espnet.nets.pytorch_backend.frontends.beamformer.get_mvdr_vector

source

espnet.nets.pytorch_backend.frontends.beamformer.get_mvdr_vector(psd_s: ComplexTensor, psd_n: ComplexTensor, reference_vector: Tensor, eps: float = 1e-15) → ComplexTensor

Return the MVDR(Minimum Variance Distortionless Response) vector:

h = (Npsd^-1 @ Spsd) / (Tr(Npsd^-1 @ Spsd)) @ u

Reference: : On optimal frequency-domain multichannel linear filtering for noise reduction; M. Souden et al., 2010; https://ieeexplore.ieee.org/document/5089420

• Parameters:
  • psd_s (ComplexTensor) – (…, F, C, C)
  • psd_n (ComplexTensor) – (…, F, C, C)
  • reference_vector (torch.Tensor) – (…, C)
  • eps (float)
• Returns: (…, F, C)
• Return type: beamform_vector (ComplexTensor)r