Perceptual discriminations between complex sounds such as those of speech and music are cued by differences along multiple dimensions, including pitch, loudness, duration, and timbre; these differences may occur simultaneously or sequentially, and may be distributed over a wide range of frequencies. While differences along some target dimensions contribute effectively to extricate the sounds, differences along other dimensions contribute little or nothing, and are therefore irrelevant. Complex-sound discriminations depend not only on the ability to attend to and integrate information arising from target dimensions, but also on the ability to ignore or filter out information arising from irrelevant dimensions. Previous research on perception of target and irrelevant dimensions has employed mainly speeded-classification tasks; the present investigation employs an adaptive, forced choice procedure. On each trial of the experimental task, a standard sound will be followed by two comparison sounds; relative to the standard, both comparisons will contain an irrelevant loudness difference, but only one randomly-chosen comparison will contain also a target pitch difference; human listeners will be asked to determine which comparison contains the target difference. Discrimination of target-irrelevant differences will be studied as a function of three variables: frequency and level separation between target and irrelevant differences, positive or negative polarity of irrelevant loudness differences, and amount of task training. The experiments are expected to shed light on the roles of perceptual grouping, relational pitch cues, and long-term memory on discrimination of target-irrelevant dimensions; in doing so, the experiments will increase our understanding of how people perceive multidimensional auditory displays.
