Study on the blind speech dereverberation method based on the Concept of Modulation Transfer Function

2008 Fiscal Year Final Research Report

• Project/Area Number: 18680017
• Research Category: Grant-in-Aid for Young Scientists (A)
• Allocation Type: Single-year Grants
• Research Field: Perception information processing/Intelligent robotics
• Research Institution: Japan Advanced Institute of Science and Technology
• Principal Investigator: UNOKI Masashi Japan Advanced Institute of Science and Technology, 情報科学研究科, 准教授 (00343187)
• Project Period (FY): 2006 – 2008
• Keywords: 変調伝達関数 / 残響音声 / 残響除去 / 音声明瞭度 / 逆フィルタ / パワーエンベロープ / キャリア / 基本周波数推定

Research Abstract

本研究では,音声明瞭度に関係づけられた変調伝達関数の概念に基づき,室内伝達特性を測定することなく,ブラインドで残響音声を回復する方法を提案した.提案法は,適切な時間・周波数領域におけるパワーエンベロープ逆フィルタ処理とキャリア再生成処理で構成された.人工・実残響環境において大規模評価実験を行った結果,提案法が残響によって著しく低下した音声明瞭度ならびに音声認識率を効果的に改善できることを明らかにした.

Research Products

• [Journal Article] 変調伝達関数に基づく音声信号処理(3)-残響環境下の基本周波数推定と残響時間ブラインド推定- (2009)
  • Author(s): 鵜木祐史
  • Journal Title: 信号処理学会論文誌(March 2009, 解説論文) Vol. 13, No. 2 Pages: 91-101
  • Peer Reviewed
• [Journal Article] A study on the noise suppression method based on the MTF concept (2009)
  • Author(s): Yutaka Yamasaki and Masashi Unoki
  • Journal Title: Proc. 2009 RISP International workshop on Nonlinear Circuits and Signal Processing (NCSP09) Pages: 157-160
  • Peer Reviewed
• [Journal Article] 変調伝達関数に基づく音声信号処理(2)-ブラインド残響音声回復法- (2009)
  • Author(s): 鵜木祐史
  • Journal Title: 信号処理学会論文誌(Jan. 2009,解説論文) Vol. 13, No. 1 Pages: 3-12
  • Peer Reviewed
• [Journal Article] A comparative evaluation of MTF-based blind restoration of the sub-band power envelopes of speech as a front-end processor for automatic speech recognition systems (2008)
  • Author(s): Xugang Lu, Masashi Unoki, and Masato Akagi
  • Journal Title: Acoustical Science & Technology Vol. 29, No. 6 Pages: 351-361
  • Peer Reviewed
• [Journal Article] 変調伝達関数に基づく音声信号処理(1)-パワーエンベロープ逆フィルタ処理の原理とその応用について- (2008)
  • Author(s): 鵜木祐史
  • Journal Title: 信号処理学会論文誌(Sept. 2008,解説論文) Vol. 12, No. 5 Pages: 339-348
  • Peer Reviewed
• [Journal Article] Robust Front End Processing for Speech Recognition in Reverberant Environments : Utilization of Speech Characteristics (2008)
  • Author(s): Rico Petrick, Xugang Lu, Masashi Unoki, Masato Akagi, and Ruediger Hoffmann
  • Journal Title: Proc. Interspeech2008 Pages: 658-661
  • Peer Reviewed
• [Journal Article] A Comprehensive Study on the Effects of Room Reverberation on Fundamental Frequency Estimation (2008)
  • Author(s): Rico Petrick, Masashi Unoki, Anish Mittal, Calros Segura, and Ruediger Hoffmann
  • Journal Title: Proc. Interspeech2008 Pages: 131-134
  • Peer Reviewed
• [Journal Article] MTF-based method of blind estimation of reverberation time in room acoustics (2008)
  • Author(s): Masashi Unoki and Sota Hiramatsu
  • Journal Title: Proc. EUSIPCO2008(CD-ROM)
  • Peer Reviewed
• [Journal Article] A Study on the Blind Estimation of Reverberation Time in Room Acoustics (2008)
  • Author(s): Sota Hiramatsu and Masashi Unoki
  • Journal Title: Journal of Signal Processing Vol. 12, No. 4 Pages: 323-326
  • Peer Reviewed
• [Journal Article] An MTF-based blind restoration of temporal power envelopes as a front-end processor for automatic speech recognition systems in reverberant environments (2008)
