EUROSPEECH 2003 - INTERSPEECH 2003
8th European Conference on Speech Communication and Technology

Geneva, Switzerland
September 1-4, 2003

        

Integration of Noise Reduction Algorithms for Aurora2 Task

Takeshi Yamada (1), Jiro Okada (1), Kazuya Takeda (2), Norihide Kitaoka (3), Masakiyo Fujimoto (4), Shingo Kuroiwa (5), Kazumasa Yamamoto (6), Takanobu Nishiura (7), Mitsunori Mizumachi (8), Satoshi Nakamura (8)

(1) University of Tsukuba, Japan
(2) Nagoya University, Japan
(3) Toyohashi University of Technology, Japan
(4) Ryukoku University, Japan
(5) University of Tokushima, Japan
(6) Shinshu University, Japan
(7) Wakayama University, Japan
(8) ATR-SLT, Japan

To achieve high recognition performance for a wide variety of noise and for a wide range of signal-to-noise ratios, this paper presents the integration of four noise reduction algorithms: spectral subtraction with smoothing of time direction, temporal domain SVD-based speech enhancement, GMM-based speech estimation and KLT-based comb-filtering. Recognition results on the Aurora2 task show that the effectiveness of these algorithms and their combinations strongly depends on noise conditions, and excessive noise reduction tends to degrade recognition performance in multicondition training.

Full Paper

Bibliographic reference.  Yamada, Takeshi / Okada, Jiro / Takeda, Kazuya / Kitaoka, Norihide / Fujimoto, Masakiyo / Kuroiwa, Shingo / Yamamoto, Kazumasa / Nishiura, Takanobu / Mizumachi, Mitsunori / Nakamura, Satoshi (2003): "Integration of noise reduction algorithms for Aurora2 task", In EUROSPEECH-2003, 1769-1772.