學術產出-Theses
Article View/Open
Publication Export
-
題名 運用學習式適應評估以樂句為單位的自動音樂重組
Phrase-based Automatic Music Recombination Using Learning-based Fitness Evaluation作者 鄭詠儒
Jeng, Yung-Ru貢獻者 沈錳坤
Shan, Man-Kwan
鄭詠儒
Jeng, Yung-Ru關鍵詞 音樂重組
基因演算法
深度學習
Music Recombination
Genetic Algorithms
Deep Learning日期 2023 上傳時間 1-Dec-2023 10:33:26 (UTC+8) 摘要 音樂自動作曲在電腦音樂領域發展行之有年，音樂自動作曲常見的作曲方法 如 Markov Chain、 Evolutionary Algorithm、 Neural Network等。音樂重組是電腦音樂自動作曲的一種方式。輸入兩首歌曲，音樂重組會輸出一首重新組合的歌曲，且希望保留兩首輸入歌曲的風格。 基因演算法是常見的音樂自動作曲方法。現有透過基因演算法的音樂自動重 組研究是以音符作為基本單位的來進行演化。但是以音符為單位的重組可能會破壞音樂的結構。此外，基因演算法往往必須設計適應函數，量化子代的品質以模擬演化的天則。隨著機器學習技術的發展，近年來已有運用學習式適應評估於電腦作曲的研究，以取代適應函數的設計。但針對音樂重組，尚未有學習式適應評估的研究。 本研究旨在提出運用學習式適應評估且以樂句為單位的自動音樂重組。為確保樂曲風格的保留以及音樂的美感，本論文針對輸入的兩首歌曲，判斷樂曲結構並擷取出樂句並以樂句為單位進行基因演算法的演化來生成新的歌曲。此外，本論文運用深度學習中的長短期記憶模型，提出音樂重組的適應度評估模型以確保曲風的遺傳和美感的傳承。本論文邀請測試者評估生成樂曲，以驗證本論文所提出的方法之效果。
Automatic music composition has a long-standing history in the field of computer music, with common composition methods such as Markov Chains, Evolutionary Algorithms, Neural Networks. Music recombination is one approach within computer music composition where two songs are input, and music recombination generates a newly composed song while aiming to preserve the styles of the two input songs. Genetic algorithm has been a prominent method for automatic music composition. Existing research on music recombination using genetic algorithms typically evolves music based on individual musical notes. However, recombining at the note level may disrupt the musical structure. Moreover, with the development of machine learning techniques, recent studies in music composition have explored the use of learning-based fitness evaluation to replace the manual design of fitness functions. However, no research has been done on the learning-based fitness evaluation for music recombination. This thesis aims to propose a phrase-based music recombination approach that utilizes learning-based fitness evaluation. To ensure the preservation of musical styles and aesthetic qualities, this thesis analyzes the structure of music, extracts musical phrases, and employs a genetic algorithm that evolves new songs at the phrase level. Furthermore, this research proposes the fitness evaluation model for music recombination by utilizing Long Short-Term Memory (LSTM) models to ensuring the inheritance of musical style and aesthetics. The experiments evaluated by subjects are performed to show the effectiveness of the proposed approach.參考文獻 [1] M. Majidi & R.M. Toroghi, A Combination of Multi-objective Genetic Algorithm and Deep Learning for Music Harmony Generation, Multimedia Tools and Applications, Vol. 82, 2022. [2] D. Cope, Recombinant Music Using the Computer to Explore Musical Style, IEEE Computer, Vol. 24, No. 7, 1991. [3] S. Majumder & B. D. Smith, Music Recombination Using a Genetic Algorithm, In Proceedings of International Computer Music Conference International Computer Music Association, 2018. [4] Y.-W. Wen & C.-K. Ting, Recent Advances of Computational Intelligence Techniques for Composing Music, IEEE Transactions on Emerging Topics in Computational Intelligence, Vol. 7, No. 2, 2022. [5] A. Horner & D. E. Goldberg., Genetic Algorithms and Computer-assisted Music Composition, In Proceedings of the Fourth International Conference on Genetic Algorithms, 1991. [6] B. L. Jacob, Composing with Genetic Algorithms, In Proceedings of the International Computer Music Conference, 1995. [7] D. Matic´, A Genetic Algorithm for Composing Music, Yugoslav Journal of Operations Research, Vol. 20, No. 1, 2010. [8] M. Marques, V. Oliveira, S. Vieira, & A. C. Rosa, Music Composition Using Genetic Evolutionary Algorithms, In Proceedings of IEEE Conference on Evolutionary Computation, 2000. [9] R.Waschka II, Composing with Genetic Algorithms: GenDash, In Evolutionary Computer Music. Springer, 2007. [10] C.-H. Liu and C.-K. Ting, Fusing Flamenco and Argentine Tango by Evolutionary composition, In Proceedings of IEEE Conference on Evolutionary Computation, 2017. [11] F. H. C. Alvarado, W.