Title :
A semi-blind approach to the separation of real world speech mixtures
Author :
Tordini, F. ; Piazza, F.
Author_Institution :
RSC Dept., Faital S.p.A., Donato, Italy
fDate :
6/24/1905 12:00:00 AM
Abstract :
The possibility of introducing a-priori information into multichannel blind deconvolution algorithms is investigated. The maximum likelihood (ML) approach allows one to introduce an important feature of the voice, namely the pitch, naturally into the ´blind´ model, removing the nonlinearity and showing the advantages of productive contaminations by such related research fields as computer-aided sound analysis (CASA) and Bayesian theory
Keywords :
Bayes methods; deconvolution; maximum likelihood estimation; speech processing; Bayesian theory; a-priori information; blind source separation; computer-aided sound analysis; maximum likelihood method; multichannel blind deconvolution algorithms; nonlinearity removal; productive contaminations; semi-blind approach; speech mixture separation; voice pitch; Audio recording; Bayesian methods; Contamination; Deconvolution; Decorrelation; Equations; Frequency domain analysis; Information geometry; MIMO; Speech;
Conference_Titel :
Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
0-7803-7278-6
DOI :
10.1109/IJCNN.2002.1007681
