Multimedia content published on UCLA websites should be accessible to everyone, including people with disabilities. These guidelines outline accessibility best practices for videos, audio recordings, and embedded media to help meet ADA and WCAG 2.1 AA standards.
1. Video Captions
Any video uploaded MUST be accompanied by a caption file (.srt or .vtt)
Do:
- Use a tool like YouTube, Otter.ai, or ChatGPT to generate a first draft of your captions.
- Review the draft before publishing. At minimum, check that spelling is accurate, especially names and technical terms.
- Include descriptions of meaningful sounds where possible, e.g., [audience applauds] or [music fades out].
- Always play the video with captions on to confirm they’re in sync with the audio.
Don't:
- Upload Zoom’s exported .vtt file as-is. It’s unreviewed and Zoom isn't accurate enough.
- Write captions from scratch by hand. Start from a generated draft and edit that instead.
- Publish auto-generated captions without a human review.
- Cram too much text on screen. Keep each caption to two short lines.
2. Audio Transcription
Any audio uploaded MUST be accompanied by a transcript file (.txt)
Do:
- Tools like YouTube, Otter.ai, or ChatGPT can also generate a transcript at the same time.
- Label speakers consistently, for example “Jane:” or “Interviewer:”.
- Describe relevant visual or audio context that a reader would otherwise miss.
Don't:
- Never ship raw auto-generated output. Always have a person review it, especially names, jargon, acronyms, and numbers.
- Leave speakers unlabeled or label them inconsistently, e.g., switching between “Jane,” “JD,” and “Speaker 2.”
- Guess at words you can’t make out. Mark them as [inaudible] or [unclear], ideally with a timestamp.
- Paste the text as one big block. Break it into paragraphs at speaker changes and topic shifts.
- Leave caption formatting in the .txt file, such as cue numbers or timestamp lines copied from an .srt.
Published: Monday, October 26, 2026
