Captions vs Subtitles vs SDH: What Is the Difference and Why It Matters
Captions, subtitles, SDH and open captions are not the same thing. Learn the differences, what accessibility rules expect, and how to prepare accurate caption files.
When you add subtitles to a video, you face one decision before any other: should the text be a separate, switchable layer, or should it be printed into the picture itself? These two choices are called soft subtitles and hardcoded subtitles (also known as burned-in subtitles or open captions), and they behave very differently for viewers, for editors, and for the platforms that play your video.
This guide explains how each one works, what you gain and lose with each, and how to decide. It also covers how to create both, so you can choose with confidence.
Soft subtitles are stored apart from the picture. The video frames are untouched, and the text lives either in its own file (a sidecar file, such as film.srt next to film.mp4) or as a subtitle stream inside the video container, alongside the video and audio streams.
When you play the video, the player reads the subtitle data and draws the text on top of the picture. Because the player draws it, the viewer can usually turn it on or off, switch between languages, and often change the size, colour, and position.
Soft subtitles can come in three common forms:
.srt or .vtt, loaded automatically or manually by the player.Hardcoded subtitles are rendered into the video frames. The text becomes pixels, exactly like the actors and the scenery. Once the video is exported, the text is a permanent part of the picture. The viewer cannot turn it off, change the language, or resize it, and software cannot read it as text.
Hardcoding happens during export from a video editor, or with a tool that re-encodes the video and draws the text over it. Because the video is re-encoded, you also have to think about quality loss.
| Question | Soft subtitles | Hardcoded subtitles |
|---|---|---|
| Can the viewer turn them off? | Yes | No |
| Multiple languages in one video? | Yes, one track per language | No, one version per file |
| Can they be edited later? | Yes, edit the text file | No, re-export the video |
| Does the video need re-encoding? | No | Yes |
| Searchable and readable by software? | Yes | No, they are just pixels |
| Look identical on every device? | No, depends on the player | Yes, exactly as you designed |
| Works on every platform? | Mostly, with some exceptions | Always |
| Supports advanced styling? | Depends on format and player | Fully, anything you can render |
Flexibility for viewers. Some people need subtitles, others find them distracting. Soft subtitles let each viewer choose. They also let viewers pick the language they understand and adjust the size and colour to their eyes, which is important for accessibility.
Easy corrections. If you find a typo a week after publishing, you fix one text file and upload it again. You do not touch the video.
No quality loss. The video is not re-encoded, so you keep the original picture quality and avoid long export times.
Better for search and accessibility. Text can be indexed, translated, and read by assistive technology. Platforms that accept caption uploads can use the text to support discovery and automatic translation features.
One video, many markets. You can ship one master video and a folder of subtitle files, one per language.
Guaranteed appearance. Whatever you see in your editor is exactly what the viewer sees: the font, the position, the colours, even animations. There is no dependency on a player's rendering quirks.
Works everywhere. Some platforms, older televisions, messaging apps, and social feeds do not support subtitle tracks at all, or hide them by default. Burned-in text always shows.
Silent autoplay. Videos in social feeds usually start muted. Text that is part of the picture is visible immediately, which is why short social videos commonly use burned-in captions.
Styled storytelling. Kinetic text, speaker colours, and creative placement are much easier to control when the subtitles are part of the design.
For soft subtitles, the main risks are inconsistent rendering (fonts and positions can differ between players), no support on some platforms, and the chance that viewers simply never turn them on. For hardcoded subtitles, the main costs are that nothing can be changed without a re-export, one version is needed per language, viewers who dislike subtitles cannot hide them, and each re-encode may cost some visual quality.
Ask these questions in order.
A practical approach used by many creators is to do both: publish a clean version with soft subtitles for the main platform, and export a second burned-in version for social media.
The simplest soft subtitle is a text file. Save your subtitles as .srt, name it like the video, and keep both in the same folder; most desktop players will pick it up automatically. See adding subtitles in VLC and other players for naming rules.
To embed the subtitles in the video container, so you ship one file, you can use a tool such as FFmpeg. MKV can hold SRT streams directly, and MP4 uses its own text format that FFmpeg can convert to. Our FFmpeg subtitle commands guide gives the exact commands.
For online platforms, upload the SRT next to the video. Our YouTube guide shows how.
There are two main routes.
In a video editor. Import the SRT as a subtitle or caption track, style it, and export with the option to burn the captions into the video. The workflow in Premiere Pro and DaVinci Resolve is covered separately.
With a converter. Tools such as HandBrake and FFmpeg can draw an SRT onto the picture during re-encoding. They let you choose the font size and style. Preview a short section first, because the text may appear larger or smaller than you expect, depending on the video resolution.
Before burning in, finish your editing. Merge competing translations, fix timing, and proof every line, because you cannot change anything afterwards without re-exporting. A browser tool like the SRT Merger helps you settle the final text first.
Burning in before the text is final. A typo in a hardcoded subtitle means re-rendering the whole video.
Forgetting the safe area. Hardcoded text too close to the bottom edge may be cut off on some screens or covered by platform interface elements. Keep a comfortable margin.
Making text too small. Text that is readable on a big monitor may be tiny on a phone. Check on the smallest screen your audience uses.
Relying only on a sidecar file. If your viewers might download only the video file, the sidecar travels separately and may be lost. Embed it or provide clear instructions.
Assuming every player renders identically. Test soft subtitles in at least two players.
Soft subtitles are flexible, editable, accessible, and lossless, but they depend on the player. Hardcoded subtitles look identical everywhere and survive any platform, but they are permanent and cost you re-encoding. For most projects, use soft subtitles as the default, and create a hardcoded copy only for places that need one. Whichever you pick, finalise and proofread the text first, and keep your SRT file as the editable source.
Merge your own subtitle files
Compare two .srt files side by side and save the best lines as one file. It runs in your browser, so nothing is uploaded.