Why Burned-In Subtitles Are Usually the Wrong Choice

Burned-in subtitles, sometimes called open captions, are text that is permanently part of the video image. They became fashionable with short social videos, where big bold words bounce across the screen in time with speech. They look modern, they guarantee the text is always visible, and for a lot of projects they are the wrong choice.

I have spent years writing subtitles for broadcasters and game studios, and the question I get most from video makers is whether they should burn them in. Usually my answer is no, or at least not only. Here is why, and when the exception applies.

What burned-in subtitles take away

When subtitles are a separate file, such as SRT or WebVTT, the viewer and the player stay in control. The text can be turned on or off, resized, recoloured, given a background box or moved. That flexibility is not a nice extra. For many viewers it is the difference between readable and unreadable.

Someone with low vision may need text twice the size you chose. Someone with dyslexia may read more easily with a particular font or a solid background. A viewer watching on a phone in sunlight needs higher contrast than one watching on a laptop at night. With a separate caption file, platforms like YouTube, Vimeo and most streaming services let the viewer adjust all of this. Once text is burned in, nobody can change it.

Language and translation

Burned-in subtitles lock a video into one language. If you later want a Spanish or Indonesian version, you need to re-export the whole video, and you cannot easily remove the English text that is already part of the picture. With separate files you simply add another track, and viewers choose the one they want.

Separate captions also feed search engines and platforms with a text version of your content. Platforms can index the words, auto-translate them and use them to understand what the video is about. Text baked into pixels is, for most of these systems, just picture.

Accessibility standards expect real captions

Accessibility guidelines such as WCAG treat captions as a text alternative synchronised with media. Separate caption tracks fulfil that role cleanly. Burned-in text can also count, but only if it is complete and accurate, including relevant sounds such as "door slams" or "music fades", which social-style word pop-ups almost never include.

There is also a difference between subtitles and captions that matters here. Subtitles usually assume the viewer can hear and only translate or transcribe dialogue. Captions for deaf and hard of hearing viewers include speaker names when they are not obvious and the sounds that carry meaning. Most burned-in styles are designed for silent autoplay feeds, not for deaf viewers, and they show it.

When burning in is the right call

I am not against open captions entirely. They make sense in a few clear cases:

• Platforms with poor caption support. Some social apps still handle uploaded caption files badly or not at all.
• Silent autoplay feeds. Most people scroll with sound off, and an animated text style can be part of the content's identity.
• Translated signs and on-screen text. Translating a sign in the footage is often best done as a forced, burned-in subtitle.
• Screens you do not control. Videos for trade show displays or public screens where nobody can switch captions on.

Even then, the best approach is often both: burn in a styled version for the feed, and also upload a proper caption file where the platform allows it.

The "everyone watches on mute so burn everything in" argument

The most common argument I hear for burning in is that most viewers watch on mute, so captions must always be visible. The statistic is real for social feeds, but the conclusion does not follow for every video.

On YouTube, streaming platforms and websites, many viewers watch with sound, and those who need captions already know how to turn them on. Forcing text onto everyone covers part of the picture, distracts from visual storytelling and cannot be removed by people who find it distracting, including some neurodivergent viewers. A better rule is to match captions to the platform: styled open captions in short social feeds, real caption tracks everywhere else.

Practical tips for good caption files

If you are switching to separate files, a few habits make them much better:

1. Keep lines to around 32 to 42 characters, and no more than two lines on screen at once.
2. Leave each caption up long enough to read: a common guide is at least one second and no faster than about 17 characters per second.
3. Break lines at natural phrase boundaries, not in the middle of a name or between an article and its noun.
4. Include speaker labels and important sounds in square brackets.
5. Check automatic transcripts carefully. They are a good start but often miss names, jargon and accents.

Games face the same choices at a much larger scale, with dialogue in dozens of languages and players who need very different text sizes. We look at how they handle it in why subtitles matter so much in online game stories. And if you stream live, captioning becomes a different challenge again, covered in why online game streams need captions for every viewer.

Burned-in subtitles are a style choice with real costs. Used deliberately, they have their place. Used by default, they quietly shut out the people captions were meant to help. More in our Captions and Subtitles section.