Research Seminars

Modeling Expression: An Empirical Analysis of Recordings of Brahms’ Cello Sonatas

Ju-Lee Hong
Dept. of Music, Goldsmiths, University of London

Wednesday 6 May 2009, 15:00, Room 105

Abstract

An empirical analysis of musical expression makes discoveries about music as sound reflecting the performer as a re-creator of the musical work. At the same time, recorded music provides the best possible evidence to music history on one hand, and many collaborative possibilities with music information retrieval (MIR) on the other. In this seminar, I present a comparative analysis of expressive timing in each of twelve selected commercial recordings of the two Brahms cello sonatas to identify the influence of pedagogical relationships and reception history in performance styles. In order to clarify why critical considerations were given to placing the boundary between cultural norms and individual expression in my modeling method, the talk will begin with the pros and cons of other approaches of modeling expression, including the KTH system, machine learning, principal component analysis, the time-alignment routine of Sonic Visualiser and the CHARM timescape. Findings based on the similarity rates of individual expression normalised through my modeling method suggest that reception history could have played a crucial role in influencing the expressive timing of the E minor sonata. However, in performing the F major sonata, a fair dissimilarity of timing style is identified between artists, whether the two selected performances are by artists from the same pedagogical relationships or by artists chosen by some other criterion.

Bio

Ju-Lee Hong is currently completing her PhD research on Cello Performing Traditions on Record: An Empirical Analysis at Goldsmiths College, University of London. Her dissertation is co-supervised by Stephen Cottrell (Dept. of Music) and Tim Crawford (Dept. of Computing; Computational Musicology) and she also consults with Christophe Rhodes on statistics. Her research interests lie in “empirical musicology”, with a particular focus on the analysis of expression by measurement, the reception history of recorded music and digital application to music research. Her empirical findings have been published and reported in the major interdisciplinary scientific conferences of music and she also contributes book reviews for the academici; recent items include How to Display Data (Blackwell, 2008) and Music Studies (CUP, 2009).