Applying a Single-notch Metric to Image-guided Head-related Transfer Function Selection for Improved Vertical Localization

This paper proposes an image-guided HRTF selection procedure that exploits the relation between features of the pinna shape and HRTF notches. Using a 2D image of a user's pinna, the procedure selects from a database the HRTF set that best fits the anthropometry of that user. The proposed procedure is designed to be quickly applied and easy to use for a user without previous knowledge on binaural audio technologies. The entire process is evaluated by means of (i) an auditory model for sound localization in the mid-sagittal plane available from previous literature, and (ii) a short localization test in virtual reality. Using both virtual and real subjects from a HRTF database, predictions and the experimental evaluation aimed to assess the vertical localization performance with HRTF sets selected by the proposed procedure. Our results report a statistically significant improvement in predictions of the auditory model for localization performance with selected HRTFs compared to KEMAR HRTFs, which is a commercial standard in many binaural audio solutions. Moreover, the proposed localization test with human listeners reflect the model's predictions, further supporting the applicability of our perceptually-motivated metrics with anthropometric data extracted by pinna images.