From artificial neural networks to deep learning for music generation: history, concepts and trends
- PDF / 4,948,532 Bytes
- 27 Pages / 595.276 x 790.866 pts Page_size
- 73 Downloads / 252 Views
(0123456789().,-volV)(0123456789(). ,- volV)
S. I : NEURAL NETWORKS IN ART, SOUND AND DESIGN
From artificial neural networks to deep learning for music generation: history, concepts and trends Jean-Pierre Briot1,2 Received: 10 April 2020 / Accepted: 24 September 2020 Ó Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract The current wave of deep learning (the hyper-vitamined return of artificial neural networks) applies not only to traditional statistical machine learning tasks: prediction and classification (e.g., for weather prediction and pattern recognition), but has already conquered other areas, such as translation. A growing area of application is the generation of creative content, notably the case of music, the topic of this article. The motivation is in using the capacity of modern deep learning techniques to automatically learn musical styles from arbitrary musical corpora and then to generate musical samples from the estimated distribution, with some degree of control over the generation. This article provides a tutorial on music generation based on deep learning techniques. After a short introduction to the topic illustrated by a recent example, the article analyzes some early works from the late 1980s using artificial neural networks for music generation and how their pioneering contributions foreshadowed current techniques. Then, we introduce some conceptual framework to analyze various concepts and dimensions involved. Various examples of recent systems are introduced and analyzed to illustrate the variety of concerns and of techniques. Keywords Artificial neural networks Deep learning Music Generation Tutorial Concepts History Trends
1 Introduction Since the mid-2010s,1 deep learning has been producing striking successes and is now used routinely for classification and prediction tasks, such as image recognition, voice recognition or translation. It continues conquering new domains, for instance, source separation2 [10] and text-to-speech synthesis [47]. A growing area of application of deep learning techniques is the generation of content, notably music, the focus of this article. The motivation is in using widely available various musical corpora to automatically learn musical styles and to generate new musical content based on them. Since a few years, there are a large number of scientific papers about deep learning architectures and
& Jean-Pierre Briot [email protected] 1
CNRS, LIP6, Sorbonne Universite´, 75005 Paris, France
2
UNIRIO, Rio de Janeiro, RJ 22290-250, Brazil
experiments to generate music, as witnessed in [3]. The objective of this article is to explain some fundamentals as well as various achievements of this stream of research.
1.1 Related work and organization This article takes some inspiration from the comprehensive survey and analysis proposed by the recent book [3], but with a different organization and material, and it also includes an original historical retrospective analysis. Another related article [4] is an analysis focusing
Data Loading...
