Standards · WCAG 2.2

SC 1.2.7 Level AAA WCAG 2.0

Extended Audio Description (Prerecorded)

When pauses in dialogue aren't long enough to fit a standard audio description, the video must pause to let extended description play — so blind users get the full visual context even in dense, fast-paced content.

What it asks

Standard audio description (1.2.5) only works when there’s space between the original dialogue. For dialogue-heavy or fast-cut content, there isn’t. Extended audio description solves this by pausing the video itself when needed, letting a longer narration play, then resuming. The result is a longer total runtime but a fuller experience for users who can’t see the screen.

How to meet it

  • Use a player that supports programmatic pause/resume around description cues — HTML5 <video> plus a JavaScript controller works.
  • Author a description script with two passes: one for in-gap narration, one for extended pauses.
  • Mark extended pause points in the description track (.vtt with cue settings or a custom JSON manifest).
  • Offer the extended-description version as an opt-in alternative — viewers who don’t need it can play the original.
  • Time pauses to scene boundaries where possible to minimize disruption.
  • Test with screen-reader users — extended description that interrupts mid-thought is worse than none.

Common failures

  • Producing standard audio description that talks over critical dialogue because there isn’t room.
  • Pauses inserted at arbitrary points instead of scene boundaries, breaking comprehension.
  • No way for users to skip back over an extended description if they missed something.
  • Extended description authored but only playable on a custom player that fails on iOS / Android.

Why it matters

AAA-level, and rarely required outside broadcasting and government public-information content. Most teams meet 1.2.5 with a standalone described version and never reach 1.2.7. Worth noting on contracts where the SOW says “WCAG AAA where applicable” — flag it as out of scope unless the content is dialogue-dense.