What Are Video Captions? Types, File Formats and How to Use Them in Ads (2026)

Video captions are on-screen text, timed to the video, that shows what is said and the sounds that matter, such as music, laughter or a door slamming. They let people follow a video without hearing it: viewers who are deaf or hard of hearing, and anyone watching on mute in an office, on the metro or in bed.

So what are video captions for a brand? They are part of the creative, not an extra. In this guide we explain the types of captions, how they differ from subtitles, how caption files work, what each platform supports, how to style captions for vertical ads, and how to handle Arabic. We write from the production side: we caption video ads and brand films in Arabic and English every week, so the rules here are the ones we use on real projects.

In this guide

• Video captions, defined
• Captions vs subtitles vs SDH vs transcripts
• Open vs closed captions
• How caption files work
• Where captions work on each platform
• Why captions matter for brands
• How to write and style captions for ads
• Arabic and bilingual captions
• How to add captions to a video, step by step
• Common caption mistakes
• Frequently Asked Questions
• Get your videos captioned properly

Video captions, defined

A caption is a short block of text that appears on screen for a set time and then disappears. Each block is tied to a start time and an end time, so the words show up as they are spoken. Good captions carry three kinds of information:

• Speech. The words people say, written as they say them, not paraphrased.
• Speaker identification. Who is talking, when it isn't obvious from the picture, for example a voiceover or someone off camera.
• Meaningful sounds. Music, sound effects and tone, usually in square brackets: [upbeat music], [phone buzzes], [whispering].

The key word is meaningful. You don't caption every rustle. You caption the sounds a viewer needs to understand the scene, such as a doorbell that makes everyone turn around.

Captions are older than online video

Captions started on television so that deaf and hard-of-hearing viewers could follow programs. Online video carried the idea forward, and phones made it mainstream. Feeds start videos on mute, so captions now serve everyone who scrolls with the sound off, not only the people who need them for access.

Captions vs subtitles vs SDH vs transcripts

These four terms get mixed up all the time, and some platforms use "subtitles" for everything. The difference matters when you brief an editor or a translator, because each one assumes something different about the viewer.

Here is how the four compare side by side.

In plain terms: captions assume you can't hear, subtitles assume you can't understand the language, and SDH covers both. A transcript isn't on the video at all. It is the same text published as a page or a download, which is useful for long webinars and training videos.

Open vs closed captions

This is the first real decision you make when you caption a video, and it changes how the video is edited and exported.

Closed captions live in a separate file that travels with the video. The player shows them only when the viewer turns them on, usually with a CC button. The text is drawn by the player, so viewers can often change its size and color, and you can swap or fix the file without re-exporting the video.

Open captions are burned into the picture during export. They are part of the pixels, so everyone sees them, on every player, exactly as designed. You can't turn them off, and fixing a typo means a new export.

When to use each

The right choice depends on where the video runs and whether you control the player.

For most brands the honest answer is both. We burn styled captions into the short ad versions and deliver a clean caption file for the long versions and for YouTube, so search and accessibility are covered too. If you use video on landing pages, the same rule applies: a landing page video that autoplays on mute needs burned-in text, or the first seconds say nothing.

How caption files work

A closed caption file is a plain text file with numbered blocks, start and end times, and the words. The most common format is SRT. A single SRT block has four parts, each on its own line:

• The block number, for example 12.
• The timing, for example 00:00:14,200 --> 00:00:17,000, which means the text appears at 14.2 seconds and disappears at 17 seconds.
• One or two lines of text, for example [soft music] on one line and "We built the whole campaign in Arabic first." on the next.
• A blank line that ends the block. The player reads the file, matches each block to the video's clock and draws the text on screen.

WebVTT (.vtt) is the format built for web players. It starts with the word WEBVTT, uses a dot instead of a comma before the milliseconds, and adds options for position and simple styling. Broadcast work uses its own formats, such as SCC for the CEA-608 caption standard.

Here are the formats you will meet most often.

YouTube lists all of these and more on its page of supported subtitle and closed caption files, and notes that SRT files must be plain UTF-8 text. UTF-8 matters a lot for Arabic: save the file in another encoding and the letters turn into junk characters.

Where captions work on each platform

Every platform handles captions a little differently, and the menus move around, so check the official help pages before a big launch. These are the patterns we see in 2026.

