Channel Compensation Method for ETSI Standard DSR Front-end
-
- Tsuge Satoru
- Faculty of Engineering, the University of Tokushima
-
- Kuroiwa Singo
- Faculty of Engineering, the University of Tokushima
-
- Shishibori Masami
- Faculty of Engineering, the University of Tokushima
-
- Kita Kenji
- Center for Advanced Information Technology, the University of Tokushima
Bibliographic Information
- Other Title
-
- ＥＴＳＩ標準分散音声認識フロントエンドにおける入力系の周波数特性正規化手法
- ETSI ヒョウジュン ブンサン オンセイ ニンシキ フロントエンド ニ オケル ニュウリョクケイ ノ シュウハスウ トクセイ セイキカ シュホウ
Search this article
Description
This paper reports an evaluation of European Telecommunications Standards Institute (ETSI) standard Distributed Speech Recognition (DSR) front-end through continuous speech recognition on a Japanese speech corpus and proposes methods, the Bias Removal Methods (BRMs), that reduce the distortion between feature parameters and the VQ codebook. Experimental results show that (1) using non-quantized features in an acoustic model training procedure can improve the recognition performance of DSR front-end features and (2) broadening the analysis band can improve the recognition performance for the same bitrate. The proposed method can improve the recognition performance in DSR condition. Notably, we observed an 18% relative improvement in the error rate using the proposed method under mismatch of channel characteristic conditions.
Journal
-
- IEEJ Transactions on Electronics, Information and Systems
-
IEEJ Transactions on Electronics, Information and Systems 125 (1), 120-127, 2005
The Institute of Electrical Engineers of Japan
- Tweet
Keywords
Details 詳細情報について
-
- CRID
- 1390282679581397888
-
- NII Article ID
- 10014100435
-
- NII Book ID
- AN10065950
-
- ISSN
- 13488155
- 03854221
-
- NDL BIB ID
- 7206381
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL Search
- Crossref
- CiNii Articles
- KAKEN
- OpenAIRE
-
- Abstract License Flag
- Disallowed