With H12 these sorts of things are not too bad now and don't require writing of code (if a few expressions is what you call writing code).

I am really leaning heavily on the new Bullet glue constraint geometry and the Glue Adjacent tool to hold steady, animate and release RBD pieces.

It takes a while to build the constraint network but once done, you can lock that sop or cache to disk and work on it in DOPs with a SOP Solver.

You can use three RBD Point Constraints on just three of the RBD objects you know are not going to be released throughout the body of the sim to hold things fast. Actually you can animate the pin constraints however you wish and the entire glued constraint network goes for the ride. This means you can animate one or two points and move all the constrained RBD objects. Great for making a building fall over. Just constrain two RBD objects at the bottom and animate one single pin constraint at the top and then dissolve a few RBD objects at the middle of the tower to get things to start to fall over, all driven by keyframe animation if you so wish.
Hopefully you can figure that one out after looking at the attached file.


Have a look at the example file for one setup. The glue constraint network is blasted away inside of a SOP Solver. Read the notes I added.

Btw the Cebas Thinking Particles Huntsmen testimonial looks to be pulled from their web site.

The new Volume Break SOP in H12 is a nice proof of concept. It would have been great to have it actually useful.

digitallysane
The new Volume Break SOP in H12 is a nice proof of concept. It would have been great to have it actually useful.

I searched the bug database but couldn't find any submission by you indicating any issues reported on the Volume Break SOP.
You know the drill. I strongly urge you to submit any questions or issues with the Volume Break SOP to Side Effects.

As for Volume Break SOP, it works best when the geometry you are intersecting with the sdf volume is at a high enough resolution. The sdf volume needs edges to cut against and currently doesn't create the face edges to cut, only the cutting edges between existing edges as far as I know.

jeff
I searched the bug database but couldn't find any submission by you indicating any issues reported on the Volume Break SOP.
You know the drill. I strongly urge you to submit any questions or issues with the Volume Break SOP to Side Effects.

Fair enough
What I'd like is of course internal surfaces, would make for a nice fracturing tool.

Me too. Submit it.

jeff
Have a look at the example file for one setup.

Hi Jeff,

Thanks for setting up an example. I do however run into an error in the sopsolvers object merges expression “`stamps(”..“, ”DATAPATH“, chsop(”relnode“) + ”:Relationships/“ + chs(chsop(”relnode“) + ”/relname“) + ”/“ + chs(chsop(”relnode“) + ”/relationship/dataname“) + ”/Geometry“)`”

relnode apparently is unknown. (in 581 as well as in 642)
I know the expression in there is default sop solver stuff but would you mind taking a minute to help me understand why it's not working as expected?

Edit: just manually setting it does work.

That's a great example for taking things further. Never thought of locking the constraint network before in SOP, this is a good speed + workflow tips…and never knew you can break the constraint network with SOP solver! Thanks Jeff!

Thanks for sharing. This is great example of how to use and maniuplate glue networks. Very helpful.
One thing i`ve noticed working with glue networks is that just decreasing gluepieces SOP parameters can give us 100x+ faster results (without any connection loss). It depends entirely on scene scale. You can read this in documentation:

Points per Area

To determine proximity, points are seeded on the surface of all the objects. There must be enough points for close points to occur to detect close surfaces. This should be scaled down by the square of the geometry size. If your geometry is 10x bigger, you want 1/100 the points per area.

In this scene it tooks 26.5 sec. on my machine to produce glue netowrk with default settings and 1.8 sec with 100pts per area,0.05 search radius, 10 max search pts. Glue network is exatly the same.

So my approach is first to tun off auto update. Then decrease default settings, and then update scene manually. This is much faster way(because canceling gluenetwork cookng always craches houdini)instead of working with defaults.

Yes you are right in that the Glue Pieces SOP is difficult to set up optimally for large models.

Please submit an rfe related to the Glue Pieces SOP to support at sidefx.com being quite a challenge to set up in an optimal way.

IMHO there should be an initialize button that evaluates the incoming model size as a start.

in regards to the scattering density for the gluepieces sop:

If the density is too high it will take forever, if it is too low, it will sometime not scatter on some of the tiny pieces.

Therefore I ended up opening up that asset and looping through each piece to scatter at least one point on that piece, then merge that in with the other scatter. This allows you to lower the scatter density hugely and still have your pieces be connected.

jeff
Btw the Cebas Thinking Particles Huntsmen testimonial looks to be pulled from their web site.

nope, it`s there:

http://www.cebas.com/index.php?pid=testimonial&tid=84 [cebas.com]

your example is really good, except that long and complicated expression understandable only by programmers or similar. also it works in that particular case only, meaning, if you change something you have to rewrite expression. funny thing is that expensive software like houdini should have nodes prepared for us for everything, instead we are forced to learn programming every time we have to write expression or similar that means some nodes are missing. we need universal solution like unglue node with custom inputs or activator node or so.
here is example how you could activate objects once they reach specified distance from activator node. basically once activator node come close pieces are transferred into active group influenced by gravity. so you just wire distance node and that`s it. this question is asked 20 times already and still no solid solution. I just hope we will get solution soon and I`m also a bit worried knowing that we waited 2 years to reach fume fx quality.

Hi zs2010,

There is a lot of scene on SideFX and odforce you just need to find them.

THis is one simple scene from odforce, activate with color transfer.

Or another little example i made to use animated geometry to activate fractures.

Thanks Igor and Rudi,

but :roll: they all use expressions as well. point here is to have universal activator or unglue or similar nodes for all kind of situations with custom inputs so we could use geo or volume to activate or unglue pieces and avoid writing expressions.

zs2010, build your activator, wrap it into an asset, use henceforth.

There are no non-programmers (at least to some degree) in the Houdini world.

zs2010
(…)funny thing is that expensive software like houdini should have nodes prepared for us for everything, instead we are forced to learn programming every time we have to write expression or similar that means some nodes are missing.

You are so wrong! Expressions are not a weakness of the application, it's a strength of it, you just have to learn them, to use Houdini effectively. If that bothers you, don't hassle with comparisions between Houdini and other apps, because you're simply not justifed.

You simple can't except that nodes will do anything what expressions can, sinces expressions have endless combinations, like any decent language.

It has nothing do to with the possibibilty that Houdini misses some functionalities you like to see in it. Very probably, but the lack of some functionality has nothing to do with a likeness you will have to use expressions using Houdini!

@zs2010 , hi .

my advice : if you ever wd have to waste 1 hour , waste it with expressions in houdini .

there is contextual help , there is the forum ( 2 indeed ) with a lot of examples and great community .

once you start getting some results , go back and try to achieve the same result differently
( in houdini or in-any-other-app ) .

just try it ..

oc , you can use houdini ignoring expressions , but why limiting yourself ?

why wasting all that power ??

digitallysane

jeff
I searched the bug database but couldn't find any submission by you indicating any issues reported on the Volume Break SOP.
You know the drill. I strongly urge you to submit any questions or issues with the Volume Break SOP to Side Effects.

Fair enough
What I'd like is of course internal surfaces, would make for a nice fracturing tool.

Hi guys,

Really interesting topic. But I don't get one thing. What is the problem with inner surface? Take a look at my hip file. It's a core of my shatter tool.

https://vimeo.com/56722980 [vimeo.com]

Hi,

I tried to replicate the activation by color exemple but cannot seem to make it work.

Can someone take a look at my scene and tell me where I'm missing something ?

Color attribute is transfered to the points, I replicated the “point()” expression in the activation node but they don't activate…

Thanks

Jean-Luc