Towards speech rate independence in large vocabulary continuous speech recognition

Author

Martínez, F. ; Tapias, D. ; Álvarez, J.

Author_Institution

Speech Technol. Group, Telefonica Investigacion y Desarrollo, Madrid, Spain

Volume

2

fYear

1998

fDate

12-15 May 1998

Firstpage

725

Abstract

We present a new speech rate classifier (SRC) which is directly based on the dynamic coefficients of the feature vectors and it is suitable to be used in real time. We also report the study that has been carried out to determine what parameters of speech are the best regarding the speech rate classification problem. In this study we analyse the correlation between several speech parameters and the average speech rate of the utterance. Finally, we report a compensation technique, which is used together with the SRC. This technique provides with a word error rate (WER) reduction of a 64.1% for slow speech rate and a 32% reduction of the average WER

Keywords

correlation methods; feature extraction; parameter estimation; pattern classification; speech recognition; average WER reduction; average speech rate; compensation technique; correlation; dynamic coefficients; feature vectors; large vocabulary continuous speech recognition; real time speech rate classifier; slow speech rate; speech parameters; speech rate classification; speech rate independence; word error rate reduction; Databases; Degradation; Energy measurement; Multilayer perceptrons; Signal processing; Speech analysis; Speech processing; Speech recognition; Testing; Vocabulary;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 1998. Proceedings of the 1998 IEEE International Conference on

Conference_Location

Seattle, WA

ISSN

1520-6149

Print_ISBN

0-7803-4428-6

Type

conf

DOI

10.1109/ICASSP.1998.675367

Filename

675367