Audio description services compared, from the script to the player

What WCAG asks of a video, and what it does not

The criterion is short, and the definition behind it does most of the work. WCAG defines audio description as narration that describes the visual details a listener cannot get from the main soundtrack, and it adds that where the soundtrack already carries all of the video's information, no description is needed. A founder talking to camera with nothing on screen but a face usually needs none. A product demo where the screen does the talking needs it.

The levels below are why a transcript alone stops being enough at Level AA, the level both the Title II rule and the Section 508 Standards name. Our guide to what WCAG is explains the levels, and video caption requirements covers the captions and transcripts that sit beside description.

The DOJ's Title II web rule for state and local governments names WCAG 2.1 Level AA, which includes 1.2.5; the dates are in our Title II web rule guide. This is general information, not legal advice. For your own case, ask a lawyer.

Standard or extended: the pauses decide

Standard description lives in the gaps between speech. Extended description stops the picture, plays a longer description, then carries on. Which one a video needs is mostly a property of its edit, not of the vendor: a fast-cut ad with music under a voice-over has almost no room, a slow walkthrough has plenty.

Extended description is a Level AAA criterion, so it is never required for AA. It is still the honest answer when a video cannot be followed without it, and Video Caption Corporation's page calls it the preferred choice for online video. The catch is delivery: a longer described version cannot share one timeline with the original, so it needs a player that pauses or a second copy of the video.

Order the format your player can play

Most of the rows in the table are delivery rows, because that is where orders go wrong. A description track is useless if the player on your site cannot offer it. W3C's own guide to description sets out the order to decide in, and the picker below follows it.

If you host video on your own pages, Able Player, an accessible HTML5 media player whose code is public on GitHub, can read a WebVTT description file aloud and pause the video while it does. It can also swap to a described version, including one hosted on YouTube or Vimeo. Two vendors in this table bring their own: Verbit's Smart Player and the 3Play Plugin.

Human voice, synthetic voice, and who checks the script

WCAG asks for narration and says nothing about whose voice reads it. 3Play Media, which sells both, argues that many blind viewers are used to synthetic speech from their screen readers. A human narrator still earns its fee on a brand film, where tone is part of the message; Descriptive Video Works picks a narrator to match the genre of each video.

The script matters more than the voice. A describer who has never used your product cannot know that the button labelled “Sync” is the point of the demo. Only one product here says plainly that you edit the script before export, and it is an AI tool. For the four services, make a script review a line in the order.

Before you sign off, listen to the described version with your eyes closed, then try it with a screen reader as in our screen reader testing guide.

Spend less: plan description into the next video

W3C notes that for some videos, training videos in particular, the presenter can describe what is on screen as the video is made, and then there is no separate description to buy. “Click Export, the orange button at the top right” is a described step; “click here” is not. Buy description for the videos you cannot reshoot, and write it into the script of the next one.

The basics of how description is written and voiced are in what is audio description. If the same vendor will caption the video too, compare them in captioning services compared. Video is one line in our ADA website checklist; tick it off there with the rest of the site.