Effect of characteristics of speakers on MSA ASR performance

Author

Droua-Hamdani, G. ; Sellouani, S. ; Boudraa, Malika

Author_Institution

Speech Process. Lab. (TAP), CRSTDLA, Algiers, Algeria

fYear

2013

Firstpage

1

Lastpage

5

Abstract

The paper deals with speaker-independent Automatic Speech Recognition (ASR) system for continuous speech. The ASR system is developed for Modern Standard Arabic (MSA) using Hidden Markov Models and a phonetically balanced corpus. The paper investigates the effect of two sources of speech variability: gender of speakers and the regional accent between the northern and southern regions of Algeria. The results show that the Word Error Rate (WER) of the ASR varies significantly between different localities according to the regional accent and gender of speakers. Indeed, higher rates are obtained in regions that present a specific pronunciation of some Arabic phonemes. As regard to gender of speakers, recordings of females are more recognized than those produced by male speakers in particular in southern localities.

Keywords

gender issues; hidden Markov models; natural language processing; speaker recognition; MSA ASR performance; WER; continuous speech; hidden Markov models; modern standard Arabic; northern Algeria regions; phonetically balanced corpus; regional accent; southern Algeria regions; southern localities; speaker characteristics; speaker gender; speaker-independent automatic speech recognition system; speech variability; word error rate; Databases; Educational institutions; Error analysis; Hidden Markov models; Speech; Speech recognition; Training; ASR; MSA; gender of speakers; regional accent; word error rate (WER);

fLanguage

English

Publisher

ieee

Conference_Titel

Communications, Signal Processing, and their Applications (ICCSPA), 2013 1st International Conference on

Conference_Location

Sharjah

Print_ISBN

978-1-4673-2820-3

Type

conf

DOI

10.1109/ICCSPA.2013.6487262

Filename

6487262