✦ Summary
SRT and VTT are two files with the same cues; burned-in captions are pixels. Here is which one each player wants and when to ship all three.
SRT vs VTT vs burned-in captions: the short answer
SRT and VTT are two text formats for the same thing: numbered lines of caption text with a start and end time. Burned-in captions are different. They are drawn into the picture, so every player shows them and no viewer can turn them off.
For most training video, ship a VTT file for your LMS or intranet player, keep an SRT for anything that asks for it, and export a burned-in copy only for places that ignore caption files. This post explains the SRT vs VTT difference, when burning in is right, and how to make each one from a script.
What each format is
SRT (SubRip). The oldest and most widely accepted caption file. Each cue has a number, a time range and one or two lines of text. Timestamps use a comma before the milliseconds. There is no official styling or positioning, so players draw it however they like.
1
00:00:01,000 --> 00:00:04,200
Check the sender before you open
any link in the email.
VTT (WebVTT). The web's caption format. The file starts with the word WEBVTT, timestamps use a full stop, and a cue can carry settings such as where on screen it sits. It also supports comments and caption styling through CSS. The HTML5 <track> element reads WebVTT, which is why browser-based players and LMS players lean on it.
WEBVTT
00:00:01.000 --> 00:00:04.200 line:85%
Check the sender before you open
any link in the email.
Burned-in (open) captions. Not a file. The text is rendered into the video frames when you export, so it travels with the MP4 wherever it goes. SRT and VTT are closed captions: the viewer switches them on or off, and can pick a language if more than one file is loaded.
| SRT | VTT | Burned in | |
|---|---|---|---|
| What it is | Text file of timed cues | Text file of timed cues, web format | Text drawn into the video |
| Viewer can turn it off | Yes | Yes | No |
| Several languages on one video | Yes, one file each | Yes, one file each | One language per video file |
| Positioning and styling | Player decides | Cue settings and CSS | Fixed at export |
| Fix a typo | Edit the text file | Edit the text file | Re-export the video |
| Screen readers and search | Can read it | Can read it | Cannot read pixels |
Which players and LMSs read which
Check your own platform's help page before you commit, but these are the common cases as of September 2026:
- Browsers and HTML5 players. WebVTT through the
<track>element. An SRT file needs converting first. - Moodle. Its default media player takes subtitle and caption files in WebVTT, and the files have to be uploaded to Moodle, not linked from elsewhere.
- Microsoft Stream on SharePoint and OneDrive. Caption uploads are WebVTT, one file per language.
- YouTube. Accepts both SRT and VTT uploads.
- SCORM packages. The player inside a SCORM zip is a web page, so it wants VTT. When you render a SCORM package on Interlace servers, the captions go into the zip as VTT tracks, one per language, and the LMS player shows its own CC button.
- Video editors and many social platforms. SRT is the safe bet.
The pattern: if the video plays inside a browser, VTT. If a person or tool outside the browser is going to open the file, SRT. Since the cues are identical, keep both.
When burned-in captions are the right call
Burn captions in when the video will play somewhere that ignores caption files or starts muted:
- Social feeds and chat apps, where video autoplays without sound.
- Screens in a break room, a warehouse or a reception area.
- Older LMS players and file shares that play the MP4 with no track support.
- A short vertical video people watch on a phone between tasks.
What you give up is real. A burned-in caption cannot be switched off, so it covers part of every frame, and it can sit on top of on-screen text or a diagram label. It carries one language per file, so ten languages means ten video files. A typo means a new export. And the text is pixels, so screen readers and search cannot use it.
A good compromise for training: export one burned-in copy for the muted places, and keep a clean MP4 with VTT and SRT files for the LMS.
Caption basics for accessible training
Captions are an accessibility requirement. WCAG 2.2 success criterion 1.2.2 (Level A) asks for captions on prerecorded video with audio. A few habits make them readable:
- Two lines at most. Around 42 characters a line reads comfortably at 1080p. Vertical video and small players need shorter lines.
- Enough time on screen. A cue that flashes by for half a second is useless. Aim for one to seven seconds per cue.
- Break on meaning. Split lines at a comma or between phrases, not in the middle of a name.
- Match the narration word for word. Learners reading along notice when captions paraphrase.
- Speakers and sounds. Name the speaker when there is more than one voice, and caption an alarm or tone the learner must recognise.
- Keep a transcript. A plain text version helps people who search, skim or use a screen reader.
How to make each one
From a script, before the voiceover exists
If you have the narration text but no recording yet, time the words instead of transcribing audio. Paste the script, give the finished duration or a reading speed, and download both files. The cues are timed by their share of the words, so check them once the real voiceover is in.
From an Interlace video
Every Interlace video already has its captions, because the narration is generated with word timings, in any of the 28 narration languages. Open the video, click Export, and look at two rows:
- Captions. Off exports a clean video. Burn in draws captions into the MP4 or SCORM package. With Burn in on, you choose the caption language (the video's own, or any subtitle translation you have made) and a style: Clean, Boxed or Outline, at the Bottom or Top, in size S, M or L, with your own text and box colours. Put captions at the top when a scene has labels along the bottom.
- Subtitle file. The .srt and .vtt buttons download the caption file, timed to the narration.
The captions are cut from the narration timings, wrapped at 42 characters a line and two lines at most, so the file and the burned-in version always match.
Burned in, for a video you already have
If the video came from somewhere else, you can still burn in captions without an editor. Load the MP4 and an SRT or VTT file into the burn subtitles into video tool, pick a style, and it renders the captions into the picture in your browser.
Translated
For a subtitle file you already have, the subtitle translator keeps every timecode and translates only the cue text, then gives you SRT and VTT per language.
Inside Interlace, click Translate on a finished video and choose Subtitles only. That adds translated captions to the original video and leaves the narration and artwork as they are. Each language then shows up in the Export menu's caption language list and on the share page, where viewers can switch caption language themselves. For a full voiceover in another language, read our multilingual training video review checklist first.
Pick once, then ship all three
The SRT vs VTT question matters less than it looks: generate both from the same cues and hand each platform the one it asks for. The decision that matters is whether a burned-in copy is needed for places that play muted, and that depends on where your learners watch. If you are captioning a script today, make SRT and VTT files from your script with the free generator.