IR64-CAR: A Dataset of Impulse Responses Captured with a Spherical Microphone Array for Automotive Acoustic Research

IR64-CAR: A Dataset of Impulse Responses Captured with a Spherical Microphone Array for Automotive Acoustic Research

The IR64-CAR dataset provides high-resolution impulse responses captured inside a passenger vehicle (Peugeot 208) using an Eigenmike EM64 spherical microphone array (64 capsules). Measurements were taken with a single external loudspeaker positioned at eight different azimuth angles around the car (0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°), simulating external noise sources at ear-height level.All recordings were conducted with the car stationary, windows and doors closed, and in a low-noise outdoor urban environment.

Contents:8 WAV files, one per source direction.Each WAV file contains 64 channels (one per microphone capsule).Sample rate: 48 kHzBit depth: 24 bitsFile Naming Convention: IR64CAR_X.wavWhere X is the azimuthal angle of the source position in degrees (e.g., IR64CAR_0.wav, IR64CAR_45.wav, etc.).

Recording Setup:Microphone: Eigenmike EM64 (spherical microphone array)Array position: Geometric center of car cabin, 1.2 m height (ear level)Source type: Loudspeaker emitting a logarithmic sine sweepSource distance: 1.5 meters from vehicle boundaryEnvironment: Open urban courtyard, low background noise (<35 dBA)

Calibration and Processing:Loudspeaker pre-equalized.Impulse responses extracted via deconvolution of the recorded sweeps.Signals checked for consistency, trimmed to 1 second.

Usage Notes:The dataset is ideal for automotive acoustics research, including spatial audio processing, source localization, beamforming, and in-cabin sound simulation.Users should cite the dataset and original publication if used in academic or commercial projects.

Citation: Please cite the dataset as follows:J. Lopez-Ballester, G. Greco, S. Roger, J. J. Lopez, F. J. Ferri, M. Cobos. "IR64-CAR: A Dataset of Impulse Responses Captured with a Spherical Microphone Array for Automotive Acoustic Research," 11th Convention of the European Acoustics Association, Malaga, Spain, 2025.

This work has been supported by Grant TED2021-131003B-C21 funded by MCIN/AEI/10.13039/501100011033 and by the “EU Union NextGenerationEU/PRTR”, as well as by Grant PID2022-137048OB-C41 funded by MICIU/AEI/10.13039/501100011033 and “ERDF A way of making Europe”. Finally, the authors acknowledge as well the Artemisa computer resources funded by the EU ERDF and Comunitat Valenciana, and the technical support of IFIC (CSIC-UV).