A Model of Expressive Timing in Tonal Music

During a performance, a pianist has direct control over only two variables, duration and intensity (Seashore, 1938). Other factors such as pitch and timbre are determined largely by the composer and the mechanics of the instrument. Thus expressiveness imparted to a performance lies in the departures from metrical rigidity and constant intensity. In this article, the first of the two variables is considered and it is shown how a duration structure can be generated, corresponding to the rubato in a performance, from the musical structure. The main input to the model is the time-span reduction of Lerdahl and Jackendoff's theory (1977, 1983). Also shown is an interesting analogy between this model and the algorithms of Grosjean, Grosjean, and Lane (1979). Thus the hypothesis that expression is largely determined by musical structure, and the formal parallel between time-span reduction and prosodic structure are given empirical support.