  • Author(s): Xugang Lu, Masashi Unoki, and Masato Akagi
  • Journal Title: Proc. Acoustics 08 Paris Pages: 1419-1424
• [Journal Article] Blind estimation method of reverberation time based on concept of modulation transfer function (2008)
  • Author(s): Masashi Unoki and Sota Hiramatsu
  • Journal Title: Proc. Acoustics 08 Paris Pages: 4489-4494
• [Journal Article] Comparative evaluations of robust and accurate F0 estimates in reverberant environments (2008)
  • Author(s): Masashi Unoki, Toshihiro Hosorogiya, and Yuichi Ishimoto
  • Journal Title: Proc. 2008 International Conference on Acoustics, Speech, and Signal Processing(ICASSP2008) Pages: 4569-4572
  • Peer Reviewed
• [Journal Article] Estimation of fundamental frequency of reverberant speech by utilizing complex cepstrum analysis (2008)
  • Author(s): Masashi Unoki and Toshihiro Hosorogiya
  • Journal Title: Journal of Signal Processing vol. 12, no. 1 Pages: 31-44
  • Peer Reviewed
• [Journal Article] Comparative evaluation of MTF-based feature extraction for speech recognition in reverberant environments (2007)
  • Author(s): Xugang Lu, Masashi Unoki, and Masato Akagi
  • Journal Title: Proc. XII International Conference Speech and Computer (SPECOM2007) Pages: 124-133
• [Journal Article] A study on robust and accurate F0 estimation in reverberant environments by utilizing complex cepstrum analysis (2007)
  • Author(s): Masashi Unoki and Toshihiro Hosorogiya
  • Journal Title: Proc. XII International Conference Speech and Computer (SPECOM2007) Pages: 222-232
• [Journal Article] A study on restoration methods for temporal envelope of reverberant speech based on the modulation transfer function concept (2007)
  • Author(s): Xugang Lu and Masashi Unoki
  • Journal Title: Proc. the Japan-China Joint Conference on Acoustics Pages: P-2-11
• [Journal Article] Estimation of fundamental frequency of reverberant speech by utilizing complex cepstrum analysis (2007)
  • Author(s): Masashi Unoki and Toshihiro Hosorogiya
  • Journal Title: JAIST Research Report IS-RR-2007-008
• [Journal Article] Dereverberation models for temporal envelope restoration based on the modulation transfer function concept (2007)
  • Author(s): Xugang Lu and Masashi Unoki
  • Journal Title: Proc. 2007 RISP International workshop on Nonlinear Circuits and Signal Processing(NCSP07) Pages: 394-400
  • Peer Reviewed
• [Journal Article] A study on the estimation method of the fundamental frequency from a reverberant speech (2007)
  • Author(s): Toshihiro Hosorogiya and Masashi Unoki
  • Journal Title: Proc. 2007 RISP International workshop on Nonlinear Circuits and Signal Processing (NCSP07) Pages: 575-578
  • Peer Reviewed
• [Journal Article] An improved carrier regeneration process in the MTF-based speech dereverberation method using instantaneous amplitude and phase (2007)
  • Author(s): Yohei Shibano and Masashi Unoki
  • Journal Title: Proc. 2007 RISP International workshop on Nonlinear Circuits and Signal Processing (NCSP07) Pages: 221-224
  • Peer Reviewed
• [Journal Article] A robust feature extraction based on the MTF concept for speech recognition in reverberant environment (2006)
  • Author(s): Xugang Lu, Masashi Unoki, and Masato Akagi
  • Journal Title: Proc. Interspeech 2006 Pages: 2546-2549
  • Peer Reviewed
• [Journal Article] Refinement of an MTF-based speech dereverberation method using optimal inverse-MTF filter (2006)
