Faculty Profiles - NISIMURA Ryuichi

Misc - NISIMURA Ryuichi

Division display241 - 260 of about 368／ All the affair displays >>

Evaluation and optimization of F0-adaptive spectral envelope estimation based on spectral smoothing with peak emphasis

Hayato Akagiri, Masanori Morise, Ryuichi Nisimura, Toshio Irino, Hideki Kawahara

Proc. ICA2010 (20th International Congress on Acoustics) ( International Commission for Acoustics ) Paper-ID 794 2010.08 [Refereed]

単語重要度を用いたN-gram補完手法が与える音声認識性能の調査

島田敏明, 西村竜一, 河原英紀, 入野俊夫

情報処理学会研究報告 ( 情報処理学会 ) 2010-SLP-82-20 2010.07

音源構造抽出法の初期推定値のバイアス除去と高速化について

河原英紀, 和田芳佳, 森勢将雅, 西村竜一, 入野俊夫

日本音響学会聴覚研究会資料 ( 日本音響学会 ) 40 ( 6 ) vol.40, no.6, H-2010-87, pp.477-482 - 482 2010.07

スペクトル距離に基づく声道長正規化のための周波数帯域の選定について

岡本恵里香, 浅香佳希, 西村竜一, 入野俊夫, 河原英紀

電子情報通信学会技術研究報告 ( 電子情報通信学会 ) 110 ( 71 ) vol.110, no.71, EA2010-36, pp.83-88 - 88 2010.06

Normalization of speaker dependent spectral differences is an important issue in speech applications, such as automatic speech recognition and high-quality voice morphing. Individual spectral differences are primarily dependent on vocal tract length differences. They are also dependent on glottal source signal and the shape of pyriform fossa. This article investigates effects of frequency range selection on spectral distance-based vocal tract length normalization (VTLN). It is based on an idea that the best VTLN performance can be attained by selecting frequency region where spectral differences are virtually exclusively determined by differences of vocal tract length. All combination of utterances spoken by 28 subjects were used to calculate estimates of their relative vocal tract lengths, which are used as the tentative "true" lengths to evaluates deviation of each VTL ratio estimation based on spectral distances. The test results revealed that the best performance is yielded by selecting frequency region spanning from 400 Hz to 4000 Hz, using an integrated logarithmic spectral distance using outputs of MFCC filter bank and their frequency derivatives.

擬似音声信号を用いた評価による音源構造抽出法の最適化について

和田芳佳, 板垣英恵, 森勢将雅, 西村竜一, 入野俊夫, 河原英紀

電子情報通信学会技術研究報告 ( 電子情報通信学会 ) vol.110, no.71, EA2010-35, pp.77-82 2010.06

Google DBを用いたトピック特化型N-gramモデル補完の検討

島田敏明, 鈴田健太郎, 永井裕貴, 西村竜一, 河原英紀, 入野俊夫

日本音響学会2010年春季研究発表会講演論文集 ( 日本音響学会 ) pp.177-178 2010.03

ウェブ収集発話を対象とした若年者判別の検討

宮森翔子, 西村竜一, 入野俊夫, 河原英紀

情報処理学会創立50周年記念(第72回)全国大会講演論文集 ( 情報処理学会 ) vol.2 pp.285-286 2010.03

高品質分析合成のための有声音の非周期成分の表現と推定について

河原英紀, 森勢将雅, 高橋徹, 坂野秀樹, 西村竜一, 入野俊夫

電子情報通信学会技術研究報告 ( 電子情報通信学会 ) 40 ( 3 ) vol.109, no.451, SP2010-165, pp.99-104 - 236 2010.03

音声からの複数の周期成分抽出および歌唱音声の周期構造分析への応用

和田芳佳, 板垣英恵, 森勢将雅, 西村竜一, 入野俊夫, 河原英紀

日本音響学会2010年春季研究発表会講演論文集 ( 日本音響学会 ) pp.505-506 2010.03

尖度に基づく音響的イベントの検出と音声分析変換合成システムへの応用について

河原英紀, 森勢将雅, 高橋徹, 坂野秀樹, 西村竜一, 入野俊夫

日本音響学会2010年春季研究発表会講演論文集 ( 日本音響学会 ) pp.315-316 2010.03

時変モーフィングに基づく歌唱音声の操作と声質および歌い回しの評価について

岡本恵里香, 和田芳佳, 森勢将雅, 西村竜一, 入野俊夫, 河原英紀

日本音響学会2010年春季研究発表会講演論文集 ( 日本音響学会 ) pp.463-464 2010.03

Simplification and extension of non-periodic excitation source representations for high-quality speech manipulation systems

Hideki Kawahara, Masanori Morise, Toru Takahashi, Hideki Banno, Ryuichi Nisimura, Toshio Irino

11TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2010 (INTERSPEECH 2010), VOLS 1-2 ( ISCA-INT SPEECH COMMUNICATION ASSOC ) 38 - + 2010 [Refereed]

A systematic framework for non-periodic excitation source representation is proposed for high-quality speech manipulation systems such as TANDEM-STRAIGHT, which is basically a channel VOCODER. The proposed method consists of two subsystems for non-periodic components; a colored noise source and an event analyzer/generator. The colored noise source is represented by using a sigmoid model with non-linear level conversion. Two model parameters, boundary frequency and slope parameters, are estimated based on pitch range linear prediction combined with F0 adaptive temporal axis warping and those on the original temporal axis. The event subsystem detects events based on kurtosis of filtered speech signals. The proposed framework provides significant quality improvement for high-quality recorded speech materials.

HIGH-QUALITY AND LIGHT-WEIGHT VOICE TRANSFORMATION ENABLING EXTRAPOLATION WITHOUT PERCEPTUAL AND OBJECTIVE BREAKDOWN

H. Kawahara, R. Nisimura, T. Irino, M. Morise, T. Takahashi, H. Banno

2010 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING ( IEEE ) 4818 - 4821 2010 [Refereed]

A voice transformation method that only relies on vowel information is proposed. The method is based on empirical cumulative distributions of perceptually relevant spectral distances, which are used to design mapping functions from distance to proximity. A set of operators are optimized in the design phase to implement on-the-fly compilation of executable transformations used in the transformation phase. Proximity of the current input parameters to the speaker's own vowel templates is used in this compilation. The proposed method deforms the source speaker's parameter space using a set of monotonic and continuous mapping functions. This smooth and topology-preserving mapping yields high-quality modification of existing speech resources.

圧縮型ガンマチャープ適合による聴覚フィルタの周波数特性と圧縮特性の推定

下塩望, 入野俊夫, 河原英紀, 西村竜一