Year: 2017
Type: conference-paper
Abstract: The goal of this project is to train a machine to compose Baroque Fugue/Canon by using Long-Short Term Memory architecture (LSTM). LSTM is a type of artificial recursive neural network (RNN), which excels at learning patterns at both long and short time periods. By limiting to particular “styles” of structures and patterns, the problem becomes more tractable. In our study, we focus on the Baroque Fugue/Canons as they are polyphonic music with standard rules and structures to regulate the composing process. A 2-layer bi-directional LSTM network has been designed. With training data of midi files of Fugue/Canon primarily composed by J.S Bach, have been collected from internet and translated into a text file. Length, frequency, intensity and timing will be considered as training features. The checkpoint file with best validation loss will be used to generate output file, and converted back into midi file. The result will be announced at the conference.
Authors: Yihe Chen, Toshiro Kubota
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2775582498