Enabling Individualized Virtual Auditory Space using Morphological Measurments

Virtual auditory space (VAS) refers to the synthesis and simulation of spatial hearing using earphones or a speaker system. High-fidelity VAS requires the use of individualized head-related transfer functions (HRTFs) which describe the acoutic filtering properties of the listener’s external auditory periphery. Because HRTFs are unique for each individual (“the auditory thumbprint”), a primary hurdle in establishing high-fidelity VAS for multimedia systems requires finding a technology that can reliably generate HRTFs matched to an individual listener in an economical fashion. In this work, a generative statistical model of HRTFs for a population of 36 people was employed in a psychoacoustical experiment to examine the sensitivity of human sound localization performance (a measure of VAS fidelity), to individual differences in HRTFs. Additionally, the relationship within the population between an individual's HRTF and the morphology of the individual's external auditory periphery was examined using methods of statistical analysis. The results indicate that the subjects were sensitive to approximately 60% of the individual variations in the population HRTFs and that their localization performance was remarkably accurate even when accounting for only 30% of individual variations. It is also shown that a functional model relating morphological measurements (of the external ear and head) to HRTFs is capable of reliably producing high-fidelity spatial hearing in VAS.