You’ve seen both words used for the same white text at the bottom of a video, sometimes in the same app. YouTube has a CC button that opens a menu labelled subtitles. A tutorial tells you to add subtitles to a TikTok, and TikTok’s own switch calls them captions. If you’re trying to work out which one you need, or which word to search for, the mix-up is fair. The two terms do mean different things, but only in some countries, and short-form video has mostly stopped caring.
The short answer. Captions are written for viewers who can’t hear the audio, so they include sounds and speaker names as well as speech. Subtitles are written for viewers who can hear but don’t understand the language, so they translate the dialogue and nothing else. In the UK, “subtitles” usually covers both. On social video the two words are used interchangeably for on-screen text of what’s being said.
The strict difference
The distinction comes down to who the text is for.
Captions assume the viewer can’t hear
Captions transcribe everything in the audio that matters, not only the dialogue. That means speech, but also lines like [door slams], [phone buzzing] or [upbeat music], and labels that say who’s talking when the picture doesn’t make it obvious. The captions are in the same language as the audio. A deaf viewer watching an English film with English captions should get the same information as someone listening.
Subtitles assume the viewer can’t follow the language
Subtitles translate the dialogue for a viewer who can hear the soundtrack. They hear the door slam and the music swell, so subtitles leave sounds out. They usually leave out speaker labels too, because the viewer can tell whose voice it is. A Korean film shown with English subtitles is the classic case.
One more term is worth knowing. SDH, short for subtitles for the deaf and hard of hearing, is what streaming services and disc releases call same-language text that includes sounds and speaker names. It does the job of captions in the format of subtitles. If you’ve seen “English (SDH)” in a language menu, that’s what it is.
It depends where you are
The strict split above is mostly an American one. In the US, closed captions became the standard name for accessibility text on TV, and subtitles kept their meaning of translation.
British English never made that split. The BBC calls its accessibility text subtitles, and so do most UK broadcasters. A British viewer who switches on subtitles for an English programme gets what an American would call captions, sound cues included.
So when someone asks for subtitles, they might mean a translation, or they might mean exactly what an American calls captions. Context usually settles it. If the request comes from a client or an accessibility team, ask whether they need sound descriptions and speaker labels. That’s the part that actually changes the work.
Open, closed and the CC button
A second pair of terms describes where the text lives, not who it’s for.
Closed captions can be switched on and off. They travel as a separate track or file alongside the video, and the player draws them on top. That’s what the CC button controls.
Open captions are part of the picture. They’re rendered into the frames, so every viewer sees them and nobody can turn them off. On social video this is usually called burning them in. The guide to burn subtitles into a video on iPhone compares the two side by side, including where each one is accepted.
The two pairs are independent of each other. You can have closed subtitles, like a translation track you pick from a menu, or open captions, like same-language text with sound cues burned into a film. Common file formats don’t take sides either. An SRT file holds numbered lines of text with start and end times, and nothing in it says whether those lines are captions or subtitles.
The web standard does draw the line. HTML’s <track> element, which attaches timed text to a video on a web page, has a kind attribute where subtitles and captions are separate values. The spec describes subtitles as suited to when the sound is available but not understood, and captions as including sound effects and musical cues for when the sound isn’t available. That’s the American split, written into the browser.
How social video blurs the line
Short-form creators mostly use captions in a sense that fits neither strict definition. The text is a transcript of the speech, in the language spoken. It’s usually burned in, so technically it’s open captions. But it rarely includes sound cues or speaker names, and the people it’s mainly for are viewers scrolling with the sound off, not only deaf viewers.
The platforms add their own labels on top. At the time of writing, TikTok and Instagram both call their automatic speech-to-text feature captions, and on Instagram the word caption also means the text under a post. YouTube’s player groups everything under a Subtitles/CC menu. None of these follow broadcast rules, and they change between app versions.
For your own videos, the word you use matters less than what goes in the text. A viewer on mute needs the speech. A deaf or hard of hearing viewer may also need the sounds that carry meaning. When you compare tools, an auto subtitle app for iPhone and an auto caption app are almost always the same thing sold under different names. Capster is one of them: it transcribes the speech in the language being spoken and burns the text into the exported file, which makes the result open captions in the strict sense.
Which one your video needs
Decide by who’s watching first, then by where the video is going.
| Situation | What goes in the text | Where it lives |
|---|---|---|
| Short vertical video for a muted feed | The speech, in the language spoken | Burned in |
| A clip where a sound is part of the point | The speech plus that sound, in brackets | Burned in |
| Viewers who speak another language | A translation of the speech | Burned into a separate version |
| YouTube long-form, Vimeo or your own site | Speech plus sound cues | A file viewers can switch on |
| A client or broadcaster with accessibility rules | Captions or SDH to their spec | Usually a separate file, so ask |
A few habits help whatever you call the text:
- Add a cue when the sound is the joke. A clip whose punchline is a crash or a burst of laughter makes no sense on mute without [crash] or [laughter] on screen.
- Name speakers when the picture doesn’t. In an interview shot from one side, a short name label or a different colour per speaker saves the viewer guessing.
- Keep one language per version. If you want a translation as well, export it as its own video rather than stacking two languages on a narrow vertical frame.
- Keep chunks short. Sound cues take room too, so the advice on how many words each caption should hold applies to them as much as to speech.
Sound cues are the one thing automatic captioning usually leaves out, because speech-to-text is built to write down words. In Capster you can tap any caption line and type one in before you export.