hey kerro, glad it's working for you

about the Mass that's not initialized in the verlet integration compound. that's basically ok, cause it gets caught by that first valid node. BUT you could of course initialize the mass yourself (best do that in a non simulated ice tree, or just use the one that's there already). you can even build something so that the tips of your strands are heavier/lighter than the roots - can give interesting effects.

@ rray
it definitely is!
but to be honest, apart from some initial doodeling, i haven't really used it (heck, i didn't even open zbrush the last couple of months )


Go for it!!

flo

Please don't go with it in a next few hours , as I already have some setups of this kind. Just need to cleanup them a bit. Even I'm not a great optimist when it come to modifiers that rely on deformation vectors, like curls or so - old story about getting a consistent ortho vector from curve, when underlying Nurbs surface doesn't exist.

So here is setup "from any strand", able to hold some of kH modifiers. Most of them should work, at least on 'guides'. Best chance is for "kH Cut By Geo Volume", "kH Hair Length", "Kh Grouping". "Kh Curls" , "kh Bend", "kH Turbulence local" and so on, will do something relative to re-construction of strand up vector. Anything with "chunk" in name, will do nothing useful.

"Clump" and "locks" are built in hair filler, on second point cloud.

Method for re-creating up vectors is some kind of direction constraint, using strand tangent and emitter's point normal, all that in local orientation, defined by first tangent and point normal. It flips at higher angles ( let's say, strand tangent is out of hemisphere defined by point normal, or like...), it also flips when tangent is parallel to point normal. Some flipping proof solution would be walking through strand, like in mesh extrusion nodes in newer SI versions. But this is still "smooth flip", imho not so usable ( nice words....) for complex styling. For comparison, NURBs surface always provides even visual feedback of "up" and "down".

Hair filler on second point cloud is adaptation of factory one - with additional ability to create hair interpolation groups (similar to ones in XSI hair), but this time based on angle between first tangents on guides. It searching for more than three closest guides, choosing the ones with lowest angle, relative to closest one. Btw I don't know what zBrush is doing about, I think it should have something similar.

For now, nothing smart when it comes to movement of all that. Generally, in some classic ICE "pairs" setup, I'd try to store deltas of second guides and second "faked emitter" in "init hair filler" on guides, then just call and apply them in hair filler. This should prevent searching for closest guides in animation - when emitter is deformed, there's a good chance for different set of closest guides.

good luck

Thanks for that. Brilliant. Will need some time to process and hopefully be able to report back with some pics.

Works nicely .. although long hair with fibermesh still is quite a pain..

i use ice strands as a hair from zbrush fibermesh, but strands thickness increase as it's far away from world origin, what should i do? it's make me crazy and i can't find anything useful, thank you

wow, my hat goes off to you, just got this in working in a scene i have been having some trouble with for a long time, and it looks like my life is going to get a lot easier now. Seriously great work.