriptions of on-screen text and actions) so that reading it is a genuine substitute for watching.
Check discoverability too: a transcript buried on an unlinked page is not "provided" in any meaningful sense. And check the exception before failing anything: if the video is itself an alternative to adjacent text content (a video version of the article beside it) and is clearly labelled as such, 1.2.3 does not apply.
iOS
The European standard applies this requirement to apps as well as web pages, so grade prerecorded video inside iOS apps the same way. AVFoundation-based players support alternate audio tracks (media selection options), which is the native route for audio description: open the player's audio selection control and look for a described track, often labelled "Audio Description" or "AD". Confirm it plays and narrates the visual content.
If there is no described track, look for the transcript route inside the app: a transcript screen, expandable text view, or link presented alongside the video. Verify with VoiceOver that the transcript screen is itself readable (it is the accessibility path, so it must work with the screen reader). As on the web, first establish whether the video has important visual-only information; promotional clips and talking-head explainers frequently pass on their own soundtrack. Common iOS failure spots: onboarding and tutorial videos that demonstrate gestures silently over music, and help videos that show where to tap without saying it.
Where the audit covers a closed system (a kiosk or embedded device where users cannot attach their own assistive technology), the European standard expects the visual information in video to be available as speech output from the system itself, since a downloadable transcript is not usable there.
Android
The same applies on Android: the European standard covers software, so in-app prerecorded video is in scope. ExoPlayer/Media3-based players expose alternate audio tracks through track selection; open the player's audio or settings control and check for a described audio track, then play it to confirm the narration covers the visual content. Some apps instead ship a separate "described version" of the video, which is equally acceptable.
For the transcript route, look for a transcript link, bottom sheet, or dedicated screen near the player, and verify it includes visual information rather than dialogue alone. Confirm with TalkBack that the transcript is reachable and readable. As elsewhere, grade the video's content first: if the soundtrack already conveys everything important, pass it and note why. Common Android failure spots: video-first tutorial flows that demonstrate UI steps with no narration, and marketing videos with heavy on-screen text over a music bed and no transcript anywhere in the app.
Pass and fail examples
Passes:
- A product demo video with a "Described version" option in the player whose alternate audio track narrates the on-screen actions during pauses in the dialogue.
- A recorded webinar with a linked descriptive transcript that includes the dialogue plus descriptions of each slide and demo step: the text alternative route satisfies Level A even with no audio description track.
- A talking-head interview where both participants are introduced by name in the dialogue and nothing important appears only on screen: the existing soundtrack already conveys the visual information.
- A video version of a written tutorial, presented beside the full text and clearly labelled "Video alternative to the guide above": media alternative for text, exception applies.
- An in-app tutorial video where the narrator speaks every step being demonstrated ("now tap the blue Save button in the top right"), leaving no important visual-only content.
Fails:
- A tutorial video that silently demonstrates gestures over background music, with no described track and no transcript: the entire instructional content is visual-only.
- A promotional video whose key statistics appear only as on-screen text overlays, accompanied by a transcript that reproduces the voiceover but omits the overlays: the transcript is not a full alternative.
- A recorded presentation where the speaker repeatedly says "as you can see here" about slides that are never read aloud, with captions provided but no audio description or descriptive transcript.
- An app help video with a described audio track that exists in the player menu but fails to play, or plays the identical undescribed soundtrack: the mechanism is present but does not deliver the content.
Not a fail under this criterion:
- A video with sound but no captions: captions serve people who cannot hear the audio and belong under SC 1.2.2 Captions (Prerecorded).
- A silent, video-only animation with no audio track: that is not synchronised media, so it files under SC 1.2.1 Audio-only and Video-only (Prerecorded).
- A video with a conforming descriptive transcript but no audio description track, in a Level A audit: the transcript route satisfies 1.2.3. In an AA audit, record the missing audio description as a fail of SC 1.2.5, not of this criterion.
- Live streams: prerecorded criteria do not apply; live captioning is SC 1.2.4 territory.
Commonly confused with
- Audio Description (Prerecorded), SC 1.2.5. The Level AA sibling covers the same videos but removes the transcript option: audio description becomes required. Audit the two together on each video, then file the result correctly: transcript-only content passes 1.2.3 and fails 1.2.5. Citing 1.2.3 for a missing audio description when a good descriptive transcript exists is the most common misfiling.
- Captions (Prerecorded), SC 1.2.2. Captions make the audio available to people who cannot hear; audio description makes the video available to people who cannot see. A fully captioned video can still fail 1.2.3, and a described video can still fail 1.2.2. One video, two independent findings.
- Audio-only and Video-only (Prerecorded), SC 1.2.1. This criterion covers synchronised media (video plus audio together). A media file that is video-only or audio-only belongs under 1.2.1. Check what the file actually contains before choosing the criterion.
- Transcript quality versus transcript presence. A dialogue-only transcript is evidence for captioning workflows, not a text alternative for time-based media. To satisfy 1.2.3 by transcript, it must convey the important visual information as well as the speech. Grade the content of the transcript, not the existence of a "Transcript" link.
How AUDITSU tests this
AUDITSU's audit walkthrough flags every screen containing prerecorded video with sound and asks a short sequence: does the video carry important visual-only information; if so, is there an audio description track or a described version; if not that, is there a descriptive transcript that covers the visual content? The walkthrough prompts you to watch the video audio-only first, so the "what would a blind user miss" judgement is made before you look for mechanisms. Each screen records a pass, fail, or not applicable result with your evidence attached (what was visual-only, which mechanism was found, or why none was needed).
Because the Level AA sibling covers the same media, the walkthrough pairs this question with SC 1.2.5 on the same screen, so one viewing produces both grades, filed under the correct criterion. For the full guided workflow, see the audit platform.