Open captions are burned into the video image itself, so every viewer sees them and nobody can turn them off. Closed captions live in a separate track or file that the viewer switches on or off, and that the platform can index, restyle and offer in several languages. For most YouTube uploads, closed captions are the better default; for Shorts, Reels and TikTok, where many people watch muted and designed text is part of the style, open captions are common, and many creators publish both.
This guide explains the practical differences, the trade-offs for accessibility and reach, and the FFmpeg commands for producing each.
The difference in one table
| Open captions (burned-in) | Closed captions (soft) | |
|---|---|---|
| Where the text lives | Pixels in the video frame | A separate file or stream (SRT, VTT, etc.) |
| Viewer can turn off | No | Yes |
| Viewer can resize or restyle | No | Often, depending on the player |
| Multiple languages | One export per language | Many tracks on one video |
| Fixing a typo | Re-export and re-upload the video | Replace the caption file |
| Styling control | Complete: any font, colour, animation | Limited to what the player supports |
| Shows on every platform | Yes | Only where the platform supports caption files |
| Readable by search and platform systems | No (it is just image) | Yes, the text is machine-readable |
"Subtitles" and "captions" are often used interchangeably. Strictly, captions (and subtitles for the deaf and hard of hearing, or SDH) include non-speech information such as [door slams] or [upbeat music], while subtitles assume the viewer can hear and only need the dialogue, often in another language. Open and closed describe how the text is delivered, not what it contains.
When open captions make sense
- Short-form vertical video. Shorts, Reels and TikTok are frequently watched with the sound off, and the punchy word-by-word caption style is part of the format. Our guide on short-form captions covers the styling.
- Platforms without caption upload. Some social feeds, ad placements, in-venue screens and messaging apps simply ignore caption files. Burning them in is the only way to guarantee they appear.
- Designed text as part of the edit. Highlighted keywords, emoji, animated reveals or captions positioned to avoid a product on screen cannot be reproduced by a standard caption track.
- Clips shared out of context. When a video will be downloaded and re-posted by others, soft caption files usually get lost along the way.
When closed captions make sense
- Long-form YouTube videos. Viewers can turn them off, pick a language, and on many devices adjust size and background for readability.
- Accessibility. Closed captions let deaf and hard-of-hearing viewers use the player's accessibility settings, such as larger text or a solid background. The captions and accessibility guide for UK creators goes into this in more detail.
- Multiple languages. One video can carry English, Spanish and German tracks without separate exports. See translating subtitles for the workflow.
- Corrections after publishing. Spot a misspelt name a week later? Replace the file. With burned-in text you would have to re-upload, losing the original video's views and comments on YouTube.
Using both at once
Many creators burn in stylised captions for Shorts and also upload a closed caption file. On long-form, a sensible pattern is to keep the main video clean, upload an accurate SRT, and burn in only short designed callouts where they add something.
The one combination to avoid is burned-in captions plus auto-generated closed captions switched on by default for the same language: two sets of words on screen compete and cover more of the picture. If you have burned in English captions, you do not need to force English closed captions on, though providing a caption file is still useful for viewers who rely on player settings.
How to burn in (open) captions with FFmpeg
FFmpeg's subtitles filter renders an SRT, VTT or ASS file onto the picture. It requires an FFmpeg build with libass, which most standard builds include.
ffmpeg -i input.mp4 -vf "subtitles=captions.srt" -c:v libx264 -crf 18 -preset medium -c:a copy output_burned.mp4
Burning in means the video must be re-encoded, so choose sensible quality settings. The audio can be copied unchanged with -c:a copy.
To change the look of an SRT without converting it to ASS, use force_style:
ffmpeg -i input.mp4 -vf "subtitles=captions.srt:force_style='FontName=Arial,FontSize=22,Outline=2,Shadow=0,MarginV=40'" -c:v libx264 -crf 18 -c:a copy output_burned.mp4
A few things that trip people up:
- Paths on Windows. The filter treats
:and\specially. Put the SRT in the same folder as your command prompt and refer to it by name only, or escape the drive colon (C\:/videos/captions.srt). - Font size scales oddly. For SRT input, sizes are interpreted relative to a default script resolution rather than your video's pixel height, so test on a short clip first. For precise control, write an ASS file and set its
PlayResXandPlayResYto match the video. - Vertical video safe areas. On 9:16 video, platform buttons and descriptions cover the bottom and right edges. Raise the captions with
MarginVand keep them away from the right-hand side. Our video aspect ratios guide lists the common sizes.
The >FFmpeg filters documentation lists every option for the subtitles filter.
How to add closed captions
On YouTube
You normally do not attach the captions to the video file at all. Upload the video, then add an SRT or VTT in YouTube Studio under Subtitles, choosing the language first. YouTube can also generate automatic captions, but these need checking, especially for names and technical terms. Menu names change occasionally, so check YouTube Help's "Add your own subtitles and captions" page if the steps look different.
As a soft track inside the file
For files you hand to a client or play locally, you can mux the captions in as a selectable stream without re-encoding anything:
ffmpeg -i input.mp4 -i captions.srt -map 0 -map 1 -c copy -c:s mov_text -metadata:s:s:0 language=eng output_soft.mp4
MP4 uses the mov_text subtitle codec. For MKV you can keep SRT as it is:
ffmpeg -i input.mp4 -i captions.srt -map 0 -map 1 -c copy -c:s srt output_soft.mkv
Player support for embedded tracks varies. VLC and most desktop players will show them; some web players and social platforms ignore them. If a platform accepts a separate caption upload, use that instead.
As a sidecar file
The simplest option of all: deliver video.mp4 alongside video.en.srt with the same base name. Many desktop players load the sidecar automatically.
Quality matters either way
Whichever route you choose, the captions still need to be accurate, in sync, and readable at a comfortable speed. With open captions, mistakes are permanent, so check the file thoroughly before burning it in. Run it through the caption speed checker, and if the timing is off by a fixed amount, correct it with the SRT shifter first.
Key takeaways
- Open captions are part of the picture: guaranteed to show, fully styleable, but permanent and single-language.
- Closed captions are a separate track: switchable, adjustable by the viewer, multilingual and easy to correct.
- Default to closed captions for long-form YouTube; burn in for short-form, muted-autoplay feeds and platforms without caption support.
- Burning in re-encodes the video, so set quality deliberately and check the caption file first.
- Muxing a soft track or uploading a sidecar file needs no re-encode.
FAQ
Are burned-in captions good enough for accessibility?
They help, and they are far better than nothing, but viewers cannot enlarge them, change the contrast, or hide them. Where the platform supports it, also provide a closed caption file so people can use their own accessibility settings.
Do burned-in captions help my video get found in search?
Not directly, because the text is part of the image rather than readable data. A closed caption file gives platforms machine-readable text. Do not rely on either as a ranking trick; accurate captions are about viewers first.
Can I remove burned-in captions from a video later?
Not cleanly. The original pixels underneath were replaced when the captions were rendered. Always keep a clean, caption-free master export in your project folder.
Why does my embedded subtitle track not appear after uploading?
Most upload platforms ignore soft subtitle tracks muxed into the video file, such as the mov_text track above. (YouTube can read some broadcast-style embedded captions, but not ordinary text subtitle streams like this.) Upload the caption file separately through the platform's own captions or subtitles section.