Evaluating vertical localization performance of 3D sound rendering models with a perceptual metric

The head-related transfer functions (HRTFs) describe individual acoustic transformation that sound sources undergo due to human anatomy before arriving at the left and right tympanic membranes. The resulting spectral modifications are the main localization cues for elevation detection in space. In this paper, synthetic HRTF models able to render the vertical spatial dimension in virtual auditory displays, are evaluated via auditory models. Perceptually-motivated metrics describe the output of 4 virtual experiments that numerically simulate real listening experiments for 20 virtual subjects. The current implementation considers a limited set of parameters for a structural model of the pinna acting as a proof-of-concept of such approach. Accordingly, results confirm that the research framework is a flexible tool for systematic evaluation of different instances of structural model.