Kristinka C

Mathaeus · Post by **Mathaeus** » 05 Sep 2009, 14:24

Guys,
I've just uploaded a smaller version that doesn't require external addons.
It' now about 12 mb.
Link:
http://www.matkovic.com/anto/kristinka- ... 9-nomt.rar

Cheers

Dixon3d · Post by **Dixon3d** » 05 Sep 2009, 21:11

Hmm, well now I'm rightly confused. All of the models included in "kristinka-hair-c-29aug09-nomt" from "Basic-Setup" through "Turbulence-Static" still don't generate strands and the ice trees are still red, same with the Simulation-PhilPack-xsibase-08aug09-no-MT.emdl as well.

All the previous hair models from the may release function as expected.

HOWEVER: on a different computer running vista 32 bit with a fresh install of 7.01 32 bit everything works great (compounds and applicable addons pulled directly from this computer).

My guess at this point would be that the kristinka compounds require xsi 7.01 or greater, and that separately there's a problem with the MT_Strands addons potentially with the 64 bit version of 7.0. More checking to do.

Thanks for all the help!

Mathaeus · Post by **Mathaeus** » 05 Sep 2009, 21:32

Thanks for response,
for now, the only difference I found between my nodes and mt_ice tutorials is that I using point clouds inside the model hierarchy (something that I really wouldn't like to change). This release also has a much more connections between point clouds.

I sow something similar once when I changed model name outside of XSI - and I was been able to fix that by supplying geometries again, even both new and old set data nodes were looking the same.

I did all that stuff with XSI 7.01.

Dixon3d · Post by **Dixon3d** » 05 Sep 2009, 23:09

Problem resolved:

Switched from 7.0-64 to 7.01-32 and everything works beautifully, (Including the MT_Strands addons).

That's some pretty amazing stuff you have there! Thanks for allowing access - so much learning to do now!

-edit-

Oh and it appears the ~32.dll / ~64.dll error message was the "other" one being rejected by XSI.
the 64 bit version rejects the 32 and causes the error message, the 32 does the opposite. Other than the 64 bit one crashing on model load (unknown reason), the fact that the 32 bit works perfectly makes me inclined to point a problem at 7.0 64 bit rather than anything else.

Hirazi Blue · Post by **Hirazi Blue** » 16 Sep 2009, 15:02

A slight (unpleasant) update: the StrandNodes Addon doesn't seem to work under SI2010 Vista64. Could have something to do with the new C++ rules explained in the "what's new" section of the SDKdocs... However, even the Dynamics stuff from the Kristinka Samples that use "Phil's ICE Pack"-compounds do not seem to work under SI2010 Vista 64, while the samples that came with "Phil's ICE Pack" seem to work just fine... I don't know what to think of this as I was looking forward experimenting with the StrandNodes and especially the Kristinka Hair System...

Hopefully there's some easy fix, that so far eludes me...

Mathaeus · Post by **Mathaeus** » 16 Sep 2009, 17:57

Actually, the only one simulation sample doesn't use mt_strands addon, it's 'simulation-PhilPack-xsibase-08aug09', just the one in '...no-mt' pack. Only this one should work without mt_strands addon.

Other than that, people reported problems when importing models instead of opening scenes (not only my models), workaround seems to be: first open the scene, then import model, something like that.... here I talking what I heard...

Maybe I'll have a time to check the scenes with 2010 trial - maybe it's better to wait. While ago, XSI 6.0 elegantly destroyed all my rigs, because I used simplified 'remove animation' command in synoptics. I changed the all synoptics, then became XSI 6.01 that allowed previous way...

I'll see when I'll have a time.

Cheers

scaron · Post by **scaron** » 16 Sep 2009, 19:43

helge will need to recompile those custom ice nodes for softimage 2010.

Mathaeus · Post by **Mathaeus** » 17 Sep 2009, 18:15

Hi,

From what I know for now, it seems that everything that doesn't include MT_strandsNodes addon, working fine with 2010. Except one model ("turbulence-static"), where I left the one node that calls MT_Strands addon. I uploaded a new "no-mt" version this morning.

So for now, pack that works with Softimage 7.01, 7.5, 2010, without extra addons, is available for download here:
http://www.matkovic.com/anto/kristinka- ... 9-nomt.rar

Support files for "Follow NURBS" tutorial should working too, with all mentioned versions.

Other than that, some people had a problems with importing models into a new, no-saved scene, but this was solved by opening an existing scene, then importing model. If something so important like ICE nodes, is working properly with one OS, this time Win XP 32, I assume it will with others too. If it doesn't, I really don't feel this as a my problem.

There are two hard coded ICE nodes with MT_strandsNodes addon: mt_HairSolver, used only once in a "big" pack, also mt_StrandFitting node - the last one is used many times. Here, this node allows even distribution of strands segments along strand's length. Generally, strand generators in my nodes don't care about even segment's distribution, strands are just multi-segmented Bezier curves, or sampled UVs of NURBS surfaces.

