Solid Metaball?

Ive been expiramenting with the simple functions of the metaball. Since it is only in a 2d state is there anyway of making a 3d metaball drivien by the same points used to make the 2d?

Attached is a range of planes making what looks like 3d metaballs but i cannot seem to make them a solid surface... no lofting or curve networking fetures work on rhino.

Attachments:

Metaballxyz.ghx, 195 KB

Replies to This Discussion

Permalink Reply by David Rutten on April 8, 2010 at 1:42pm

Hi Derek,

a single 3D Metaball surface can have any genus, and thus you will never be able to loft or patch a single nurbs surface through a set of metaball sections.

Metaballs are practically always represented as polygonal meshes or Voxel objects. You could divide your sections into points, bake those points and then run the MeshFromPoints command.

--
David Rutten
david@mcneel.com
Poprad, Slovakia

Permalink Reply by Daniel Piker on April 8, 2010 at 2:16pm

Vicente Soler actually posted an all gh definition for a marching cubes type approach ages ago on the old forum.
It was slow, but very impressive to be done without any scripting.

I've messed around on and off with scripting a 'marching tetrahedra' approach for a while, but never got around to finishing it. I can't see myself having the time to finish this off myself at any point over the next month, but maybe some sort of open collaboration is possible. I don't know exactly how this might work, but it could be interesting to try. Anyway, I'll post some of my ideas on this here in case anyone is interested in taking it further.

The best intro to this I know is Paul Bourke's page

My personal preference is for marching tetrahedra over marching cubes because it avoids an awkward ambiguous case and just seems generally much simpler with only 8 possibilities for the way the surface can pass through 1 cell instead of 256.
Using midpoints of edges gives a very chunky looking result, whereas linear interpolation should be good enough for most purposes and is pretty straightforward to do.
For anything other than the coarsest mesh, I think checking for intersections in every single sub-cube would be really slow, so I guess one should use an octree approach to narrow it down.

Anyway, just my 2c

Permalink Reply by Lara Behmoaram on December 7, 2011 at 2:42pm