Personal tools
Document Actions

abstract_hirata

Click here to get the file

Size 2.0 kB - File type text/plain

File contents

Demo Title:
Application of Generalized Spectral Subtraction to
Blind Speech Extraction with ICA-Based Noise Estimation

Authors:
- Nobuhisa Hirata (e-mail: nobuhisa-h@is.naist.jp)
- Hiroshi Saruwatari (e-mail: sawatari@is.naist.jp)
- Takayuki Inoue
- Kiyohiro Shikano
(Hirata and Saruwatari are corresponding persons.)

Affiliation:
- Nara Institute of Science and Technology, JAPAN

----------------------------------------------
Abstract:

We newly apply a generalized spectral subtraction to
robust blind speech extraction under realistic noise
environments. First, Infomax-type ICA-based noise estimator
can cancel out the target point-source speech and extract
the noise estimate with high accuracy [1]. Next, we apply
channel-wise denoising processing to the observed signals
based on the estimated noise, in which we introduce
"Generalized Spectral Subtraction (GSS) [2]" with the
subtraction domain exponent of 0.001; such a
compressed-domain subtraction leads to less musical noise.
DOA is estimated via the outputs of channel-wise GSS based
on minimum variance DOA estimator. Finally, based on the
estimated DOA, we summarize the channel-wise GSS outputs
via DS. Such a "channel-wise SS with post-DS" structure leads
to less musical noise [3].

References:
[1] Y. Takahashi, T. Takatani, K. Osako, H. Saruwatari, and
K. Shikano, ``Blind spatial subtraction array for speech
enhancement in noisy environment,'' IEEE Transactions on Audio,
Speech and Language Processing, vol.17, no.4, pp.650--664, 2009.

[2] T. Inoue, Y. Takahashi, H. Saruwatari, K. Shikano, K. Kondo,
"Theoretical Analysis of Musical Noise in Generalized Spectral
Subtraction: Why Should not Use Power/Amplitude Subtraction?,"
Proc. EUSIPCO2010 (in printing).

[3] Y. Takahashi, Y. Uemura, H. Saruwatari, K. Shikano, K. Kondo,
``Musical noise analysis based on higher order statistics for
microphone array and nonlinear signal processing,"
Proc. IEEE International Conference on Acoustics, Speech, and
Signal Processing (ICASSP2009), pp.229--232, 2009.