Parametric spatial audio codec based on Higher-Order Directional Audio Coding (HO-DirAC).
Companion page http://research.spa.aalto.fi/publications/papers/hoac/.
The encoder extracts a set of transport audio channels and metadata. The decoder reconstructs low orders, and resynthesizes high orders from the input parameterization.
The audio transport channel coding uses Opus in discrete channel mode (implemented also here and here). It is also possible to use other audio-codecs.
The codec is currently prototyped for 5th order HOA (and higher), at a total bit rate of ~512 kbit/s for 'low', ~768 kbit/s for 'med', and ~1280kbit for 'high' profiles.
Reference implementation for papers:
[C. Hold, L. McCormack, A. Politis and V. Pulkki, "Perceptually-Motivated Spatial Audio Codec for Higher-Order Ambisonics Compression", IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024.]
[C. Hold, L. McCormack, A. Politis and V. Pulkki, "Optimizing Higher-Order Directional Audio Coding with Adaptive Mixing and Energy Matching for Ambisonic Compression and Upmixing," 2023 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA), 2023.]
[C. Hold, V. Pulkki, A. Politis and L. McCormack, "Compression of Higher-Order Ambisonic Signals Using Directional Audio Coding," in IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 32, 2024.]
Clone the repository
git clone https://github.com/chris-hld/hoac.git
It's a good idea to use conda for the python environment and package management
conda create --name hoac python=3.11 numpy scipy matplotlib portaudio cffi
To use the new environment
conda activate hoac
Then download and install spaudiopy and safpy. For spaudiopy, in the conda environment, you can use for example
pip install spaudiopy
For safpy you need to follow the build instructions in its safpy-README.
Opus can be installed from source here, or through most system package managers. By default, HOAC uses these Python bindings which can be installed for example with
pip install git+https://github.com/chris-hld/opuslib.git@master
If all dependencies are met, an example encoder and decoder shows the basic functionality.
If you would like to explore beyond the provided interfaces please reach out! All function docstrings can be shown e.g. with
In [1]: import hoac
In [2]: hoac.grouped_sector_parameters??
Signature: hoac.grouped_sector_parameters(x_nm, A_wxyz_c, M_grouper, TRANSPOSE=False)
Call signature: hoac.grouped_sector_parameters(*args, **kwargs)
Type: cython_function_or_method
String form: <cyfunction grouped_sector_parameters at 0x7f9d83048fb0>
Docstring:
Sector S parameters from SH signals L, frequency K band G grouped.
Parameters
----------
x_nm : np.ndarray
L x K.
A_wxyz_c : np.ndarray, complex
4*S x L.
M_grouper : np.ndarray
K x G.
TRANSPOSE : np.ndarray, optional
The default is False.
Returns
-------
azi_s, zen_s, dif_s, ene_s, int_s : np.ndarray
S x G, or 3*S x G, or transposedTested against these forks/branches: https://github.com/chris-hld/opus/tree/update_ambi_map3, https://github.com/chris-hld/opusfile/tree/channel-mapping-2-and-3, https://github.com/chris-hld/opus-tools/tree/channels-individual