Enhancing vertical localization with image-guided selection of non-individual head-related transfer functions

Abstract

A novel approach to the selection of generic head-related transfer functions (HRTFs) for binaural audio rendering through headphones is formalized and described in this paper. A reflection model applied to the user's ear picture facilitates extraction of the relevant anthropometric cues that are used for selecting two HRTF sets in a database fitting that user, whose localization performances are evaluated in a complete psychoacoustic experiment. The proposed selection increases the average elevation performances of 17\% (with a peak of 34\%) with respect to generic HRTFs from an anthropomorphic mannequin. It also significantly enhances externalization and reduces the number of up/down reversals

The Impact of an Accurate Vertical Localization with HRTFs on Short Explorations of Immersive Virtual Reality Scenarios

“…We provided users with personalized HRTFs that were individually selected based on anthropometric data of the external ear (also known as pinna) [12,16] in order to provide reliable elevation cues for a subsequent free VR exploration. Based on screening results, we defined a criterion which allowed to cluster users into good and bad localizers; we assumed that our personalization method provided reliable spectral cues in the acoustic domain for all users, thus confining the cause of poor/good localization performances in the non-acoustic domain.…”

A Review of the Capabilities of Current Low-Cost Virtual Reality Technology and Its Potential to Enhance the Design Process

“…Since headphones produce sound directly at the ear, all localization cues must be reproduced virtually. While most of the cues are relatively generic, the HRTF is unique to each person and using a poorly matched HRTF to reproduce the localization cues can cause trouble localizing sounds for the participants [109][110][111]. Thus, for accurately creating virtual sounds getting an accurate HRTF is critical.…”

“…This process is time consuming and unfortunately not very scalable for widespread use [111]; however, research into this area is ongoing. One study has suggested that it may be possible to pick a HRTF that is close enough from a database of known HRTFs based on a picture of the user's outer ear [110,113]. Another research group has been studying the inverse of the standard method, whereby speakers are placed in the user's ears and microphones are placed at various locations around the room.…”

Surround by Sound: A Review of Spatial Audio Recording and Reproduction

“…Due to the fact that HRTF individualization is strongly related to the anthropometry of a person, methods have been proposed for HRTF personalization by choosing a small set of anthropometry features with a pre-trained model [49][50][51][52][53][54]. The training was established based on a direct linear or nonlinear relationship between the anthropometric data and the HRTFs, where the first step is to reduce the HRTF data dimensionality.…”