Studying short-term memory within the framework of the working memory model and its associated paradigms (Baddeley, 2000; Baddeley & Hitch, 1974) offers the chance to compare similarities and differences between the way that verbal and tonal materials are processed. This study examined amateur musicians' short-term memory using a newly adapted version of the visual-auditory (V-A) recognition method (Schendel & Palmer, 2007) within the framework of an irrelevant sound paradigm. We report evidence for a modality specific irrelevant sound effect: irrelevant tones disrupted memory for sequences of tones, whilst only irrelevant speech disrupted memory for sequences of letters. These preliminary results suggest that the adapted V-A recognition method will be useful for future parallel investigations of short-term memory for verbal and tonal materials. (Contains 1 table, 4 figures and 3 notes.)