Install in Python¶

compas_acoustics needs Python 3.9 or newer; the Binaural and Ambisonics methods need 3.11 or newer. Create an environment, for example with conda:

conda create -n acoustics python=3.12
conda activate acoustics

Then install it from PyPI:

pip install compas_acoustics

To install the latest main instead, install it from the git repository:

pip install "compas_acoustics @ git+https://github.com/gramaziokohler/compas_acoustics.git"

Binaural and ambisonics methods¶

The Binaural and Ambisonics methods share one hybrid engine (image sources, edge diffraction and a directional feedback delay network) and differ only in how they render it. They need pyFDN, spharpy and multislope, which compas_acoustics installs on Python 3.11 and newer; on Python 3.9 and 3.10 (Rhino 8's Python) it leaves them out and the methods are not available.

pyFDN brings flamo and with it PyTorch: about 0.5 GB on macOS, several GB with CUDA on Linux. Only pyFDN's training modules use it; the methods run without them.

Check the installation¶

python -c "import compas_acoustics; print(compas_acoustics.__version__)"

Then follow the tutorial.

Development¶

To work on compas_acoustics, install it from a local clone with the "dev" requirements:

git clone https://github.com/gramaziokohler/compas_acoustics.git
cd compas_acoustics
pip install -e ".[dev]"

See Contributing to the documentation for building the documentation.