Audio Description (Prerecorded)

What it asks

At Level AA, the “transcript only” option from 1.2.3 disappears. Prerecorded video with audio must have an audio description track — a narrator’s voice inserted into gaps in the original soundtrack to describe what’s happening on screen: scene changes, actions, on-screen text, facial expressions, charts, anything visually material that the dialogue doesn’t already convey.

How to meet it

• Write a description script that fits within natural pauses in the dialogue — don’t talk over the speaker.
• Record the narration with a clear voice distinct from the original speakers, mixed at a level slightly above the background.
• Deliver the description as a second audio track switchable via the player UI (HTML5 supports <track kind="descriptions">; YouTube allows alternate audio tracks).
• For short corporate videos, an alternate “described version” can be uploaded alongside the original instead of a switchable track.
• Describe on-screen text exactly: “Slide reads: Q3 revenue, twelve point four million.”
• Skip the obvious — don’t describe what the dialogue already states.

Common failures

• Marketing videos with rapid visual cuts and a music-only soundtrack, with no described version produced.
• Training videos that rely entirely on screen-recorded UI actions, narrated only with “click here” and “now this.”
• Described version that interrupts the dialogue because the writer didn’t time the gaps.
• Description track present but not labelled in the player, so users don’t know it exists.
• A vendor-provided described version that exists somewhere but is not linked from the page hosting the original video.

Why it matters

Audio description is the most-skipped 1.2.x requirement after transcripts — most organizations don’t budget for it. The fastest path to AA is to plan for description at the storyboard stage: leave deliberate pauses for narration, and the production cost drops by half compared to retrofitting later.