Page 45 - Spring2022

conducted at this time would lead to a dramatic change in thinking about the factors that affect auditory masking.
Spatial Attributes of Sound
As discussed, Bill views sound source perception as the primary function of our sense of hearing. He has argued that identifying the sources of sound in our environment is paramount to survival (Yost, 2008). From an evolu- tionary perspective, the job of the perceptual system is to make sense of the world so that the organism can interact effectively with it. More specifically, identifica- tion is required for organisms to discriminate predators, prey, and potential mates so that they can act accordingly to survive. But using pitch, temporal, and other cues to deduce that a sound source is a potential predator, for example, would not be especially helpful if we were not able to also identify its location and then avoid it.
Sound source localization arises from our ability to process relatively small differences in the auditory signals between the two ears. A sound coming from the left of a listener will arrive at the left ear sooner in time than it will at the right ear. The sound will also generally be louder at the left ear than the right due to the head shadow. These interaural differences of time (ITDs) and level (ILDs) are the cues used to localize sound sources in the horizontal plane. Bill has contributed a wealth of information to our understanding of these spatial cues in numerous papers spanning over 40 years (e.g., Green and Yost, 1975; Yost and Pastore, 2019). For example, thanks to Bill’s efforts, we better understand ILD and ITD sensitiv- ity across frequency (e.g., Yost and Dye, 1988), by cochlear implant users (e.g., Doorman et al., 2014), and in the presence of time-varying amplitude fluctuations (e.g., Yost et al., 1989).
In addition to facilitating sound source localization, spa- tial cues can also provide additional benefits for detection, discrimination, and identification tasks that occur in the presence of one or more additional concurrent, spatially separated sound sources or maskers. When the task is speech perception, it is often described as solving the
“cocktail party problem,” a term coined by Cherry (1953). There are other terms for the general perceptual bene- fits that arise from spatial separation of sound sources,
including spatial release from masking (SRM) and the masking level difference (MLD).
Bill has made significant contributions to the litera- ture characterizing the MLD, which was first described
nearly simultaneously by both Licklider (1948) and Hirsh (1948). If the same tone is presented to both ears using headphones (described as “homophasic” because the tone has the same phase to both ears) and then one adds a homophasic noise, the signal-to-noise ratio (SNR) can be manipulated so that the listener can just detect the tone. If the phase of the tone is changed in one ear so that it is different than that in the other ear (an “antiphasic” condition), the perceived location of the tone in the lis- tener’s head will change because of the interaural phase delay. Interestingly, the level of the noise will have to be increased to produce the same amount of masking. In this example, the difference in SNR between the homo- phasic and antiphasic conditions is the MLD.
Figure 5 depicts an even simpler and more striking example. In Figure 5, top, the tone and noise are both homophasic and the sad face indicates that the listener is having difficulty detecting the tone. In Figure 5, bottom, the tone has been turned off in one ear and the noise remains homophasic, a condition in which the amount of masking (the MLD) is reduced, as indicated by the happy face. To summarize, simply eliminating the tone in one ear made the tone more easily detected!
The MLD is a particularly elegant example of taking a complex phenomenon (the perceptual benefits of spatially separated targets and maskers) and reducing the problem to its essence so that it can be studied systematically. Since
Figure 5. Conditions in a typical masking level difference (MLD) experiment. Top: both the tone and noise are diotic or homophasic. Bottom: the noise is homophasic while the tone is presented monaurally to the left ear only, or antiphasic. Adapted from Green and Yost (1975).
Spring 2022 • Acoustics Today 45