Getting Particle Instancer Objects To Collide

[krank]

Hi. Is there any way to get instancer objects from an emitter to collide? I tried making the instanced objects rigid bodies, but no luck. I want the collision to use the vertices of the geometry and not the particle itself.

Much thanks.

[gmask]

I have seen some tutorials around about interparticle collisions.. keep in mind the the new nParticles system in 2009 has this feature built in.

[krank]

Thanks, I've seen some tutorials but they often refer to using a radial field to push particles apart. I was looking for a way to emit rigid body geometry (& I'm guessing that's by using instancing).

I've checked out nParticles features and I'm not sure that's what I'm looking for since it only applies to the particles themselves.

[gmask]

QUOTE(Krank @ 11/21/08, 05:53 PM) [snapback]295721[/snapback]
I've checked out nParticles features and I'm not sure that's what I'm looking for since it only applies to the particles themselves.

yes but you'd have use the "bounding box" of the particle to generally fit whatever object you are planning to instance because the instancer does not deal with rigids at all.

I mean if you want large numbers of complex rigid bodies colliding it will be extremely slow.. and I mean glacier slow

[joojaa]

QUOTE(Krank @ 11/22/08, 12:35 AM) [snapback]295714[/snapback]
Hi. Is there any way to get instancer objects from an emitter to collide?

You don't. Term particle refeers to the fact that you make a approximation that the thing you model is sufficently small so that you dont need to account for its shape. So yes you can make the particles collide NOT the objects because you just saifd you dont care what their shape is.

Use rigid bodies instead if you want to use the whole of he object. Yes you loose the advantage of speed but you dont loose the shape in the process (that is you trade in the shape for speed with particle representation).

QUOTE
I mean if you want large numbers of complex rigid bodies colliding it will be extremely slow..

Well yes but it can be accelerated beyond what the default maya rigid bodies give you.

[gwelch]

If the number of particles are low enough...you could bake the instances into animated geometery using "nimble tools" uninstancer plug-in, then looking thru your shoot camera(s), only make the geometery that collides into rigid bodies.

Consider cutting to a close up shot for the detailed collisions (if you are allowed)...and then use the radial particle method for medium to far shots, and also for far away background collisions in close up shots.

[krank]

Thanks all for the suggestions and workarounds. It's actually not for a practical shot but more a proof of concept/technical demo. I will check some of the ideas out!

[joojaa]

you can jsut solev the issue with drawing order also. By comping in different passes thisway objects apper on top even if they arent. IT often looks good enough.

[rfx74]

Look at this:
http://www.tronicstudio.com/media/mov/sony\_hdna.mov

How would you do it?

[joojaa]

QUOTE(rfx74 @ 12/12/08, 08:30 PM) [snapback]296969[/snapback]
How would you do it?

I would animate 3-6 different objects doing the jumps then i would just snapshot those animations then assign each particle with a different subslice of object to cycle trough.

Then bounce them with a offset surface and then constrain to a goal object thats designed to fit them. And offcourse the detail one i would do by hand. But yes no need for accurate collisions untill at the edn where i would just un them tough a rigdbody sim tough not with mayas nativerigdbody. For the closeups and then just let the rst colldie on comp plate.

Alternatively id make use of several particle objects wich are as blobby particles and then jsut let them hit offset surfaces. THs would work out even for the last part-

Anyway theres lot less actual accuracy needed for thsi to seem correct.