The effect of language on speaker identification

Author

Barger, Peter ; Sridharan, Sridha ; Moody, Miles

Author_Institution

Queensland Univ. of Technol., Brisbane, Qld., Australia

Volume

1

fYear

1997

fDate

4-4 Dec. 1997

Firstpage

141

Abstract

The effect of cross-language testing on automatic speaker identification systems is examined specifically for Mandarin and English speech and a Gaussian mixture model (GMM) based classifier. The analysis is based on examination of sorted lists of "most-likely" matches to a given model from a set of test sequences. Cross-language testing is shown to cause the distribution of speakers in a ranked list to depart significantly from the uniform distribution expected if there were no cross-language testing effects.

Keywords

Gaussian distribution; natural languages; pattern classification; speaker recognition; English; GMM based classifier; Gaussian mixture model based classifier; Mandarin; cross-language testing; most-likely matches; ranked list; sorted lists; speaker identification; test sequences; Acoustic testing; Australia; Automatic testing; Databases; Forensics; Loudspeakers; Natural languages; Signal processing; Speech processing; System testing;

fLanguage

English

Publisher

ieee

Conference_Titel

TENCON '97. IEEE Region 10 Annual Conference. Speech and Image Technologies for Computing and Telecommunications., Proceedings of IEEE

Conference_Location

Brisbane, Qld., Australia

Print_ISBN

0-7803-4365-4

Type

conf

DOI

10.1109/TENCON.1997.647277

Filename

647277