Illustration for: Accessible video: captions, transcripts, audio description
Home › Guides › Accessible video: captions, transcripts,…
Guide

Accessible video: captions, transcripts, audio description

A practical guide to the three things that make a video usable by more people, with the standards that apply and a checklist to run before delivery.

In short: Accessible video needs accurate captions, a transcript and, for video with important visuals, audio description. WCAG 2.2 requires captions at Level A and audio description at Level AA for prerecorded video. Ofcom sets access targets for broadcasters and streamers, and they pass down to producers as delivery requirements. Plan all three at script stage.

Accessible video means four things: accurate captions, a transcript, audio description where the pictures carry meaning, and a player people can operate. The cheapest time to plan them is before you shoot, because description written into the script costs far less than description added afterwards.

What each one is

ItemWhat it isWho it helpsWCAG 2.2 level
CaptionsTimed text of speech and the non-speech sounds needed to follow the contentDeaf and hard-of-hearing viewers, and anyone watching with the sound offPrerecorded: A (1.2.2). Live: AA (1.2.4)
TranscriptA text version of the speech and sounds, on the page or linked beside the videoDeaf viewers, people who prefer to read, search enginesAudio-only: A (1.2.1). With video: AAA (1.2.8)
Audio descriptionSpoken description of the visual information in a videoBlind and partially sighted viewersPrerecorded: AA (1.2.5)
Descriptive transcriptA transcript that also describes the visualsViewers who are both deaf and blindCan meet the Level A alternative (1.2.3)

WCAG is the technical standard. Whether it is a legal requirement for you depends on who you are and who commissions you, so ask your client's compliance lead. The levels above tell you what a conformance claim needs.

Captions

Accuracy comes first. A missing "not" reverses a sentence. Automatic captions are a draft. Correct names, brands, numbers and jargon by hand, and watch punctuation, because it carries sense.

Include the sounds that matter. Identify speakers when they are off screen or when it is not obvious who is talking. Add sounds that change the meaning, such as a phone ringing, laughter or a door closing. The W3C calls choosing which sounds to include "more art than science", so ask whether a deaf viewer would miss something without it.

Keep captions in time with the speech. The W3C describes captions as synchronised with the audio, so check timing again after any re-edit.

Choose open or closed. Closed captions can be switched on or off by the viewer. Open captions are part of the picture and cannot be turned off. Use closed captions on your own site and where the platform supports them. Use open captions for social video that autoplays on mute. Platform support for separate caption files differs, so check before you plan around one.

Know the file format. WebVTT is the most common on the web. SRT and TTML are also used. Ask the platform which it prefers before you export.

Live captions. For live events, WCAG expects captions at Level AA. These are usually made by professional real-time captioners. If you post a live event afterwards, the recording usually needs a light edit for accuracy.

Transcripts

A basic transcript covers speech and important sounds. A descriptive transcript adds the visual information, such as on-screen text, speaker names and what is shown. For video, a descriptive transcript can replace the separate basic one.

For a podcast, a transcript with speaker names in plain paragraphs is enough. Put it on the same page as the media, under the player, or link to it right beside the player. If the video is hosted elsewhere, give the transcript its own page and make it easy to find from the video.

Transcripts made straight from captions usually lack the visual details, because captions assume you can see the picture. Add the on-screen text and any action the speech does not explain.

Audio description

Audio description is spoken or written description of what the picture shows. It is not needed for audio-only content such as a podcast. For video, WCAG 2.2 requires it at Level AA for prerecorded content.

There are three ways to deliver it:

  • Integrated. The speakers describe the visuals as part of their script. This works well for explainers and training films.
  • Alternative version. A second version of the video carries a description track, usually voiced by a different person.
  • Separate file. A timed text file or synced audio track is supplied, which the player must support.

Plan it early. Integrated description has to be written into the script before filming, so a presenter can say "this chart shows" instead of "as you can see here". Read on-screen titles aloud or describe them. Leave short pauses where description may be needed. Where the pauses are too short, WCAG's AAA criterion calls for extended description, but the W3C says pausing the video to fit description in is not suggested in most cases.

