Welcome to Phase Processing Lab conducted by Pejman Mowlaee CV
Improved Phase Reconstruction in Single-Channel Speech Separation
- Audio samples -
- Phase enhancement (PE) combined with conventional source separation (NMF) -
Different Gender Mixture Scenario at SNR = 0 (dB)
Target speaker: “been blue at d 3 now”
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Masker speaker: “been blue at d 8 please”:
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |