Room Acoustics Simulation
RIR Generator
Generates multi-channel room impulse responses (RIRs) using the image source method proposed by Allen and Berkley [1]. A description of the method as well as the generator can be found in [2]. The user can control the reflection order, room dimensions, and microphone directivity. The MATLAB version is a MEX-function compiled from C++ source; the Python version is the official port and shares the same underlying core.
The Python implementation is available here with documentation and can be installed via pip install rir-generator. The MATLAB implementation is available here.
- J.B. Allen and D.A. Berkley, Image method for efficiently simulating small-room acoustics, Journal of the Acoustical Society of America, 65(4), 1979.
- E.A.P. Habets, Room Impulse Response Generator, Technical report, 2010.
DEISM — Diffraction Enhanced Image Source Method
Simulates room transfer functions between transducers mounted on audio devices (e.g., smart speakers) using a modified image source method that incorporates local diffraction and scattering effects around the transducers [3]. Supports arbitrary source and receiver directivities expressed as spherical-harmonic coefficients, angle-dependent reflection coefficients, and convex room shapes (DEISM-ARG) beyond the standard shoebox model.
The Python implementation is available here with documentation and can be installed via pip install deism.
- Z. Xu, A. Herzog, A. Lodermeyer, E.A.P. Habets, A.G. Prinn, Simulating room transfer functions between transducers mounted on audio devices using a modified image source method, Journal of the Acoustical Society of America, 155(1):343–357, 2024.
SMIR Generator (SMIRgen)
Simulates the sound pressure signals observed on a spherical microphone array in a reverberant enclosure [4]. The method employs spherical harmonic decomposition and accounts for scattering from a rigid sphere. The decomposition error can be made arbitrarily small by increasing the number of spherical harmonics used.
The MATLAB implementation is available here.
- D.P. Jarrett, E.A.P. Habets, M.R.P. Thomas and P.A. Naylor, Rigid sphere room impulse response simulation: algorithm and applications, Journal of the Acoustical Society of America, 132(3):1462–1472, 2012.
Monte Carlo RIR Simulation
Simulates room impulse responses for various microphone array architectures — uniform linear arrays, uniform circular arrays, and uniform spherical arrays — placed at randomized positions and orientations inside a room. Uses quaternion-based rotations to ensure uniform angular coverage. Suitable for generating large RIR datasets for training and evaluating deep learning-based speech and audio processing models.
The Python implementation is available here.
Dynamic Scene Generator
Generates the response of a moving sound source or receiver in a reverberant environment. The user specifies source and receiver positions at each discrete time instance; the output is computed by convolving the anechoic source signal with the time-varying room impulse response computed via the image method. Multiple receiver positions can be specified simultaneously.
The Python implementation is available here. The MATLAB implementation is available here.