DatasetSRIRACHA

DatasetSRIRACHA is an experimental microphone array dataset generator that builds multi-source scenes by convolving synthetic source signals with measured spatial room impulse responses (SRIRs) from the `SRIRACHA`_ dataset [PKS25]. Multi-source scenes are realized by superposition of individually convolved source signals.

SRIRACHA (Shoebox Room Impulse Response Archive with Varying Absorption) provides SRIRs measured in a rectangular (“shoebox”) room with two absorption conditions. The dataset comprises eight scenarios that combine two source–receiver plane distances, two source arrangements, and two room absorption settings.

Measurement environment and room configurations

The SRIRACHA measurements were conducted in a rigid concrete shoebox room with dimensions \(6.22 \times 3.85 \times 3.07~\mathrm{m}\) (length × width × height). Two room configurations are provided:

• SR: “empty” shoebox room configuration with foam absorbers placed in two corners to control low-frequency decay.

• SRA: shoebox room with additional absorption, realized by placing three absorber walls inside the room.

A simplified geometry of the absorbent configuration is provided with the dataset as an .stl file.

Experimental setup rendering for the absorbent room configuration (SRA). The planar microphone array is mounted in front of a motorized positioning system; absorber walls are shown in brown.

Reverberation time measurement

For both room configurations, reverberation times were measured according to DIN-EN-ISO 3382-2 using an omnidirectional loudspeaker and omnidirectional microphone with an interrupted-noise procedure. The measurement setup (SRA configuration) is shown below.

Photograph of the reverberation-time measurement setup in the SRA configuration (omnidirectional dodecahedron loudspeaker, microphone array, and absorber wall).

Scenarios

SRIRACHA provides eight measurement scenarios. Naming follows:

• SR vs SRA: room configuration (empty vs absorbent),

• 1 vs 2: source–receiver plane distance (near vs far),

• -D suffix: dense local source arrangement.

In AcouPipe, the scenario is selected via the scenario parameter of acoupipe.datasets.experimental.DatasetSRIRACHA.

Available SRIRACHA scenarios (SRIR grids and meta-data)

Scenario	Room configuration	# sources	# mics	dx = dy	dz	Δx = Δy	Notes
SR1	Empty shoebox room (SR)	64 × 64 = 4096	64	146.7 cm	75.3 cm	23.3 mm	Near distance, full grid
SRA1	Absorbent room (SRA)	64 × 64 = 4096	64	146.7 cm	75.3 cm	23.3 mm	Near distance, full grid
SR1-D	Empty shoebox room (SR)	33 × 33 = 1089	64	16.0 cm	75.3 cm	5.0 mm	Near distance, dense local grid
SRA1-D	Absorbent room (SRA)	33 × 33 = 1089	64	16.0 cm	75.3 cm	5.0 mm	Near distance, dense local grid
SR2	Empty shoebox room (SR)	64 × 64 = 4096	64	146.7 cm	147.9 cm	23.3 mm	Far distance, full grid
SRA2	Absorbent room (SRA)	64 × 64 = 4096	64	146.7 cm	148.0 cm	23.3 mm	Far distance, full grid
SR2-D	Empty shoebox room (SR)	33 × 33 = 1089	64	16.0 cm	147.9 cm	5.0 mm	Far distance, dense local grid
SRA2-D	Absorbent room (SRA)	33 × 33 = 1089	64	16.0 cm	148.0 cm	5.0 mm	Far distance, dense local grid

Signal acquisition and post-processing

SRIRACHA SRIRs were acquired using a planar 64-channel microphone array (Vogel spiral, aperture 1.47 m) and a motorized 2D positioning system to place an excitation loudspeaker on a source plane parallel to the array. Excitation used an exponential sine sweep (3 s) covering 100 Hz to 16 kHz; recordings were 6 s long at 51.2 kHz to capture the room decay. The processed signals are resampled to 32 kHz and deconvolved to obtain impulse responses. To reduce storage requirements, SRIRs are truncated based on a global decay criterion (instantaneous average energy decay of at least 60 dB).

Environmental parameters (temperature and humidity) are recorded and used to compute the speed of sound. Source locations are provided as both nominal and corrected coordinates.

Data format and coordinate systems

The dataset is distributed as HDF5 files. Conceptually, each scenario contains an impulse-response tensor of shape (n_sources, n_mics, n_samples) plus geometry and meta-data:

• /data/impulse_response: float32 SRIR array

• /location/receiver: float64 microphone coordinates (shape (n_mics, 3))

• /location/source: float64 corrected source coordinates (shape (n_sources, 3))

• /location/source_raw: float32 nominal source coordinates (shape (n_sources, 3))

• /metadata: speed of sound, temperature, humidity, sampling rate, etc.

Note

Large scenarios (e.g., SR1, SR2, SRA1, SRA2) may be split into multiple HDF5 chunks (e.g., SR1-C1.h5 … SR1-C4.h5) to keep file sizes manageable.

The measurement coordinates are provided in a left-handed coordinate system for compatibility with MIRACLE. The accompanying .stl geometry is defined in a right-handed coordinate system; when using the .stl as reference, mirror the x-axis of the provided measurement coordinates.

Default FFT parameters

The underlying default FFT parameters are:

FFT Parameters

Sampling Rate	fs = 32,000 Hz
Block size	256 samples
Block overlap	50 %
Windowing	von Hann / Hanning

Randomized properties

Several properties are randomized for each generated source case when building multi-source scenes from SRIRACHA. By default, source positions are sampled from the discrete SRIR grid of the selected scenario.

Randomized properties (defaults)

No. of sources	Poisson distributed (\(\\lambda = 3\))
Source positions	Bivariate normal distributed (\(\\sigma = 0.1688 d_a\))
Source strength (\([{Pa}^2]\) at reference position)	Rayleigh distributed (\(\\sigma_R = 5\))
Relative noise variance	Uniform distributed (\(10^{-6}\), \(0.1\))

Example

from acoupipe.datasets.experimental import DatasetSRIRACHA

ds = DatasetSRIRACHA(scenario="SRA2-D")

License and access

SRIRACHA is distributed under CC BY-NC-SA 4.0 (non-commercial). Please ensure your intended use complies with the license terms.