What is it about?
Music generation based on a previous piece of music faces the problem of not keeping the structure. This structure includes repeats of similar music patterns. Graph-based mutation is a method to keep the structure more consistent than direct mutation. Graph-based mutation is a method which mutates a graph representing the melodic structures in the existing piece of music, rather than mutating the notes directly which can destroy music relationships within the music. In this study we show how graph-based mutation works better and needs less number of mutations to have enough changes in Irish folk tunes.
Featured Image
Photo by Techivation on Unsplash
Why is it important?
This study can help improve music generation systems by generating music with a more real structure. Structures created by these methods are similar to those created by humans more than the other method.
Perspectives
This article has been an absolute pleasure to write as it can have an indirect impact on AI music education systems, data augmentation, musicological analysis, and so on. It would be great to see those effects in future works.
Maziar Kanani
National University of Ireland - Galway
Read the Original
This page is a summary of: Graph-Based Mutations for Music Generation, July 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3583133.3596318.
You can read the full text:
Contributors
The following have contributed to this page
