Technical University of Denmark
Browse
.H5
Jabra_measurements.h5 (66.42 MB)
IMAGE
Jabra_measurements.png (181.38 kB)
TEXT
README.md (2.27 kB)
1/0
3 files

Sparsely sampled planar sound field in a furnished room (Jabra - GN Audio)

dataset
posted on 2023-09-27, 08:23 authored by Xenofon Karakonstantis

This dataset consists of 136 Room Impulse Responses (RIRs) measured at Jabra offices in Ballerup, Denmark. These RIRs were obtained using a 1/4" GRAS microphone and a single KEF LS50 loudspeaker. The positions of both the microphone and loudspeaker were tracked using a Vicon motion capture system. The Vicon system is mounted on a rack close to the ceiling on a square rack. The tracking area depends on the camera orientation on the side of the tracked object. There are 8 cameras that are placed at the corners of the square and in the middle of the sides. The room was lightly damped and used for meetings and speech communication experiments. It was fully furnished during the measurement process, and the measurement positions were chosen randomly within an approximate 2.6$\times$2.3 m$^2$ space at the end of the room with both microphone and loudspeaker placed at the same height. The loudspeaker was placed in the middle of the room. From the 136 RIRs, 6 were measured in the nearfield of the loudspeaker. These positions are denoted as `ref_mic_grid' in the dataset. The layout of the data set can be seen in the figure 'Jabra_measurements.png'.


The room is of size 8m × 5m × 2.5m with an average reverberation time (T60) of 0.55 s.


The data are available as H5 (Hierarchical Data Formats) with the following structure:

Dataset "Jabra data - HDF5 datastructure"

Dataset.attributes
   └── fs (1) 32000 int
  ├── loudspeaker_position (3x1) float64

   ├── mic_grid (130x3) float64

   ├── ref_mic_grid (6x3) float64

   ├── ref_rirs (6x64000) float64

   └── rirs(130x64000) float64


Use with Matlab: https://mathworks.com/help/matlab/hdf5-files.html 

Use with Python: https://docs.h5py.org/en/stable/

This dataset is part of the project Danish Sound Cluster project titled "Physics-informed  Neural Networks for Sound Field Reconstruction" [url]. Find more datasets here: Project Page

Funding

Ministry of Education and Research (UFS)

Danish Agency for Higher Education and Science

History

ORCID for corresponding depositor