Curly kale

Comment

Comment by Nick Tyrer on June 24, 2015 at 5:47am

Ok, good to know. I was expecting it, i just created a mobius with your method. But after all that unfortunately end result just didn't look very good....

Comment by Daniel Piker on June 24, 2015 at 5:15am

Hi Nick,

Nice helicoid!

Yes it makes sense that MeshMachine will fail on any non-orientable surface. Since it uses the Plankton half-edge mesh representation, which relies on consistent face orientations. You could probably get around this by splitting the surface in 2, and setting some fixed points to make sure the vertices still match along the boundary, then rejoining them

Comment by Nick Tyrer on June 24, 2015 at 4:58am

Very fun definition, i've just tried some different forms, Helicoid is a cool effect (yes, in Nervous colours). On a side note David, i was trying to use meshmachine on a mobius strip, but it kept failing. Would you expect that behaviour? I wasn't sure if it was to do with mesh normals etc...

Comment by Simon Schleicher on June 23, 2015 at 10:13pm

this is great Daniel!

Comment by Daniel Piker on June 23, 2015 at 5:08pm

Thanks Roger.

Though it would be possible to just make the mesh in the 1st step even finer toward the edge to get more intricate end results, I do think that actually extending / subdividing / remeshing the topology during the relaxation could be a better way. I suspect that something like this is happing in both nervous system's work and this video that was being discussed earlier.

To do this effectively would probably require scripting - both the Plankton and Kangaroo libraries can be called from scripts, so it is possible, and I think would be an interesting direction of investigation. David Stasiuk has also done some interesting work with mesh growth in Grasshopper.

Comment by Ángel Linares on June 23, 2015 at 9:18am

I know that you, as me like nervous systems work :) Awesome implementation.

Comment by Daniel Piker on June 22, 2015 at 5:58pm

Comment by Daniel Piker on June 22, 2015 at 5:41pm

@Jacek - their work was definitely an inspiration. I've also played around with these ideas a bit in the past https://vimeo.com/38706165, but that was before having the self-collision, which really limited it 

Comment by Jacek Jaskólski on June 22, 2015 at 5:29pm

great as usual!

btw. you've probably seen this:

https://vimeo.com/131242199

coincidence or collaboration?

Comment by Daniel Piker on June 22, 2015 at 5:24pm

@Dave - yes, starting with the variable edge length remeshing produces lots of valence 7 vertices which cause the hyperbolic shape once the lengths become more equal. I suppose another way to do it would be to start with regular edge lengths and topology, then increase edge lengths towards the boundary, but then the self-collision gets a bit more tricky

• ‹ Previous
• 1
• 2
• 3
• Next ›
• Page