A STATISTICAL PATTERN-RECOGNITION APPROACH TO ROBUST RECURSIVE-IDENTIFICATION OF NONSTATIONARY AR MODEL OF SPEECH PRODUCTION SYSTEM

Authors
MARKOVIC MZ KOVACEVIC BD MILOSAVLJEVIC MM
Citation
Mz. Markovic et al., A STATISTICAL PATTERN-RECOGNITION APPROACH TO ROBUST RECURSIVE-IDENTIFICATION OF NONSTATIONARY AR MODEL OF SPEECH PRODUCTION SYSTEM, IEEE transactions on speech and audio processing, 4(6), 1996, pp. 456-460
Citations number
14
Categorie Soggetti
Engineering, Eletrical & Electronic",Acoustics
Journal title
IEEE transactions on speech and audio processing → ACNP
ISSN journal
10636676
Volume
4
Issue
6
Year of publication
1996
Pages
456 - 460
Database
ISI
SICI code
1063-6676(1996)4:6<456:ASPATR>2.0.ZU;2-P
Abstract
We propose a new robust recursive procedure based an WRLS algorithm wi th VFF and frame-based quadratic classifier for identification of nons tationary AR model of speech. Two versions of the frame-based quadrati c classifier design procedure are elaborated upon. Experimental result s are obtained in analyzing speech signal on voiced and mixed excitati on frames.