Wednesday, January 9, 2013

next i needed to create control points for rbd sim

i chose to set up a simple scene with a fractured tube(using the aforementioned volume method) animating from screen right to left. Next i fed the vornoi fracture into a for each sop which contains an add node that i directed to find the centroid of each "piece group" created by the vornoi fracture sop. This will create a single point for each fractured piece of the object with a point number to match the piece group number. I then give each point a black color which i fed into an attr transfer. The opposite side of the attr transfer i animated a sphere(attatched to a trail sop) that animates down in the y axis. The sphere is given a red color. This way when the sphere comes with in the attr transfer bounds it will transfer a  red color onto the originally black tube points. This red color will can now be used to activate fractured pieces in a very controlled manor.

Next i moved into dops and began to set up my simulation.
I set up a ground plane from the shelf, then i manually created two RBD fractured objects. one for the deforming tube(which is actually just translated... not deforming) and the other for the tube object moving peices(the pieces activated by the attr transfer).
I attached an RBD state node to the tube object deforming with a "defualt operation" with a value of "set always" to specify the activation state of the tube object deforming object.
from here I allow editing of my fractured objects in dops and dive inside and set point expressions on the  activation parameters of the sopgeo nodes(see below). This takes the point color from my attr transfer that i did on the sops level on the un-animated geometry and feeds in a per point activation that correlates to the pieces of that fractured object i want to be active. 

This gives me a result like this:


  1. Hi Lee,
    I am learning Houdini and working something like you have made here, could please let me know more about your DOP configuration?


  2. Hey Babak,
    the dop config shown above, is basically 2 RBD objects one of which is modified by an RBD state node( The RBD state node uses an expression(which i think you can see in the pics I attached to the blog) that reads the point color info, i use this info to active the pieces from the fractured RBD(modifing the $OBJID), this allows some pieces to fall away from the original.
    In hind sight, a better way to achieve this effect would be to do this with a sop solver. There is a great peter quint tutorial on this ( here is the 1st of 4: with a sop solver you could run an attr transfer similar to what i did without having to have 2 RBD objects.
    hopefully this is helpful... but as is often the case with Houdini, i fear without an actual scene file it is very difficult to explain. if you prefer you can email me a scene file and I can take a look for you, as i know how it can be difficult to get started in Houdini and there is so much to learn. here is my email if you deem it helpful.