Dataset of head-related transfer functions measured with a circular loudspeaker array

Abstract

In this paper, we describe a dataset of head-related transfer functions (HRTFs) measured at the Research Institute of Electrical Communications, Tohoku University. The current dataset includes HRTFs for 105 subjects at 72 azimuths  13 elevations of spherical coordinates. Anthropometric data for 39 subjects are also included. The measurement and postprocessing methods are outlined in this paper. These data will be freely accessible for nonprofit academic purposes via the Internet. Moreover, this dataset will be included in an international joint project to gather several HRTF datasets in a unified data format.

Distance-varying filters to synthesize head-related transfer functions in the horizontal plane from circular boundary values

“…12 Overall accuracies achieved with the distance-varying filters in interaural coordinates. The angular bandwidths of filters were limited according to the frequency-dependent threshold in (11) and proposed magnitude-dependent threshold in (12). Overall accuracies were calculated using the spectral distortion (SD) in (13) and normalized circular correlation (CC) in (14).…”

“…(11) and proposed magnitudedependent threshold in (12). Overall accuracies were calculated using the spectral distortion (SD) in (13) and normalized circular correlation (CC) in (14).…”

“…. Synthesis was performed using the methods in interaural and spherical coordinates, with the frequency-dependent band-limiting threshold (FBT) in (11) and the magnitude-dependent band-limiting threshold (MBT) in (12).…”

“…Prior knowledge about the HRTFs at the target positions is not required when applying this model-based method. This is a useful result, since spherical datasets are typically obtained for only one far distance [12]. Nevertheless, optimally distributing sound sources on the sphere is a non-trivial problem [13,14], and obtaining the HRTFs for all required initial positions might not be feasible for many applications.…”

Spatial accuracy of binaural synthesis from rigid spherical microphone array recordings

“…Such HRTF datasets are typically obtained for a spherical array of sound sources at a single distance from the listener's head [6]. Because HRTFs hardly depend on distance when sources are beyond 1 to 1.5 m [7][8][9], the radius of the array is typically taken within or beyond this interval.…”