Captions and Subtitles for Short-Form Video: A Complete Guide

Most short-form video is watched on mute. The exact percentage depends on the study, but it’s never small. TikTok, Reels, Shorts—the default state is muted autoplay.

No captions means no message. You’ve lost viewers before they hear a word.

Captions vs. Subtitles

The terms get swapped constantly, but they mean different things.

Captions include dialogue, speaker identification, sound effects, and music cues. They’re designed for viewers who can’t hear the audio. Closed captions toggle on/off. Open captions are burned into the video permanently.

Subtitles show only spoken dialogue, usually for translation or clarity.

For short-form, the distinction barely matters. Most creators burn open captions directly into the video—always visible, always styled, no toggle. That’s what this guide covers.

Why They Matter

The first 1-3 seconds decide if someone keeps watching. Text on screen gives immediate context. A video that starts mid-sentence with no text? Skippable. Visible words give viewers a story to follow even before they tap for sound.

Captions also keep people watching longer. Reading along reinforces the message—especially for tutorials and explainers. And they reach people who wouldn’t otherwise watch: non-native speakers, viewers in loud environments, people with hearing loss.

Platform Specs

TikTok

Auto-captions in the native editor. Accuracy is hit-or-miss, especially with accents, slang, or fast speech. Captions appear below center, styled automatically.

Most creators skip TikTok’s native captions entirely and burn styled text into the video with third-party tools before uploading.

Instagram Reels

Caption sticker with auto-generation. Appears at bottom of frame, position adjustable. Only available for some accounts and requires manual enable per reel.

YouTube Shorts

Auto-captions via speech recognition, displayed as standard CC. Accuracy hovers around 80-90%. Edit them in YouTube Studio before publishing—the auto-generated version will get proper nouns wrong.

For all three platforms: burn open captions into your video before uploading. Native captions can be toggled off. Burned-in text is always there.

Technical Specs for Burned-In Captions

Font

Sans-serif only. Helvetica, Arial, Open Sans, Roboto. Bold weight—thin text vanishes on phone screens. Size should be readable at 480p on mobile. Test it on your phone, not your monitor.

Contrast

White text on a semi-transparent black box is the standard for a reason—it works against any background.

Text color: White (#FFFFFF), or yellow for emphasis
Background: Black (#000000) at 50-70% opacity
Stroke: 2-4px black outline as an alternative to the box

Position

Lower third of the frame, but above platform UI elements (like/comment/share buttons). Center-aligned, one to two lines max. Keep within the center 80% of frame width.

Timing

Short-form moves fast. Captions need to keep up.

Characters per line: 32-42 max
Lines per frame: 1-2 (3 only for critical moments)
Display time: 2-4 seconds per block
Reading speed: 15-20 characters per second is average—calculate your display time from there

Formatting Styles

Static

White text on a black bar. Simple, readable, professional. Works for everything.

Animated

Word-by-word or phrase-by-phrase reveals synced to audio. Popular on TikTok. CapCut, Descript, and Kapwing generate these automatically. Adds energy, but feels gimmicky if overused. Save it for personality-driven content.

Styled

Custom fonts, brand colors, word highlighting. High production value, easy to overdo. If readability suffers, your style is working against you.

Word-by-Word Highlighting

Each word lights up as it’s spoken. Common in trending audio content. Most caption tools (Captions app, CapCut, Veed) support it natively.

Tools

Native Platform Tools

TikTok: Upload → Edit → Captions → Auto-generate → Review

Reels: Create reel → Aa button → Caption → Edit text

Shorts: Upload → YouTube Studio → Subtitles → Auto-generate → Edit

Fast, but accuracy is spotty. Always review before publishing.

Caption Apps

Captions (iOS/Android): Purpose-built. Generates captions, multiple styles, word-by-word animation. Exports with text burned in.

CapCut: Auto-captions, multiple styles, free, integrates with TikTok export. The default choice for most creators.

Veed.io: Browser-based. Auto-captions, translation, batch processing.

Professional Editors

DaVinci Resolve: Manual caption tracks with full control. Precise but time-consuming for short-form.

Premiere Pro: Speech-to-text generates captions you can style and export as open or closed.

Accessibility

Accuracy First

Auto-captions average 80-95% accuracy. The gap matters. “Let’s eat, grandma” vs. “Let’s eat grandma” is a meme—real errors kill comprehension. Review auto-captions before publishing. Fix names, brands, and technical terms. Auto-captions are worst at proper nouns.

More Than Dialogue

For accessible captions, include:

Speaker identification: [Narrator], [Interviewer], [Guest]
Sound effects: [door slams], [laughter], [phone rings]
Music cues: [upbeat music plays], [music fades]

Standard brackets format. Brief but clear.

Reading Speed

Give viewers enough time to actually read before the text disappears. 15-20 characters per second is average reading speed. If a block exceeds that, shorten it or split it.

Consistency

Same font, same style, same position for the whole video. Switching styles mid-video breaks the viewer’s flow.

Common Mistakes

Text too small. What looks readable on your desktop preview disappears on a phone screen. Test at mobile size.

Poor contrast. White text on a light background. Yellow on white. Use backgrounds or outlines—always.

Too much text. If viewers have to pause to read your captions, you’ve written too much. Edit down.

Bad sync. Text appearing before or after the spoken words increases cognitive load. Viewers spend energy reconciling what they see with what they hear instead of absorbing your message.

Export

For burned-in captions:

Edit video
Generate captions (auto or manual)
Style and position within safe zones
Export with captions embedded
Upload

For closed captions:

Export video without text
Export caption file (SRT, VTT, or platform-specific)
Upload both
Enable captions in platform settings

Burned-in is more reliable for short-form. No viewer action required. The text is just there.

Captions aren’t optional and they aren’t complicated. Get the specs right, review for accuracy, test on mobile, and publish. That’s it.