Exploring Monaural Features for Classification-Based Speech Segregation

• Published in: IEEE transactions on audio, speech, and language processing speech, and language processing, 2013-02, Vol.21 (2), p.270-279
• Main Authors: Wang, Yuxuan, Han, Kun, Wang, DeLiang
• Format: Article
• Language: English
• Subjects: Applied sciences; Binary classification; Classification; computational auditory scene analysis (CASA); Exact sciences and technology; feature combination; Feature extraction; group Lasso; Information, signal and communications theory; Linear prediction; Mel frequency cepstral coefficient; Modulation, demodulation; monaural speech segregation; Natural language processing; Scene analysis; Segregations; Signal and communications theory; Signal representation. Spectral analysis; Signal to noise ratio; Signal, noise; Spectrograms; Speech; Studies; Telecommunications and information theory; Training; Transaction processing
• ISSN: 1558-7916; 2329-9290; 1558-7924; 2329-9304
• DOI: 10.1109/TASL.2012.2221459

Monaural speech segregation has been a very challenging problem for decades. By casting speech segregation as a binary classification problem, recent advances have been made in computational auditory scene analysis on segregation of both voiced and unvoiced speech. So far, pitch and amplitude modulation spectrogram have been used as two main kinds of time-frequency (T-F) unit level features in classification. In this paper, we expand T-F unit features to include gammatone frequency cepstral coefficients (GFCC), mel-frequency cepstral coefficients, relative spectral transform (RASTA) and perceptual linear prediction (PLP). Comprehensive comparisons are performed in order to identify effective features for classification-based speech segregation. Our experiments in matched and unmatched test conditions show that these newly included features significantly improve speech segregation performance. Specifically, GFCC and RASTA-PLP are the best single features in matched-noise and unmatched-noise test conditions, respectively. We also find that pitch-based features are crucial for good generalization to unseen environments. To further explore complementarity in terms of discriminative power, we propose to use a group Lasso approach to select complementary features in a principled way. The final combined feature set yields promising results in both matched and unmatched test conditions.