Now moved to GitHub:
https://github.com/CoreyShay/FFTL
Provides some relatively easier to use than normal templated FFT classes written in C++. This is currently still a work in progress, but the FFT portion (complex and real) both output the correct results, and the code executes very quickly, around only 30% or so slower than the much more difficult to compile FFTW on SSE enabled platforms. Though optimized code only runs on SSE (x86) based platforms, it's a very simple matter to port the existing code to any other platform with different SIMD intrinsics by simply implementing a small number of inlines that map to a small number of intrinsics.
The algorithm currently used is Radix-2, so there's easily room for speed improvements.
Features
- FFT