Online blind source separation and dereverberation of speech based on a joint diagonalizability constraint

Ho-Gun Yu; Do-Hui Kim; Min-Hwan Song; Hyung-Min Park

doi:10.7776/ASK.2021.40.5.503

All Issue

2021 Vol.40, Issue 5 Preview Page Next Page

Research Article

Online blind source separation and dereverberation of speech based on a joint diagonalizability constraint 공동 행렬대각화 조건 기반 온라인 음원 신호 분리 및 잔향제거

30 September 2021. pp. 503-514

PDF XML

Abstract

Reverberation in speech signals tends to significantly degrade the performance of the Blind Source Separation (BSS) system. Especially in online systems, the performance degradation becomes severe. Methods based on joint diagonalizability constraints have been recently developed to tackle the problem. To improve the quality of separated speech, in this paper, we add the proposed de-reverberation method to the online BSS algorithm based on the constraints in reverberant environments. Through experiments on the WSJCAM0 corpus, the proposed method was compared with the existing online BSS algorithm. The performance evaluation by the Signal-to-Distortion Ratio and the Perceptual Evaluation of Speech Quality demonstrated that SDR improved from 1.23 dB to 3.76 dB and PESQ improved from 1.15 to 2.12 on average.

Keywords

Online blind source separation

Online dereverberation

Independent component analysis

Joint diagonalizability constraints

신호에서의 잔향은 암묵음원분리 시스템의 성능을 크게 저하시키는 경향이 있다. 특히 온라인으로 진행되는 시스템일 때, 그 영향이 더욱 두드러진다. 최근 공동 행렬대각화를 활용하여 해당 문제를 해결하고자 하는 연구들이 이루어지고 있다. 본 논문에서는 이를 활용, 발전하여 잔향이 존재하는 환경에서의 미결정 다중 화자의 음원 분리 온라인 알고리즘에 잔향 제거 기능을 추가함으로써 분리한 음원의 품질을 개선하였다. WSJCAM0 데이터베이스에서 실험을 통해 기존에 사용되고 있는 온라인 알고리즘 성능과 비교하였다. 성능 평가는 신호 대 왜곡 비(Signal-to-Distortion Ratio, SDR)와 Perceptual Evaluation of Speech Quality(PESQ)를 통해 이루어졌고, 기존 알고리즘 대비 SDR은 평균 1.23 dB에서 3.76 dB로 향상되었고, PESQ는 1.15에서 2.12로 성능이 향상되었음을 검증하였다.

키워드

온라인 암묵음원분리

온라인 잔향제거

독립 성분 분석

공동 행렬대각화 조건

References

P. Smaragdis,"Blind separation of convolve mixtures in the frequency domain," Neurocomput. 22, 21-34 (1998). 10.1016/S0925-2312(98)00047-2

T. Kim, H. T. Attias, S.-Y. Lee, and T.-W. Lee, "Blind source separation exploiting higher order frequency dependencies," IEEE Trans. ASLP. 15, 70-79 (2007). 10.1109/TASL.2006.872618

N. Ono, "Stable and fast update rules for independent vector analysis based on auxiliary function technique," Proc. IEEE Workshop Appl. Signal Process. Audio Acoust. 189-192 (2011). 10.1109/ASPAA.2011.6082320

N. Ono and S. Miyabe, "Auxiliary-function-based independent component analysis for super-Gaussian sources," Proc. Int. Conf. Latent Variable Anal. Signal Separation, 165-172 (2010). 10.1007/978-3-642-15995-4_21

D. Kitamura, N. Ono, H. Sawada, H. Kameoka, and H. Saruwatari, "Determined blind source separation unifying independent vector analysis and nonnegative matrix factorization," IEEE/ACM Trans. ASLP. 24, 1626-1641 (2016). 10.1109/TASLP.2016.2577880

T. Nakatani, T. Yoshioka, K. Kinoshita, M. Miyoshi, and B. H. Juang, "Blind speech dereverberation with multi-channel linear prediction based on short time fourier transform representation," Proc. ICASSP. 85- 88 (2008). 10.1109/ICASSP.2008.4517552

T. Yoshioka and T. Nakatani, "Generalization of multi- channel linear prediction methods for blind MIMO impulse response shortening," IEEE Trans. Audio, Speech Lang. Process. 20, 2707-2720 (2012). 10.1109/TASL.2012.2210879