-H. Lee, Y.-H. Huang & Y.-S. Chen, Melody Similarity and Tempo Diversity as Evolutionary Factors for Music Variations by Genetic Algorithms, In Proceedings of the 11th International Conference on Computational Creativity, 2020. [12] M. Takano & Y. Osana, Automatic Composition System using Genetic Algorithm and N-Gram Model Considering Melody Blocks, In Proceedings of IEEE Conference on Evolutionary Computation, 2012. [13] C.-K. Ting, C.-L. Wu & C.-H. Liu, A Novel Automatic Composition System Using Evolutionary Algorithm and Phrase Imitation, IEEE Systems Journal, Vol. 11, No. 3, 2017. [14] C. Sulyok, C. Harte & Z. Bodó, On the impact of domain-specific knowledge in evolutionary music composition, In Proceedings of the Genetic and Evolutionary Computation Conference, 2019. [15] D. Eck & J. Schmidhuber, A First Look at Music Composition Using LSTM Recurrent Neural Networks, Technical Report, Istituto Dalle Molle Di Studi Sull Intelligenza Artificiale, IDSIA-07-02, 2002. [16] H.-W. Dong, W.-Y. Hsiao, L.-C. Yang, & Y.-H. Yang, MuseGAN: Multi-track Sequential Generative Adversarial Networks for Symbolic Music Generation and Accompaniment, In Proceedings of AAAI Conference on Artificial Intelligence, 2018. [17] H.-W. Dong & Y.-H. Yang, Convolutional Generative Adversarial Networks with Binary Neurons for Polyphonic Music Generation, In Proceedings of International Society for Music Information Retrieval Conference, 2018. [18] A. Roberts, J. Engel, C. Raffel, C. Hawthorne & D. Eck, A Hierarchical Latent Vector Model for Learning Long-term Structure in Music, In Proceedings of the 35th International Conference on Machine Learning, PMLR, Vol. 80, 2018. [19] Z. Wang et al., Pianotree VAE: Structured Representation Learning for Polyphonic Music, In Proceedings of International Society for Music Information Retrieval Conference, 2020. [20] W.-Y. Hsiao, J.-Y. Liu, Y.-C. Yeh & Y.-H. Yang, Compound Word Transformer: Learning to Compose Full-song Music over Dynamic Directed Hypergraphs, In Proceedings of AAAI Conference on Artificial Intelligence, 2021. [21] A. Uitdenbogerd & J. Zobel, Melodic Matching Techniques for Large Music Databases, In Proceedings of the 7th ACM International Multimedia Conference, 1999. [22] Y.-J. Chen, A Fast Repeating Pattern Finding Algorithm for Music Data: A Human Perceptive Approach, Master Thesis, Department of Computer Science, National Cheng Kung University, 2004. [23] E. Cambouropoulos, The Local Boundary Detection Model (LBDM) and Its Application in the Study of Expressive Timing, In Proceedings of the International Computer Music Conference, 2001. [24] PJP de León, JM Inesta, MIREX 2005: Symbolic Genre Classification with an Ensemble of Parametric and Lazy Classifiers, In Proceedings of Music Information Retrieval Evaluation eXchange, 2005. [25] Q.-Q. Kong, B.-C. Li, J.-T. Chen & Y.-X. Wang, GiantMIDI-Piano: A Large-Scale MIDI Dataset for Classical Piano Music, In Proceedings of International Society for Music Information Retrieval Conference, 2022. [26] Z. Wang, K. Chen, J. Jiang, Y. Zhang, M. Xu, S. Dai, X. Gu, G. Xia , POP909: A Pop-song Dataset for Music Arrangement Generation, In Proceedings of International Society for Music Information Retrieval Conference, 2020. 描述 碩士
國立政治大學
資訊科學系
110753126資料來源 http://thesis.lib.nccu.edu.tw/record/#G0110753126 資料類型 thesis dc.contributor.advisor 沈錳坤 zh_TW dc.contributor.advisor Shan, Man-Kwan en_US dc.contributor.author (Authors) 鄭詠儒 zh_TW dc.contributor.author (Authors) Jeng, Yung-Ru en_US dc.creator (作者) 鄭詠儒 zh_TW dc.creator (作者) Jeng, Yung-Ru en_US dc.date (日期) 2023 en_US dc.date.accessioned 1-Dec-2023 10:33:26 (UTC+8) - dc.date.available 1-Dec-2023 10:33:26 (UTC+8) - dc.date.issued (上傳時間) 1-Dec-2023 10:33:26 (UTC+8) - dc.identifier (Other Identifiers) G0110753126 en_US dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/148473 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系 zh_TW dc.description (描述) 110753126 zh_TW dc.description.abstract (摘要) 音樂自動作曲在電腦音樂領域發展行之有年，音樂自動作曲常見的作曲方法 如 Markov Chain、 Evolutionary Algorithm、 Neural Network等。