Bug / Crash Obi cloth does not properly work for models stitched to an existing skeleton.

[josemendez]

> As long as your cloth and your character meshes behave fine when animated (meaning no collision or z-fighting happens), skin constraints will respect this. So in a sense they do prevent the cloth from colliding with the character mesh. They prevent penetration with any mesh that is skinned to the same skeleton as the cloth (and the character is) so the tutorial text is accurate.

That does not seem to work in my case:


The shirt has pinned vertices on the shoulders while all the unpinned vertices have mass == 5, zero backdrop, very high backdrop radius and a desired skin radius.

And I made sure there is no Z-Fighting between my skinned meshes. Moreover, I've just added a collision layer into the solver's list that has these colliders:



And that did nothing at all. Are colliders working for skinned meshes?



> Why you'd want to save values to an external file is not clear to me

That is mostly because I have lost my progress (particle editing) several times during my experiments and I am afraid to loose it again. So far I have not tried reproducing the progress loss bug, but I will try it again in the future in a test project.


One more question, if I may: my cloth simulation feels like a shaky Christmas tree when I start walking from a full stop: it gains speed way too fast. I've tried removing this effect but I cannot remove it while keeping this effect that I really like:



Is there any way to do it? I tried playing with absolutely everything: mass, skin radius, all the parameters really, even the ones on other components - nothing seems to work. I even tried playing with damping on the solver - nothing seems to work. Is there any way to influence the acceleration? I want it to be real small, but to have a high potential amplitude.

Increase particle backstop radius, it might be a bit too small for your character. Skin constraints create a "collision sphere" at each skinned vertex position, if the spheres are too small the cloth might slip past them. Mass has no effect at all on the simulation if all particles have the same, it only affects collisions with rigidbodies.

Colliders do work with skinned meshes. Just make sure the particles and the collider have different phases, and that collisions are enabled in the solver.

Regarding the shaky simulation, this is due to how character movement influences cloth. You can control how much world-space velocity affects the character cloth ()as long as your solver is set to simulate in local space) by tweaking the cloth component "world velocity scale".