Performance Estimation of Reverberant Speech Recognition Based on Reverberant Criteria RSR-Dn with Acoustic Parameters

Takahiro FUKUMORI  Masanori MORISE  Takanobu NISHIURA  Yoichi YAMASHITA 

Publication
D - Abstracts of IEICE TRANSACTIONS on Information and Systems (Japanese Edition)  Vol.J94-D  No.4  pp.712-720
Publication Date: 2011/04/01
Online ISSN: 1881-0225
Print ISSN: 1880-4535
Type of Manuscript: PAPER
Category: 
Keyword: 
reverberant speech recognition,  performance estimation,  acoustic parameters,  reverberation time, 

Full Text(in Japanese): PDF(1.1MB)

Summary: 
As no common reverberation criteria for the recognition of reverberant-speech have been proposed, it has been difficult to estimate this. We thus proposed new reverberation criteria, RSR-Dn (Reverberant Speech Recognition criteria with Dn) based on ISO3382 acoustic parameters. We first designed suitable reverberation criteria with the relation between ISO3382 Annex A acoustic parameters and speech recognition performance. We then estimated speech recognition performance based on our designed reverberation criteria, RSR-Dn. As a result of evaluation experiments, we could confirm that the recognition performance could be accurately and robustly estimated with proposed RSR-D20.