A talking head with no visuals that matter may not need separate description. A video of a speaker pointing at slides does. For help with how it is made, see what audio description is.

What Ofcom requires, and where producers come in

Ofcom's rules bind broadcasters and platforms, not independent producers directly. They still shape what you are asked to deliver.

Broadcast television. Ofcom's list of channels required to provide access services in 2027, published 14 July 2026, includes 82 domestic channels. The subtitling target is 80% for most channels, 90% for Channel 3 services and Channel 4, and 100% for BBC channels. The statutory audio description target is 10% and the signing target is 5%, both applying from a channel's tenth anniversary. Channels with smaller audience shares are exempt from the signing target and have a minimum sign-presented amount or a contribution instead. The Code on Television Access Services, last updated on 15 July 2024, sets the underlying framework.

Streaming. Ofcom published a draft Tier 1 Accessibility Code on 14 May 2026. It proposes that services with more than 500,000 UK users provide at least 80% of their catalogue with subtitles, 10% with audio description and 5% with signing. Poor-quality access features would not count toward the targets. The consultation closed on 7 August 2026, and Ofcom said final codes are expected later in 2026. Check Ofcom's accessibility pages for the final version before you rely on any number here. The detail for producers is in Ofcom's streaming access code explained.

What this means in practice. A commissioner with a target to meet will put access services in the delivery schedule. As one example, the BFI NETWORK short film fund requires open captions and audio description to be delivered. Ask for the delivery spec before the edit, so the files and the budget match.

Build it into production

  1. Script. Write description into the narration where you can. Keep names and technical terms consistent so captions stay accurate.
  2. Shoot. Record clean speech. Poor sound produces poor automatic captions. Frame on-screen text so it stays legible and give it time to be read.
  3. Edit. Allow time for a caption pass. Avoid content that flashes more than three times in any one second above WCAG's flash thresholds, which is a Level A failure (2.3.1).
  4. Playback. Do not autoplay audio for more than three seconds without a way to stop it (1.4.2). Give viewers a way to pause, stop or hide moving content that runs on its own for more than five seconds (2.2.2).
  5. Delivery. Supply the caption file, the transcript and any description track with the master.

Delivery checklist

  • Captions are reviewed by a person and cover speech, speaker changes and meaningful sounds.
  • Caption timing matches the audio on the final edit, not an earlier cut.
  • A transcript sits on the page or beside the player.
  • Visual information is described in the audio, in a description track or in a descriptive transcript.
  • No flashing content above the WCAG threshold.
  • The player supports captions, keyboard control and any description track you are supplying.
  • File names, formats and languages match the commissioner's delivery spec.

Where it fits in a shoot

If you need a description track for a finished video, Granary's audio description service explains how it is made and who is required to provide it.

Sources

Rather we made it for you? A Content Studio day plans, shoots and edits a batch of content in one visit. Content Studio day →

The picture at the top of this page is an illustration, not a photograph of a client or a real job.

Questions we get asked

What is the difference between captions, subtitles and a transcript?

Captions and subtitles are timed text on the picture. Captions include important non-speech sounds as well as speech, and subtitles in UK usage often mean the same thing for deaf viewers. A transcript is a separate text version of the speech and sounds, which a descriptive transcript extends with descriptions of the visuals.

Does my video need audio description?

WCAG 2.2 requires audio description for prerecorded video at Level AA. It is not needed for audio-only content such as a podcast, and it may add little where the speaker already says everything that matters on screen. Broadcast and streaming commissioners often ask for it, so check your delivery spec.

Are automatic captions good enough?

Not on their own. The W3C says automatic captions do not meet accessibility requirements unless confirmed fully accurate, and that they usually need significant editing. Use them as a first draft, then have a person correct names, jargon, punctuation and any word that could reverse the meaning.

10% OFF YOUR FIRST HIRE

From the Barn

What shot here this month, new kit, masterclass dates and availability. Join the list and get 10% off your first studio hire.

You're on the list. Mention it when you book and we'll take 10% off your first hire.

No spam, roughly one email a month, unsubscribe whenever.

Get a quote