Skip to content

Vocal Isolation

ScratchTrack includes AI-powered vocal isolation that can separate vocals from an audio track, creating a new track containing only the vocal part.

How It Works

On-device processing

Vocal isolation uses a CoreML neural network that runs entirely on your Mac. No audio is sent to the cloud — processing happens on-device using your CPU and GPU.

Isolating Vocals

  1. Select the audio track you want to process.
  2. Choose Isolate Vocals from the track options.
  3. A progress indicator appears in the toolbar showing the operation status (0–100%).
  4. When complete, a new track is created containing the isolated vocals.

During Processing

While vocal isolation is running:

  • A progress bar shows the current percentage
  • You can cancel the operation at any time
  • The audio is processed in 10-second chunks
  • Audio is automatically resampled to 44.1 kHz if needed

Processing time depends on the length of the track and your Mac's hardware. GPU acceleration speeds up the process significantly.

Use Cases

  • Remove vocals from a track to create an instrumental backing
  • Extract vocals for remixing or applying separate effects
  • Isolate a vocal take from a rough recording with bleed

Tips

Tip

  • Works best on stereo recordings with clear vocal presence
  • Results vary depending on the complexity of the mix
  • For best quality, use the isolated vocal track alongside the original — you can mute/solo to compare
  • The isolated track is a new audio track — your original recording is never modified

Next Steps