  • Author(s): Masashi Unoki, Masato Toi, and Masato Akagi
  • Journal Title: Proc. XI International Conference Speech and Computer (SPECOM2006) Pages: 323-326
• [Presentation] Effects of Room Reverberation on Robust and Accurate F0 Estimates (3009)
  • Author(s): Masashi Unoki, Rico Petrick, Anish Mittal, and Ruediger Hoffmann
  • Organizer: IEICE Tech. Rep.(SP2008-43, pp.1-6)
  • Place of Presentation: 岩手県立大学アイーナキャンパス
  • Year and Date: 3009-07-18
• [Presentation] Robust Front End Processing for Speech Recognition in Reverberant Environments : Utilization of Speech Properties (2008)
  • Author(s): Rico Petrick, Xugang Lu, Masashi Unoki, Masato Akagi, and Ruediger Hoffmann
  • Organizer: IEICE Tech.(Rep. SP2008-44, pp.7-12)
  • Place of Presentation: 岩手県立大アイーナキャンパス
  • Year and Date: 2008-07-18
• [Presentation] 変調伝達関数に基づいた残響時間のブラインド推定 (2008)
  • Author(s): 平松壮太, 鵜木祐史
  • Organizer: 日本音響学会春季研究発表会(2-6-10, pp.683-684)
  • Place of Presentation: 千葉工業大学
  • Year and Date: 2008-03-18
• [Presentation] Comparative evaluations of modulation transfer function based dereverberation for robust speech recognition (2008)
  • Author(s): Lu Xugang, Masashi Unoki, and Masato Akagi
  • Organizer: 日本音響学会春季研究発表会講演論文(1-10-12, pp.39-40)
  • Place of Presentation: 千葉工業大学
  • Year and Date: 2008-03-17
• [Presentation] 残響環境下におけるロバストで正確な基本周波数推定法の比較評価 (2008)
  • Author(s): 鵜木祐史, 細呂木谷敏弘, 石本祐一
  • Organizer: 電子情報通信学会技術報告(SP2007-168, pp.7-12, TL2007-73, WIT2007-73)
  • Place of Presentation: 島根大学
  • Year and Date: 2008-01-25
• [Presentation] 変調伝達関数に基づいた残響時間のブラインド推定法の検討 (2008)
  • Author(s): 平松壮太, 鵜木祐史
  • Organizer: 日本音響学会聴覚研究会資料(vol.37, no.11, H-2007-149)
  • Place of Presentation: 京都大学
  • Year and Date: 2008-01-23
• [Presentation] Robust and accurate F0 estimation for reverberant speech by utilizing complex cepstrum analysis (2007)
  • Author(s): Masashi Unoki and Toshihiro Hosorogiya
  • Organizer: J. Acoust. Soc. Am.(Vol.122, No.5, Pt. 2.2, pp.3021)
  • Place of Presentation: New Orleans, Luisiana, USA
  • Year and Date: 2007-11-29
• [Presentation] 複素ケプストラム分析を利用した残響音声の基本周波数推定法 (2007)
  • Author(s): 鵜木祐史, 細呂木谷敏弘
  • Organizer: 電子情報通信学会技術報告(Vol.107, No.186, pp.31-36, EA2007-44)
  • Place of Presentation: 東北大学
  • Year and Date: 2007-08-09
• [Presentation] 残響環境における音声の基本周波数推定法に関する検討 (2007)
  • Author(s): 細呂木谷敏弘, 鵜木祐史
  • Organizer: 日本音響学会春季研究発表会(1-Q-1, pp.265-266)
  • Place of Presentation: 千葉工業大学
  • Year and Date: 2007-03-17
• [Presentation] Suppression of speech intelligibility loss through an MTF-based speech dereverberation method (2006)
  • Author(s): Masashi Unoki, Masato Toi, Yohei Shibano, and Masato Akagi
  • Organizer: J. Acoust. Soc. Am.(vol.120, no.5, Pt. 2, 5pAA12, pp. 3360)
  • Place of Presentation: Honolulu, Hawaii, USA
  • Year and Date: 2006-12-02
• [Remarks] 平成20年度「良いシーズをつなぐ知の連携システム(つなぐしくみ)」支援課題
  • URL: http://www.jst.go.jp/pr/info/info624/besshi1.html
• [Remarks] J-STORE:
  • URL: http://jstore.jst.go.jp/cgi-bin/prompt/detail.cgi?prompt_id=6800
• [Remarks]
  • URL: http://jstshingi.jp/abst/jst/090619/program.html#6
• [Patent(Industrial Property Rights)] 残響時間推定装置及び残響時間推定方法 (2008)
  • Inventor(s): 鵜木祐史, 平松壮太
  • Industrial Property Rights Holder: 北陸先端科学技術大学院大学
  • Industrial Property Number: 特願,番号:2008-95540
  • Filing Date: 2008-03-04