Comparison of subband cepstrum and Mel cepstrum for open set speaker classification

Author

Patil, Hemant A. ; Basu, T.K.

Author_Institution

Dept. of Electr. Eng., IIT, Kharagpur, India

fYear

2004

fDate

20-22 Dec. 2004

Firstpage

35

Lastpage

40

Abstract

In this paper, a new method of feature extraction based on multirate signal processing and wavelet analysis using a polynomial classifier has been described. Dialectal zone based speaker classification in Marathi language has been attempted in the open set mode. The method consists of dividing the speech signal into nonuniform subbands in approximate Mel-scale using an admissible wavelet packet interbank and modeling each dialectal zone with the 2^nd order polynomial expansion of feature vector.

Keywords

cepstral analysis; feature extraction; natural languages; polynomial approximation; signal classification; speaker recognition; wavelet transforms; 2^nd order polynomial expansion; Marathi language; Mel cepstrum; Mel-scale approximation; admissible wavelet packet interbank; dialectal zone modeling; feature extraction; multirate signal processing; open set speaker classification; subband cepstrum; Cepstral analysis; Cepstrum; Feature extraction; Natural languages; Polynomials; Signal analysis; Signal processing; Speech; Wavelet analysis; Wavelet packets;

fLanguage

English

Publisher

ieee

Conference_Titel

India Annual Conference, 2004. Proceedings of the IEEE INDICON 2004. First

Print_ISBN

0-7803-8909-3

Type

conf

DOI

10.1109/INDICO.2004.1497701

Filename

1497701