Exporting
A captioned video
Section titled “A captioned video”Select Download, and choose:
| Choices | |
|---|---|
| Format | MP4 (H.264), MP4 (H.265), WebM (VP9), MOV (ProRes) |
| Size | 720p, 1080p or 4K (the short side), with the video’s own size named by its number. It may be above the original’s. |
| Frame rate | 30 or 60 fps, with the video’s own named by its number. |
| Background | Video (the captions on the picture), or Green screen |
There is no quality slider. A download is a second encoding of your video, so each format is made as good as its codec does well, and you choose the size with the size and frame rate.
| Format | File | Notes |
|---|---|---|
| MP4 (H.264) | .mp4 | Plays everywhere. The usual choice. |
| MP4 (H.265) | .mp4 | Smaller files at the same quality, but not every player or editor reads it. |
| WebM (VP9) | .webm | For the web. Slow to encode: expect 5 to 10 times real time on a modern CPU. |
| MOV (ProRes) | .mov | For editing. The files are very large: about 220 Mbps, roughly 1 GB per 40 seconds of video. |
A higher frame rate than the original repeats the picture’s frames, but the captions are drawn at every output frame, so their animation is smoother.
The video is made on your server, and the result is kept, so asking for the same thing again is instant. Anything that changes the output (the words, the style, the timing offset, the format, the size or the frame rate) is a new file.
Captions on a green screen
Section titled “Captions on a green screen”Choose Green screen as the background to get only the captions, over solid chroma green and with no sound. Key the green out in your video editor and lay the captions over any footage. Use ProRes or H.264 for it.
Subtitle files
Section titled “Subtitle files”The Download subtitles button gives:
- SRT, one cue for each caption segment, the format nearly every player and platform reads. It has no word timing.
- VTT (WebVTT), the same cues with the timing of every word inside them, for players that can highlight the word being said.
- JSON, the full transcript with the timing and confidence of every word. See Transcript format.
These come straight from the transcript, with no processing.
From a script
Section titled “From a script”All of this is available over the HTTP API.