Closed Captions vs Subtitles What Your Video Really Needs

Almost every brief that lands on our desk asking for subtitles is really asking for two different things at once, and nobody has noticed. The closed captions vs subtitles distinction sounds like pedantry until a compliance officer rejects a finished film, or a viewer writes in to say they could follow the dialogue but had no idea the phone was ringing. The two formats look identical on screen. They are built for different people, and they are produced differently as a result.

What are closed captions, precisely

Captions were designed for viewers who cannot hear the audio. That means they carry the dialogue and everything else the soundtrack is doing: a door slamming off camera, music swelling under a scene, a speaker's tone turning sarcastic, the identity of whoever is talking when the speaker is out of frame. They are written in the same language as the audio, and a good caption file reads a little like a stage direction because that is essentially what it is.

Subtitles assume the opposite. They assume the viewer can hear perfectly well and simply does not speak the language. So they carry translated dialogue and nothing else, because the sound effects are already doing their job. Adding "door slams" to a subtitle track would be noise.

The practical consequence is that a subtitle file translated from a caption file is usually cluttered, and a caption file adapted from subtitles is usually incomplete. They are not interchangeable source material, however tempting it is to treat them that way.

Open vs closed captions, and why the choice is not cosmetic

The other pairing people confuse is open vs closed captions. Closed means the text sits in a separate track and the viewer can switch it off. Open means it is burned into the picture and cannot be removed by anyone, ever.

Burned in text wins on social platforms, where autoplay runs silently and playback controls are minimal. It loses everywhere else. You cannot search it, you cannot swap it for another language, and you cannot correct a typo without re-rendering and re-uploading the whole file. A closed track is a small structured file you can fix in ten minutes. If a video has any chance of being reused, licensed or localised later, keep the text out of the pixels.

The rules are stricter than most teams expect

Accessibility is where the closed captions vs subtitles question stops being editorial and starts being legal. Public sector bodies across the UK and Europe are already bound by accessibility regulations, and broadcasters have operated under caption quotas for years. The W3C Web Accessibility Initiative guidance on captions is the reference most auditors reach for, and it is unambiguous that captions must convey non-speech audio, not just dialogue. Automatic transcription almost never does this, which is the single most common reason an in-house caption file fails review.

For background on how the format evolved and why the terminology diverged between North America and Europe, the entry on closed captioning is a clear summary. It explains, among other things, why an American client and a British client can use the word subtitles to mean two different things in the same meeting.

Captions and subtitles solve different problems, and confusing them costs reach with the audiences who most depend on getting it right. Accessibility work like this is one of the underrated levers in the creator economy, where a large share of viewing happens with the sound off. The technical distinction turns out to be a distribution decision.

Where automatic tools genuinely help, and where they stop

Speech recognition has improved enough that a rough English transcript of clean studio audio is a reasonable starting point. It saves real time. What it cannot do reliably is handle overlapping speakers, accented speech, brand names, technical vocabulary or anything recorded in a room with hard surfaces. It also has no concept of reading speed, so it will happily produce a caption that flashes past in under a second.

Machine translation applied on top of that transcript compounds the problem, because subtitle lines are short, context poor and full of idiom. The examples collected in this piece on why Google Translate is not enough for some languages map almost exactly onto the failures we see in raw machine subtitle output. The errors are rarely obvious to someone who does not speak the target language, which is what makes them expensive.

Reading speed is the craft nobody sees

The part of subtitling that looks trivial and is not is timing. A subtitle has to appear when the line starts, disappear before the shot changes, break at a natural grammatical point, and stay on screen long enough to be read at roughly fifteen to seventeen characters per second. Meeting all four constraints at once usually means condensing the dialogue rather than transcribing it, and knowing what to cut without losing meaning is a translation skill in its own right.

This is also where cheap work reveals itself. Lines that run across cuts, three line stacks that cover faces, and subtitles that vanish mid sentence all signal a file produced without a human watching the picture.

Deciding what your video actually needs

Start with the audience. If some of your viewers are deaf or hard of hearing, or if the content is public facing and subject to accessibility rules, you need captions in the original language. If your viewers speak another language, you need subtitles, and proper subtitle translation services rather than a machine pass. Most serious video projects need both, delivered as separate closed tracks and stored as clean source files.

Budget for that from the start rather than treating it as a post production afterthought. Retrofitting captions onto a finished catalogue costs several times what it would have cost to produce them alongside the edit, and the deadline that forces the decision is almost never one you set yourself.