Extract only the outside mesh of a fluid mesh

Forums Technical Discussion Extract only the outside mesh of a fluid mesh

7789 13 2


mzigaib: Member; 978 posts; Joined: April 2008; Offline

Aug. 28, 2012 6:03 p.m.

Sorry the redundant title I don't know how to edit it.

I have a low fluid mesh and inside I have some shapes that is formed by the empty areas of the particles.
I just want to exctract the outside mesh of the fluid, does it make sense? Is there a way to do it easily?

Thanks.

https://vimeo.com/user2163076 [vimeo.com]


mzigaib: Member; 978 posts; Joined: April 2008; Offline

Aug. 28, 2012 9:30 p.m.

Since is hard to explain my problem I am trying to illustrate.

In this case I just want to extract the outside part of my fluid mesh.

I hope I could make sense now.

Attachments:
fluid_surface_graph.jpg (147.0 KB)

https://vimeo.com/user2163076 [vimeo.com]


zdimaria: Member; 390 posts; Joined: Jan. 2012; Offline

Aug. 28, 2012 9:39 p.m.

You should be able to take the surface field and delete any points that are x points away from the surface. I think scott keating covers how to do this in his waterfall tutorial.


mzigaib: Member; 978 posts; Joined: April 2008; Offline

Aug. 28, 2012 11:21 p.m.

zdimaria
You should be able to take the surface field and delete any points that are x points away from the surface. I think scott keating covers how to do this in his waterfall tutorial.

Thanks for the feedback!

I´ll take a look on the waterfall tutorial again, thanks.

The only thing that is kind of specific is that in this case is that all is part of the same surface and I didn't make any simulation this is just the result of a particle mesh from points scattered on a surface, the gap is where the points are placed so it's like a thick sheet of surface surrounding my object, I hope that makes sense.

I´ll take a look on the tutorial I´ll cry for help again if I don´t succeed on this

Cheers!

https://vimeo.com/user2163076 [vimeo.com]


zdimaria: Member; 390 posts; Joined: Jan. 2012; Offline

Aug. 29, 2012 1 a.m.

well that method wont really help in your immediate situation.. you would need to re-sim, save the surface field and then delete the points before you skin. if you can't re sim than you'll need to go about it some other way. what that is off the top of my head i have no idea lol. i will try and think though


mzigaib: Member; 978 posts; Joined: April 2008; Offline

Aug. 29, 2012 3:15 p.m.

Thanks for the help, I tried to use connectivity and delete the just the objects that I don't want and that work at certain point but I need to choose the group name myself, and since my object with the scatter points is going to move a little those groups are messed up, would be nice to just select the outside part in some way

https://vimeo.com/user2163076 [vimeo.com]


zdimaria: Member; 390 posts; Joined: Jan. 2012; Offline

Aug. 29, 2012 3:49 p.m.

i think it will be hard to do right every time procedurally. this is because even if use connectivity, as soon as you get a small splash or bubble, you have a different number of groups. if i had to get this done for a shot with a short deadline, i would probably hit ‘0’ and go frame by frame and hack away in the viewport, and then the next time i had this kind of shot, id cache out the surface field to use.


zarti: Member; 330 posts; Joined: July 2007; Offline

Aug. 29, 2012 4:47 p.m.

dont know how complex meshes are in your case , but this might help ..

or at least gives you another idea you can expand and fit .

.cheers

Attachments:
remove_inner_geo_z.hipnc (75.4 KB)

except the things that cannot be seen , nothing is like it seems .


eetu: Member; 606 posts; Joined: May 2007; Offline

Aug. 29, 2012 7:19 p.m.

That's a neat way of approaching it.

I have too had this problem, mostly when creating proper SDFs for objects.
If the source object is messily constructed, you end up having to use the “Minimum” mode and then you get all sorts of mess inside the object.

What I did was build a small “delete all but the largest continuous piece” tool, it tends to be what is wanted. (I'll send a hip tomorrow)


Pagefan: Member; 519 posts; Joined:; Offline

Aug. 30, 2012 6:44 a.m.

Instead of the for each loop you can just create a partition sop, wire that in the ray sop and then extract the class primitive attribute. Reference this one in the delete sup using prim instead of point. Nice use of the ray sop though, handy fellow…


mzigaib: Member; 978 posts; Joined: April 2008; Offline

Aug. 30, 2012 8:26 p.m.

zarti
dont know how complex meshes are in your case , but this might help ..

or at least gives you another idea you can expand and fit .

.cheers

Wow! That´s really clever!

Thanks Zarti! I think that can solve my problem

https://vimeo.com/user2163076 [vimeo.com]


zdimaria: Member; 390 posts; Joined: Jan. 2012; Offline

Aug. 30, 2012 9:06 p.m.

wow, awesome use of the ray sop! very clever indeed!


tamte: Member; 8839 posts; Joined: July 2007; Online

Aug. 30, 2012 9:33 p.m.

here is another idea
since outside shell has normals pointing outwards and for all inside bubbles they usually point inwards
when you calculate volume of individual connected pieces it will be negative for pieces with normals pointing inwards so you can delete them

Attachments:
delete_bubbles.hipnc (81.3 KB)

Tomas Slancik
FX Supervisor
Method Studios, NY


zarti: Member; 330 posts; Joined: July 2007; Offline

Aug. 31, 2012 5:54 a.m.

glad i could help and contribute .

except the things that cannot be seen , nothing is like it seems .

Quick Links

                    
                        Search links
                        Show recent posts
                        Show unanswered posts