[Paraview] Gaussian splatter filter

Celia Bremer fuwsle at hotmail.com
Tue Oct 9 01:39:28 EDT 2012



Hello,

I have point coordinates of a bunch of atoms and I'm trying to visualize the pore space - i.e., the volume that's a certain distance away from the point coordinates. As suggested here, I've been using Gaussian splatter, then isovolume, and choosing thresholds for isovolume by eye to find those that match the sizes of point sprites. (When I set point sprites as spheres, the radius is in the same units as xyz coordinates, so I know what a point sprite sphere of a given size means physically.) But now, I need to stop eyeballing it and to find the splatter values that correspond to the radii that I'm interested in.

The unscaled output of Gaussian splatter is from 0 (empty space) to 1 (point coordinate). I thought I could figure out the upper threshold by plugging e.g. .5, my desired radius, into the Gaussian distribution function:

f(x) = ScaleFactor * exp( ExponentFactor*((.5/Radius)^2) ) 

But, with default exp factor (-5) and radius (.1) that I'm using, that gives 5E-55, when the value that looks right in PV is around .9. I don't know how splat accumulation mode is factored into the calculation, but I turned off elliptical splats to make it simpler. I'd appreciate any help.

Thank you.

Celia 		 	   		  


More information about the ParaView mailing list