Glossary / Captions and transcripts
Captions and transcripts
Captions are synchronized text for speech and meaningful non-speech audio in video; transcripts are text documents representing media content, and descriptive transcripts also convey important visual information. Captions support Deaf and hard-of-hearing users and people in noisy or quiet settings. Transcripts support people who prefer or need text, searchability, and Deaf-blind access when properly descriptive and compatible with braille. They are distinct from subtitles, which may translate dialogue without identifying speakers or non-speech sounds.
Requirements and distinctions
- WCAG 1.2.2 AA requires captions for prerecorded synchronized media, subject to stated exceptions. WCAG 1.2.4 AA covers live synchronized media captions.
- Prerecorded audio-only content needs a text alternative at Level A; prerecorded video-only content needs an alternative or audio description, depending on the criterion.
- Captions should include speaker identification when unclear, meaningful sound effects, music cues when relevant, and timing that keeps text synchronized and readable.
- Transcripts are not a substitute for synchronized captions where captions are required. Audio description conveys essential visual information not present in the soundtrack.
- Web video commonly uses WebVTT via
<track kind="captions" srclang="en" label="English">; provide a usable player, selectable tracks, and a default/clear way to enable captions. - Machine speech recognition is a draft, not a quality guarantee. Human review must correct names, punctuation, timing, speaker labels, and meaningful sounds.
Production and review
Plan accessibility during recording: identify speakers, avoid talking over critical visual events, and leave pauses for description. Review captions against the actual media, verify sync and line breaks, avoid covering essential visuals, and ensure controls are keyboard and screen-reader operable. A descriptive transcript should include dialogue, significant sound, and visual events essential to understanding. Localize captions and transcripts along with the audio.
Interview questions
- Captions versus subtitles? Captions include relevant non-speech audio and speaker identification; subtitles typically render dialogue and may translate it.
- When is a transcript enough? It can satisfy some audio-only alternatives, but does not replace synchronized captions where WCAG requires them.
- What is audio description? Narration of important visual information that the main audio does not convey.
- What should captions include besides dialogue? Meaningful sounds, speaker changes, and relevant music or tone cues.
- How do you ensure quality at scale? Automate initial transcription, require human review, track language/version metadata, and validate player behavior and timing.