Head Related Transfer Functions (HRTFs) capture the binaural information required for correct identification of a sound source position in 3D space. They are individual to each user and heavily depend on the direction of arrival of the sound from the considered source. Acquisition of personalized HRTFs for all possible directions is a lengthy process requiring carefully calibrated measurements which is difficult to apply in practical situations. In this paper we propose a data-driven, machine-learning based approach for the generation of personalized, direction-dense, lowpass-filtered HRTFs computed from a mesh of the users head and from a non-uniform set of dynamic data measurements. We present the different steps of our approach and show results from laboratory experiments. Comparison with a state-of-the-art, BEM-based HRTF generation method confirms the effectiveness of the proposed solution.
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=22498
(19476KB)
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.
