DocumentCode
2458909
Title
Model-Based Eigenspectrum Estimation for Speech Enhancement
Author
Bhunjun, Vinesh ; Brookes, Mike ; Naylor, Patrick
Author_Institution
Dept. of Electr. Eng., Imperial Coll. London, London
fYear
2006
fDate
Oct. 29 2006-Nov. 1 2006
Firstpage
1331
Lastpage
1334
Abstract
The enhancement of noisy speech in the eigenspectrum domain requires an estimate of the clean speech eigenspectrum. In this paper, we develop a model for clean speech eigenspectrum for a speaker and use it to constrain the estimation process. As a result, the estimate is contained by the acoustic space defined by the model and is thus robust even for high levels of noise.
Keywords
acoustic signal processing; eigenvalues and eigenfunctions; speech enhancement; acoustic space; eigenspectrum estimation; speaker; speech eigenspectrum estimation; speech enhancement; Acoustic noise; Additive noise; Covariance matrix; Educational institutions; Eigenvalues and eigenfunctions; Hidden Markov models; Loudspeakers; Noise robustness; Speech enhancement; Speech processing;
fLanguage
English
Publisher
ieee
Conference_Titel
Signals, Systems and Computers, 2006. ACSSC '06. Fortieth Asilomar Conference on
Conference_Location
Pacific Grove, CA
ISSN
1058-6393
Print_ISBN
1-4244-0784-2
Electronic_ISBN
1058-6393
Type
conf
DOI
10.1109/ACSSC.2006.354972
Filename
4176782
Link To Document