Abstract
Expressive interpretation forms an important but complex aspect of music, particularly in Western classical music. Modeling the relation between musical expression and structural aspects of the score being performed is an ongoing line of research. Prior work has shown that some simple numerical descriptors of the score (capturing dynamics annotations and pitch) are effective for predicting expressive dynamics in classical piano performances. Nevertheless, the features have only been tested in a very simple linear regression model. In this work, we explore the potential of non-linear and temporal modeling of expressive dynamics. Using a set of descriptors that capture different types of structure in the musical score, we compare linear and different non-linear models in a large-scale evaluation on three different corpora, involving both piano and orchestral music. To the best of our knowledge, this is the first study where models of musical expression are evaluated on both types of music. We show that, in addition to being more accurate, non-linear models describe interactions between numerical descriptors that linear models do not.
Highlights
Performances of written music by humans are hardly ever exact acoustical renderings of the notes in the score, as a computer would produce
We first present the predictive accuracies, and continue with a qualitative analysis of the results, in which we use sensitivity analysis methods to reveal what relations between the basis functions and the expressive dynamics the linear basis models (LBMs) and non-linear basis models (NBMs) models have learned
Further improvements can be made by modeling temporal dependencies, but the results show that the model needs a significant capacity in order to capture the relationship between the basis functions and expressive dynamics
Summary
Performances of written music by humans are hardly ever exact acoustical renderings of the notes in the score, as a computer would produce. Becoming an expert musician takes many years of training and practice, and rather than adhering to explicit rules, achieved performance skills are to a large degree the effect of implicit, procedural knowledge This is not to say that regularities cannot be found in the way musicians perform music. Decades of empirical research have identified a number of factors that jointly determine the way a musical piece is rendered (Palmer 1996; Gabrielsson 2003). Aspects such as phrasing (Todd 1992), meter (Sloboda 1983), and intended emotions (Juslin 2001), all have an effect on expressive variations in music performances
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
