Surface Plots and FFT processing

[pmar]

Hello there,

I'd like to use imagemagick for my radar and signal processing classes, particularly its FFT functionality. I am going through the documentation and examples on imagemagick & FFTs, and everything works great.

However, I came across an alternative web-page on FFTs, that seems to be an older version of whats currently on the main site. And, I'd also like to be able to produce surface plots, given there, here's the link to an example of the plots I am thinking about: link.

Could be that I am missing something, and I know that I am referencing to the links outside the main imagemagick site, but I'd appreciate if someone could give a quick pointer, on how those surface plots are being created.

Thanks a lot & Cheers,
Petar

[fmw42]

The two FFT documentation pages were written first by me when I first helped develop the FFT tools in IM (see link below) and then Anthony modified it for his use on his Examples pages. There are some difference in what is emphasized in each, so it might be good to review both. At some point, I need to update my page as there are now easier ways to generate the jinc function that is used for lens defocus as it has been added directly to -fx, since I wrote my own script to compute it.

In my version, I made 3D plots using some tools in ImageJ. see http://rsbweb.nih.gov/ij/ and its plugins. The FFT examples were done in IM.

[pmar]

Thanks #1: just tried ImageJ , it works...

Thanks #2: for a great tutorial like document on ImageMagick vs FFTs.

Cheers,
\p

[anthony]

Essentially Fred's version concentrates on mathematical meanings of the FFT processing. While I tried to concentrate on the specific image processing aspects, through there is a lot of overlap.

I also have a sub-page on various attempts to do multiplication and division of FFT images using both non-HDRI, and HDRI (floating point) versions of IM, using the magnatude/phase verses complex number image pairs of the FFT forms, and using either slow FX, or faster but more steps using image composition.

Eventually these extra functions (FFT math) should be available as -fourier methods, but it is getting a programmer with time and interest to go in and work on them that has so far prevented it.