What an SRT caption file looks like
An SRT file is plain text. Each caption has a number, a time range and one or two lines of text, followed by an empty line. Here is the start of a file made from a cooking video:
1
00:00:00,480 --> 00:00:03,900
Today we are making bread
with only four ingredients.
2
00:00:04,120 --> 00:00:07,350
You need flour, water, salt
and a little dry yeast.
3
00:00:07,800 --> 00:00:09,600
Let's start with the water.The time is written as hours, minutes, seconds and milliseconds, with a comma before the milliseconds. A VTT file holds the same captions with a WEBVTT line at the top and a period instead of the comma. The SRT to VTT converter shows both side by side and switches between them.
How the captions are timed
The tool does not time captions word by word. Whisper hands it phrases, each with a start and an end time, and the tool turns those phrases into captions:
- A short phrase becomes one caption with Whisper’s own start and end.
- A long phrase is wrapped into lines and split into several captions of one or two lines, so viewers are not faced with a wall of text.
- When a phrase is split, its time is shared out by the length of each part.
The start and end of every phrase come straight from the model. Inside a long, fast phrase the split is an estimate: if a line shows up a little early or late, nudge its time by hand.
How to fix words and timing
Open the SRT in any plain text editor, such as Notepad or TextEdit in plain text mode. Change the words, keep the number and time lines as they are, and save the file as UTF-8. To move a caption, edit its two times; the end must stay after the start.
If you publish on YouTube, you can also upload the file and fix the text in YouTube’s own captions editor.
Where to use the file
- YouTube. Accepts both SRT and VTT. SRT must be plain UTF-8 and any styling in it is ignored. The full upload steps are in how to export and upload subtitles.
- Your own website. The HTML video player reads captions in VTT, so download VTT or convert the SRT.
- Video editors. Many editors can import an SRT as a caption track; look for an import captions option in yours. If you want captions drawn into the picture, that step happens in the editor when you export the video.
- A voiceover for the same video. If you also need narration, text to speech for YouTube can read your script aloud.
What this caption generator does not do
- It does not burn captions into the video. You get a separate file; the video is never re-encoded.
- It does not style captions. No fonts, colours or word-by-word animated captions.
- It does not add speaker names. Captions show only the words.
- It does not translate. The captions use the language that was spoken.
If you only need the plain text of a video, video to text gives the same result without the caption focus.