İsmail Hakkı PARLAK, Yalçın ÇEBİ, Derya BİRANT, Cihan IŞIKHAN

Deep learning for Turkish makam music composition

In this paper, we introduce a new deep-learning-based system that can compose structured Turkish makam music (TMM) in the symbolic domain. Presented artificial TMM composer (ATMMC) takes eight initial notes from a human user and completes the rest of the piece. The backbone of the composer system consists of multilayered long short-term memory (LSTM) networks. ATMMC can create pieces in Hicaz and Nihavent makams in Şarkı form, which can be viewed and played with Mus2, a notation software for microtonal music. Statistical analysis shows that pieces composed by ATMMC are approximately 84% similar to training data. ATMMC is an open-source project and can assist Turkish makam music enthusiasts with creating new pieces for professional, educational, or entertainment purposes.

PDF

___

[1] Kumar NH, Ashwin PS, Ananthakrishnan H. MellisAI - an AI generated music composer using RNN-LSTMs. International Journal of Machine Learning and Computing 2020; 10 (2): 247-252. doi: 10.18178/ijmlc.2020.10.2.927
[2] Oord A, Dieleman S, Zen H, Simonyan K, Vinyals O et al. WaveNet: a generative model for raw audio. ArXiv 2016; arXiv:1609.03499 [cs.SD].
[3] Roberts A, Engel J, Mann Y, Gillick J, Kayacik C et al. Magenta sudio: augmenting creativity with deep learning in Ableton Live. In: International Workshop on Musical Metacreation (MUME); Charlotte, USA; 2019. pp. 1-7.
[4] Huang C, Hawthorne C, Roberts A, Dinculescu M, Wexler J et al. The Bach doodle: approachable music composition with machine learning at scale. ArXiv 2019; arXiv:1907.06637 [cs.SD].
[5] Hadjeres G, Pachet F, Nielsen F. DeepBach: a steerable model for Bach chorales generation. In: Thirty-fourth International Conference on Machine Learning; Sydney, Australia; 2017. pp. 1362-1371.
[6] Deng J, Schuller B, Eyben F, Schuller D, Zhang Z et al. Exploiting time-frequency patterns with LSTM-RNNs for low-bitrate audio restoration. Neural Computing and Applications 2020; 32 (4): 1095-1107. doi: 10.1007/s00521- 019-04158-0
[7] Dieleman S, Oord A, Simonyan K. The challenge of realistic music generation: modelling raw audio at scale. In: 32nd International Conference on Neural Information Processing Systems; Montreal, Canada; 2018. pp. 8000-8010.
[8] Abidin D, Öztürk Ö, Öztürk T. Using data mining for makam recognition in Turkish traditional art music. Journal of the Faculty of Engineering and Architecture of Gazi University 2017; 32 (4): 1221–1232 (in Turkish with an abstract in English). doi: 10.17341/gazimmfd.369557
[9] Yarman O. 79-tone tuning & theory for Turkish maqam music. PhD, İstanbul Technical University, İstanbul, Turkey, 2008.
[10] Şentürk S, Serra X. Composition identification in Ottoman-Turkish makam music using transposition-invariant partial audio-score alignment. In: 13th Sound and Music Computing Conference (SMC 2016); Hamburg, Germany; 2016. pp. 434-441.
[11] Bozkurt B, Ayangil R, Holzapfel A. Computational analysis of Turkish makam music: review of state-of-the-art and challenges. Journal of New Music Research 2014; 43 (1): 3-23. doi: 10.1080/09298215.2013.865760
[12] Parncutt R, Reisinger D, Fuchs A, Kaiser F. Consonance and prevalence of sonorities in western polyphony: roughness, harmonicity, familiarity, evenness, diatonicity. Journal of New Music Research 2018; 48 (1): 1–20. doi: 10.1080/09298215.2018.1477804
[13] Kızrak MA, Bolat B. A musical information retrieval system for classical Turkish music makams. Simulation 2017; 93 (9): 749-757. doi: 10.1177/0037549717708615
[14] Karaosmanoğlu MK. A Turkish makam music symbolic database for music information retrieval: Symbtr. In: 13th International Society for Music Information Retrieval Conference (ISMIR); Porto, Portugal; 2012. pp. 223–228.
[15] Wright O, Turabi AH. Klasik Türk musikisi’nde çargah: tarih-teori çelişkisi. Marmara Üniversitesi İlahiyat Fakültesi Dergisi 2001; 21 (2): 81-104 (in Turkish).
[16] Öztürk O. How was the traditional makam theory westernized for the sake of modernization. Rast Müzikoloji Dergisi 2018; 6 (1): 1769-1787. doi: 10.12975/pp1769-1787
[17] Aydemir M. Türk Müziği Makam Rehberi. İstanbul, Turkey: Pan Yayıncılık, 2010 (in Turkish).
[18] Şentürk S, Chordia P. Modeling melodic improvisation in Turkish folk music using variable-length markov models. In: 12th International Society for Music Information Retrieval Conference (ISMIR 2011); Miami, FL, USA; 2011. pp. 269–274.
[19] Gönül M. A Look at the representation, statement and classification of Turkish music rhythms. İslam, Sanat, Tarih, Edebiyat ve Musikisi Dergisi (İSTEM) 2015; 25 (1): 31-46 (in Turkish with an abstract in English).
[20] Şenocak E. Tarihi süreç içinde Türk müziğinde şarkı formu. MSc, Haliç University, İstanbul, Turkey, 2012 (in Turkish).
[21] Tüfekçioğlu S. The new music forms added to Turkish music with Tanzimat and the theory-practice dilemma of these forms. Atatürk Üniversitesi Sosyal Bilimler Enstitüsü Dergisi 2019; 42 (1): 173–182 (in Turkish with an abstract in English). doi: 10.32547/ataunigsed.523556
[22] Hananoi S, Muraoka K, Kiyoki Y. A music composition system with time-series data for sound design in nextgeneration sonification environment. In: International Electronics Symposium (IES); Denpasar, Indonesia; 2016. pp. 380-384.
[23] Briot JP, Hadjeres G, Pachet F. Deep Learning Techniques for Music Generation. New York, NY, USA: Springer International Publishing, 2020.
[24] Xu X. LSTM networks for music generation. ArXiv 2020; arXiv:2006.09838 [eess.AS].
[25] Wu J, Hu C, Wang Y, Hu X, Zhu J. A hierarchical recurrent neural network for symbolic melody generation. IEEE Transactions on Cybernetics 2020; 50 (6): 2749-2757. doi: 10.1109/TCYB.2019.2953194
[26] Yang LC, Lerch A. On the evaluation of generative models in music. Neural Computing and Applications 2020; 32 (9): 4773-4784. doi: 10.1007/s00521-018-3849-7
[27] Briot JP, Hadjeres G, Pachet FD. Deep learning techniques for music generation - a survey. ArXiv 2019; arXiv:1709.01620 [cs.SD].
[28] Karaman S. Problems in performing and naming 9 beats tempos in classical Turkish music. Türkiyat Araştırmaları Dergisi 2014; 35 (1): 461-478 (in Turkish with an abstract in English). doi: 10.21563/sutad.187109
[29] Kumar H, Ravindran B. Polyphonic music composition with LSTM neural networks and reinforcement learning. ArXiv 2019; arXiv:1902.01973 [cs.SD].
[30] Weiss K, Khoshgoftaar TM, Wang D. A survey of transfer learning. Journal of Big Data 2016; 3 (1): 1-40. doi: 10.1186/s40537-016-0043-6