HOMULA-RIR: A Room Impulse Response Dataset for Teleconferencing and Spatial Audio Applications Acquired Through Higher-Order Microphones and Uniform Linear Microphone Arrays

Official repository of the HOMULA-RIR dataset and the related article.

HOMULA-RIR is a dataset of room impulse responses (RIRs) acquired using both higher-order microphones (HOMs) and a uniform linear array (ULA), in order to model a remote attendance teleconferencing scenario. Specifically, measurements were performed in a seminar room, where a 64-microphone ULA (Eventide-Polimi eStick) was used as a multichannel audio acquisition system in the proximity of 2 speakers (Genelec 8020), while HOMs (Voyage Audio Spatial Mic) were used to model 25 attendees actually present in the seminar room.

• Read the paper 📄
• Download the dataset ⬇️

Dataset at a glance

• seminar-room measurements for a remote-attendance teleconferencing scenario
• 2 sources, identified as S1 and S2
• 1 64-microphone ULA and 25 HOM positions arranged over 5 rows
• multichannel RIRs at 48 kHz, truncated to 1 s, stored as 32 bit wav files
• calibrated microphone and source positions provided as csv files

RIRs have been recorded at a sample rate of fs = 48 kHz and truncated to a duration of 1 s. They are provided as multichannel wav files, saved at 32 bit per sample. Together with RIRs, we also release acoustically calibrated positions of each microphone capsule.

Sample code

The repository contains two sample scripts for reading and visualizing both RIRs and microphone/source positions:

• homula_rir.py: Python example script
• homula_rir.m: MATLAB example script

The dataset should be downloaded separately from Zenodo. By default, both scripts assume that the dataset is available in HOMULA-RIR; this path can be adjusted directly in the scripts if needed.

Naming convention

File names follow a structured, hyphen-separated convention in which each token conveys specific information about the file content and the associated measurement configuration.

The dataset adopts the following file name patterns:

rir-S<source>-ULA.wav
rir-S<source>-R<row>-HOM<hom>.wav

with the following meaning:

• rir denotes a room impulse response measurement
• S<source> identifies the source, with <source> equal to 1 or 2
• ULA identifies the 64-microphone Eventide-Polimi eStick array
• R<row>-HOM<hom> identifies a Spatial Mic position, where:
  • <row> is the row index, ranging from 1 to 5
  • <hom> is the microphone index within the corresponding row, ranging from 1 to 5

Position files follow the analogous conventions pos-<array>.csv for microphone positions and pos-sources.csv for source positions.

Examples:

• rir-S1-ULA.wav denotes the RIR measured from source S1 to the ULA
• rir-S1-R3-HOM2.wav denotes the RIR measured from source S1 to the Spatial Mic located at row R3, position HOM2

Authors

Federico Miotello
Paolo Ostan
Mirco Pezzoli
Luca Comanducci
Alberto Bernardini
Fabio Antonacci
Augusto Sarti

Politecnico di Milano

Citation

If you use this dataset, please cite:

F. Miotello et al., "HOMULA-RIR: A Room Impulse Response Dataset for Teleconferencing and Spatial Audio Applications Acquired through Higher-Order Microphones and Uniform Linear Microphone Arrays," 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW), Seoul, Korea, Republic of, 2024, pp. 795-799, doi: 10.1109/ICASSPW62465.2024.10626753.

BibTeX:

@inproceedings{miotello2024homula,
  title={HOMULA-RIR: A Room Impulse Response Dataset for Teleconferencing and Spatial Audio Applications Acquired through Higher-Order Microphones and Uniform Linear Microphone Arrays},
  author={Miotello, Federico and Ostan, Paolo and Pezzoli, Mirco and Comanducci, Luca and Bernardini, Alberto and Antonacci, Fabio and Sarti, Augusto},
  booktitle={2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)},
  pages={795--799},
  year={2024},
  organization={IEEE}
}