The Multiple-Model High Resolution Head-Related Impulse Response
Database (MMHR-HRIR)

Joachim Thiemann, University of Oldenburg, 2017


This directory contains the files comprising the MMHR-HRIR database,
a set of recordings of the acoustic path between loudspeakers which
are positioned roughly 1.7 m from a Head-and-Torso (HATS) simulator.
The loudspeakers were mounted on a movable platform, and recordings
were made form points on a sphere with 2 degree spacing in azimuth
and elevation.  Above 80 degree elevation, the resolution is reduced
since the arc between points becomes smaller.  The lowest elevation
measured (due to the physical setup) is -64 degrees from the equator.

There are four HATS that were used for recording impulse responses.
- Bruel & Kjaer Type 4128C  (BKwHA)
- HEAD Acoustics HMSII.2 (Head)
- G.R.A.S.~KEMAR (KEMAR)
- AG SigProc Custom (DADEC)

For all HATS, the responses of the in-ear microphone was recorded,
but for the Bruel & Kjaer Type 4128C 6 aditional channels of a
behind-the-ear hearing aid were also recorded.

Each set of recordings comes in a "full" and "short" version.  The
short version is trucated at 6.66 ms to remove a reflection caused
by the recording setup.  The shortened recordings are denoted by 
"_s" in the filename.

The data is distributed as SOFA (Spatially Oriented Format for 
Acoustics, https://www.sofaconventions.org/) files, with the extension 
".sofa".  We also provide ".h5" files, which is the same data in a
generic HDF5 format.  The "loadHRIRnear.m" file is a MATLAB script to
read from this format.  The HDF5 format is considered deprecated, and
should not be used to read the MMHR-HRIR datbase.

For citing this data, please use "A multiple model high-resolution
head-related impulse response database for aided and unaided ears",
by J. Thiemann and S. van de Par, EURASIP Journal on Advances in Signal
Processing, 2019:9, doi:10.1186/s13634-019-0604-x.  The first
description of the database can be found in "Multiple model
high-spatial resolution HRTF measurements", by J. Thiemann and S. van 
de Par, in Proc. DAGA 2015.

The data and this README are published under the Creative Commons 
(CC-BY 4.0) license.