A generalized Stein's estimation approach to speech enhancement based on perceptual criteria


Autoria(s): Krishnan, Sunder Ram; Seelamantula, Chandra Sekhar
Data(s)

2012

Resumo

We address the problem of speech enhancement using a risk- estimation approach. In particular, we propose the use the Stein’s unbiased risk estimator (SURE) for solving the problem. The need for a suitable finite-sample risk estimator arises because the actual risks invariably depend on the unknown ground truth. We consider the popular mean-squared error (MSE) criterion first, and then compare it against the perceptually-motivated Itakura-Saito (IS) distortion, by deriving unbiased estimators of the corresponding risks. We use a generalized SURE (GSURE) development, recently proposed by Eldar for MSE. We consider dependent observation models from the exponential family with an additive noise model,and derive an unbiased estimator for the risk corresponding to the IS distortion, which is non-quadratic. This serves to address the speech enhancement problem in a more general setting. Experimental results illustrate that the IS metric is efficient in suppressing musical noise, which affects the MSE-enhanced speech. However, in terms of global signal-to-noise ratio (SNR), the minimum MSE solution gives better results.

Formato

application/pdf

Identificador

http://eprints.iisc.ernet.in/46540/1/Sta_Per_Aud_1_2012.pdf

Krishnan, Sunder Ram and Seelamantula, Chandra Sekhar (2012) A generalized Stein's estimation approach to speech enhancement based on perceptual criteria. In: Workshop on Statistical and Perceptual Audition (SAPA) - Speech Communication with Adaptive Learning (SCALE), September 11, 2012, Bangalore, Karnataka, India.

Publicador

International Computer Science Institute (ICSI)

Relação

http://www.icsi.berkeley.edu/icsi/events/2012/09/generalized-steins-estimation

http://eprints.iisc.ernet.in/46540/

Palavras-Chave #Electrical Engineering
Tipo

Conference Paper

PeerReviewed