compute_speaker_embedding

Compute speaker embedding from audio

Description

Compute speaker embedding from audio

Usage

1compute_speaker_embedding(model, audio, sr, overlap = 0.5)

Arguments

  • model: Voice encoder model
  • audio: Audio samples (numeric vector or tensor)
  • sr: Sample rate of audio
  • overlap: Overlap between partials (default 0.5)

Value

Speaker embedding tensor (1, 256)