Generalization of head-related transfer function database using tensor-singular value decomposition

Researches on three-dimensional multimedia have been performed actively in recent years. Virtual 3D sound corresponding to virtual image should be provided to implement 3D multimedia with high quality. Head-related transfer function (HRTF) plays a key role in this research area. HRTFs measured in various azimuth, elevation, and distance for each subject are necessary for generating ideal solution. However, it is unpractical to measure all subjects' HRTFs, so various HRTF databases have been built by many researchers. Because HRTF vary considerably from subject to subject, HRTF of dummy head has been used for generic usage. However, mannequin's HRTF showed much worse performance comparing with individual case so this solution may not be regarded as common HRTF. Therefore, this research proposed HRTF generalization based on tensor-singular value decomposition method as one of the HRTF averaging methods. Also, verification with subjective listening test for four subjects is accomplished. Based on the listening test result, vertical perception performance of virtual sounds generated by the proposed method is improved to some extent. HRTF data used in this paper are extracted from Korean HRTF database which is built by the authors.