Help with Slippery Static Terrain.

[TactfulWaggle]

Hello, I've got a working implementation of Bepu in unity with a space and everything working except drawers (which is making my life working with this actual hell).

I make terrain using procedural generating and using that data I turn it into a Terrain class from a TerrainShape instance, when my dynamic entities interact with the terrain they keep sliding and spinning around no matter how high i put their friction value to, same goes with the terrain itself, i've tried changing its material's friction to super high amounts despite knowing they shouldn't change much since its static.

For more information here are the variables from my physics component for characters

I've tried locking rotation using both velocity clamping and constraints, velocity clamping didn't seem to affect it too much it just cushioned it, and the constraint worked better but the sliding stayed, I'm also looking for maximum performance so having to use constraints isn't ideal.

[RossNordby]

Sounds like the actual pair material isn't ending up with your expected friction values. The default is multiplicative, so if one if the two materials has 0 friction, the pair will have zero friction no matter what finite friction value the other material has:

bepuphysics1/BEPUphysics/Materials/MaterialManager.cs

Line 77 in 88243f8

public static void DefaultMaterialBlender(Material a, Material b, out InteractionProperties properties)

You can also specify arbitrary material blenders by changing the callback:

bepuphysics1/BEPUphysics/Materials/MaterialManager.cs

Line 22 in 88243f8

public static MaterialBlender MaterialBlender = DefaultMaterialBlender;

[TactfulWaggle]

Thanks for the answer, I will try messing with the material blender since I do want it to take in the average at times with like bounciness and stuff, I didn't know that was possible to change.

For friction however, I did have both materials set up, it just for some reason doesn't use the material of the terrain, dynamic entities material works as expected with everything.

I have the friction of the terrain set to really high values for testing and it's like I haven't set anything, that's what's puzzling me.

[RossNordby]

Yup, something doesn't add up; if the material blender is outputting correct friction values for the constraints, it shouldn't be sliding around. Probably some oddness afoot. May want to try reproducing it in the demos to narrow things down.

[TactfulWaggle]

I'll try to debug it further and get back if I find out anything else that I can't figure out, thanks for the help, I highly appreciate it, this has been the one issue halting the development of my game for months now, and now I know more about where to look.

[TactfulWaggle]

I just remembered some important info, I don't really know the best way to clamp the velocities so this is how I attempted to achieve it

For the constraint version I've got this snippet, but the constraint has a different bug where it starts tilting sideways, then tries to pick itself back up and ends up picking up speed like a top and spinning away.

I'm not used to working with libraries much, I've tried searching the forums and see if there's anything about this but I couldn't find many resources, I did find one comment from years ago saying clamping velocities is cheaper than constraints.

Hope I'm not too bothersome with these posts.

[RossNordby]

A couple of notes:

1. Clamping AngularVelocity and AngularMomentum is redundant; can just do AngularVelocity alone.
2. You can set rows/columns of the entity's LocalInertiaTensorInverse to zeros to make the local space axes corresponding to those rows/columns have effectively infinite inertia. No extra clamping, no constraints, costs nothing.