Open vs closed captions: which should you use?

The distinction is simple: closed captions can be turned off, open captions cannot. What follows from it is less obvious, and it decides how you export every video you make.

Closed captions

A separate text track — an SRT or VTT file — delivered alongside the video. The player draws the text; the video file itself contains none of it.

Advantages

• The viewer chooses. Someone who does not want captions can hide them.
• The text is machine-readable, so platforms can index it, translate it, and feed it to screen readers.
• You can attach several languages to the same video.
• Fixing a typo means re-uploading a small text file, not re-rendering the video.
• Quality is untouched — no re-encode.

Disadvantages

• Off by default nearly everywhere. Most viewers never turn them on.
• Styling is the player’s, not yours.
• Not supported at all on some platforms.

Open captions

Burned into the picture during export. Every frame contains the text as pixels.

Advantages

• Always visible. No setting to find, nothing to enable.
• You control the look completely — font, colour, animation, position, per-word highlighting.
• They work everywhere a video works, including platforms with no caption support at all.
• They survive re-uploads, downloads and reposts.

Disadvantages

• Cannot be turned off, ever.
• Not machine-readable — to a platform they are just pixels.
• Fixing a typo means re-exporting.
• One language per file.
• They can be covered by platform interface elements if positioned carelessly.

Which should you use, and where?

TikTok, Reels, Shorts → open captions. These platforms are watched muted, at speed, mostly in public. A viewer scrolling past will not tap a CC button they never noticed. If the words are not already on screen, they are not being read. None of these platforms accepts a subtitle file for a normal post anyway, so the decision is made for you.

Long-form YouTube → closed captions, usually. YouTube shows them properly, indexes the text, and can auto-translate it. Burn them in only if you specifically want the styled look.

YouTube Shorts → both. This is the one case where doing two things is clearly right. Burn captions in so the muted scroller reads them, and upload an SRT so YouTube can index and translate. They come from the same transcript, so it costs one extra export.

Your own website → closed captions. Use VTT with the HTML5 <track> element. Accessible, indexable, and the viewer decides.

LinkedIn, X, Facebook → open captions. Caption file support is inconsistent and autoplay is muted. Burn them in.

Anything with an accessibility obligation → closed captions. Screen readers and assistive technology need real text. Burned-in pixels are invisible to them, and open captions alone will not satisfy a compliance requirement.

The trap

Open captions are permanent. If you burn in a typo, or a name spelled wrong, the only fix is exporting again — and if the video is already posted and gathering views, you are choosing between leaving the mistake up and losing the engagement.

So: read the transcript before you export. Automatic transcription is strong on ordinary speech and weak on exactly the words that matter most — names, brands, jargon, acronyms. Those are also the ones your audience will notice.

Keep the source project rather than only the exported file. Re-exporting from a project takes a minute; rebuilding captions from scratch does not.

Where should captions sit on screen?

Open captions sit in the frame permanently, so they have to work with the platform’s interface rather than fight it. TikTok covers the bottom of the screen with the caption text and CTA, plus a column of buttons down the right. Instagram and YouTube cover different regions.

Text that looked perfectly centred while editing can end up half-hidden behind a username once posted. Use a safe-zone overlay that shows those regions before you export, rather than discovering the problem on the live post.