The FFT operator
The fft operator perform a Fast Fourier Transform on the data.
Usage:
p0y = fft(q0)
The fft operator only works on real float scalar .
The output of the fft operator is array/spectra .
To see the fft Dialog window
mark the fft with the left mouse button in the
input text line in the igr calculator window.
Then press Edit>Operator. Now the fft Dialog window appears.
As default the selection of parameters is done automatically,
in such a way as to use all the available data with as few
FFT's as possible (one or two).
mode
The user can select Auto or Manual .
In automatic mode you are free to select : mode , window and output .
If mode is changed to Manual you must also select: fft size, real samples and shift .
 fft_size

This is the size of the fft. It has to be a power of two (32, 64, 128, 256, 512 or 1024) .
 real samples

Here you may specify how many of the points used in the fft should be real sample's.
(The real samples are the data from the input quantity.)
The rest of the data points in the fft will be zeros.
You should not have more real sample's than the size of the fft. The advantage with
this method is better frequency resolution in the output spectra.
(The default is to have fft size equal to real samples .)
 shift

Here you may specify how many points in the real
data you should shift for each new fft done.
 window

Here you may specify which, if any, data window to apply to the real sample's.
When applied you will see the shape of the data window
plotted in the white bar in the fft Dialog window .
The data windows can be (,see signal processing documentation for explanation) :
 Rectangular (no data window)
 Hamming
 Blackmann Harris
 Gaussian
 output

The default and most useful is PSD , Power Spectral Density .

 Spectra

The default and most useful is All, i.e. the fft returns all spectra .

In the status line at the bottom of the fft Dialog window we can see how many
input data points we are using, and more important how many Spectra is returned.
Very useful in the igr client is to display the output from the fft operator
as spectrogram.
Updated June 28, 2006 by Bjørn Lybekk