Modelling musical dynamics

Abstract

This paper deals with dynamics, i.e. loudness, in music. We developed models to describe and recreate dynamics properties of human musicians' expressive performances within a performance system. Their diversity and variability bears a particular challenge to the flexibility of such models. Therefore, our approach is based on a multi-layered discrimination of different dynamics aspects. Each of which includes certain degrees of freedom to flexibly define different shapings. For evaluation we compare these models with human performed dynamics.