Verlet simulation in Control Rig doesn't work well with large delta times

Development Character & Animation
, ,
1

Hey!

I’m trying to use Control Rig to drive some secondary animations, and all seems well, but if I minimise the Editor while I am running the project with the Verlet node in the control rig, when I tab back I can see it is way off position. If I keep it minimised for long enough eventually the Editor crashes with a NaN in FPoseLink::Evaluate

The Control Rig that exhibits this behaviour is quite simple:

Screenshot 2021-08-22 183647
Screenshot 2021-08-22 183647963×367 35.5 KB

Am I doing something wrong? Is Control Rig not handing large delta times well? It seems like it would be good if it used substepping, or if it let me cap the delta for the control rig.

I can’t use set delta on the control rig because I don’t have a reference to the control rig from this actor’s blueprint - I use different control rigs depening on actor state, chosen from the Animation BP.

Any advice is appreciated! Thanks

1 Like
2

Hi!
Have you found the solution?
I’m having the same issue with verlet simulation. In my project i have fully procedural animation, and control rig is driven from animation blueprint

3

Hey folks, I think this happens because the algorithm is based on feedback and is not frame rate independent because of the way it uses motion data from previous frames. One solution could be to use different a simulation method. Check out this video, it explains the concept well.

1 Like
4

Does anyone know a good simulation node that’s not framerate independent? I’m surprised Verlet is not