MMSE BASED NOISE PSD TRACKING WITH LOW COMPLEXITY

被引：174

作者：

Hendriks, Richard C. ^{[1
]}

Heusdens, Richard ^{[1
]}

Jensen, Jesper ^{[2
]}

机构：

[1] Delft Univ Technol, NL-2600 AA Delft, Netherlands

[2] Oticon AS, Smorum, Denmark

来源：

2010 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING | 2010年

关键词：

Noise PSD estimation; speech enhancement; SPEECH ENHANCEMENT; POWER ESTIMATION; MODEL;

D O I：

10.1109/ICASSP.2010.5495680

中图分类号：

O42 [声学];

学科分类号：

070206 ; 082403 ;

摘要：

Most speech enhancement algorithms heavily depend on the noise power spectral density (PSD). Because this quantity is unknown in practice, estimation from the noisy data is necessary. We present a low complexity method for noise PSD estimation. The algorithm is based on a minimum mean-squared error estimator of the noise magnitude-squared DFT coefficients. Compared to minimum statistics based noise tracking, segmental SNR and PESQ are improved for non-stationary noise sources with 1 dB and 0.25 MOS points, respectively. Compared to recently published algorithms, similar good noise tracking performance is obtained, but at a computational complexity that is in the order of a factor 40 lower.

引用

页码：4266 / 4269

页数：4

共 14 条

[1]

[Anonymous], 2000, P862 ITUT

[2] SPEECH ENHANCEMENT USING A MINIMUM MEAN-SQUARE ERROR SHORT-TIME SPECTRAL AMPLITUDE ESTIMATOR [J].

EPHRAIM, Y ;

MALAH, D .

IEEE TRANSACTIONS ON ACOUSTICS SPEECH AND SIGNAL PROCESSING, 1984, 32 (06) :1109-1121

[3] Tracking of nonstationary noise based on data-driven recursive noise power estimation [J].

Erkelens, Jan S. ;

Heusdens, Richard .

IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2008, 16 (06) :1112-1123

[4] Minimum mean-square error estimation of discrete fourier coefficients with generalized gamma priors [J].

Erkelens, Jan S. ;

Hendriks, Richard C. ;

Heusdens, Richard ;

Jensen, Jesper .

IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2007, 15 (06) :1741-1752

[5]

Gradshteyn S., 2014, Table of Integrals, Series, and Products, V8th

[6] Noise tracking using DFT domain subspace decompositions [J].

Hendriks, Richard C. ;

Jensen, Jesper ;

Heusdens, Richard .

IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2008, 16 (03) :541-553

[7] FAST NOISE PSD ESTIMATION WITH LOW COMPLEXITY [J].

Hendriks, Richard C. ;

Heusdens, Richard ;

Jensen, Jesper ;

Kjems, Ulrik .

2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS, 2009, :3881-+

[8]

Kato M., 2001, IWAENC

[9] Speech enhancement by MAP spectral amplitude estimation using a super-Gaussian speech model [J].

Lotter, T ;

Vary, P .

EURASIP JOURNAL ON APPLIED SIGNAL PROCESSING, 2005, 2005 (07) :1110-1126

[10] Noise power spectral density estimation based on optimal smoothing and minimum statistics [J].

Martin, R .

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 2001, 9 (05) :504-512

← 1 2 →