[PD] fft on control data

[Thomas Grill]

Hi Charles,

> does anyone know how to take a  real time fft on control data stored in
> a table, as oppose to fft of signal ?
> I tried the vasp.cfft object, but I can't understand why it takes 2
> buffer-name arguments (is the second one assumed to be zero)

vasp.cfft does a complex fft of non-interleaved data (this means there are
separate buffers for real and imaginary points).
If you want to do a real fft you can either set the second buffer to zero
or, more efficient, try vasp.rfft where the transformed output is packed in
the following way: zero-frequency, real part (increasing frequencies),
Nyquist frequency, imaginary part (increasing frequencies). See the
vasp.rfft help file.

best greetings,
Thomas