T. Nakatani, C. Boeddeker, K. kinoshita, R. Ikeshita, M. Delcroix, and R. Haeb-Umbach, "Jointly optimal denoising, dereverberation, and source separation," IEEE/ACM Trans. ASLP. 28, 2276-2282 (2020). 10.1109/TASLP.2020.3013118

R. Ikeshitam N. Ito, Nakatani, and H. Sawada, "A unifying framework for blind source separation based on a joint diagonalizability constraint," Proc. Eur. Signal Process. Conf. 1-5 (2019). 10.23919/EUSIPCO.2019.890308730625339

R. Ikeshita, N. Ito, T.Nakatani, and H. Sawada, "Independent low-rank matrix analysis with decorrelation learning," Proc. IEEE WASPAA. 288-292 (2019). 10.1109/WASPAA.2019.8937171

K. Sekiguchi, Y. Bando, A. Nugraha, K. Yoshiim, and T. Kawahara, "Fast multichannel nonnegative matrix factorization with directivity-aware jointly-diagonalizable spatial covariance matrices for blind source separation," IEEE/ACM Trans. ASLP. 28, 2610-2625 (2020). 10.1109/TASLP.2020.3019181

M. T. Akhtar, T.-P. Jung, S. Makeig, and G. Cauwenberghs, "Recursive independent component analysis for online blind source separation," IEEE Int. Symp. Circuits Syst. 6, 2813-2816 (2012). 10.1109/ISCAS.2012.6271896

T. Taniguchi, N. Ono, A. Kawamata, and S. Sagayama, "An auxiliary-function approach to online independent vector analysis for real-time blind source separation," Proc. HSCMA. 107-111 (2014). 10.1109/HSCMA.2014.6843261PMC4490172

S.-H. Hsu, T. Mullen, T.-P. Jung, and G. Cauwenberghs, "Online recursive independent component analysis for real-time source separation of high-density EEG," Proc. IEEE Eng. Med. Biol. Soc. Conf. 3845-3848 (2014).

T. Yoshioka and T. Nakatani, "Dereverberation for reverberation-robust microphone arrays," Proc. Eur. Signal Process. Conf. 1-5 (2013).

T. Nakatani and K. Kinoshita, "A unified convolutional beamformer for simultaneous denoising and dereverberation," IEEE Signal Processing Letters, 26, 903- 907 (2019). 10.1109/LSP.2019.2911179

S.-I. Amari, A. Cichocki, and H. H. Yang, "A new learning algorithm for blind signal separation," Adv. Neural Inf. Process. Syst. 8, 752-763 (1996).

M. Woodbury, "Inverting modified matrices," Memorandum Rep. 42, MR0038136 (1950).

E. Vincent, R. Gribonval, and C. Févotte, "Performance measurement in blind audio source," IEEE Trans. Audio, Speech, and Lang. Process. 14, 1462-1469 (2006). 10.1109/TSA.2005.858005

A. W. Rix, J. G. Beerends, M. P. Hollier, and A. P. Hekstra, "Perceptual evaluation of speech quality (PESQ)-A new method for speech quality assessment of telephone networks and codecs," Proc. IEEE Int. Conf. Acoust., Speech, Signal Process. 2, 749-752 (2001).

T. Robinson, J. Fransen, D. Pye, J. Foote, and S. Renals, "WSJCAM0: A british english speech corpus for large vocabulary continuous speech recognition," Proc. ICASSP. 81-84 (1995).

J. B. Allen and D. A. Berkley, "Image method for efficiently simulating small-room acoustics," J. Acoust. Soc. Am. 65, 943-950 (1979). 10.1121/1.382599

S. Bradley, H. Sato, and M. Picard, "On the importance of early reflecꠓtions for speech in rooms," J. Acoust. Soc. Am. 113, 3233-3244 (2003). 10.1121/1.157043912822796

T. Nishiura, Y. Hirano, Y. Denda, and M. Nakayama, "Investigations into early and late reflections on distant-talking speech recognition toward suitable reverberation criteria," Proc. Interspeech, 1082-1085 (2007). 10.21437/Interspeech.2007-109

Information

Publisher :The Acoustical Society of Korea
Publisher(Ko) :한국음향학회
Journal Title :The Journal of the Acoustical Society of Korea
Journal Title(Ko) :한국음향학회지
Volume : 40
No :5
Pages :503-514
Received Date : 2021-07-20
Revised Date : 2021-09-07
Accepted Date : 2021-09-14
DOI :https://doi.org/10.7776/ASK.2021.40.5.503

The Journal of the Acoustical Society of Korea ISSN:1225-4428(Print) 2287-3775(Online) 한국음향학회지

All Issue