What you get
- An audio-described version (.mp4) of the video, with the descriptions voiced in the gaps between speech. Post it next to the original.
- A description track (.vtt) for players that read descriptions aloud.
- A description script (.txt), the text of every description with its time.
- A [descriptive transcript](/features/transcripts): what's said and what's on screen, in order, for people who read rather than listen.
How the descriptions are written
Ferndean finds the pauses in speech and the scene changes, looks at frames from each moment, and writes a short description of what can be seen and matters: who appears, what happens, and text on screen like titles, slides and scores. It doesn't repeat what the speakers say, and it doesn't guess who a person is. Each line is sized to fit its pause, so the description never talks over anyone.
Most school and health center video needs less description than people expect. A talking-head message where everything is said aloud needs none. A game where the score is only on screen, a slideshow with music, or a meeting with a vote by show of hands does. The two-question checker tells you for one video, and the guide to when audio description is required covers the rule.