Extended Audio Description (Prerecorded)
Some videos leave no room for description. The gaps between lines of dialogue are too short to say what is on screen, so extended audio description pauses the video, lets the narration finish, then starts it again. The rule only bites where those gaps really are too short. If your existing silences are long enough to hold standard description, you already meet it.
Why it matters
Standard audio description has to fit inside the silences the edit already has. A dense tutorial or a fast-cut demo has almost none, so the description gets trimmed until a blind viewer hears half of what the picture is showing. Pausing removes that constraint. There is also a case where none of this applies, and it is the one that saves the most work. Where the audio already says everything the picture shows, no description is needed at all, at any level.
Who this rule protects
Blind viewers and people with low vision who cannot see the screen feel this first, and W3C names a third group beside them. People with cognitive disabilities who find it hard to work out what is happening visually.
How to check it yourself
- Watch the video with your eyes closed and note where you lose the thread of what is happening.
- Check whether the audio already says everything the picture shows. If it does, no description is needed at any level and you are finished.
- Where description is needed, time the gaps in the dialogue and ask whether they are long enough to hold it.
- If they are not, check whether an extended version exists. Then check whether a viewer who does not need the pausing can get past it, either with a switch or with the plain version published alongside. That is the pattern W3C describes rather than a condition of passing, and it is worth doing anyway.
Failures we see most often
- Description gets crammed into gaps too short for it, so the narration talks over the dialogue.
- The description gets trimmed until it no longer says what the screen is showing.
- Text on screen goes undescribed, which is the visual detail teams forget most often.
- The extended version is the only version, with no way past the pausing for a viewer who does not need it. That is a usability problem rather than a failure of this rule, and it is the one W3C warns about.
Who this one is for
Read from this rule's own note above, so the grouping and the note cannot disagree.
- Blind and screen reader userspeople who cannot see the screen
- Low visionpeople who can see the screen but not easily
- Cognitive and learningpeople for whom the difficulty is understanding, remembering, or staying with it
How to fix it
- Produce an extended version for any content where the picture carries the teaching.
- Offer it as a switchable track, or publish two versions and link them clearly, which is the route W3C names where the player cannot switch.
- Plan the pauses into the edit at the points where you already know description will be needed.
Step-by-step fix guides (4)
Passes vs. fails
Passes
An extended version pauses on each screen while the narration names exactly what changed.
Fails
A software demo where the narrator says nothing but 'and then this happens' over four rapid screen changes.
In audits and lawsuits
This is Level AAA and the least implemented of the four AAA media criteria, for a reason you can go and check. The only sufficient technique W3C documents runs on SMIL, which is not a live web technology in any mainstream player. The HTML track element is listed as advisory, which means W3C does not treat it as proven on its own. So the criterion matters most in training and educational video, where the picture carries the teaching, and that is also where the tooling is thinnest.