The challenge of inverse-E: the RASTA-PLP method

Author

Hermansky, Hynek ; Morgan, Nelson ; Bayya, Aruna ; Kohn, Phil

Author_Institution

US West Advanced Technologies, Boulder, CO, USA

fYear

1991

fDate

4-6 Nov 1991

Firstpage

800

Abstract

A novel technique for estimating a robust time-varying spectrum, RASTA-PLP (relative spectral-perceptual linear predictive), is described. A large test was conducted on a speaker-independent telephone digit recognition task using speech that had been corrupted with convolutional noise. Results from this test show an order-of-magnitude improvement in error rate over conventional spectral estimation techniques such as LPC (linear prediction coding) or PLP. Results from similar tests with large vocabulary continuous speech recognition show a halving of error rate with no tuning for the new problem

Keywords

filtering and prediction theory; noise; spectral analysis; speech analysis and processing; speech recognition; RASTA-PLP; continuous speech recognition; convolutional noise; error rate; inverse-E; large vocabulary; relative spectral-perceptual linear predictive; robust time-varying spectrum; speaker-independent telephone digit recognition; speech; speech analysis; speech processing; Auditory system; Cepstral analysis; Drives; Error analysis; Frequency estimation; Humans; Robustness; Speech analysis; Speech processing; Steady-state;

fLanguage

English

Publisher

ieee

Conference_Titel

Signals, Systems and Computers, 1991. 1991 Conference Record of the Twenty-Fifth Asilomar Conference on

Conference_Location

Pacific Grove, CA

ISSN

1058-6393

Print_ISBN

0-8186-2470-1

Type

conf

DOI

10.1109/ACSSC.1991.186557

Filename

186557