Publication Details

Title: Deployment of RASTA-PLP with the Siemens ZT Speech Recognition System
Author: M. L. Shire
Group: ICSI Technical Reports
Date: December 1997
PDF: ftp://ftp.icsi.berkeley.edu/pub/techreports/1997/tr-97-057.pdf

Overview:
RelAtive SpecTral Analysis - Perceptual Linear Prediction (RASTA-PLP) is the standard speech feature extraction method used at the International Computer Science Institute. There it has been used primarily in conjunction with a hybrid Artificial Neural Network (ANN) and Hidden Markov Model (HMM) speech recognition system. this work explores the viability of the RASTA-PLP as a candidate feature extraction method in the Siemens ZT recognition system. Experiments with a basic RASTA-PLP setup confirm that it provides good performance and is a potentially useful tool which merits further research and experimentation.

Bibliographic Information:
ICSI Technical Report TR-97-057

Bibliographic Reference:
M. L. Shire. Deployment of RASTA-PLP with the Siemens ZT Speech Recognition System. ICSI Technical Report TR-97-057, December 1997