SRT, VTT or burned-in: choosing a caption export
By Captionwale · 31 October 2026
A caption file is text the player lays over a video, and viewers can turn it off. Burned-in captions belong to the image. What each one keeps and what it loses.
Beside the video, or inside it
There are two forms captions can take when they leave an editor. One is a caption file. An SRT or VTT file holds plain text and timestamps and travels next to the video, and the viewer's player chooses how to display it. The other is burned-in captions, drawn into the frames themselves and no more separable from the picture than the speaker's face.
Nearly everything else follows from that split: who sets the look, whether viewers are able to hide the text, and what a typo costs you once the video is up.
SRT and VTT explained
The name SRT comes from SubRip, a program that pulled subtitles out of video files. The format is bare: a series of numbered blocks, each giving the text with its start and end times, and the milliseconds follow a comma. DaVinci Resolve and Premiere Pro both import SRT, YouTube accepts it as an uploaded subtitle, and VLC and similar players load it beside a video. Fonts, colours and placement have no agreed place in it.
VTT is short for WebVTT and was built for the web. Its files begin with a WEBVTT header, and a full stop sits where SRT puts its comma. It is what the HTML track element feeds to a browser's own video player. The spec allows a degree of styling and placement; whether a viewer sees any of it is up to the player.
Burning in: what you gain and what you give up
Burn captions in and the template's entire design comes with them: the typeface, faces made for Indian scripts included, colours, outlines, word-by-word highlights, the entrance animation, and the spot you placed them in for that frame, which the Reels and Shorts safe-area guide explains. Every viewer in every app sees them, with nothing to switch on. Control is what you trade away:
- No off switch. Anyone who also enables an app's own automatic captions ends up with two sets of words.
- No restyling. The YouTube player, for instance, lets people make a caption track bigger or change its colours, and pixels ignore all of that.
- Mistakes stick. Correcting one means a new export and a new post, as most apps won't swap the video beneath a published post.
- No text to read. Both the transcript panel on YouTube and its automatic caption translation depend on a caption track, and here there is none.
- One frame shape. Recrop the video later and the captions may be sliced off.
How Captionwale makes each format
Click SRT or VTT and the file is built right there in the browser from the captions as grouped at that instant. Each caption becomes one cue, with all its words on one line, and each cue spans exactly the time its caption is on screen. The two files match word for word and time for time. No fonts, colours, emphasis or animation from the template go into either, and the VTT skips its format's positioning options entirely.
Your own device renders the burned-in MP4, and your video never goes to Captionwale's servers for this step, so leave the tab open until it completes. Original keeps the source resolution; 2K, 1080p and 720p hold the frame's shorter side to 1440, 1080 or 720 pixels, and a smaller source is never enlarged. In the MP4, every caption starts roughly a tenth of a second ahead of its first spoken word, so the entrance is underway by the time the word sounds. SRT and VTT files leave this lead out.
Decide by where it will play
Usually the destination makes the call:
- Short vertical video, Reels and Shorts included: burned-in captions. There, the look of the captions belongs to the edit itself, and some apps offer no way to add a caption file anyway.
- Landscape videos on YouTube: upload a caption file, SRT or VTT, through YouTube Studio, which lets viewers switch captions on, change their size or leave them off.
- A player embedded in your own site: VTT.
- Handing footage to an editor working in Resolve or Premiere Pro: SRT, so they can style the captions in that software.
- A single recording bound for several platforms: correct the captions once, then export the MP4 and a caption file from them, so both carry every fix.
Where this advice stops
What caption files an app accepts, and how it displays them, changes over time. Look at the upload options in the app you're actually posting to; don't trust any list, including this one.
Captionwale has its own boundaries. Downloading SRT and VTT is part of the paid plans; free accounts see a lock on those buttons, and their exports carry a Captionwale badge. You can't open an existing caption file in the editor. And if a browser has no video encoder of its own, a two-minute clip needs about two minutes of rendering, because the video must be recorded live during playback.
Questions
Why does each cue in my SRT hold just one word?
Downloads mirror the editor's grouping at that moment. With a one-word template active, each cue gets a single word. Before downloading, move to a single-line template and choose a value for Max words per caption in Caption settings; once the file is saved, go back to your original template.
Can a video with burned-in captions also get an SRT?
It can, but viewers who enable captions will read every line twice. On a burned-in video, a track earns its place only if it adds something the picture lacks, an English translation for example.
Does the VTT from Captionwale carry my template's colours?
No. It matches the SRT in text and timing and holds nothing beyond that. The template's look survives only in the burned-in MP4.
Is SRT or VTT better for YouTube?
Either is fine: YouTube takes both, and the two Captionwale files contain the same captions. If the captions will play in a video player on a web page, go with VTT.