We are excited to share our new paper appearing in the journal Royal Society Open Science, entitled “Rhythmic Qualities of Jazz Improvisation Predict Performer Identity and Style in Source-Separated Audio Recordings”. This was completed by Huw Cheston during his PhD at the CMS, and builds from two earlier publications appearing both in Transactions of the International Society for Music Information Retrieval and Music Perception.
Great musicians have a unique style which makes them immediately recognisable. But how do we know what to listen for when making these judgements? We have developed a machine learning model that automatically extracts data relating to the use of rhythm in jazz recordings, and uses this to identify the performer who is playing. Our model was able to correctly identify the jazz pianist playing on approximately 60% of recordings, solely using the rhythmic qualities of their playing. Looking at how the model makes predictions can additionally help us understand what distinguishes one musician from another. Our work showcases how machine learning can bridge the gap between intuition and quantifiable analysis of art, presenting a paradigm applicable across many domains.
See the paper for more information!
Cheston, H., Schlichting, J. L., Cross, I., &; Harrison, P. M. C. (2024). Rhythmic Qualities of Jazz Improvisation Predict Performer Identity and Style in Source-Separated Audio Recordings. Royal Society Open Science, 11(11). https://doi.org/10.1098/rsos.240920
We have also released an interactive map, enabling the predictions of the model to be explored in the browser: https://huwcheston.github.io/Jazz-Trio-Database/_static/prediction-app.html