What is it about?
Musical scores help us visually comprehend music. We design a new representation of musical scores that is familiar to AI. This allows transformer models, which are excellent at handling languages and images, also to handle musical scores. Take a look at the project site on the right!
Featured Image
Photo by Gerold Hinzen on Unsplash
Why is it important?
Musical scores need various musical symbols (e.g., key signatures) and attributes (e.g., voice). Our model handles these elements comprehensively and also has a surprisingly high ability to estimate them. This means that AI can now handle music in the form of scores as well.
Perspectives
What AI can handle is expanding. I am very pleased to add musical scores to the list, as a music researcher. Many extensions are possible, such as transcription of music audio into notation, full-scratch score generation, or music analysis based on scores. I am looking forward to the future development of the research and would like to expand it further myself.
Masahiro Suzuki
Read the Original
This page is a summary of: Score Transformer: Generating Musical Score from Note-level Representation, December 2021, ACM (Association for Computing Machinery),
DOI: 10.1145/3469877.3490612.
You can read the full text:
Resources
Contributors
The following have contributed to this page
