Head-related transfer functions (HRTFs) describe how an ear receives sound from various directions in space. HRTFs are unique to each listener and are an important part of audio systems in virtual and augmented reality. HRTFs are usually measured acoustically, for which a sophisticated laboratory equipment and long measurement time are required. However, HRTFs can also be numerically calculated from a surface mesh of listeners head and pinna. These meshes require geometric details of the pinna which are difficult to capture. Additionally, meshes obtained by scanners can be corrupted by noise and outliers. Parametric pinna models (PPM) have emerged to tackle this problem by adapting a high-resolution clean generic geometry to the actual listener geometry, however, evaluation of such PPMs requires a database of pinna-only geometries of real human. In this work, we present a pinna-only dataset for both ears generated for 200 subjects from the head-and-torso meshes of the SONICOM HRTF database. The meshes were aligned on the interaural axis and cut-off. These pinna-only meshes were then restitched on heads having various levels of detail and HRTFs were calculated to evaluate the effect of the head details in HRTFs. Our results will be discussed in the light of geometric head details and their effects on the accuracy and efficiency of HRTF individualization.
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=22562
(929KB)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
