Spatial hearing advantages in everyday environments Barbara

Spatial auditory cues are not a dominant factor in human auditory scene analysis (i.e., in " parsing " the sound reaching the ears to determine the number and spectral content of competing sound sources) [1-5]. However, spatial hearing is very important for understanding a target source in an environment that has multiple sound sources [6-9]. Resolving this apparent paradox is critical for understanding how human listeners operate in difficult conditions, for instance when there is a heavy workload and there are competing demands on attention. Such knowledge is very important for designing effective auditory displays and other human-machine interfaces. Traditional views of the benefits of spatial hearing [10] fail to explain this contradiction as well as other observations, such as the 1) relatively poor ability of hearing impaired listeners to parse and understand speech in situations with competing sources and/or reverberation [11], 2) relatively large inter-subject differences in performance on tasks involving " informational masking " compared to tasks in which the masker is dissimilar from the target [11-13], and 3) very large improvements in speech intelligibility that can arise when similar competing talkers arise from different locations compared to when they are in the same location [6-9, 14, 15]. This short paper provides a preliminary conceptual framework that unifies these seemingly contradictory results by isolating and identifying multiple ways in which spatial hearing impacts the ability to listen to competing, simultaneous sound sources. It has long been known that spatial separation of a target from an interfering source (a masker) improves a listener's ability to detect and understand the content of the target (a phenomenon known as " spatial unmasking; " e.g., see [16-18]; recent reviews include [10, 19]). Much of this improvement can be attributed to simple acoustic effects: spatially separating the target and masker generally increases the target-to-masker energy ratio (TMR) at one of the two ears. Because speech intelligibility improves with TMR, the improvement in TMR in one ear leads directly to an improvement in performance. The acoustic TMR varies with frequency because the acoustic interaction of the head and body of the listener with an impinging sound wave varies with the sound wavelength. Thus, the TMR changes more with spatial location of target and masker at high frequencies than at low frequencies. In the most extreme cases (i.e., when one of the sources is very close to the listener), the TMR at moderate frequencies important …