Differential Growth in Curves

Hi all,

I m trying to do differential growing curves and couldn't come up with any ideas to get started.Any strategies will be helpful .

Thank u

Replies to This Discussion

Permalink Reply by David Stasiuk on June 22, 2015 at 8:34am

Do you have any reference images for what you're thinking about?

Permalink Reply by Shridhar Mamidalaa on June 22, 2015 at 9:12am

yes,David much like this from 0:31

Permalink Reply by Daniel González Abalde on June 22, 2015 at 10:11am

Wow, really awesome video, thanks for share.

When you speak of differential, you mean differential equations or think about recursivity or something else?

If you really want to do something like video you will need numerical calculation, differential equations (with Laplacians) and external conditions I think, as RD or Laplacian growth.

Once you have a value field, you can transform/represent them on curves, meshes, colors...

Permalink Reply by Shridhar Mamidalaa on June 22, 2015 at 10:19am

David Stasiuk has made something close to the "Laplacian growth" using Anemone soon after its release. I m not sure of the approach to make Daniel (for the Differential Growth). I always did stuff using Logic.

Permalink Reply by Shridhar Mamidalaa on June 22, 2015 at 10:22am

Dani, u did good with Reaction Diffusion using Differential equations ... would u give this also a try ...???

Permalink Reply by David Stasiuk on June 22, 2015 at 11:39am

There's a really nice description of the process that Nervous System uses in their edge-growth floraform system, which behaves in many ways very similar to the one by Deskriptiv:

https://vimeo.com/130977932

All this looks like it would be very achievable with a number of tools available here (Daniel Piker's Kangaroo2, and his and Will Pearson's MeshMachine (https://github.com/meshmash)). So for sure a scripting approach appears desirable, as the topology will constantly be changing.

Permalink Reply by Shridhar Mamidalaa on June 22, 2015 at 12:35pm

Tough way David.... Thank u :D

Permalink Reply by David Stasiuk on June 22, 2015 at 1:36pm

Yeah, this approach is fairly tricky, but I'd bet that Deskriptiv relies on physical simulation as well. I think the rule sets they use for the geometric transformations are probably (fairly) simple. The nice thing about the Nervous System method is that they only have to adjust a few parameters to get such a great variety of results...edge stiffness, shell stiffness, and then growth along the naked edge lengths.

It would be fun to try it out for sure!

1. set base mesh, evenly triangulated at desired resolution for target edge length

2. set base goals

3. lengthen each naked edge

4. re-mesh with original target edge length

5. register any changes to the goal topology (new/changed edges and vertices)

6. execute physical simulation

6. repeat from 3

Permalink Reply by Daniel Piker on June 22, 2015 at 2:11pm