音樂重組是電腦音樂自動作曲的一種方式。輸入兩首歌曲，音樂重組會輸出一首重新組合的歌曲，且希望保留兩首輸入歌曲的風格。 基因演算法是常見的音樂自動作曲方法。現有透過基因演算法的音樂自動重 組研究是以音符作為基本單位的來進行演化。但是以音符為單位的重組可能會破壞音樂的結構。此外，基因演算法往往必須設計適應函數，量化子代的品質以模擬演化的天則。隨著機器學習技術的發展，近年來已有運用學習式適應評估於電腦作曲的研究，以取代適應函數的設計。但針對音樂重組，尚未有學習式適應評估的研究。 本研究旨在提出運用學習式適應評估且以樂句為單位的自動音樂重組。為確保樂曲風格的保留以及音樂的美感，本論文針對輸入的兩首歌曲，判斷樂曲結構並擷取出樂句並以樂句為單位進行基因演算法的演化來生成新的歌曲。此外，本論文運用深度學習中的長短期記憶模型，提出音樂重組的適應度評估模型以確保曲風的遺傳和美感的傳承。本論文邀請測試者評估生成樂曲，以驗證本論文所提出的方法之效果。 zh_TW dc.description.abstract (摘要) Automatic music composition has a long-standing history in the field of computer music, with common composition methods such as Markov Chains, Evolutionary Algorithms, Neural Networks. Music recombination is one approach within computer music composition where two songs are input, and music recombination generates a newly composed song while aiming to preserve the styles of the two input songs. Genetic algorithm has been a prominent method for automatic music composition. Existing research on music recombination using genetic algorithms typically evolves music based on individual musical notes. However, recombining at the note level may disrupt the musical structure. Moreover, with the development of machine learning techniques, recent studies in music composition have explored the use of learning-based fitness evaluation to replace the manual design of fitness functions. However, no research has been done on the learning-based fitness evaluation for music recombination. This thesis aims to propose a phrase-based music recombination approach that utilizes learning-based fitness evaluation. To ensure the preservation of musical styles and aesthetic qualities, this thesis analyzes the structure of music, extracts musical phrases, and employs a genetic algorithm that evolves new songs at the phrase level. Furthermore, this research proposes the fitness evaluation model for music recombination by utilizing Long Short-Term Memory (LSTM) models to ensuring the inheritance of musical style and aesthetics. The experiments evaluated by subjects are performed to show the effectiveness of the proposed approach. en_US dc.description.tableofcontents 第一章 緒論 9 第二章 相關研究 11 2.1 音樂重組 11 2.2 基因演算法的音樂作曲研究 11 2.3 神經網絡的音樂作曲研究 13 第三章 研究方法與步驟 14 3.1 研究架構 14 3.2 資料處理 16 3.3 初始群眾Initial Population 20 3.4 交配Crossover 21 3.5 突變Mutation 22 3.6 適應評估Fitness Evaluation 24 3.6.1 適應評估指標 24 3.6.2 適應評估模型 25 3.7 選擇Selection 31 第四章 實驗 32 4.1 實驗資料 32 4.2 音樂重組實驗 33 4.2.1 實驗設計 33 4.2.2 實驗結果 34 4.3 適應評估模型實驗 67 4.3.1 實驗設計 67 4.3.2 實驗結果 68 第五章 結論 70 參考資料 71 zh_TW dc.format.extent 3489516 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0110753126 en_US dc.subject (關鍵詞) 音樂重組 zh_TW dc.subject (關鍵詞) 基因演算法 zh_TW dc.subject (關鍵詞) 深度學習 zh_TW dc.subject (關鍵詞) Music Recombination en_US dc.subject (關鍵詞) Genetic Algorithms en_US dc.subject (關鍵詞) Deep Learning en_US dc.title (題名) 運用學習式適應評估以樂句為單位的自動音樂重組 zh_TW dc.title (題名) Phrase-based Automatic Music Recombination Using Learning-based Fitness Evaluation en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] M. Majidi & R.M. Toroghi, A Combination of Multi-objective Genetic Algorithm and Deep Learning for Music Harmony Generation, Multimedia Tools and Applications, Vol. 82, 2022. [2] D. Cope, Recombinant Music Using the Computer to Explore Musical Style, IEEE Computer, Vol. 24, No. 7, 1991. [3] S. Majumder & B. D. Smith, Music Recombination Using a Genetic Algorithm, In Proceedings of International Computer Music Conference International Computer Music Association, 2018. [4] Y.-W. Wen & C.