Transcription Models

For users of the native application, downloading and enabling one of the provided transcription models will enable the following new actions in the Editor screen:

Transcribe

Selecting this action from any passage's action menu will use the selected transcription model to transcribe the corresponding audio segment to a new Passage, replacing any tokens currently contained by the passage. The new passage will include tokenization and timestamps derived from the machine transcription.

This functionality is useful in situations in which no existing text accompaniment is available for the source material, which is often this is the case for TV shows, movies, etc.

Auto-Sync Tokens

Selecting this action from the chapter title's action menu will transcribe the audio for the chapter (indicated by the duration between the first and last audio position), and adjust the timestamp for each existing token to match the transcription output.

This functionality is useful in situations where text matching the audio is already available and timestamps for each token are desired. This is often the case for audiobooks.