To allow a general-purpose processor and dedicated tone generator device to execute an ensemble with their tone generating timing properly matched with each other, tone synthesizing processing based on performance information is assigned to either the general-purpose processor or the dedicated tone generator device. The general-purpose processor, after completion of the tone synthesizing processing thereby, performs control to cause the dedicated tone generator device to execute the tone synthesizing processing assigned to the device. By the time the dedicated tone generator device starts the tone synthesizing processing, a time necessary for the general-purpose processor to execute the processing has already passed away, so that tones synthesized by the processor and tone generator device can be generated at properly matched timing.