Real-Time Physics Simulation Forum

Your screenshots look like 'zig-zag' to me. I used 'diamond' but I think zig-zag offers better coverage for less vertices (at least for graphics LODs, not really sure regarding physics quality), plus is better suited to very steep terrain, though I haven't tested behaviour except for (actual?) diamo...

Might be worth considering heightfield patches for this. There are options for diamond subdivision and zigzag subdivision, though they appear to be named transposed?

Hi

I've seen this happen with spheres and heightfields when the btDbvtBroadphase is used. Instead, try a btAxisSweep3 for your overlapping pair cache. Now even extremely low density spheres collide perfectly even with dense dynamic stuff also on a heightfield.

Try (for example) the btNNCGConstraintSolver, it solidly solves all sorts of contraptions.

*though I'm recently trying to get 'rolling friction' to work again, but iirc something was conflicting with the above solver :/ (settling for no rolling friction here)

Perhaps try with a skin of 0.0 on both rigids.

Failing that I think perpetual bouncing is such a rare case that it might be ok to very slightly throttle a force on contact started/ended (for example).

The forum's search doesn't always help even when we know what to look for!
This will get you started http://bulletphysics.org/Bullet/phpBB3/ ... f=9&t=7739

Nvm, figured it using this: http://www.bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=11061&p=37118&hilit=compound+ray#p37118 I subclassed btCollisionWorld::ClosestRayResultCallback and put a field to store an in that is the index of what I hit. Hi, there's a rare case nullptr bug ...