Without mt_StrandFitting node, you can notice irregular look of curls or turbulences, or anything other that rely on strand lenght. In styling phase, this can be somewhat compensated by modeling. "Pure" simulation with PhilPack nodes will working without any problems, because simulation engine takes care about even segments distribution. Also you can apply mentioned modifiers in post-simulation region, just as I did in samples. But, taking static strands that aren't resampled, into the strand's simulation, generally won't looking nice - that's a reason that two 'blend' simulation samples aren't included.

Cheers

Hirazi Blue · Post by **Hirazi Blue** » 18 Sep 2009, 07:04

Thanks again for the effort...
Now someone should maybe try to convince Helge Mathee to update his Addon...

;)

Letterbox · Post by **Letterbox** » 19 Sep 2009, 10:15

Hirazi Blue wrote:Thanks again for the effort...
Now someone should maybe try to convince Helge Mathee to update his Addon... ;)

I dont think any of them exist any more... I was looking for one, but mindthink.de is gone.

Hirazi Blue · Post by **Hirazi Blue** » 19 Sep 2009, 11:01

Many luckily are still around, but you'd have to look for them...
The mt_spring addon for instance still can be found on this Asian site.
I seriously doubt, Helge Mathee has wiped them all from his hard drive...
After he's settled into his new job, I guess some gentle and friendly begging

might get him to find a new host for his many goodies...
(And I happen to know of one site that would happily provide such a service)

edit: the following tools still live at their respective "local backup" at rray.de:
SurfaceTextureSupport / ImageToWM / Motiongraphics Addon 2.0 / FormulaXSI / ScatterXSI / Solidify2 and RayLimiter ;)

Mathaeus · Post by **Mathaeus** » 13 Dec 2009, 23:48

Hi all,

just to let you know, I have a new version. This one is designed by following experiences of artists that used compounds for their jobs - all five including me

of them that gave me the usable response. Generally, it's more 'compact', allowing simple ICE trees when you don't want to use simulation. Also, some nice styling features from first version, such as 'loose strands', various randomizing options, are there again.

So what's new:

- much better interpolation, splitting into locks-clumps, multiresolution styling (curled curls). Now it using method described here, where it creating strand's groups regarding of "discontinuities" in emitter's mesh UVs.

- non-simulated hair can be created from single point cloud with single ICE tree, just as in first version.

- it's possible to create hairs, together with modifiers, over strands generated with factory compounds, using them as a hair guides. That includes a famous 'flow around surface' simulations.

- Fur: complete procedurally generated fur. By mentionig 'fur' I think of strands that takes their orientation from emitter's surface. It also using emitter's UVs as an initial orientation, so you can get something usable in a really short time - of course, everything can be modified later. Strand profile is generated by ICE compounds - something like LM Gradient Hair Comb, but with all tools that ICE allows. All hair modifiers will work with fur too.

- Simple, one-point fur dynamics, that still allows bending of strands.

I expect to release it sometimes around (this

) Christmas - New Year. It will be free. Possibly, I'll add more options to fur, such as taking deformations directly from underlying skeleton, external up vectors, so on. Only factory nodes were used, so it should work nicely with all versions from 7.01 to the last one.

Cheers

rray · Post by **rray** » 15 Dec 2009, 19:39

Brilliant - sounds very exciting for lazy people like me, simpler setup for still renders, yet more control options.

Thank you for all your work on this, will definately use it. I have a lot of heads lying around that need a little hairdo

Squizmek · Post by **Squizmek** » 16 Dec 2009, 13:37

rray wrote:I have a lot of heads lying around that need a little hairdo

heads lying around ->

Thanks for sharing your compounds mathaeus! Very nice work!

druitre · Post by **druitre** » 16 Dec 2009, 15:49

appreciated! (BTW, the previous versions have always worked flawlessly on my 7.0.1 vista 64/ now win7 64. This after reading all the posts regarding problems.)

Mathaeus · Post by **Mathaeus** » 22 Dec 2009, 01:22

Thank you guys

So what new we have here, is a fur 'vector field'.
Basically, fur takes it's orientation from emitters UVs, but direction is modified by rig of nulls. That gives a much smoother interpolation than taking the whole orientation directly from surface, also it helps in avoiding visible UV seams.
As you can notice, UVs that I've used is just an "ordinary" UV layout. Fur shape is modeled by ICE compound, something similar to factory strands compounds. Here I've used only one shape, but it is possible to combine many of them.

It's a single point cloud with single ICE tree - and just another one that creates necessary attributes on emitter.

Next time, I hope I'll show fur dynamics. It really isn't ground breaking, but it should be very simple for use.

Cheers

And of course, here is draft render - for hair, I've used existing "Follow Nurbs" modifier (yes I wrote tutorial for that

)

Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C

Re: Kristinka C