Hello,

Been looking at some of the tuts like Jeff Laits recent one on vellum…

I was wondering if it's probably necessary to use a ‘Collision’ solver in conjunction with the vellum solver for cloth to prevent collisions/crossover.

I've create a grid with spheres around the perimeter with a boolean union.

I used this as the ‘source’ for the vellum cloth solver.

In the solver itself I increased the substeps to 30, and also the Collision Passes and Post Collision Passes to the same values of 30 each.

But I still find that the spheres penetrate through the cloth at points in the simulation.

Maybe because the collision algorithms for vellum cloth are only designed for ‘flat’ planar objects as source?

And I have too find a way of using it with a collision solver additionally?

Feedback is appreciated, thanks.

maybe it's because your spheres are set to primitives and not polys?

I originally thought doing it with primitives might be a problem, but for some reason thought running it through the boolean as union would ‘convert’ it to all polygons; But this doesn't happen.

However, changing it to polygons has the drawback that the surface of the spheres are now acting as cloth and deforming too.

I initially tried some vexpression wrangling in the vellum solver with a vellumcontraint property node with previously setup attributes e.g. changing different attributes like stiffness only on i@Frill == 1(which are the spheres).

But I noticed something that may have contributed to that not working. Once the geo passes the first cloth constraint node, the attribute i@Frill == 1, and i@Frill == 0, is only seen coming out of the geometry output(grey). Only i@Frill == 0 is seen coming out the second constraint output(pink).

To change that, I input the geometry into the second input as well, which resolves it in that both values of the attribute show up after…but I'm not sure how it might affect adversely other things since its a grey going into a pink.(mismatch? - although no complaint messages).

So instead I tried putting the spheres into a group rather than as attribute, and put down another constraint node and specified the spheres as a group to apply it too only.

There I tried all manners of parameters like upping up the stiffness values, and with different constraint types - Cloth, Bend Across Angles, Distance Along Edges, etc…..

Was not able to change/prevent the spheres from deforming.

Thought I would try a different approach.

Add spheres to cloth edge after vellum sim.

In geo node import cloth with object merge and do same in another for the spheres.

Then set up both with deforming geometry shelf tool.

Even with sub steps set to 50 (surely enough?), still getting intersections with cloth, and between the spheres too.

Anyone have an idea as to what kind of set up to go with? ….. I mean, I could probably make my own vex ‘solver’ for this >.<

my simple approach…..left substeps at default 1

I prepared 2 groups, the hankerchief and the pearls anticipating I would need to use them later..but turned out they weren't even needed

(sorry but I couldn't decipher all the complexities in your scene, so I could be on the wrong track, apologies if I am)