How styled captionswork on YouTube

Most people assume captions on YouTube are white text on a grey box, and that anything fancier has to be burned into the video. Neither is true. YouTube's own caption format carries colours, fonts, sizes, positions, outlines and per-word timing — the styling lives in the caption track, and the video file is never touched.

Two kinds of captions

Burned in

The text is rendered into the pixels of the video in an editor. It looks identical everywhere, including on platforms that have no captions at all — and it is permanent. One language, no way to switch it off, and a typo means exporting and uploading the video again.

A caption track

A separate file attached to the video. Viewers turn it on and off, pick a language, and see it in the player's own caption menu. Because it is text, YouTube can read it; because it is separate, you can replace it in seconds.

The two are not exclusive: channels often burn in short-form captions and publish caption tracks on long-form videos, where the audience is more international and more likely to be watching with captions on.

What a caption track can actually do

YouTube's timed-text format — the one that gets saved as a .ytt file, called format 3 or SRV3 inside YouTube — supports far more than SRT:

• Colour and opacity for the text and for the background box, independently.
• Seven fonts, including monospace, serif and a small-caps face — a fixed list, not your own font files.
• Size, relative to YouTube's default caption size.
• Bold, italic, underline.
• Edges — hard shadow, bevel, glow or soft shadow, in a colour you choose. A dark glow is how outlines are made.
• Position and alignment, anywhere on the frame rather than only at the bottom.
• Per-word timing, which is what makes word-by-word captions possible.

Everything else — animation, custom fonts, images, gradients — is not in the format. Effects like zoom or roll-up exist only as many caption entries written one after another.

The catch: the viewer has the last word

Caption appearance is an accessibility setting, so YouTube lets viewers override it. Anyone who has set a font, size, colour or background in the player's caption options sees their own settings instead of your styling. Rendering also differs between the web player, the mobile apps and TV devices — the same track can look slightly different, and some effects are simply not played back everywhere.

This is the honest trade: burned-in captions are pixel-identical for everyone and unusable by anyone who needs something different; styled caption tracks look right for most viewers and respect the ones who need their own settings.

Why styling a caption track is awkward without a tool

The format is XML, and it is unforgiving in small ways: opacity stops at 254 rather than 255, the size value moves four times faster than the size you mean, positions are squeezed into a safe area, a line that starts at 0 ms is ignored by some Android players, and multi-styled lines need invisible padding characters to keep their styling on mobile. Hand-writing that is possible; maintaining it across twelve languages is not. We documented the details in the .ytt format reference.

A practical setup

1. Start from a correct transcript. Styling a bad transcript just makes the mistakes bigger.
2. Pick one style and stick to it across the channel: weight, outline, one accent colour.
3. Keep lines short — two lines at most, broken where the speaker pauses.
4. Stay clear of the lower edge, where the player controls appear, and of faces.
5. Publish the styled track per language rather than relying on YouTube's on-the-fly translation, which drops styling entirely.
6. Watch it back on a phone before you call it done.

Questions

Can YouTube captions be coloured?

Yes. The caption format supports text colour, background colour, opacity for both, and an edge colour for outlines and shadows. SRT files carry none of this — the styling has to be in a timed-text track.

Do styled captions work on phones and TVs?

They play back in the web player and in the mobile apps. Support varies on TV devices, and viewers who changed their caption settings see those instead. Test on a phone, since that is where most of the audience is.

Will styled captions replace my existing subtitles?

A styled track is uploaded like any other caption track for a language, and replaces the track for that language. The video file and every other language stay as they are.