TY - GEN
T1 - Simultaneous Declipping and Beamforming via Alternating Direction Method of Multipliers
AU - Masuyama, Yoshiki
AU - Tanaka, Tomoro
AU - Yatabe, Kohei
AU - Kusano, Tsubasa
AU - Oikawa, Yasuhiro
N1 - Publisher Copyright:
© 2021 European Signal Processing Conference. All rights reserved.
PY - 2021
Y1 - 2021
N2 - Audio declipping aims to reconstruct the original signal from a clipped observed signal. Clipping has a negative effect on subsequent multichannel processing such as beamforming. This is because the channel-wise nonlinear distortion deteriorates the spatial covariance matrix in the time-frequency domain. Hence, it is important to consider multichannel audio declipping that explicitly takes into account the subsequent audio signal processing. However, most of the existing methods have been developed for the single-channel case. In this paper, we propose a joint optimization method for declipping and beamforming. The multichannel declipped signal and the spatial filter are jointly optimized to sparsify the output of beamforming. Our experimental results show the effectiveness of the proposed method compared to the beamforming with prior declipping.
AB - Audio declipping aims to reconstruct the original signal from a clipped observed signal. Clipping has a negative effect on subsequent multichannel processing such as beamforming. This is because the channel-wise nonlinear distortion deteriorates the spatial covariance matrix in the time-frequency domain. Hence, it is important to consider multichannel audio declipping that explicitly takes into account the subsequent audio signal processing. However, most of the existing methods have been developed for the single-channel case. In this paper, we propose a joint optimization method for declipping and beamforming. The multichannel declipped signal and the spatial filter are jointly optimized to sparsify the output of beamforming. Our experimental results show the effectiveness of the proposed method compared to the beamforming with prior declipping.
KW - Multichannel audio declipping
KW - Nonconvex optimization
KW - Source separation
KW - Sparsity
KW - Weighted ℓ norm
UR - http://www.scopus.com/inward/record.url?scp=85123208963&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85123208963&partnerID=8YFLogxK
U2 - 10.23919/EUSIPCO54536.2021.9616089
DO - 10.23919/EUSIPCO54536.2021.9616089
M3 - Conference contribution
AN - SCOPUS:85123208963
T3 - European Signal Processing Conference
SP - 316
EP - 320
BT - 29th European Signal Processing Conference, EUSIPCO 2021 - Proceedings
PB - European Signal Processing Conference, EUSIPCO
T2 - 29th European Signal Processing Conference, EUSIPCO 2021
Y2 - 23 August 2021 through 27 August 2021
ER -