Chord-based stepwise Korean Trot music generation technique using RNN-GAN

Seo-Rim Hwang; Young-Cheol Park

doi:10.7776/ASK.2020.39.6.622

All Issue

2020 Vol.39, Issue 6 Preview Page

Research Article

Chord-based stepwise Korean Trot music generation technique using RNN-GAN RNN-GAN을 이용한 코드 기반의 단계적 트로트 음악 생성 기법

30 November 2020. pp. 622-628

PDF XML

Abstract

This paper proposes a music generation technique that automatically generates trot music using a Generative Adversarial Network (GAN) model composed of a Recurrent Neural Network (RNN). The proposed method uses a method of creating a chord as a skeleton of the music, creating a melody and bass in stages based on the chord progression made, and attaching it to the corresponding chord to complete the structured piece. Also, a new chorus chord progression is created from the verse chord progression by applying the characteristics of a trot song that repeats the structure divided into an individual section, such as intro, verse, and chorus. And it extends the length of the created trot. The quality of the generated music was specified using subjective evaluation and objective evaluation methods. It was confirmed that the generated music has similar characteristics to the existing trot.

Keywords

Music generation

Chord-based generation

Generative Adversarial Network (GAN)

Bidirectional Long Short-Term Memory (BLSTM)

본 논문은 순환 신경망(Recurrent Neural Network, RNN)으로 구성된 적대적 생성 신경망(Generative Adversarial Network, GAN) 모델을 사용하여 자동으로 트로트 음악을 생성하는 음악생성 기법을 제안한다. 제안된 방법은 음악의 뼈대를 담당하는 코드를 만들고, 만들어진 코드 열을 기반으로 멜로디와 베이스(bass)를 단계적으로 생성한 뒤, 해당 코드에 붙임으로써 구조화된 음악을 완성하는 방법을 사용한다. 또한 인트로나 벌스, 코러스 등과 같이 일정 구간으로 나뉘어 구조가 반복되는 트로트 가요의 특징을 적용하여 벌스의 코드 진행으로부터 새로운 코러스 코드 진행을 만들어내고, 다시 해당 코드로부터 멜로디와 베이스를 단계적으로 생성하여 초기에 만들어진 트로트의 길이를 확장한다. 주관적 평가와 객관적 평가방법을 사용하여 생성된 음악의 품질을 측정하였으며, 기존의 트로트가 갖고 있는 음악적 특성과 유사한 음악을 생성함으로 확인하였다.

키워드

음악 생성

코드 기반 생성

적대적 생성 신경망

Bidirectional Long Short-Term Memory (BLSTM)

References

J. Wu, C. Hu, Y. Wang, X. Hu, and J. Zhu, "A hierarchical recurrent neural network for symbolic melody generation," IEEE Trans. on Cybernetics, 50, 2749-2757 (2019). 10.1109/TCYB.2019.295319431796422

Y. Yu and S. Canales, "Conditional lstm-gan for melody generation from lyrics," arXiv preprint arXiv:1908. 05551 (2019).

N. Zhang, "Learning adversarial transformer for symbolic music generation," IEEE Trans. on Neural Networks and Learning Systems, 1-10 (2020). 10.1109/TNNLS.2020.299074632614773

H.-W. Dong, W.-Y. Hsiao, L.-C. Yang, and Y.-H. Yang, "Musegan: Multi-track sequential generative adversarial networks for symbolic music generation and accompaniment," Proc. the AAAI Conf. on Artificial Intelligence, 34-41 (2018).

S. Hochreiter and J. Schmidhuber, "Long short-term m emory," Neural Computation, 9, 1735-1780 (1997). 10.1162/neco.1997.9.8.17359377276

M. Mirza and S. Osindero, "Conditional generative adversa rial nets," arXiv:1411.1784 (2014).

K. Chen, W. Zhang, S. Dubnov, G. Xia, and W. Li, "The effect of explicit structure encoding of deep neural networks for symbolic music generation," Proc. MMRP. 23-24 (2019). 10.1109/MMRP.2019.00022

Information

Publisher :The Acoustical Society of Korea
Publisher(Ko) :한국음향학회
Journal Title :The Journal of the Acoustical Society of Korea
Journal Title(Ko) :한국음향학회지
Volume : 39
No :6
Pages :622-628
Received Date : 2020-10-01
Revised Date : 2020-11-02
Accepted Date : 2020-11-11
DOI :https://doi.org/10.7776/ASK.2020.39.6.622

The Journal of the Acoustical Society of Korea ISSN:1225-4428(Print) 2287-3775(Online) 한국음향학회지

All Issue