Skip to content

Anysub 4.3.0

Latest

Choose a tag to compare

@github-actions github-actions released this 17 Nov 09:43
· 1 commit to main since this release
v4.3.0
f3f1a8f

This release lets you pick the audio track to transcribe in files with several, keeps several transcripts per project, and gives the studio its own player controls.

Added

  • A prompt under Advanced options when you transcribe: names and terms the model should spell the way you write them. It works with local models and with providers, and the API takes it as prompt.
  • LRC export, for synced lyrics in music players.
  • An Audio track choice when you transcribe a file with more than one, such as a film with a dub. Each transcript plays the track it came from, and the audio menu under the video can play another track while you edit. The API takes it as audioTrack.
  • Several transcripts per project. New transcript replaces Re-transcribe and keeps the transcript you had, and a menu above the subtitle list switches between them. You can keep editing while a new one runs. The API lists each transcript's audioTrack.
  • Playback speed, volume and full screen in the bar under the video, which replaces the browser's own player controls.
  • Media details, from the studio's title bar, shows the file's tracks and the space each prepared audio track takes, and can rebuild or remove one.

Changed

  • The transcribe form keeps Start transcription at the bottom of the panel, and word alignment moved to Advanced options.
  • Borders and dividers are easier to see in the dark theme.
  • Transcripts from earlier Re-transcribe runs, which the studio used to hide, now show in the transcript menu, so you can open or delete them.
  • Deleting a transcript or a project asks in a dialog that names it. The studio's menu no longer lists the individual waveform, thumbnail and preview files.
  • The worker protocol is now 1.1. If you run a custom adapter image, rebuild it from the new worker image, because the application refuses workers on protocol 1.0. Adapters should transcribe the audio track a job asks for, as Adapters describes.

Fixed

  • In a file with several audio tracks, transcription could use a different track from the one the studio played.