| Inverse 3D FFT [message #44601] |
Tue, 05 July 2005 12:39 |
ntigris@gmail.com
Messages: 6
Registered: July 2005
|
Junior Member |
|
|
Hello,
We need to fourier-transform a complex velocity field in wavenumber
space into a real velocity field in physical space. So i'm not sure
about the proper input to the inverse FFT function.
We'll have a cube in k-space, 128 on a side, filled with complex V
[velocity] values. Since we're dealing with a real velocity field in
the physical space, it looks like, in the k-space, we should have
another cube, such that v(-Kx, -Ky, -Kz) = v(Kx, Ky, Kz)* [a property
of the Fourier Transform]. So we end up with a centrally symmetric
system of two cubes in k-space.
Is that right so far?
Now, for the FFT algorithm we need to have all the complex conjugates
(corresponding to k = -127...0) in positions k = 127...255. So our
input to the FFT seems to be two cubes:
(1) kx = 0...127, ky = 0...127, kz = 0...127 with V's
AND
(2) kx = 127...255, ky = 127...255, kz = 127...255 with complex
conjugates of the respective V's.
Would this input produce the correct results?
Thanks a lot!!
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
