Page 35 - 《应用声学》2022年第1期

第 41 卷 第 1 期 韩哲等： 分布式无线声传感网加权预测误差语声去混响方法 31

2019, 44(5): 874–886. [26] Yoshioka T, Tachibana H, Nakatani T, et al. Adaptive
[17] Nakatani T, Yoshioka T, Kinoshita K, et al. Blind dereverberation of speech signals with speaker-position
speech dereverberation with multi-channel linear pre- change detection[C]//Acoustics, Speech and Signal Pro-
diction based on short time fourier transform rep- cessing (ICASSP), 2009 IEEE International Conference
resentation[C]//Acoustics, Speech and Signal Process- on. IEEE, 2009: 3733–3736.
ing (ICASSP), 2008 IEEE International Conference on. [27] Juki A. Sparse multi-channel linear prediction for blind
IEEE, 2008: 85–88. speech dereverberation[D]. Oldenburg: University of Old-
[18] Drude L, Boeddeker C, Heymann J, et al. Integrating neu- enburg, 2017.
ral network based beamforming and weighted prediction [28] Haykin S. Adaptive filter theory[M]. United States: Pren-
error dereverberation[C]// Interspeech, 2018: 3043–3047. tice Hall, 2002.
[29] Drude L, Heymann J, Boeddeker C, et al. NARA-WPE:
[19] Wung J, Jukic A, Malik S, et al. Robust multichannel
a Python package for weighted prediction error derever-
linear prediction for online speech dereverberation using
beration in Numpy and Tensorflow for online and of-
weighted householder least squares lattice adaptive fil-
fline processing[C]// Speech Communication 13th ITG-
ter[J]. IEEE Transactions on Signal Processing, 2020, 68:
Symposium. VDE, 2018: 1–5.
3559–3574.
[30] Raphael H. Floating point operations in matrix-vector cal-
[20] Gergen S, Nagathil A, Martin R. Audio signal classifica-
culus[R]. Technische Universität München, 2007.
tion in reverberant environments based on fuzzy-clustered
[31] Alien J B, Berkley D A. Image method for efficiently sim-
ad-hoc microphone arrays[C]//Acoustics, Speech and Sig-
ulating small-room acoustics[J]. The Journal of the Acous-
nal Processing (ICASSP), 2013 IEEE International Con-
tical Society of America, 1976: 65(4): 943–950.
ference on. IEEE, 2013: 3692–3696.
[32] Rix A W, Beerends J G, Hollier M P, et al. Per-
[21] Pasha S, Ritz C. Clustered multi-channel dereverbera-
ceptual evaluation of speech quality (PESQ): a new
tion for ad-hoc microphone arrays[C]//Asia-pacific Signal
method for speech quality assessment of telephone net-
& Information Processing Association Summit & Confer-
works and codecs[C]// Acoustics, Speech and Signal Pro-
ence. IEEE, 2015: 274–278.
cessing (ICASSP), 2001 IEEE International Conference
[22] Abed-Meraim K, Moulines E. Prediction error method for
on. IEEE, 2001: 749–752.
second-order blind identification[J]. IEEE Transactions on
[33] Taal C H, Hendriks R C, Heusdens R, et al. An algorithm
Signal Processing, 1997, 45(3): 694–705.
for intelligibility prediction of time–frequency weighted
[23] Yoshioka T, Nakatani T, Kinoshita K, et al. Speech noisy speech[J]. IEEE Transactions on Audio, Speech, and
dereverberation and denoising based on time vary- Language Processing, 2011, 19(7): 2125–2136.
ing speech model and autoregressive reverberation [34] Santos J F, Senoussaoui M, Falk T H. An improved non-
model[M]//Speech Processing in Modern Communication. intrusive intelligibility metric for noisy and reverberant
Berlin Heidelberg: Springer, 2010: 151–182. speech[C]// International Workshop on Acoustic Signal
[24] Jukić A, van Waterschoot T, Gerkmann T, et al. Multi- Enhancement. IEEE, 2014: 55–59.
channel linear prediction-based speech dereverberation [35] Habets E A P, Gannot S. Generating sensor signals in
with sparse priors[J]. IEEE/ACM Transactions on Au- isotropic noise fields[J]. The Journal of the Acoustical So-
dio, Speech, and Language Processing, 2015, 23(9): ciety of America, 2007, 122(6): 3464–3470.
1509–1520. [36] Breithaupt C, Gerkmann T, Martin R. Cepstral smooth-
[25] Caroselli J, Shafran I, Narayanan A, et al. Adaptive mul- ing of spectral filter gains for speech enhancement without
tichannel dereverberation for automatic speech recogni- musical noise[J]. IEEE Signal Processing Letters, 2007,
tion[C]// Interspeech, 2017: 3877–3881. 14(12): 1036–1039.