From: Hans on
Hi all

Is there any script or code around to perform a f-k filtering or velocity filtering ?

Ok, probably it's based on the fft2 command. But may be there is already an existing filter routine allowing to insert velocity intervals for filtering ?

hans
From: Rune Allnor on
On 11 Mar, 16:23, "Hans " <te...(a)gmx.com> wrote:
> Hi all
>
> Is there any script or code around to perform a f-k filtering or velocity filtering ?
>
> Ok, probably it's based on the fft2 command. But may be there is already an existing filter routine allowing to insert velocity intervals for filtering ?

It's probably both a lot faster and simpler if you hack
this up yourself. It's only a matter of manipulating
the filter weights in fk domain, based on the simple
v -> k mapping I can't remember off the top of my head.

Rune
From: Hans on
That would mean for me the need to look deeper into the way, how a 2-D Filter is defined in Matlab.
From: Rune Allnor on
On 11 Mar, 16:47, "Hans " <te...(a)gmx.com> wrote:
> That would mean for me the need to look deeper into the way, how a 2-D Filter is defined in Matlab.

There is no such thing as a '2D filter' in matlab. Only
matrices. You are looking for something as simple as

y = real(ifft2(fft2(x).*H));

where x is your time-space data, H is your fk filter,
and y is the filtered data.

The 'hard' part is to fill in the numbers in H. You can
do that in a couple of minutes, once you have some sort
of idea what you want the filter to do.

Coming up with that idea is hard, though, as it requires
you to actually contemplate what you are up to, rather than
fiddling with ready-made software routines.

Rune
From: Hans on
Jep, Rune You are completely right, in your suspicion, what the hard part is for me :-)

I have an idea about the "dip" in the signals I want to filter out in my data if we look at them in the non transformed data. So it's a linear relation between f and k or x and y, depending how You call the 2 dimensions.

How can I create a filter from my linear equations in the f-k domain ?