Effective Phase Encoding for End-To-End Speaker Verification

Effective Phase Encoding for End-To-End Speaker Verification

Stať ve sborníku v databázi WoS či Scopus

The widely used magnitude spectrum based features have shown their superiority in the field of speech processing. Incontrast, the importance of phase spectrum is always ignored.This is because the patterns hidden in phase cannot be intuitivelymodelled and interpreted, due to phase wrapping phenomenon.In this paper, we explore novel phase spectrum basedfeatures, named Learnable Group Delay (LearnGD), to captureuseful information in speech signals. Specifically, firstly, thenegative of the spectral derivative of the phase spectrum, calledgroup delay (GD), is used to unwrap the phase. Then, to suppressthe spiky nature of GD, which is caused by its roots closeto the unit circle in the Z domain, a carefully designed light convolutionalsmoothing layer is employed to reconstruct the GD.Finally, an exponential hyper-parameter is introduced to reconstructGD features to restore the spectrum range and generateLearnGD features. For performance evaluation, speaker verificationexperiments are conducted on the VoxCeleb2 corpus.Compared to the traditional acoustic feature derived from themagnitude spectrum, the proposed phase-based features reacha 27.8% relative improvement in terms of EER. Furthermore,experimental results on TIMIT phoneme recognition task alsodemonstrate the effectiveness of our proposed phase-based features.

The widely used magnitude spectrum based features have shown their superiority in the field of speech processing. Incontrast, the importance of phase spectrum is always ignored.This is because the patterns hidden in phase cannot be intuitivelymodelled and interpreted, due to phase wrapping phenomenon.In this paper, we explore novel phase spectrum basedfeatures, named Learnable Group Delay (LearnGD), to captureuseful information in speech signals. Specifically, firstly, thenegative of the spectral derivative of the phase spectrum, calledgroup delay (GD), is used to unwrap the phase. Then, to suppressthe spiky nature of GD, which is caused by its roots closeto the unit circle in the Z domain, a carefully designed light convolutionalsmoothing layer is employed to reconstruct the GD.Finally, an exponential hyper-parameter is introduced to reconstructGD features to restore the spectrum range and generateLearnGD features. For performance evaluation, speaker verificationexperiments are conducted on the VoxCeleb2 corpus.Compared to the traditional acoustic feature derived from themagnitude spectrum, the proposed phase-based features reacha 27.8% relative improvement in terms of EER. Furthermore,experimental results on TIMIT phoneme recognition task alsodemonstrate the effectiveness of our proposed phase-based features.

end-to-end speaker verification, phase information,group delay, on-the-fly

end-to-end speaker verification, phase information,group delay, on-the-fly

PENG, J.; QU, X.; GU, R.; WANG, J.; XIAO, J.; BURGET, L.; ČERNOCKÝ, J.

2022

30.08.2021

International Speech Communication Association

Brno

Proceedings Interspeech 2021

1990-9772

Proceedings of Interspeech

2021

8

Francouzská republika

2366

2370

5

https://www.isca-speech.org/archive/interspeech_2021/peng21c_interspeech.html

http://hdl.handle.net/

peng21c_interspeech