New script idea, Multi-Mesh skinning

Don’t shoot me Theeth This is a new request. :smiley: :smiley: :smiley: :smiley: :smiley:

I have been working with Zbrush, In Zbrush it has a feature called Unified Skinning. When two or more meshes are placed near each other and the featured is enacted it produces a mesh that joins all of the peices together. So in sort in is like Meta balls. But it does not work in live form as meta balls. And it makes the mesh kinda soft . It also takes a bit of time to compute. So it is some sort of different method.

As explained in the manual it forms a mesh over the other meshes then deletes the meshes inside.


It could be used to join hands and arms and other stuff. Also it could make new primitives.


Soft meshes. Python devs could eaisly remedy it with tite control .

High mesh. Again python control stuff

Bad mesh. That is also the problem of meta balls. I have no clue if that could be controlled. You would have to find out. But the knife tool would help create loops so that would not to much of an issue.

Usse: It could be used as a builder of a simple mesh outline, then the mesh could be worked on from there as normal.

Ok now I ask. Does that sound like a good feature that could be done with python ?

I can’t tell you about python, but I think some form of a script that works along the lines of boolean operations. So if 4 or 5 spheres intersect each other, they will be reduced to one mesh with the intersections removed.

Should be possible, not that I can help I’m afraid, I’m useless with python. :expressionless:


It’s not that it doesn’t sound like a good feature, but if nobody is willing to do it, it won’t happen. Personnally, without any clue as to how such a feature might work (I mean technically) and with the current features that I need to finish, there’s a really slim chance that I’ll work on that.

There’s a really higher chance that this will grab the coders’ attention if you have some technicalities, not just stuff we can get from a PR guy (no offense meant).


a sort of hack i have used to get something somewhat like this is to use low poly models, boolean them together then subsurf them, though the boolean isnt very clean with some objects, so it can produce some weird faces.

Well at least it is approved as an fesable idea.

I will post some examples in a bit. As I ‘can’ do this all ready so it does work, but it is in another program and I think it could be done even better in blender.

That sounds like it would be quite interesting to implement.
I’ve kinda toyed with this idea in the past (although not related to Blender - I’m fairly new to Blender), and I’d be interested in having a go at this at some point, if I ever have the time :frowning:

I’m looking forward to seeing the examples.

Could be done by making section cuts of the meshes on all axis and reconstructing the final mesh from that. Really dirty method though.


Okay here are the files.

And here I have some notes:

I do not think that it is Boolean at all but instead a new mix of meatball math.

What do I know !?! .
Ok see the meshes ? They intersect true. But they also mush together like a thick clay was poured over them as a cast. So that is the method that it is doing. It pours a mesh over and around the meshes.

This is proven by the outcome mesh. The mesh is brand new and the vertices are evenly distributed instead of the complicated and never clean boolean math.

From looking at your bunny.blend example, it seems to work exactly like I described it with an added smooting algorithm.

The basis for that is a stupid poly/Box intersection algorithm, the only real “problem” would be the smoothing algorithm.

But like I said, I’m too busy with much more interesting and useful stuff.

A word of advice though. Instead of using techniques like this which create an aweful lot of useless vertice, it would have been so much simplier to model the wings as part of the body to start with. Especially if you want to animate later (Wrong edge loop structure, no repartition of the vertices around the joints, …).


I know I know. Everything that you said is true for the mesh. But this tool could be used for making new primitives and just random art methods. Not for presice modeling methods. And I just thought that with python at the helm, somebody could control the way that it is all created, so instead of the smooth mesh they could make it behave and create loops instead and some way to retail fine details.

I know that you are a very busy coeder. So is anybody else intrigued with the idea ? I saw that artificer was most curious in trying it…

This program, has alsorts of other stufff that blender could use.