Title :
Analysis and improvement of post-processing for BSS aiming to suppress recognizable crosstalk
Author :
Nakagawa, Kazuya ; Takahashi, Kota
Author_Institution :
Univ. of Electro-Commun., Tokyo
Abstract :
The performance of the blind signal separation (BSS) must be evaluated using not only quality of separated signals but also absolute level of residual noise signals (crosstalk components). The purpose of our study is to suppress the recognizable crosstalk components and to extract only the target speech using a post-processing unit. In this paper, the proposed methods are analyzed in detail. Additionally, the efficient post-processing is newly proposed. The experimental results show that the proposed method achieves preventing the recognition of the crosstalk components without sacrificing hearing of the target speech.
Keywords :
blind source separation; crosstalk; speech processing; BSS; blind signal separation; recognizable crosstalk suppression; residual noise signals; Auditory system; Blind source separation; Crosstalk; Delay estimation; Independent component analysis; Nonlinear filters; Signal to noise ratio; Source separation; Speech recognition; Target recognition; Blind Source Separation (BSS); Independent Component Analysis (ICA); cocktail party problem; linear prediction analysis; residual crosstalk suppression;
Conference_Titel :
SICE, 2007 Annual Conference
Conference_Location :
Takamatsu
Print_ISBN :
978-4-907764-27-2
Electronic_ISBN :
978-4-907764-27-2
DOI :
10.1109/SICE.2007.4421203