Events

Keeping Score: Computational Methods for the Analysis of Encoded “Symbolic” Musical Scores

Centre for Multimodal AI 
Image:

Date: 3 June 2026   Time: 11:30 - 12:30

Location: Engineering G2

This talk presents work in progress from a large project combining the book Keeping Score with an associated software library, AMADS.

The project focuses on computational methods for the analysis of encoded, or "symbolic", musical scores. For the C4DM audience, the talk will highlight new algorithms and ideas from the project's "Step Further" sections, including musically informed approaches to reconstructing fractional values from floating-point errors, a definition of syncopation based on limited opportunities for resolution, computationally robust yet culturally flexible methods for encoding metre, contour reduction of melodic and bass lines, applications of "keyscape" visualisation to other musical parameters, and new categories for harmonic progression analysis.
Many of these methods are already available through the public book, the AMADS code repository, and accompanying notebooks, which provide accessible introductions to the techniques.

Links
Book: doi.org/10.5281/zenodo.14938027
Code library: github.com/music-computing/amads
Python package: pip install amads

Updated by: Iran Roman