Home / Guides
Audio description file formats
In short: there are three ways to deliver audio description. A described video file with the narration mixed in, a separate audio track the viewer can switch to, or a timed text file (usually WebVTT) that a player or screen reader voices.
1. Described video (MP4)
A copy of the video with descriptions mixed into the soundtrack. It plays everywhere and is the only practical way to deliver extended description, because the picture itself is paused to make room. Publish it next to the original and label it clearly.
2. Description audio track
An audio file the same length as the video: the original soundtrack plus descriptions. Platforms with multiple audio tracks, such as YouTube descriptive audio, let viewers switch to it. It only works for standard description, because the video timing cannot change.
3. WebVTT text description track
A WebVTT file whose cues contain the description text, timed to the video. W3C notes descriptions can be provided as a timed text file and read aloud by a media player that supports it (W3C WAI). In HTML5 it is attached with <track kind="descriptions">:
<video controls src="lecture.mp4">
<track kind="captions" src="captions.vtt" srclang="en" label="English">
<track kind="descriptions" src="descriptions.vtt" srclang="en" label="Audio description">
</video>WEBVTT
00:00:12.500 --> 00:00:16.000
A bar chart titled Enrollment shows 2019 at 4,000 and 2024 at 9,500.Browser support for voicing description tracks is limited, so use an accessible player that reads them, or also provide one of the audio formats.
Which one to choose
| Where you publish | Best format |
|---|---|
| YouTube (with descriptive audio access) | Description audio track |
| YouTube, Vimeo, LMS without audio tracks | Described MP4, labeled |
| Your own site or app | WebVTT descriptions plus a described MP4 |
| Lectures with non-stop narration | Extended described MP4 |
What Seen Aloud delivers
Every video comes back with all of them: described-video.mp4, audio-description-track.mp3 and .wav, descriptions.vtt, captions as SRT and WebVTT, the script as DOCX and TXT, and a compliance note. Each file is checked automatically before delivery, for example that cues do not overlap and the audio track matches the video length.