More information on Tree branching contraint setup

[dokipen3d]

Hi all :)

I've been reading "Position and Orientation based Cosserat rods" as well as "Direct Position-Based Solver for Stiff Rods" and I can't seem to find in the papers how the constraints for the trees were set up.

I've dug into houdini that seems to have it's own way of setting up branching constraints (looks like its a distance contraint of zero length and another bend twist constraint).

Are any of the authors of the papers able to go into more detail on how constraints at branches are configured?

Thanks

[dokipen3d]

In the Direct Rods paper, it says... "The first three rows connect a point p1 on segment 1 with a point p2 on segment 2."

Does this mean that we simply add a second contraint to the point at which it branches? In the Related work section it says "In contrast our model supports branching that is independent of the parent branch discretization." That made me think that there would be no break between segments where a new branch begins.

That is actually an approach I took with houdini where a new branch could begin anywhere along the branch and would be attached by two distance constraints at the beginning and end of the parent branch. Although that ignores rotation so isn't a great solution.

What I am wondering is if something like the following is doable...

Tree model with straight rods for each branch

Skeleton (where the roots of the branches are further out in space due to the thickness of the parent branch...

Pin constraints that rigidly maintain their orientation relative to the parent branch, but let the child branch bend between the two segments...