• YouTube. You can upload caption files in many formats, type captions in YouTube Studio, or start from automatic captions. YouTube's own help page on automatic captions says their quality may vary with accents, dialects and background noise, and asks creators to review them. Closed captions work well here because viewers expect the CC button.
• Meta (Facebook and Instagram ads). Ads Manager lets you add captions to a video ad, either from an SRT file or, where it is available, generated automatically for you to review. Captions you attach follow the video into every ad that uses it. Because people scroll feeds on mute, we still burn captions into most feed and Reels ads.
• Instagram Reels (organic). Instagram can generate auto captions that viewers switch on. Many never do, so for brand Reels the captions belong in the edit.
• TikTok. Viewers can turn on auto captions, and the TikTok Ads Manager video editor can generate captions from a voiceover or from a script you paste in, then let you change the font, color and alignment. Most performance ads still use captions designed in the edit.
• LinkedIn. Video posts and ads accept an SRT file. B2B feeds are watched in offices on mute, so captions there are close to required. Our LinkedIn video ads playbook covers the rest of the creative.
• Your website. HTML video players read WebVTT files through a track element. For hero videos that autoplay without sound, use burned-in captions instead.

Here is the short version for each platform.

Why captions matter for brands

Accessibility is the first reason and it is enough on its own. Millions of people are deaf or hard of hearing, and a video without captions shuts them out. Accessibility standards for the web, such as the Web Content Accessibility Guidelines (WCAG), list captions for pre-recorded video among the most basic requirements, and many public bodies and large companies build that into their supplier rules.

The business reasons stack on top:

1. Sound-off viewing. Feeds play video on mute by default. Without text, the first seconds of a talking-head ad are a silent face, and silent faces get scrolled past.
2. Comprehension. People remember more when they see and hear the message. This matters for product names, offers, addresses and anything with numbers.
3. Non-native speakers. In Dubai, a large share of any audience watches in their second or third language. Captions help them catch fast speech and accents.
4. Search. Caption files and transcripts give platforms text to index, so a video can be found for the words spoken in it.
5. Ad performance. Captions are one of the cheapest changes you can make to an ad with a weak hold. If you are unsure, run a captioned and an uncaptioned version of the same cut against each other on your own account rather than assuming.

We won't quote a sound-off percentage here, because the widely shared numbers are rarely sourced. Check the share of muted views in your own platform reports; it is usually high enough to settle the question.

How to write and style captions for ads

Captions are read, not heard, and reading is slower than listening. The best-captioned ads are planned for it from the script onward.

Start in the script

A sentence that sounds fine spoken can make a terrible caption: too long, full of filler, with the key word at the end. When we write a video ad script, we read it as captions too. Short sentences, one idea each, and the brand or offer named in the first lines. If a line has to be cut to fit on screen, the script was too wordy.

Line length and timing

Common subtitling practice keeps each caption to one or two lines of roughly 32 to 42 characters, on screen for at least a second and rarely more than six or seven. For vertical ads we go shorter: two to five words per block, shown in time with the speech. Short blocks are easier to read on a phone and leave room for the picture.

• Break lines at natural pauses, never in the middle of a name or a number.
• Keep the speaker's exact words. Clean up "um" and false starts, but don't rewrite meaning.
• Write numbers the way they are easiest to read: "3 days", not "three days".

Placement and safe zones

On Reels, Stories, Shorts and TikTok, the bottom of the screen is covered by the caption, buttons and the call-to-action bar, and the right edge has like and share icons. Captions placed too low disappear under the interface. We set captions in the middle third of the frame, clear of faces and products, and check every export in the real app before it goes live.

Style that stays readable

• Size. Bigger than you think. Captions should be readable on a small phone at arm's length.
• Contrast. White or brand-color text with an outline, shadow or box behind it, so it reads over any background.
• Emphasis. Highlight one key word per block, such as the product or the offer, in a brand color. Highlighting everything highlights nothing.
• Consistency. Same font, size and position across a campaign. Animated "word pop" styles work for hooks, but keep them calm through the rest of the ad.

Animated captions and type are a craft of their own. Our motion graphics team builds caption styles into a brand's templates so every new ad matches without starting over.

Arabic and bilingual captions

Arabic captions are where most auto tools and many editors break. If you advertise in the UAE and the Gulf, check these points on every video.

• Right to left and letter joining. Arabic letters change shape depending on their position in a word. Use editing software and fonts with full Arabic support, or the letters appear disconnected and in the wrong order. Always check the final export, not only the timeline.
• Encoding. Save SRT and VTT files as UTF-8. Other encodings turn Arabic text into question marks or random symbols.
• Dialect. Auto captions often guess Modern Standard Arabic spelling for Gulf or Levantine speech, or fail on dialect words. A native speaker should review every Arabic caption file.
• Size and length. Arabic needs a slightly larger font size than English to stay legible on a phone. Keep blocks short.
• One language per version. Stacking Arabic and English captions on one ad crowds the frame and halves reading time for both. We make separate Arabic and English versions and let targeting do the rest.

