Youka uses advanced AI models to separate vocals and instrumentals from songs. By default, the separation is optimized for instrumental quality, which excludes choirs (background vocals). If you’d like to preserve the choirs, follow these steps:

1. Access the Audio Menu:
   • Click on the Audio menu located below the player.
2. Select the Desired Model:
   • Choose MDX-Net (with background vocals) from the list.
3. Process the Audio:
   • This will start a new job to create audio stems using the selected model.
   • You can monitor the progress in the Jobs tab.
4. Play the New Stems:
   • Once the job is complete, go to the Jobs tab and click Play to use the new stems.
   • (Note: Stems refer to the separated components of the audio file, such as vocals, instrumentals, etc.)
5. Verify the Selection:
   • Open the Audio menu again to confirm that the new stems are selected.
6. Set as Default (Optional):
   • To make this model the default for all future karaoke generations:
     • Go to Settings.
     • Change the Separate Model to MDX-Net (with background vocals).

Following these steps ensures that choirs or background vocals are retained in your karaoke tracks.