Closed Captions vs Subtitles vs SDH: The Difference

Quick answer — Subtitles translate dialogue for viewers who can hear the audio. Closed captions describe all significant sound for viewers who cannot, including speaker changes and effects. SDH is subtitle-styled captioning delivered as a subtitle file.

Vitra.ai Universe handles the dub, the captions and the artwork around them.

Three things, routinely confused

The distinction is not stylistic. A subtitle track delivered where captions were required is an accessibility failure, and on most platforms it is a compliance failure as well.

What captions carry that subtitles do not

Non-speech information, and it is most of the difference.

Speaker identification when the speaker is off screen. Sound effects that carry meaning — a door closing, a phone ringing, a gunshot. Music cues, including whether lyrics matter. Tone indicators where sarcasm or a whisper changes the sense.

A caption track that lists only dialogue is not a caption track. It is a transcript with timings, and a deaf viewer watching a thriller with it learns nothing about the footsteps behind the door.

Why SDH exists

Closed captions in the broadcast sense are carried in the signal itself, in formats built for that purpose. Streaming delivery generally uses subtitle files instead. SDH is the bridge: caption content, subtitle delivery. It is what most platforms actually want when they ask for accessible text, and it is why a delivery spec can say "subtitles" and mean something with sound effects in it.

Read the spec rather than the word.

Doing both without doing it twice

A translated programme usually needs both — a translation subtitle track per language, and an accessible track at least in the original.

The efficient order is to build the caption track first, because it is the superset. Stripping non-speech elements out to make a clean subtitle track is mechanical; adding them afterwards means watching the programme again.

Video dubbing generates the caption and subtitle tracks alongside the audio from the same script, and a check over the output verifies reading speed and overlap, which are the two things a manual pass consistently misses.

Where a series runs long, keep the sound-effect vocabulary consistent with a translation memory. The same doorbell should not be three different words across a season.

Which container these ship in is a separate decision, covered in subtitle file formats.

FAQ

What is the difference between subtitles and closed captions? Subtitles assume the viewer can hear and translate dialogue only. Closed captions assume they cannot and include speaker identification, sound effects, music cues and tone, because those carry meaning a deaf viewer would otherwise lose.

What does SDH mean? Subtitles for the Deaf and Hard of Hearing: caption content delivered as a subtitle file. It exists because streaming platforms deliver text as subtitle files rather than in the broadcast caption signal.

Is delivering subtitles instead of captions a problem? Yes. It is an accessibility failure and on most platforms a compliance failure, because the track omits the non-speech information that captions exist to provide.

Should captions or subtitles be produced first? Captions, because they are the superset. Removing non-speech elements to produce a clean subtitle track is mechanical, whereas adding them later means reviewing the programme again.