A total of 180 participants, namely three groups of children (8, 10, 12 years old) and a group of young adults participated in a two session experiment in which timing accuracy was tested under single- and dual task conditions. Half of the children received formal training at a musical instrument, the others did not play an instrument. Young adults were either untrained, amateurs, or professional musicians (mostly drummers, N= 20 in each group). Timing tasks required isochronous drumming on digital equipment at target durations of 450, 600, and 750 ms. The cognitive task was a spatial serial memory task, the difficulty level was individually adjusted to a level of 80% accuracy in Session 1. Timing variability increased linearly as a function of target duration with steeper increases for novices compared to amateurs and best performance in experts. In addition timing accuracy increased with age. Effects of musical training became LARGER with age as would be expected from a perspective of accumulated training and resulting skill benefits. Under dual-task conditions, memory performance deteriorated (i.e. significant dual-task costs), however, effects were similar across age groups and skill-levels, most likely because of the individual adjustment of task-difficulty. Timing accuracy was also lower during simultaneous memory performance and this effect was pronounced at slower tempos suggesting that timing becomes increasingly ‘cognitive’ (i.e. involves higher-level processes) at longer durations.