Spherical Harmonics Based HRTF Datasets: Implementation and Evaluation for Real-Time Auralization
Jan-Gerrit Richter
Martin Pollow
Frank Wefers
Janina Fels
Abstract
Head-related transfer functions (HRTFs) are commonly used to simulate sound events from arbitrary directions in binaural synthesis applications. These HRTFs are usually either measured or simulated with a discrete angular resolution. Using this data directly has several drawbacks,
mainly due to the necessary interpolation between these discrete sampling points as well as simplifications in the distance dependence of the near-field HRTF. This paper describes how spherical harmonics (SH) can be used to process HRTF data with the use in real-time capable auralization
environments. During the generation of the SH coefficients, the optimal transformation center point is used to achieve a compact transformation. Several strategies are introduced to achieve fast filter reconstruction suitable for real-time auralization.