Captions, transcripts & audio description
Sound-off viewing is the default on social, but captions aren't a nicety. They are how Deaf and hard-of-hearing people follow your video at all. Pair them with transcripts and, when the picture tells its own story, audio description.
Captions vs. subtitles vs. transcripts
- Captions convey all meaningful audio, so dialogue plus sound
cues like
[phone rings]
or[tense music]
, for viewers who can't hear it. - Subtitles assume you can hear and just translate the dialogue into another language. They skip sound effects, so they're not a substitute for captions.
- SDH, meaning subtitles for the deaf and hard of hearing, is the term to watch for. It looks like a kind of subtitle and it behaves like captions, because it does carry the sound cues. British and European broadcasters use this wording a lot, and American ones tend to say captions for the same thing. If a brief asks for subtitles, ask which of the two they actually mean before you quote for it.
- Transcripts are the full text of a video or audio clip in one block. They are essential for podcasts and audio-only posts, and handy for anyone who'd rather read.
Open vs. closed captions
Closed captions can be toggled on or off and are styled by the player
(uploaded as an .srt or .vtt file). Open captions
are burned into the video pixels, so they are always visible and never dependent on player support.
On platforms with reliable caption support (YouTube, LinkedIn), prefer closed captions. Where support is patchy or captions are easily missed, open captions guarantee everyone sees them, which is why they are a common choice for Reels, TikTok and Stories.
Auto-captions are a draft, not a deliverable
Automatic captions are a brilliant head start and a terrible finish. They routinely
mangle names, technical terms, punctuation and homophones, so A11y.mv
becomes
a lily M V
, WCAG
becomes w cag
. Always open the editor and fix them.
When you correct captions, also:
- Add punctuation and sentence breaks so the text is readable.
- Label speakers when more than one person talks
(
Irshad:
…Guest:
…). - Note meaningful sound in brackets, so music, laughter or a doorbell.
- Keep timing tight so captions match what's on screen.
Transcripts for audio-only posts
A voice note, a podcast clip, an audio Space. If there's no video, captions have nowhere to live. Post a transcript in the caption, a linked doc, or a thread. It's the only way the content reaches Deaf followers, and it makes your words searchable and quotable for everyone.
Audio description: for what the eyes catch
Captions cover what you can hear. Audio description covers what you can only see, such as an on-screen phone number, a silent gesture, or text that appears without being read aloud. If a blind viewer would miss key information with their eyes closed, you have a gap to fill. Two easy fixes:
- Narrate as you film. Say the thing you're pointing at:
the button's in the top-right corner
instead oftap here
. - Describe it in the caption or add a described version of the clip.
Don't bury text in fast frames
Text that flashes on screen for half a second, whether a punchline, a stat or a website, is unreadable for many and invisible to screen readers. Keep on-screen text on long enough to read comfortably, ensure it has strong contrast against the footage, and repeat the important bits in your captions and post text.
Per-platform notes
- YouTube: upload a caption file or correct the auto-captions in Studio, then add a transcript in the description for long videos.
- LinkedIn / X / Facebook: upload an
.srtfile with the video rather than trusting auto-captions. - TikTok / Reels / Stories: enable captions, then edit them. Open captions are the safest bet for reach.
- Instagram: auto-captions can be toggled and edited before posting, so always review them.
Captions are one habit. See the rest, including alt text, hashtags and emoji, on the accessible social media guide.