Abstract:
Personalized head-related transfer functions (HRTFs) are essential for presenting authentic spatial audio through binaural rendering. However, measuring personalized HRTF...Show MoreMetadata
Abstract:
Personalized head-related transfer functions (HRTFs) are essential for presenting authentic spatial audio through binaural rendering. However, measuring personalized HRTFs for every user is a tedious task and requires specialized equipment. This paper presents an easy and efficient method for obtaining personalized magnitude response of HRTFs. It treats the problem of HRTF synthesis as finding the sparse representation of the anthropometric features of the new listener with respect to anthropometric features of the user set in the CIPIC database. Unlike the previous sparse representation methods, our method assigns different weights to different anthropometric features depending on their relevance. We compared our approach with state of the art sparse representation and closest-match based approaches. The results show that our approach outperforms the previous approaches resulting an average spectral distortion value of 5.53 dBs between synthesized and actual HRTFs for all users present in the CIPIC database.
Published in: 2017 International Conference on 3D Immersion (IC3D)
Date of Conference: 11-12 December 2017
Date Added to IEEE Xplore: 11 January 2018
ISBN Information:
Electronic ISSN: 2379-1780