Re: fft in spherical coordinates
- From: glen herrmannsfeldt <gah@xxxxxxxxxxxxxxxx>
- Date: Mon, 24 Aug 2009 09:04:02 +0000 (UTC)
Henrietta Denoue <henrietta@xxxxxxxxxxxx> wrote:
< I have a set of data in spherical coordinates (phi, theta, r) and I need
< to subject these to a fft operation.
Why do you want to subject them to an fft? (It matters)
< When I search the literature, I am
< being warned of regular FFT since the spherical data are not defined on
< a regular grid having well-defined equidistant grid cells (uniformely
< sampled), but an irregular grid where (to simplify take a 2-dim. polar
< grid) those with smaller 'r' are clustered together more densely than
< those with higher 'r'.
(snip)
What are the boundary conditions on the data?
My first thought would be to expand in spherical harmonics for
the theta and phi part.
I know for cylindrical coordinates there is a Fourier-Bessel
transform, with sine/cos on the Z axis and theta axis, and
bessel functions on the R axis.
As with a recent discussion here, it depends on the problem
you are trying to solve. If you want to find the vibrational
modes for a solid or hollow sphere, you have to have the right
boundary conditions.
-- glen
.
- Follow-Ups:
- Re: fft in spherical coordinates
- From: Steven G. Johnson
- Re: fft in spherical coordinates
- References:
- fft in spherical coordinates
- From: Henrietta Denoue
- fft in spherical coordinates
- Prev by Date: Re: fft in spherical coordinates
- Next by Date: Re: fft in spherical coordinates
- Previous by thread: Re: fft in spherical coordinates
- Next by thread: Re: fft in spherical coordinates
- Index(es):
Relevant Pages
|