Versatile music processing by combining statistical signal processing and music theory

结合统计信号处理和音乐理论的多功能音乐处理

• 批准号: 23240021
• 负责人: SAGAYAMA Shigeki
• 金额: $ 24.79万
• 依托单位: National Institute of Informatics (2013)The University of Tokyo (2011-2012)
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (A)
• 财政年份: 2011
• 资助国家: 日本
• 起止时间: 2011-04-01 至 2014-03-31
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-23240021/
• 关键词: 音楽信号処理; 音楽情報処理; 楽音分離; 自動作曲; 自動伴奏; ジャンル認識; リズム解析; 自動編曲; 自動採譜; 多重音解析; 国際研究者交流（フランス）; 楽譜追跡; 自動ギター運指決定; 自動演奏表情付け; 楽音分離加工; 和声モデル; 多声部リズムモデル; 音楽構造解析; 自動演奏

We investigated signal and information processing of music, acoustic and speech by integrating statistical signal processing and mathematical models of music theory. Just as integration of acoustic and linguistic processing was the key technology in speech recognition, integrating signal processing and music theory is essential in music processing. More precisely, we developed A: analysis, conversion, separation and detection of music signals based on mathematical models and statistical learning, B: chord recognition, rhythm analysis, segmentation, structural analysis, genre recognition based on mathematical formulation of music theory, and C: automatic music rendering, automatic music composition, automatic music accompaniment and automatic music arrangement based on machine learning and mathematical optimization.

结合统计信号处理和音乐理论的数学模型，研究了音乐、声学和语音的信号和信息处理。就像语音识别的关键技术是声学和语言处理的结合一样，音乐处理的关键技术是信号处理和音乐理论的结合。更准确地说，我们开发了A：基于数学模型和统计学习的音乐信号分析、转换、分离和检测，B：基于音乐理论数学公式的和弦识别、节奏分析、分割、结构分析、体裁识别，以及C：基于机器学习和数学优化的自动音乐渲染、自动音乐作曲、自动音乐伴奏和自动音乐编曲。

项目成果

Polyhymnia: An Automatic Piano Performance System with Statistical Modeling of Polyphonic Expression and Musical Symbol Interpretation

• DOI: 10.5281/zenodo.1178069
• 发表时间: 2011-06
• 作者: Tae Hun Kim;Satoru Fukayama;T. Nishimoto;S. Sagayama

Singing Voice Enhancement in Monaural Music Signals Based on Two-stage Harmonic/Percussive Sound Separation on Multiple Resolution Spectrograms

• DOI: 10.1109/taslp.2013.2287052
• 发表时间: 2014
• 期刊: IEEE/ACM Transactions on Audio, Speech, and Language Processing
• 作者: Hideyuki Tachibana;Nobutaka Ono;S. Sagayama

Acoustic score following to musical performance with errors and arbitrary repeats and skips for automatic accompaniment

音乐演奏后的原声乐谱存在错误以及自动伴奏的任意重复和跳过

• 发表时间: 2013
• 期刊: Proc. SMC
• 作者: Tomohiko Nakamura;Eita Nakamura;Shigeki Sagayama
• 通讯作者: Shigeki Sagayama

Generative modeling of speech F0 contours

语音 F0 轮廓的生成建模

• 发表时间: 2013
• 期刊: Proceedings of Interspeech 2013
• 作者: Hirokazu Kameoka;Kota Yoshizato;Tatsuma Ishihara;Yasunori Ohishi;Kunio Kashino;Shigeki Sagayama
• 通讯作者: Shigeki Sagayama

音声合成モデル学習装置、方法、及びプログラム

语音合成模型学习装置、方法及程序

• 发表时间: 2013