Abstract
This study examines the timing of grace notes by separating the timing variation caused by the grace note from the expressive timing profile of the "plain" musical excerpt. In addition, scaling behavior of the two kinds of timing variations are examined over a wide, but musical, range of tempi. In an experiment, eight professional pianists performed three fragments of three bars from a Theme by Beethoven. They repeated each fragment eight times, performed it with and without grace notes and in seven different tempi. The tempi were indicated by a metronome before each tempo block and by a mechanical introduction to each repetition of a fragment. Preliminary results show that (for this piece of music) there is a limited number of (consistent) ways to time a grace note. Short grace notes do not scale with tempo, but have a fixed (short) duration, while longer grace notes do scale with tempo. The scaling of longer grace notes is however never proportional to tempo. Interestingly, the timing variation due to absence or presence of a grace note is much more consistent between pianists than the expressive timing patterns are. Also, though the scaling of expressive timing patterns with tempo is complex and non-linear, their scaling behavior is unaffected by the presence or absence of a grace note. These observations suppose that there is some kind of modularity within the timing of notes, such that temporal changes at one level do not necessarily influence all other aspects of timing.
