Reprocess voice trainer

What is this?

With Reprocess voice trainer you have the AI regenerate the individual voice tracks — based on the existing source file, without having to swap it. This is useful when:

• Chorilo has released a new, better AI version
• The old tracks contain small errors that a new render could remove
• You have changed settings that need a re-render (such as a different voice line-up)
• A previous processing was aborted and you need to restart it

How to start a reprocessing

1. Open the piece and switch to the voice trainer.
2. Open the settings.
3. Click Reprocess.
4. Confirm the action in the dialog. Any cost is shown here.
5. Wait for the notification as soon as the new tracks are ready (1–5 minutes).

While processing runs, the old version remains usable — practice is not interrupted.

When reprocessing is worth it

• After an AI model update — Chorilo announces it in the release notes when a new version is available
• When members report bad separation (such as "the tenor is too quiet")
• When the old tracks have artefacts that disturb practising
• After changing the voice line-up in the settings

What changes, what stays

Regenerated:

• The audio tracks for each voice
• The waveform display in the player
• Possible improvements thanks to the current AI model

Preserved:

• Source file and timbre
• Loop points and jump markers
• Personal volume mixes of all members
• Tempo and transposition settings

Permission

To trigger a reprocessing you need the sheet_music.edit permission. Because each processing run consumes server resources, the action is intentionally restricted to editors.

Tips

• Do not reprocess every piece over and over again — usually the first tracks are already very good.
• Wait for a new AI version before re-rendering all pieces — otherwise you will get the same sound as before.
• For really bad tracks, replacing the source file is often more effective than re-rendering with the same source.
• Do not schedule reprocessing right before important rehearsals — members get used to a particular sound.