Computational Models of Expressive Music Performance: The State of the Art

Abstract

This contribution gives an overview of the state of the art in the field of computational modeling of expressive music performance. The notion of predictive computational model is briefly discussed, and a number of quantitative models of various aspects of expressive performance are briefly reviewed. Four selected computational models are reviewed in some detail. Their basic principles and assumptions are explained and, wherever possible, empirical evaluations of the models on real performance data are reported. In addition to these models, which focus on general, common principles of performance, currently ongoing research on the formal characterisation of differences in individual performance style are briefly presented.

Transcribing Debussy's Syrinx dynamics through Linguistic Description: The MUDELD algorithm

“…It is precisely by this shaping that the music does sound alive and expressive rather that dull and plain [1,2]. However, those deviations are just vaguely indicated in the music sheet -and in many occasions, they are not even indicated at all-and therefore it is the performer's duty to employ them according to her own aesthetic criteria and experience.…”

“…Widmer and Goebl [1] precisely defined expressive music performance as "the deliberate shaping of the music by the performer, in the moment of playing, by means of continuous variations of parameters such as timing, loudness or articulation". Advances in computational power have enabled the management and analysis of audio recordings through computers, and therefore, they have propitiated the computational modeling of music performance by means of analyzing music recordings.…”

An evaluation of linear and non-linear models of expressive dynamics in classical piano and symphonic music

“…A broader overview of computational models of expressive music performance can be found in Widmer and Goebl (2004).…”

“…On the one hand, there is an interest in understanding the cognitive principles that determine the way a musical piece is performed (Clarke 1988;Palmer 1997) such as the effects of musical imagery in the anticipation and monitoring of the performance of musical dynamics (Bishop et al 2014). On the other hand, computational models of expressive music performance attempt to investigate the relationships between certain properties of the musical score and performance context with the actual performance of the score (Widmer and Goebl 2004). These models can serve mainly analytical purposes (Widmer 2002;Windsor and Clarke 1997), mainly generative purposes (Teramura et al 2008), or both (Grindlay and Helmbold 2006;De Poli et al 2001;Grachten and Widmer 2012).…”

Measuring Ensemble Synchrony through Violin Performance Parameters: A Preliminary Progress Report

“…The problem of expressive performance analysis can be stated as follows: given a score S 1 and a recorded performance of that score E 1 , performance analysis is carried out by measuring the deviations between S 1 and E 1 in terms of timing (onset times and note durations), dynamics, articulation (vibrato, tremolo, overall timbre etc.) and, depending on the instrument with which the piece is performed, intonation.…”