-K. Ting, Recent Advances of Computational Intelligence Techniques for Composing Music, IEEE Transactions on Emerging Topics in Computational Intelligence, Vol. 7, No. 2, 2022. [5] A. Horner & D. E. Goldberg., Genetic Algorithms and Computer-assisted Music Composition, In Proceedings of the Fourth International Conference on Genetic Algorithms, 1991. [6] B. L. Jacob, Composing with Genetic Algorithms, In Proceedings of the International Computer Music Conference, 1995. [7] D. Matic´, A Genetic Algorithm for Composing Music, Yugoslav Journal of Operations Research, Vol. 20, No. 1, 2010. [8] M. Marques, V. Oliveira, S. Vieira, & A. C. Rosa, Music Composition Using Genetic Evolutionary Algorithms, In Proceedings of IEEE Conference on Evolutionary Computation, 2000. [9] R.Waschka II, Composing with Genetic Algorithms: GenDash, In Evolutionary Computer Music. Springer, 2007. [10] C.-H. Liu and C.-K. Ting, Fusing Flamenco and Argentine Tango by Evolutionary composition, In Proceedings of IEEE Conference on Evolutionary Computation, 2017. [11] F. H. C. Alvarado, W.-H. Lee, Y.-H. Huang & Y.-S. Chen, Melody Similarity and Tempo Diversity as Evolutionary Factors for Music Variations by Genetic Algorithms, In Proceedings of the 11th International Conference on Computational Creativity, 2020. [12] M. Takano & Y. Osana, Automatic Composition System using Genetic Algorithm and N-Gram Model Considering Melody Blocks, In Proceedings of IEEE Conference on Evolutionary Computation, 2012. [13] C.-K. Ting, C.-L. Wu & C.-H. Liu, A Novel Automatic Composition System Using Evolutionary Algorithm and Phrase Imitation, IEEE Systems Journal, Vol. 11, No. 3, 2017. [14] C. Sulyok, C. Harte & Z. Bodó, On the impact of domain-specific knowledge in evolutionary music composition, In Proceedings of the Genetic and Evolutionary Computation Conference, 2019. [15] D. Eck & J. Schmidhuber, A First Look at Music Composition Using LSTM Recurrent Neural Networks, Technical Report, Istituto Dalle Molle Di Studi Sull Intelligenza Artificiale, IDSIA-07-02, 2002. [16] H.-W. Dong, W.-Y. Hsiao, L.-C. Yang, & Y.-H. Yang, MuseGAN: Multi-track Sequential Generative Adversarial Networks for Symbolic Music Generation and Accompaniment, In Proceedings of AAAI Conference on Artificial Intelligence, 2018. [17] H.-W. Dong & Y.-H. Yang, Convolutional Generative Adversarial Networks with Binary Neurons for Polyphonic Music Generation, In Proceedings of International Society for Music Information Retrieval Conference, 2018. [18] A. Roberts, J. Engel, C. Raffel, C. Hawthorne & D. Eck, A Hierarchical Latent Vector Model for Learning Long-term Structure in Music, In Proceedings of the 35th International Conference on Machine Learning, PMLR, Vol. 80, 2018. [19] Z. Wang et al., Pianotree VAE: Structured Representation Learning for Polyphonic Music, In Proceedings of International Society for Music Information Retrieval Conference, 2020. [20] W.-Y. Hsiao, J.-Y. Liu, Y.-C. Yeh & Y.-H. Yang, Compound Word Transformer: Learning to Compose Full-song Music over Dynamic Directed Hypergraphs, In Proceedings of AAAI Conference on Artificial Intelligence, 2021. [21] A. Uitdenbogerd & J. Zobel, Melodic Matching Techniques for Large Music Databases, In Proceedings of the 7th ACM International Multimedia Conference, 1999. [22] Y.-J. Chen, A Fast Repeating Pattern Finding Algorithm for Music Data: A Human Perceptive Approach, Master Thesis, Department of Computer Science, National Cheng Kung University, 2004. [23] E. Cambouropoulos, The Local Boundary Detection Model (LBDM) and Its Application in the Study of Expressive Timing, In Proceedings of the International Computer Music Conference, 2001. [24] PJP de León, JM Inesta, MIREX 2005: Symbolic Genre Classification with an Ensemble of Parametric and Lazy Classifiers, In Proceedings of Music Information Retrieval Evaluation eXchange, 2005. [25] Q.-Q. Kong, B.-C. Li, J.-T. Chen & Y.-X. Wang, GiantMIDI-Piano: A Large-Scale MIDI Dataset for Classical Piano Music, In Proceedings of International Society for Music Information Retrieval Conference, 2022. [26] Z. Wang, K. Chen, J. Jiang, Y. Zhang, M. Xu, S. Dai, X. Gu, G. Xia , POP909: A Pop-song Dataset for Music Arrangement Generation, In Proceedings of International Society for Music Information Retrieval Conference, 2020. zh_TW
