heya,

I've been fighting for the last few days with some surfacing of a fluid animation I'm working on.

the animation I'm working on is to archieve an effect very close to this one:
http://vimeo.com/16052801 [vimeo.com]
so you've got a frame of reference.

my main problem now is the surfacing stage.

I can't get rid of the “overlapping sampling radius” (or whatever is it properly called) .. which is shown in the attached png file for a better explaination.
basically once the animation goes on and particles move far away from each other that “in between” geometry will fade off, but I don't want it there in first place.

my simulation has to be very slow, so every problem really pops out. (using time scale 0.1).


any help would be very appreciated! thanks

A part of the problem may be the Average Position method on the surfacer. Try using Metaballs to see if that gets rid of those “diamond” that form in the gaps. Submit a bug for the intermediate “diamond” shapes that form between the lobes if you find it an issue.

You can try to reduce the pscale size for particles that form the tendrils which is real challenging but reducing the pscale will make finer surfaces and reduce the chance of fusing to other neighbouring points. Use a Pointcloud to count nearest neighbour points and threshold that to drive pscale reduction?


This is a challenging problem as two points close together form/fuse to build the surface…

hey Jeff,

thanks for your reply.
I tried calculating the density of surrounding points but I'm not getting easy results, I either loose lot of surface or get back the overlapping radius surface.
Metaball method seems pretty much giving me a similar results in fact, where instead I either have no surface (cause of lack of density) or too buly one ..

basically I'm very happy when surfacing “gluey” style liquids but what's the best advices/experiences tips for “water sheets”/flats type of fluids surfaces? i mean, thin type of sheets …

I guess I will submit a bug, but as you said, its a challenging problem.

Hi,

Depend of your mesh and scene you can try PeakSOP, and scale mesh after meshing. Sometime this can help a lot.

Thanks,

yeah I'm peaking my surface using a custom sop that consider the curvature of the mesh, work pretty well for creating thin surface … but that doesn't make those overlapping sampled geometry disappear, actually they look even worst if I peak my geometry, unfortunately.

In the screenshot I posted there are yellow dots that would represent particles position.. the geometry created between “streams” of particles is the problem I'm talking about.

EDIT: it seems the problem lays in the “filtering” tab.

I'm trying other approaches and I just found out that the _previous_sim_position attribute has the same exact coordinates as the current point positions, shouldn't that attribute contain .. erm, the previous simulation frame positions? that's what I'd assumed.