Guitar Hero III and similar games potentially offer a vehicle for improvement of musical rhythmic accuracy with training delivered in both visual and auditory formats and by use of its novel guitar-shaped interface; however, some theories regarding multimedia learning suggest sound is a possible source of extraneous cognitive load while playing so players may score higher with sound turned off. Also, existing studies have shown that differences in the physical format of interfaces affect learning outcomes. This study sought to determine whether (a) the game’s audio content affects rhythmic accuracy, and (b) the type of game controller used affects learning of rhythmic accuracy. One hundred participants were randomly assigned in approximately equal numbers (ns = 25) to the four cells of a 2x2 between-subjects design. The first variable was the audio content of the game with two levels: on or off. The second variable was the type of game controller: the standard guitar-style controller or tablet interface. Participants across all conditions completed a pre- and post-test with a system that required them to tap along with repeated rhythmic patterns on an electronic drum pad. Statistical evidence showed better outcomes with a tablet controller with respect to input time error, reduction of extra notes played, and reduction of missed notes; however, the guitar-style controller produced superior outcomes in terms of avoiding missed notes and was associated with higher satisfaction by participants. When audio was present better outcomes were achieved at multiple factor-levels of reduction of missed responses, but superior outcomes in input time error were seen without audio. There was no evidence to suggest an interaction between controller type and the presence or absence of audio.