Binaural localization in multiple-source reverberant environments is limited inherently by the ambiguities of the acoustic field at the two ears. Without additional assumptions, there is no unique placement of sources and their reflections in the acoustic environment. This fundamental ambiguity limits the utility of generalizations from single-source localization experiments to the problem of localization. We propose to study binaural localization in complex acoustic fields by utilizing the theoretical tools previously developed in other sensing disciplines, e.g., radar, which are also subject to the same data ambiguities. The principles of multitarget tracking radar will be generalized to the binaural localization problem. Experiments are proposed that will measure the quantities relevant to this mathematical model. These experiments will determine the extent to which localization can be appropriately studied using this mathematical approach. We also propose to study binaural localization in complex acoustic fields by designing tools for rapidly assessing and accurately synthesizing arbitrary complex acoustic fields over earphones. Experiments are proposed both in the free-field (anechoic chamber) and over earphones. With such a design tool, technology will no longer limit our ability to ask questions about localization in the normal auditory environment. The contributions of this research lie in expanding our understanding of what conditions give rise to the formation or termination of a perceived acoustic source, developing a theoretical framework within which to study binaural localization in complex fields, and refining the technology of binaural field synthesis.
