Parameter optimization for vocal tract length normalization

Author

Dognin, Pierre ; El-Jaroudi, Amro ; Billa, Jayadev

Author_Institution

Dept. of Electr. Eng., Pittsburgh Univ., PA, USA

Volume

3

fYear

2000

fDate

2000

Firstpage

1767

Abstract

This paper focuses on the optimization of model parameters for vocal tract length normalization (VTLN). For maximum likelihood (ML) based normalization techniques, the complexity of the VTL-models is a source of variation in system performance. An optimal complexity for the VTL-model that ensures best global word error rate is proposed. The choice of frequency warping factor also depends on the signal processing step of VTLN. A best set of parameters for the VTLN signal processing stage is proposed with extensive results for an optimal frequency range

Keywords

computational complexity; error statistics; physiology; speech; VTL-models; VTLN; complexity; frequency warping factor; global word error rate; maximum likelihood based normalization; parameter optimization; signal processing step; system performance; vocal tract length normalization; Electronic mail; Error analysis; Frequency domain analysis; Frequency estimation; Loudspeakers; Performance loss; Signal processing; Speech processing; Speech recognition; System performance;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 2000. ICASSP '00. Proceedings. 2000 IEEE International Conference on

Conference_Location

Istanbul

ISSN

1520-6149

Print_ISBN

0-7803-6293-4

Type

conf

DOI

10.1109/ICASSP.2000.862095

Filename

862095