Computational rhythm analysis from audio recordings
By M. Rocamora

Martín Rocamora talks to ADASP about his recent research on computational rhythm analysis from audio recordings.


Most of the research conducted on information technologies applied to music has been oriented towards mainstream popular music in the so-called “Western” tradition. Although it proved to be effective for various music styles and repertoires, new approaches are needed to deal with other music traditions, such as those from Africa, China, India, or the Arab world. Fortunately, over the last few years there have been increasing efforts devoted to the study of traditional, folk or ethnic music. The computational analysis of rhythm from audio signals remains a challenging task in several cases, for instance, for syncopated or poly-rhythmic music.

This talk offers an overview of the research we conducted over the last few years on computational rhythm analysis from audio recordings, considering the Afro-Uruguayan candombe drumming as a case study. It comprises the creation of datasets, the discovery and analysis of rhythmic patterns, the study of micro-timing and the development of algorithms for beat and downbeat tracking. Besides, it also discusses our current efforts to improve and extend the methods to other music traditions, in particular, to Afro-Brazilian Samba.


Martín Rocamora received the B.Sc., M.Sc. and D.Sc. degrees in Electrical Engineering from the School of Engineering at Universidad de la República, Uruguay. In 2005 he started working as lecturer at the same university, in the School of Engineering and in the School of Music. He worked as Teaching Assistant in Music Technology at the School of Music until 2016. He currently holds a position as Assistant Professor (full-time) in Signal Processing with the Electrical Engineering Department (Universidad de la República). His recent work focus on music information retrieval and computational musicology. His research interests also include digital audio signal processing and machine learning. He is a member of the IEEE (Institute of Electrical and Electronics Engineers) and the AES (Audio Engineering Society).