ICE strand dynamics

[cnever]

The ICE strand dynamics does not seem to be as stable as wished. If the collision offset is not set to 0.8 or higher it crashes, also don't dare to move anything while simulation is moving also crashes.

Is there a way to optimize this or am I doing something wrong?

Thanks

[Mathaeus]

The ICE strand dynamics does not seem to be as stable as wished. If the collision offset is not set to 0.8 or higher it crashes, also don't dare to move anything while simulation is moving also crashes.

Is there a way to optimize this or am I doing something wrong?

Thanks

Hi,

*maybe* you'll be able to fix the 'collision offset' problem by setting 'cutoff distance' to more than any possible distance between strands and collider, probably to disable 'cutoff distance' at all. If 'cutoff distance' isn't exposed, it's under 'get closest location' node in 'test collision' node.
Another possible problem might be a 'hair filler' node, it doesa few 'risky' calculations on possible very short distance between guide/hair roots (divide by scalar, square root..). This produces NaNs, sometimes somehow solved by ICE, but not MR. Sometimes you'll get crash, sometimes MR won't render some hairs - looks like effect of some Radiation disease.

Afaik most problems become from mentioned calculations on small numbers. Generally not big deal to avoid by some creativity, if you're already able to build this stuff. Imho these strand dynamic nodes are very basic, much more an example for creating your own stuff, instead of ready-to-use solution.

Cheers