The best results come when the Arabic is written first rather than translated at the end, which is how our Arabic video production team works: script, voiceover and right-to-left captions created in Arabic from the start. For the bigger picture, see our guide to Arabic video ads that perform in the UAE and Saudi Arabia.

How to add captions to a video, step by step

This is the workflow we use for brand and ad videos. It works whether you caption in an editing app, a platform tool or a dedicated captioning tool.

1. Lock the edit. Captions are timed to frames, so finish the cut first. Changing the edit afterwards shifts every block.
2. Get a transcript. Start from the script or an automatic transcript. AI transcription saves time, and modern AI video editing tools do this in minutes, but treat the result as a first draft.
3. Correct every word. Fix brand names, product names, people's names, numbers and dialect words. These are exactly the words auto captions get wrong, and the ones that matter most to you.
4. Add speakers and sounds. Label off-screen speakers and add the sounds that carry meaning.
5. Break and time the blocks. Split into short, readable blocks and check each one appears with the speech, not before or after.
6. Style or export. For open captions, apply the brand caption style and burn them in. For closed captions, export a UTF-8 SRT or VTT file.
7. Check on a real phone. Watch every placement on mute, in the actual app, and look for text hidden under buttons or covering faces.
8. Upload and keep the files. Upload caption files with the video and store them with the master files, so you can reuse them for cut-downs and translations.

When we handle video production for ads, captions are one stage of the whole job: we plan captions at the brief, write for them in the script, and add them before final export.

Common caption mistakes

Most caption problems we fix for clients come from the same short list.

• Publishing auto captions unchecked. One wrong brand name in the first line does more damage than no captions at all.
• Captions under the interface. Text placed at the very bottom of a vertical video gets covered by the caption and buttons.
• Blocks that are too long. Three or four lines of text flashing past can't be read before they disappear.
• Bad timing. Text that runs ahead of the speaker spoils punchlines and reveals.
• Covering the product. Captions over a face, a logo or the product shot hide the thing you paid to film.
• Only captioning the ad version. The long YouTube version and the website version need captions too.

Talking-head and presenter formats suffer most from these mistakes, because the speech carries the whole message. A Dubai property broker's AI presenter videos aimed at UK buyers are a good example of the format: one person, one idea per video, speaking to camera. Real estate ads in general lean on presenters and walk-throughs, which is why captions are standard in our real estate video ads. The same goes for internal and training films in corporate video production, where accessibility rules often make captions a requirement.

Frequently Asked Questions

What are video captions?

Video captions are on-screen text, timed to the video, that shows the spoken words, who is speaking and important sounds such as music or a doorbell. They let people follow a video without hearing it, including viewers who are deaf or hard of hearing and anyone watching on mute. Captions can be closed, so viewers switch them on, or open, burned into the picture.

What is the difference between captions and subtitles?

Captions assume the viewer can't hear the audio, so they show speech plus speaker names and meaningful sounds, in the same language as the video. Subtitles assume the viewer can hear but doesn't understand the language, so they translate the dialogue only. SDH, subtitles for the deaf and hard of hearing, combines both: translated text with sounds and speaker labels.

Should I burn captions into my video ads?

For feed ads on Reels, Stories, TikTok and Shorts, yes in most cases. Those videos play on mute and most viewers never switch closed captions on, so burned-in captions are the only text they see. For long YouTube videos, webinars and films with several language versions, upload a caption file instead, so viewers can choose and you can fix mistakes easily.

What is an SRT file?

An SRT file is a plain text caption file. It contains numbered blocks, each with a start time, an end time and one or two lines of text. Video players and platforms such as YouTube, LinkedIn and Meta Ads Manager read the file and show each block at the right moment. Save SRT files as UTF-8, especially for Arabic.

Are automatic captions good enough?

Automatic captions are a useful first draft, not a finished product. They often get brand names, product names, numbers, accents and dialect words wrong, and YouTube itself says their quality may vary. Always have a person review and correct them before publishing, and give Arabic captions to a native speaker, because auto tools often struggle with Gulf and Levantine speech.

How long should each caption be?

Common subtitling practice is one or two lines of roughly 32 to 42 characters, shown for at least a second. For vertical ads, shorter works better: two to five words per block, timed to the speech and placed in the middle third of the frame. Break lines at natural pauses and never split a name or a number across two blocks.

Get your videos captioned properly

Captions decide whether a muted viewer gets your message or scrolls past it. Done well, they are planned in the script, styled for each placement, written in the right language and dialect, and checked by a person on a real phone before launch.

That is how we deliver every ad and brand film, in Arabic and English, for Dubai and global markets. If you want videos that work with the sound off and pass an accessibility check, book a strategy call with XMA and get a quote for your next campaign.