Core Guidance
Do This First
-
Caption and transcript first. Create accurate captions for video and a transcript for audio or video when users need the full spoken content in text.
-
Describe visual meaning. If key actions or visuals are not explained in the soundtrack, add audio description or provide a clear text alternative.
-
Check playback experience. Make sure controls work with a keyboard, contrast is strong, and media does not cause seizure risk or surprise audio.
Essential Checks
Caption Quality
- Review auto-captions and fix names, technical terms, timing, and punctuation.
- Include relevant non-speech sounds such as applause, music changes, or alerts when they affect meaning.
- Keep captions in sync and readable instead of crowding each frame with text.
Transcripts and Description
- Provide transcripts for podcasts, audio clips, and longer media.
- Add audio description when important visual information is not spoken aloud.
- Identify speakers when there are multiple voices or when off-screen narration matters.
Playback and Visuals
- Provide pause, stop, and volume controls for media that plays longer than three seconds.
- Keep burned-in text large enough, high-contrast, and on screen long enough to read.
- Avoid flashing more than three times per second and reduce unnecessary motion when possible.
Common Problems to Catch
- Publishing video with auto-captions that were never edited.
- Depending on a visual demonstration that has no narration or descriptive equivalent.
- Using tiny lower-third text or low-contrast captions over video.
- Auto-playing audio that users cannot pause or mute quickly.
Key WCAG 2.1 AA Checkpoints
| Success Criterion | What to confirm |
|---|---|
| SC 1.2.2 Captions (Prerecorded) | Prerecorded video with audio needs captions. |
| SC 1.2.3 Audio Description or Media Alternative | At Level A, prerecorded video needs audio description or a full text alternative. |
| SC 1.2.5 Audio Description (Prerecorded) | At WCAG 2.1 AA, give audio description when visuals carry essential information. |
| SC 1.4.3 Contrast (Minimum) | Text overlays and controls need enough contrast to stay readable. |
| SC 2.3.1 Three Flashes or Below Threshold | Avoid flashing content that could trigger seizures. |