Using the shape parameter in high_pass_filter

17 views
Skip to first unread message

Makrand Ajay Khanwale

unread,
Aug 22, 2025, 8:31:26 PMAug 22
to Dedalus Users
Hello,

I want to use the high_pass_filter to filter out scales only in x and y direction (Fourier Periodic), and not filter in z direction (Chebyshev with no slip) for incompressible Navier Stokes solution.  I am wondering if the shape input parameter is the way to achieve it, if yes, could you please help me on how to use this parameter to filter? 

Thank you so much. 

With sincere regards,
Makrand
 

Keaton Burns

unread,
Aug 23, 2025, 10:58:59 AMAug 23
to dedalu...@googlegroups.com
Hi Makrand,

Yes as the docstring shows here, the filtering functions let you pass tuples of scale factors for filtering anisotropically.

Best,
-Keaton

--
You received this message because you are subscribed to the Google Groups "Dedalus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dedalus-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dedalus-users/95f6d063-0bc1-47a3-a651-4bf2f6923497n%40googlegroups.com.

Makrand Ajay Khanwale

unread,
Aug 25, 2025, 5:35:20 PMAug 25
to Dedalus Users
Hello Keaton,

Thanks for your reply.  I am trying to understand the precise meaning of the numbers that will be provided in the tuple. 

Let us, for example, say I want to do high pass filtering in x and y, and no filtering in z, and my (nx, ny, nz = 64,64,128) and I want to cutoff all scales larger than half of the wavenumbers in x and y, would the appropriate input be [32, 32, 128] or [32, 32, 0]? Apologies if this sounds silly, I am just trying to understand how the numbers are interpreted by the function.

Thanks again for your help.

With sincere regards,
Makrand
Reply all
Reply to author
Forward
0 new messages