When I use an object that is animated at the object level as a collider for a cloth mesh, I get much less interpenetration during fast movements than when I use the same object, but animated via a bone that it is skinned to (identical animation).
I understand that it's probably easier to calculate the vertex positions for the first case, but I would expect the calculation to simply take longer rather than being considerably worse in the second case. I can remedy this somewhat by increasing the time steps, but that regularly changes the simulation and affects all other operators.
Is this the expected behavior for collisions with deforming meshes?
I understand that it's probably easier to calculate the vertex positions for the first case, but I would expect the calculation to simply take longer rather than being considerably worse in the second case. I can remedy this somewhat by increasing the time steps, but that regularly changes the simulation and affects all other operators.
Is this the expected behavior for collisions with deforming meshes?