Speech enhancement using non-linear prediction

Author

Le, T.T. ; Mason, J.S.

Author_Institution

Dept. of Electr. & Electron. Eng., Univ. Coll. of Swansea, UK

Volume

3

fYear

1993

fDate

19-21 Oct. 1993

Firstpage

306

Abstract

Describes a method for speech enhancement using multi-layer perceptrons (MLP) configured for time domain non-linear prediction (NLP) as originally proposed by Kaouri and Lin (1988) and Tamura and Waibel (1988). In addition, the authors present details of the influence of the non-linear operations compared directly with the linear case. Experimental results show that nonlinear MLPs provide better performances than linear equivalent systems. Results also show that MLPs with a multi-output structure give improved performances over MLPs with single-output structure and normalisation can affect MLPs performance.<>

Keywords

feedforward neural nets; filtering and prediction theory; speech analysis and processing; time-domain analysis; MLP; multilayer perceptrons; multioutput structure; normalisation; performance; speech enhancement; time domain nonlinear prediction; Databases; Degradation; Educational institutions; Linearity; Multilayer perceptrons; Nonlinear filters; Predictive models; Signal mapping; Speaker recognition; Speech enhancement;

fLanguage

English

Publisher

ieee

Conference_Titel

TENCON '93. Proceedings. Computer, Communication, Control and Power Engineering.1993 IEEE Region 10 Conference on

Conference_Location

Beijing, China

Print_ISBN

0-7803-1233-3

Type

conf

DOI

10.1109/TENCON.1993.327983

Filename

327983