The MATLAB script render_sma_to_ambisonics.m demonstrates how to compute ambisonic signals from the signals that are captured by the microphones of a classical spherical microphone array with a rigid spherical baffle.
The underlying concept was presented in the literature in many locations in different variants. The present implementation uses definitions of the involved quantities such that the resulting ambisonic signals ares compatible with software tools like SPARTA and the IEM Plugin Suite. The mathematical formulation is summarized in
J. Ahrens, "Ambisonic Encoding of Signals From Spherical Microphone Arrays," Technical note v. 1, Chalmers University of Technology, Aug. 2022 [pdf].
If you execute the script then the microphone signals from this recording will be encoded into ambisonics. Additionally, a binaural preview will be computed, which is a binaural rendering of those ambisonic signals while assuming that the listener is looking straight. Note that the result of this computation is already included in the repository (out_ambisonics.wav and out_sma_binaural.wav) (Actually, by default, out_ambisonics.wav contains the encoding from the EMA presented here below. That one is much higher order, and I like it even more). You'll need to download the employed HRIRs from here and store them in the subfolder resources (The MATLAB script is going to do that automatically for you.) as well as the SOFA MATLAB API from here for being able to compute the binaural preview yourself. The ambisonic encoding works either way.
The repository comprises also the Reaper project binaural_rendering.rpp that reads the ambisonic signals and renders them binaurally in realtime. This allows for using head tracking if you happen to have a tracker available. The Reaper project requires the IEM Plugin Suite to be installed. Note that an evaluation version of Reaper is sufficient for this. If you use a Supperware headtracker, you’ll need the Bridgehead, which you can obtain from the manufacturer’s website. Go to Bridge settings and set the Profile to IEM Scene Rotator.
In Reaper,
- Go to
Preferences -> Control/OSC/web -> Add - Set the
Control surface modetoOSC (Open Sound Control) - Set
ModetoLocal port [receive only] - Set
Local listen port:8000
Note that the example implementation of the rendering that we provide does not comprise any form of equalization, which is usually useful to mitigate the effects of spherical harmonic order truncation and spatial aliasing. Thamas Deppisch provides a MATLAB implementation of eMagLS on his GitHub repository that does exactly this.
The MATLAB script render_ema_to_ambisonics.m demonstrates how to compute ambisonic signals from the signals that are captured by the microphones of an equatorial microphone array (EMA). The underlying concept was originally presented in
J. Ahrens, H. Helmholz, D. L. Alon, S. V. Amengual Garí, “Spherical Harmonic Decomposition of a Sound Field Based on Observations Along the Equator of a Rigid Spherical Scatterer” in J. Acoust. Soc. Am. 150(2), 2021 [pdf].
The MATLAB script uses a reformulation of the equatorial array solution that is again compatible with software tools like SPARTA and the IEM Plugin Suite. The details are described in
J. Ahrens, "Ambisonic Encoding of Signals From Equatorial Microphone Arrays," Technical note v. 1, Chalmers University of Technology, Aug. 2022 [pdf].
If you execute the scripts then the microphone signals from this recording will be encoded into 7th-order ambisonics and stored in the file out_ambisonics.wav, which is the same file that is mentioned above (it is getting overwritten). Start Reaper after computing the signals so that it loads the updated ones. The binaural preview will be stored in out_ema_binaural.wav.
We will shortly provide a MATLAB implementation of eMagLS for EMAs.
We thank Reality Labs for funding the initial work on the EMA concept.
The content of this repository is licensed under the terms of the MIT license. Please consult the file LICENSE for more information about this license.