Tutorials Accessibility

Audio Description for Video: What It Is and How to Write It

Intermediate · ~30 min

Overview

Audio description (AD) is narrated commentary, placed in the pauses between dialogue, that describes the visual content a blind or low-vision viewer can't see, actions, settings, on-screen text, who just entered the room. It's the least familiar of the core accessibility features to most creators, and increasingly a formal requirement for published video in several jurisdictions. This guide covers the craft: deciding what to describe, writing to fit gaps, and recording a track that sits properly in the mix.

What You Need

Your finished video with a locked edit (AD is written against final timing)
A quiet recording setup and a clear, neutral narration voice
A timeline pass to map the available dialogue gaps

Steps

Understand what audio description is

AD narrates the visuals that matter, not a transcript of the audio (that's a caption's job) but a voice describing what's happening on screen for someone who can't see it. Watch anything with AD switched on for ten minutes before writing your first script. The conventions teach themselves fast.

Decide what to describe and what to skip

Describe what the story needs: actions, scene changes, on-screen text, who's present, meaningful expressions. Skip what the soundtrack already conveys and decoration that doesn't affect understanding. The discipline is prioritization, gaps are short, and describing everything means describing nothing well.

Write to fit the gaps

Map your dialogue gaps first, then write descriptions to fit them, present tense, concrete verbs, no interpretation ("she frowns" not "she seems annoyed"). Read each line aloud against the gap with a stopwatch. A description that overruns into dialogue is worse than a shorter one that breathes.

Record and mix the AD track

Record with the same care as any narration, consistent mic position and a neutral, even read that doesn't perform over the content. In the mix, the description sits clearly above the program audio during gaps. Gentle ducking of the program under AD lines keeps both intelligible.

Use extended AD when gaps don't exist

Fast-cut or wall-to-wall-dialogue content sometimes has no usable gaps. Extended AD pauses the video briefly to make room for necessary description, heavier-handed, but the accepted answer when the alternative is a blind viewer losing the plot entirely.

Deliver AD as a track or a mix

Two delivery shapes: a separate audio track the player can switch to (preferred where the platform supports multiple audio tracks), or a fully mixed "described version" published alongside the original. Check what your distribution platforms actually support and deliver accordingly.

Pro Tips

Name characters once identified ("Maya pours the coffee," not "the woman pours the coffee" forever). It's the single most common beginner AD mistake.
Write the AD script into your transcript document with timestamps: one master document then feeds captions, transcript, and AD.
Have someone listen to the described version without watching, if they can follow the whole piece, the AD works.

Knowledge Base

AD Is Editorial Work, Not Just Narration

The recording is the easy half. The real craft is deciding, within a four-second gap, which of the six visible things on screen the story actually needs. The same prioritization an editor does when cutting. That's why good AD writing tends to come from people who understand story structure, and why fully automatic description still needs a human editorial pass.

The Regulatory Context, In General Terms

Audio description has moved from broadcast-only requirements toward broader digital-service rules in several jurisdictions. Most prominently the EU's Accessibility Act, applying since June 2025, with library deadlines extending to 2030 for some services. Scope and specifics vary by country and by what kind of service you operate. Verify what applies to you, and treat this page as craft guidance, not legal advice.

Where This Fits

This guide covers one specific part of captions and access. The wider picture, caption formats, reading speed, speaker identification, what automatic captioning still gets wrong, and a practical QA pass, is in Beyond Auto-Captions: Caption Quality, Styling, and Readability, which frames the discipline as a whole and links out to the detailed guides underneath it, including this one. If you are starting from scratch rather than solving a specific problem, read that first and come back here.

FAQ

Q: Does every video legally need audio description?
A: It depends on where and how you publish, requirements vary by jurisdiction and service type. In the EU, the Accessibility Act (in force since June 2025) requires AD for in-scope audiovisual services. Other regions have their own broadcast and streaming rules. Verify what applies to your specific situation. This is general guidance, not legal advice. Practically: even where it isn't required, AD opens your content to an audience that otherwise can't use it.

Q: Can AI generate usable audio description?
A: As a drafting aid, increasingly. AI can propose descriptions and a synthetic voice can read them. Its honest limits are visual judgment (knowing which detail matters to the story) and timing discipline (fitting gaps without stepping on dialogue). A human review pass on what gets described, and when, is still what separates usable AD from noise.