I’m building SPARK (Signal Processing Algorithms, Routines, and Kernels), an open-source library of modular, efficient DSP components for low-power embedded audio systems.
The goal is to make it straightforward to design and deploy small, composable audio graphs that fit on MCUs and similar hardware. The project is in its infancy, so there’s plenty of room for experimentation and contributions.
I've been clumsily exploring the space of what kind of DSP you can do without multipliers, because (in hardware) multiplication is one or two orders of magnitude more expensive than addition; see https://nbviewer.org/url/canonical.org/~kragen/sw/dev3/spars... for one example.
The goal is to make it straightforward to design and deploy small, composable audio graphs that fit on MCUs and similar hardware. The project is in its infancy, so there’s plenty of room for experimentation and contributions.
https://github.com/Colahall/SPARK