Home/ Questions/Q 8222425
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-07T14:16:33+00:00

I’m using FFTW to analyze elevation data, but I started out just by looking

  • 0

I’m using FFTW to analyze elevation data, but I started out just by looking at a simple 1D fft to make sure I had everything working. It seems to work except for the normalization (scale) factors, which I can’t figure out from the documentation, mostly because my numbers are ridiculously large. For example, when I tried to just take an fft of a cosine wave

1 6.12323e-017 -1 -1.83697e-016 1 3.06162e-016 -1 -4.28626e-016 1 5.51091e-016 -1      -2.44991e-015 etc...

I got the transform to be 

4.33743e-013 1.06036e+267 3.25205e+265 1.05963e+267 6.49675e+265 1.05743e+267 9.72679e+265 1.05379e+267 1.29349e+266    1.04869e+267 1.61138e+266 1.04216e+267 etc...

eg, ridiculously huge numbers.

Any help would be appreciated.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    You are FFT-ing complex numbers, but you initialized your array as normal (without imaginary part) array.

    I went ahead onto their side, and found out that there are NEW functions that provide double to complex and complex to double fft-ing. Neat-o!

    Here: http://www.fftw.org/doc/New_002darray-Execute-Functions.html

    • 0