Speaker normalization for speech recognition

Author

Huang, Xuedong

Author_Institution

Sch. of Comput. Sci., Carnegie Mellon Univ., Pittsburgh, PA, USA

Volume

1

fYear

1992

fDate

23-26 Mar 1992

Firstpage

465

Abstract

A codeword-dependent neural network (CDNN) is presented for the study of speaker adaptation. The CDNN is used as a nonlinear mapping function to transform speech data between two speakers. The mapping function is characterized by a number of important properties. First, the assembly of mapping functions enhances overall mapping quality. Second, multiple input vectors are used simultaneously in the transformation. This not only makes full use of dynamic information but also alleviates possible errors in the supervision data. Finally, the mapping function is derived from training data, with the quality dependent on the available amount of training data. Based on speaker-dependent models, performance evaluation showed that speaker normalization significantly reduced the error rate from 41.9% to 5.0%

Keywords

learning (artificial intelligence); neural nets; speech recognition; codeword-dependent neural network; dynamic information; multiple input vectors; nonlinear mapping function; overall mapping quality; performance evaluation; speaker adaptation; speaker normalization; speech data transformation; supervision data; training data; Assembly; Computer science; Databases; Error analysis; Loudspeakers; Management training; Neural networks; Speech recognition; Speech synthesis; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1992. ICASSP-92., 1992 IEEE International Conference on

Conference_Location

San Francisco, CA

ISSN

1520-6149

Print_ISBN

0-7803-0532-9

Type

conf

DOI

10.1109/ICASSP.1992.225871

Filename

225871