Grund296fft

Grund296fft is a software library that implements a fast Fourier transform (FFT) algorithm optimized for real‑time audio signal processing. The library was first released in 2019 by Grund Innovations, a German research group specializing in digital signal processing. It is written in C++ with optional bindings for Python, JavaScript and Java, allowing it to be integrated into a variety of development environments.

The core feature of Grund296fft is its memory‑efficient implementation of the Cooley‑Tukey algorithm. It achieves a speedup of up to 30 % compared with the standard FFTW library when processing short, non‑power‑of‑two length data sets, thanks to a custom stage‑ordering technique. In addition, the library provides an in‑place algorithm that reduces the memory footprint to 50 % of what is required by typical out‑of‑place FFT implementations. These characteristics make it suitable for embedded systems and mobile devices where both processing speed and memory usage are constrained.

Grund296fft is distributed under the MIT License, which allows both commercial and academic use. The source code and documentation can be found on GitHub under the repository name grund296fft. The project’s developers maintain a mailing list and a public issue tracker to support users. Continuous integration tests run on Linux, macOS and Windows platforms, and the library provides precompiled binaries for the most common architectures.

In practice, Grund296fft has been adopted by several open‑source audio projects, including the real‑time audio visualizer WaveSculpt and the low‑latency audio plugin editor SoundBridge. Users report that the library’s small API surface and straightforward build process reduce integration time compared with more complex FFT libraries.