stiffness

[super gnu]

ive been playing with the clustering and getting some nice chunks of grains going on... however, even with stiffness set to 1, the chunks are still kinda floppy and springy..  

how would i get more rigid chunks, or plastic deformation?  it seems possible as you demonstrated it brilliantly in your car crash demo. 

im guessing i have to use the physx operators for more complex stuff like that? 



i was also hoping to get an effect i had in houdini a while back, which was rigid chunks surrounded by loose particles.   in that case i used a procedural texture to split the particles into two groups with different attributes, 

or, possibly, i used an inside/outside geometry test on the particles to achieve the same thing,  and drove it with some chunks of geometry..   id be interested to know if either of those things are possible.. 

edit:  guess i can use "surface test" to split particles which are inside geometry. 


-if textures only could be used, then vraydistancetex, could provide the geometry masking quite simply.

[insertmesh]

ive been playing with the clustering and getting some nice chunks of grains going on... however, even with stiffness set to 1, the chunks are still kinda floppy and springy..  

how would i get more rigid chunks, or plastic deformation?  it seems possible as you demonstrated it brilliantly in your car crash demo. 

im guessing i have to use the physx operators for more complex stuff like that? 



i was also hoping to get an effect i had in houdini a while back, which was rigid chunks surrounded by loose particles.   in that case i used a procedural texture to split the particles into two groups with different attributes, 

or, possibly, i used an inside/outside geometry test on the particles to achieve the same thing,  and drove it with some chunks of geometry..   id be interested to know if either of those things are possible.. 

edit:  guess i can use "surface test" to split particles which are inside geometry. 


-if textures only could be used, then vraydistancetex, could provide the geometry masking quite simply.

I did some similar r&d with my snowman scene.
You can have different particle binds with different settings. And to get really stiff grain, you need to both up your particle bind solver steps AND lower the integration step.

[tyFlow]

PhysX is ideal for the most rigid particle bindings, however, you can achieve reasonably stiff results with the Particle Bind Solver, by simply lower the overall sim timsteps and increasing the Particle Bind Solver steps, exactly as insertmesh explained.

[gizzer]

Yes this works fine. But the sim time becomes much longer. Is there any other way to get stiffer results?

[tyFlow]

Within the bind solver, stiffness and timestep are inextricably linked. There is no way to reduce timestep and maintain the same stiffness.

[tyFlow]

Within the bind solver, stiffness and timestep are linked together. One thing you can try is to switch your bindings to "constraned" mode, which introduces an algorithm that attempts to make stiffness somewhat independent of timestep, but it only works if your stiffness values are less than 1.