The Muscle deformer also offers automatically calculated Self Collision, which is based on collision area groupings that you define. Enabling Self Collision can be slower, since it runs over the entire mesh grouping you create (although Self Collision weights can be painted to cull areas you do not want calculated). The benefit of this approach is that there is less setup work and fewer attributes to deal with than Smart Collision, and it functions even in areas that are not connected by two joints. The computation automatically finds, resolves, and corrects the areas that you define as self-colliding.
Self Collision is not time-based, so you can apply the Muscle deformer to a static model and have it resolve collisions. There are three paintable Self Collision weights types available in the Muscle Paint window:
It is important to note that Self Collision is related to the detail resolution of your model. The more detailed the model, the slower collision is, but the more accurate. Collision is partially point-based, so it cannot occur when the areas colliding are smaller than a polygon on the mesh. In general, collision appears smoother and more accurate when you use a more detailed model.
In addition, the collision occurs in point groupings that you define. This is similar to SmartCollide nodes where you paint an A and B set of points. For Self Collision, you use the Self Collision Grouping window ( Muscle > Self/Multi Collision > Self Collision Grouping) to mark pairs of regions you want to collide. For example, if you want the upper leg and lower leg to collide, you create a grouping defining the points in each area, and store that to the cMuscleSystem node.
