SRT, VTT or burned-in captions: which to export

Captions leave an editor in one of two forms. A caption file, SRT or VTT, is plain text with timestamps that travels alongside the video, and the player on the viewer's screen decides how it looks. Burned-in captions are drawn into the frames themselves, as much a part of the picture as the speaker's face.

That difference settles most of what follows: who controls the look, whether a viewer can hide the words, and what it takes to fix a typo once the video is posted.

SRT takes its name from SubRip, a program for extracting subtitles from video, and the format is minimal: numbered blocks, each holding a start time, an end time and the text, with a comma before the milliseconds. Premiere Pro and DaVinci Resolve import it, YouTube accepts it as a subtitle upload, and players such as VLC open it beside a video file. It has no agreed way to set a font, a colour or a position.

VTT, short for WebVTT, was designed for the web. A VTT file opens with a WEBVTT header, puts a full stop where SRT has a comma, and is the format a browser's own video player reads through the HTML track element. The specification allows some positioning and styling, but how much of it appears depends on the player.

Burned-in captions carry a template's whole design: its typeface, including faces drawn for Indian scripts, its colours and outlines, word-by-word highlighting, entrance animation, and the position you chose for that frame, which the guide to Reels and Shorts margins covers. Every viewer sees them in every app, with nothing to switch on. What you give up is control:

The SRT and VTT buttons build their file in the browser the moment you click, from the captions exactly as the editor groups them at that point: one cue per caption, its words on a single line, timed to that caption's start and end. The two files hold the same words and times. Neither includes a template's fonts, colours, emphasis or animation, and the VTT uses none of its format's positioning features.

The burned-in MP4 is rendered on your own device, so the tab has to stay open until it finishes, and the video file isn't uploaded to AutoCap's servers to do it. You can keep the source's resolution with Original, or choose 2K, 1080p or 720p, which limit the shorter side of the frame to 1440, 1080 or 720 pixels; nothing is scaled up. Each caption also appears about a tenth of a second before its first word is heard, so its entrance animation is already moving when the word lands. The caption files leave that lead out.

Where the video will be watched usually decides it:

Apps change which caption files they accept and how they display them, so check the upload screen of the app you post to rather than trusting any list, including this one.

AutoCap has limits of its own. SRT and VTT downloads come with the paid plans; on the free plan those buttons show a lock, and free exports carry an AutoCap badge. The editor can't open an SRT or VTT file you already have. And where a browser has no built-in video encoder, AutoCap has to capture the video in real time while it plays, so a two-minute clip needs roughly two minutes to render.