The high-accuracy calculation of head-related transfer functions (HRTFs) using the voxel-based finite differential time domain (FDTD) method requires a relatively fine grid size on a staggered-structure grid. Consequently, it leads to intensive computational resource utilization and impractical time requirements. This paper presents a boundary indexing method that offers improved memory efficiency for storing model information during calculations. Compared to the full-indexing FDTD method with ??(??3) memory complexity, the proposed method achieves an ??(??2) memory complexity. The proposed boundary-based indexing method is considerably faster than the full indexing method. Combined implementation with the nonuniform-structured grid is discussed as a suitable method for HRTFs calculation, offering additional reductions in memory usage and calculation time. The implementation results demonstrate that the integrated approach can successfully produce HRTFs using the FDTD method with relatively low computational costs.

Authors: Hong, Zhenxiang; Kusano, Tsubasa; Takahashi, Koki; Sun, Chang; Okubo, Kan
Affiliations: Tokyo Metropolitan University; Huawei Technologies Japan K.K.; Tokyo Metropolitan University; Tokyo Metropolitan University; Tokyo Metropolitan University(See document for exact affiliation information.)
AES Convention: 155 (October 2023) Paper Number: 10676
Publication Date: October 25, 2023 Import into BibTeX
Subject: Immersive & Spatial Audio
Permalink: https://www.aes.org/e-lib/browse.cfm?elib=22257

Download Now (913 KB)

This paper is Open Access which means you can download it for free.

Learn more about the AES E-Library

E-Library Location:   /conv/155/10676.pdf

Start a discussion about this Immersive & Spatial Audio!