Accuracy of head-related transfer functions synthesized with spherical microphone arrays

Abstract

The spherical harmonic decomposition can be applied to present realistically localized sound sources over headphones. The acoustic field, measured by a spherical microphone array, is first decomposed into a weighted sum of spherical harmonics evaluated at the microphone positions. The resulting decomposition is used to generate a set of virtual sources at various angles. The virtual sources are thus binaurally presented by applying the corresponding head-related transfer functions (HRTF). Reproduction accuracy is heavily dependent on the spatial distribution of microphones and virtual sources. Nearly uniform sphere samplings are used in positioning the microphones so as to improve spatial accuracy. However, no previous studies have looked into the optimal arrangement for the virtual sources. We evaluate the effects of the virtual source distribution on the accuracy of the synthesized HRTF. Furthermore, our study considers the impact of spatial aliasing for a 252-channel spherical microphone array. The microphone's body is modeled as a human-head-sized rigid sphere. We evaluate the synthesis error by comparison with the target HRTF using the logarithmic spectral distance. Our study finds that 362 virtual sources, distributed on an icosahedral grid, can synthesize the HRTF in the horizontal plane up to 9 kHz with a log-spectral distance below 5 dB.

Design theory for binaural synthesis: Combining microphone array recordings and head-related transfer function datasets

“…The entries of the matrix associated to the linear system are acoustic transfer functions from the positions of microphones to the positions used to obtain the HRTF dataset. Examples of binaural systems based on this approach are the SENZI system [35][36][37][38][39], the virtual artificial head (VAH) [41][42][43][44], binaural beamforming systems [34,45,47,49,52,53], and the implementations reported in [40,56].…”

“…These modal representations enjoy popularity because they provide an encoding and decoding scheme of directional sound information at scalable resolutions [33,34,[45][46][47][48][49][50][51][52][53][54][55]. A directional resolution at a specific scale can be associated with a single parameter that is called the order, which is denoted by # in this section.…”

Implementing Continuous HRTF Measurement in Near-Field

“…The second category of approaches synthesizes NF HTRFs using numerical methods, which include the boundary element method (BEM) [17,18], finite element method (FEM) [19], and finite difference time method (FDM) [20]. These methods usually involve high computational cost and employ non-trivial acquisition processes, such as magnetic resonance imaging (MRI) [18], computed tomography (CT) [18,20,21,22], or 3D scanner [17] to obtain the human listener's geometry.…”

Spatial accuracy of binaural synthesis from rigid spherical microphone array recordings

“…During the last decade, there has been an increasing interest on methods for combining spatial information contained in HRTF datasets with recordings made with RSMAs [10][11][12][13][14][15][16][17][18][19][20]. In particular, the use of representations of HRTF datasets and RSMA recordings, in terms of solutions to the acoustic wave equation at different spatial resolutions (orders), enjoy popularity because they enable scalable encoding and simplify multichannel processing [10,11,[14][15][16]18,19].…”

“…In particular, the use of representations of HRTF datasets and RSMA recordings, in terms of solutions to the acoustic wave equation at different spatial resolutions (orders), enjoy popularity because they enable scalable encoding and simplify multichannel processing [10,11,[14][15][16]18,19].…”