Update: 08 January 2010
The Discrete Fourier Transform (DFT) translates a spatial-domain image into two frequency (or Fourier) domain images. Mathematically speaking, this a complex transformation. So, depending on the output format, the frequency domain images will be different :
With the order parameter, one can choose where frequencies are located in the resulting files :
The user may choose to process the displayed image, or to process a
file stored on a hard disk.
When processing the previously displayed image, two resulting files
will be created which names will depend on the selected format.
If the format is set to polar,
the files will be named as "modulus.fit"
and "argument.fit". If the
format is set to cartesian,
the names will be "real.fit"
and "imaginary.fit". If there were some files with the
very same names, their content will be overridden.
If the user prefers to directly work on file, he will be able to select
the names of the resulting files.
In both cases, when the processing is complete, the displayed image is
the first output image, i.e., depending on the selected format, the
real part or the modulus of the DFT operator.
The output images are encoded into unsigned 16-bits words. As the DFT result is signed, pixels values are shifted and normalised to fit the [0-32767] interval. The keywords DFT_OFFS and DFT_NORM contain respectively the shift and normalisation values.