VAD

object VAD

Voice activity detection models for use with CalibraVAD.

Functional Presets (recommended)

  • SPEECH: Optimized for speech detection (~628KB)

Algorithm Access (for reproducibility/research)

Note: VADBackend.SINGING_REALTIME uses Pitch.REALTIME (SwiftF0), not a separate VAD model.

Properties

Link copied to clipboard

All available VAD models

Link copied to clipboard

Silero VAD model for speech detection.

Link copied to clipboard

Optimized for speech detection.