SPARK is a lightweight DSP library that combines high-level math in Python with SIMD-optimized C kernels. It’s designed for developers who need real-time speed without the overhead of heavy scientific libraries.
⚠️ Work in progress: SPARK is still in early alpha. APIs may change.
- Prototype DSP algorithms quickly in Python.
- Run them at near-native speed with optimized C backends.
- Bridge research code and production kernels without rewriting everything.
- High Performance: SIMD-optimized C implementations for speed and memory efficiency.
- Pythonic API: Simple, intuitive bindings for everyday use.
- Lightweight: Minimal dependencies—easy to embed in larger projects.
- Cross-Platform: Linux and macOS supported out of the box.
- Modern Tooling: Meson + meson-python build system for reliable integration.
Requirements:
- Python ≥ 3.10
- C compiler (GCC or Clang)
- meson + ninja
Early releases will be published soon. For now, build from source.
# 1. Clone the repository
git clone https://github.com/Colahall/SPARK.git
cd SPARK
# 2. Create a virtual environment
uv venv
source .venv/bin/activate
# 3. Install in editable mode with test deps
uv pip install -e ".[test]" --no-build-isolation- Python package (
spark): GPL-3.0-or-later - C core (
libspark): MIT License
This dual-licensing model ensures the core C kernels can be embedded widely, while the Python package remains open under GPL.
- SIMD-optimized EQ and filter kernels
- Cross-platform CI and packaging
- Expanded DSP primitives (convolution, FFT wrappers, etc.)