Hi everyone,

I've got a problem with pop grain. I'm new to houdini so I'm probably missing something.

I'm emitting particles from a circular growing selection of a grid with mountain sop. (The polygon selection of the grid is made by mops.)

I wanna replicate the effect of sticky blobs of grain growing on a landscape.

In dop i do pop grain, set to the desired level of clumping, but when particles comes to a rest they keep spinning in place like crazy or the continue to jump from 1 position to another. I realized it in sops, after copying to point the dopio to an icosahedron which eventually shows the rotation of particles.

When clumping is set to 0 the problems disappear, but of course particles run separated from each other.

Upping the substeps does a tiny bit better (tiny)

I believe that in some way too much energy is involved in constraints or the scale of the animation isn't right.

Is there some material you could suggest to get the right knowledge on setting emitting clumpy grain?

Could someone take a look at the mess i made in the file?
https://www.dropbox.com/s/jexcicf4r7wtxyo/pop%20grains%20jitter.mp4?dl=0 [www.dropbox.com]


Thanx for attention

Luigi

Particles don't inherently have an orientation like RBDs do. If you try to copy to points and you don't have an orientation defined, the Copy to Points SOP is going to use velocity as a "forward" direction, and then guess at an "up" direction based on the difference between forward and +Z. It's often going to give you very weird results.

You can explicitly define an orient attribute on the scattered points if you want, and that will stay consistent throughout the sim. If you want the particles to "roll" as they move, though, that gets much more complicated and you'll need to write a custom solver in VEX or a SOP Solver DOP to handle that. Or use packed RBDs.

Thanx a lot! I'll give it a try with orient attribute and then eventually dig in RBD grains.
Do you know any tuts on this?

A couple of silly questions:

Orient attribute has to be cast on scattered point in sop. Could you explain why not with a popvop in dop?
I believe vellum behave in the same way. Is it possible to set random orient attribute which can freeze spinning on vellum grain?

Thank you Toadstorm, also for the wonderful MOPS that you gifted to the community.

You could apply orient via a POP VOP if you wanted, but it wouldn't have any effect on the simulation or be affected by the simulation unless you wrote the solver yourself. Particles have no understanding of orientation at all. It's just easier to apply it in SOPs unless you plan on trying to roll your own solver.

Vellum grains I think will work the same as other particles. I think the only Vellum type that uses orient internally is hair.

I can't speak to a specific RBD tutorial... as always your best starting point is tokeru.com/cgwiki [tokeru.com].

I'm happy you like MOPs!

Hi Toadstorm,

I take the chance to ask you a question about mops+ which i subscribed and I'am trying to understand how to bend it to my needs.
I saw the example vellum falloff from the folder Examples. What if I would do the same thing in dops. In the scene you saw upward,e at a certain point I would like to animate a linear falloff to deactivate the strenght of the constraint in order to leave the grain free to fall over the surface. I tried feeding the mop falloff shape either in the forces input of the solver, either after the vellum source, but with no luck.

Would you be so kind to illustrate me the correct way to do it.

Thanx for attention
Luigi

I'd have to spend a little more time tuning the file to get the exact effect you want, but I think the main reasons this wasn't doing what you wanted is because A.) the falloff you were animating was reversed; meaning you were animating your stiffness from 0-100 instead of 100-0 over time, and B.) your glue constraints were set to break at a fairly low threshold, and were immediately breaking after the emission stops around frame 37. There were hardly any constraints left by the time the falloff shape was being animated, so changing the glue didn't have much of an effect.

I tweaked a few other parameters... mostly just reducing the number of points generated and the number of constraints generated to speed things up for testing, and I disabled constraints averaging to prevent points from accelerating out of control.

Thanx a lot Toadstorm, that's really helpful.