Captions for Short Videos People Watch with the Sound Off

Watch your own feed for a day and count how often you have the sound on. For a lot of people the answer is: rarely, unless something on screen gives them a reason to turn it up. A short video that relies on what is said, with nothing on screen to read, is a silent clip of someone talking to that viewer.

Captions fix most of that, but only if they are readable, placed where the app won't cover them, and timed to the voice. Captions also help people who are deaf or hard of hearing, and anyone watching in a language that isn't their first. Doing them well is not extra polish; it is part of the video.

Two kinds of text on a video

It helps to separate what is shown from what is said:

• On-screen text is the headline of a scene: a few words that say what this part is about — “Step 2: grind finer”. It is there for the whole scene and works even if nobody reads anything else.
• Captions are the words being spoken, appearing as they are said. They carry the detail and the voice.

A video watched on mute should still make sense from the on-screen text alone. Captions then fill in the rest for anyone who keeps watching. If both say exactly the same thing at the same moment, you have twice the text and no extra meaning — let the headline name the point and the caption say it.

Burned-in captions or the platform's own

For short vertical video, burned-in captions are the usual choice because they look right on every app you post to. If you also have a caption file, upload it where the platform allows — YouTube does, for example — so viewers who rely on captions can use the player's own.

Keep words out of the app's way

On a tall 9:16 video, each app lays its own interface over your picture: the account name, the post text and the sound credit across the bottom, a column of buttons down the right side, and sometimes a header at the top. Anything you put there is covered for most viewers.

• Leave the bottom clear. Keep captions above roughly the bottom fifth of the frame, not at the very bottom where a film subtitle would go.
• Leave the right side clear. The like, comment and share buttons sit there; keep words in from that edge.
• Keep the top margin too. Some apps show a header or a search bar over the top of the video.
• Check on a phone. Post a test privately, or put your export in the app's preview, and look at where the buttons land before you publish.

Each app draws its interface a little differently and changes it from time to time, so a margin that is generous on all sides is safer than a layout cut to one app's exact pixels.

Pacing: show words when they are said

• Short pages. One to three words at a time for a punchy style, or up to two short lines for a calmer one. A full paragraph on screen is skipped.
• Break on phrases, not mid-phrase. “Grind finer / for a stronger cup” reads better than “Grind finer for / a stronger cup”.
• Appear just before the word. A caption that turns up a moment early reads as in sync; one that lags behind the voice feels wrong.
• Highlight the word being said if your style does that, so the eye can follow along.
• Leave lines up long enough to read. If you can't read a page out loud in the time it shows, it is too long or too fast.

Make them readable on any background

Your video changes behind the words, so caption text has to read on a bright sky and a dark room alike. White text with a dark outline or shadow does that without hiding the picture; a solid band behind the text works too, at the cost of covering more of the frame. Use a heavy, plain font at a size that's easy to read at arm's length, and avoid thin scripts, all-italic text and low-contrast colour pairs.

Write for silence from the start

1. 1Write each scene's headline firstIf someone only read the headlines, would they get the point? If not, rewrite them before recording anything.
2. 2Write the script to be read as well as heardShort sentences, plain words, no long build-up. It makes better captions and a clearer voiceover.
3. 3Put the hook on screen in the first secondThe opening promise should be readable before anyone decides whether to turn the sound up. The guide to hooks has more on that first second.
4. 4Caption, then watch it mutedPlay the finished video with the sound off, start to end. Anything confusing gets a headline or a better caption.

Doing it without editing software

Postclo makes short videos up to 60 seconds long, at 1080 × 1920 or 1080 × 1080, from your project. Each scene gets a headline on screen and a caption that is spoken by a voiceover made in your browser (four voices to choose from). Captions come in three styles, are timed to the voice word by word and stay inside the frame's safe area, and you can download them as .srt or .vtt files as well as the MP4. You edit every word before you export, and you post the video yourself.