Wednesday, January 9, 2013

deforming geometry fractures:

At this point i have the basics down and have R&D'ed the fracturing in a controlled manor. The next bit of R&D is to test a deforming piece of geometry.

The process outlined in the previous posts was replicated now with the deforming arm geometry as my fractured object.

the main difference here is in the solver type. FOR the geometry i need to stick together through out the sim(the arm before it fractures) I chose to use an RDB solver for this section due to its ability to handle more complex geometry shapes. Thus when the arm bends and shapes become distorted,  it  will have a higher likelihood of producing concave and or manifold shapes. The bullet solver(used in the tube example) although very fast and efficient, notoriously does not handle the aforementioned shapes with any real dignity. Using the bullet solver on deforming geometry, even with the penetration threshold cranked up: the piece shapes will remain rigid(not deforming per piece), and will try to maintain a minimum distance and position from each other resulting in an exploded view of the the animated geometry with unsightly gaps in what should look like a solid object.

However for the pieces that become active ( peices falling off the arm), i was able to use a second solver the bullet solver to speed simulation times, given the shot framing, ect... Another small trick to speed simulation times, is to set the merge node to "LEFT AFFECTS RIGHT" this dictates to dops that the falling pieces will be affected by the animated arm geo only.

the result of this test is deforming geometry that fractures in a controlled manor. This workflow gives the artist complete  control over which fractured pieces become active, by animating a attr transfer from say a simple sphere to static "T pose" or default pose of the geometry attr transfer as a blueprint for the deforming geometry.

here is an example of the result:

No comments:

Post a Comment