Abstract:

The ability of musicians and nonmusicians to detect a one-note change from a fixed three-voiced musical piece (the standard) was evaluated. Following previous studies, which had tested only musicians, several potentially important variables were manipulated. Musical context (polyphonic or homophonic) was manipulated between subjects. Melody location (high, middle, or low voice), and thus the standard, was varied for each subject across sessions. Within a session, changes varied in harmonic relatedness (related or unrelated), and voice location (high, middle, or low). Subjects performed a same-different task and indicated their confidence as to whether or not a change had been present. Furthermore, a signal detection analysis of the subjects' pattern of responses was performed to directly measure sensitivity to the presence of changes. As expected, musicians were more confident in their responses and revealed a higher sensitivity to changes than nonmusicians. Interestingly, the effects of musical training on sensitivity to changes did not interact with any variable except musical context. The results will be discussed in terms of general and music-specific aspects of perception and training. [Work supported by AFOSR.]