Captions

Transcribe a sequence's speech into timed captions, import or export .srt and .vtt files, and style the captions burnt into the picture.

Captions show what is said as lines over the picture. They play in the program monitor and are burnt into every export. Edit mode makes them three ways: transcribed from the sequence's own sound, imported from a subtitle file, or typed in.

Import captions, then style them · 0:13
Description

The AI trailer in Edit with the Captions tab open. An .srt file is imported: its lines fill the list and the one under the playhead shows over the picture. The style goes from Clean to Bold to Boxed at the top of the frame, and one line is fixed by typing in the list.

Transcribe

Open the Captions tab next to Timeline and Mixer, then click Transcribe the sequence. Edit mode mixes the sequence's sound as the export hears it (every audio track at its volume, muted tracks left out) and sends it a minute at a time to OpenAI's Whisper through your PopCraft AI connection. Silent minutes are skipped. The words come back with their times and are cut into captions: a new caption starts at the end of a sentence, at a pause, or before it runs past two lines or five seconds.

Transcription uses AI credits. Transcribing again replaces every caption.

Fix and time them

The list shows every caption with its times. Click the times to move the playhead there. Click the text to change it, then press Enter. Add a caption at the playhead adds a two-second line, and the bin icon removes one. Remove all captions clears the track.

Style

StyleLooks like
CleanWhite type with a soft shadow
BoxedWhite type on a dark rounded box
BoldLarge upper-case type with a black outline
Word by wordLarge outlined type where each word lights up as it is said (transcribed captions only: an imported file has no word times)

Position puts them at the bottom, the middle or the top of the frame. Size sets the type as a share of the frame's height, so a vertical video and a widescreen one both read well. Turn off Show captions to keep them for the subtitle file without burning them in.

Subtitle files

Import reads an .srt (SubRip) or .vtt (WebVTT) file and replaces the captions with its lines. Styling tags in the file are dropped. The .srt and .vtt buttons save the captions as a file, for YouTube, Vimeo, a video player or a festival's subtitles. Long lines are split over two lines.

From the agent, the CLI, the API and plugins

  • transcribe_sequence (the agent) transcribes the open sequence, or the one it names, with an optional language. It runs in the editor, where the sound can be decoded.
  • edit_sequence with captions replaces captions from words, cues or an .srt / .vtt file. It can also edit, add or remove one, set the style (preset, size, position, color), hide them, or clear them.
  • get_sequence lists the captions with their ids, times and style.

See Agent, CLI, API and plugins.