Performance oriented FFT

Worth checking out. Quoting the FFTW Home Page:

FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions, of arbitrary input size, and of both real and complex data (as well as of even/odd data, i.e. the discrete cosine/sine transforms or DCT/DST). We believe that FFTW, which is free software, should become the FFT library of choice for most applications.

Now, can it work when you have less than 16KB available RAM?


4 thoughts on “Performance oriented FFT”

  1. A couple of thoughts:
    1. I took a look at their comparisons to older methods, as I recall it only starts to be faster with very large N, in the range of 1 million or larger.
    2. I’d guess from (1) that it’s not particularly amenable to 16 KB of RAM, though not sure about that.
    3. It’s not “free” if you want to use it commercially, and other methods are less encumbered by licensing. With FFTW, you have to negotiate with MIT to get a license.

  2. Try Numerical Recipes in C (or C++, I can’t remember). They have several versions, including one for single-dimensional signals like audio, and that as I recall is fairly small. Their license allows free use, as long as you don’t expose the routine’s access externally to a user, for general use. Their book is well worth it for the documentation, and they’ve been around for a long time so their routines are well-debugged. A web search will turn up other approaches.


What do you think?

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s