"Out of Timbre Space - How to measure timbre similarity?"

13.12.2019 12:00 - 13:00

Christoph Reuter, Department of Musicology


What is timbre and how is it possible to calculate timbre similarities? Since the 19th century, philosophers, acousticians, physicists and other researchers rack their brains to attain an applicable concept of musical timbre. With the establishment of so-called timbre spaces, where timbre (dis)similarities are visualized as spatial distances, a solution seems to be found with the help of multidimensional scaling in the 1970ies.

In this talk, I will demonstrate that, when it comes to timbre research, those well-established timbre spaces are, however, not able to measure what they should measure, and that we have to search for alternatives. With the rise of signal analysis techniques and music information retrieval many tools for mathematical timbre feature description have been developed. For calculating timbre similarity the Mel Frequency Cepstral Coefficiants (MFCCs) turned out to be a very fast and effective tool for this task in the last years, but they do not shed any light on the features which enables us to detect timbre similarity in everyday life. In order to obtain a comparable but more intuitive way to deal with the calculation/visualisation of timbre similarity, I will introduce the formant concept as a more feature-orientated approach.


Christoph Reuter, born in 1968, studied musicology, German philology and educational science at the University of Cologne, where he received his doctorate in 1996 and habilitated in 2002. After scholarships (GrFG, DFG), assistance and lecturer periods (University of Cologne, University of Vienna, University of Music Franz Liszt Weimar) as well as activities in media and publishing houses (e.g. Schott Music International (Mainz), b.i.b. International College (Hanover), IAMP solutions (Cologne)) he became professor for Systematic Musicology at the University of Vienna in 2008 (since 2016 head of institute).
His research interests include musical acoustics, room acoustics and psychoacoustics, music psychology, sound analysis/synthesis and acoustical signal analysis/music information retrieval, as well as music-related internet and software projects.

Christoph Reuter

Data Science @ Uni Vienna

SR 16

Oskar-Morgenstern-Platz 1, 1090 Vienna, 3th floor