am_fft.h

A single-header library for fast fourier transformations. License: MIT

This is a work in progress!

Current limitations

• 32bit floating point only
• 1D input must be power-of-two sized
• 2D input must be power-of-two sized and square

TODO

• Find and implement a good solution for the in-place non-square matrix transpose to get rid of the square 2D input limitation