www.si-community.com

Hello Mathaeus, I have been using your hair tools a lot recently and have come across a problem I cannot seem to figure out. When strand length is controlled by a weightmap, the strands seem to get cut, and I loose the taper set by strand size in SI units. Is there any way around this?

Strand thickness depends on the absolute distance traveled from the strand's root (That is, unless I'm missing something)
So with this setup it should be ok that hair gets cut that way. Does using the normal Strand fCurve (which uses relative distance traveled) does what you want? This one shouldn't depend on the absolute distance traveled, but would set thickness depending on %age along the length of the strand.
I hope I'll also be able to use KHair in a while again!

Yes, the normal Strand fCurve seems to work fine. Thank you. For some reason I thought I had to use the fCurve in SI Units to keep the thickness constant when changing camera distance. I am still a little (very) confused on this point.

Tickness has to be in SI Units and length in percent units in your case. Luckily it works together!

Btw I found Anto's approach to keeping thickness constant produces better render results than using the built-in "AbsoluteStrandWidth" attribute

I was never able to produce strand taper using Anto's Strand Screen Size In SI Units. SEE That is why I stuck with the ShapeSizeAbsolute method. I am sure I am missing something simple. And it has been explained several times in this forum. But, I just cant seem to get it.

That's weird I just rebuilt the tree exactly like your screenshot, and the tapering works. Could it be that there's some weird multipliction factor in your fcurve?

My multiply factor is default 1. I set it to 0.1 like you, but still no taper. Can you PM me your scene file so I could take a look at it? I should be able to figure out what I am doing wrong that way.

how come that i have the exact same setup, but red nodes...??

caledonian_tartan wrote:how come that i have the exact same setup, but red nodes...??

Probably 'emitter' needs some work too. Emitter is a 'grid'. I wrote something about it in docs.It's not one step process - if you find it overcomplicated, well, all what I can say is, use some another hair system.


Izze, could you post screen shot of expanded 'kH2 Strand Length' node. There were a few versions of this node, I would like to be sure what you're using.
Also, could you be sure, is it a 'cut', or calculation is just lost - 'lost', means exactly the same thickness along whole strand, instead of array. Just point size is used.
Definitively it can happen that ICE considers this array as 'ready to kill' (especially if there is 'get array sum' before, as it is in 'get length' node), and from my little knowledge, various SI versions can behave differently, there. Anyway I *believe* there is a way to fix it 'in advance'.

Cheers

no problem. i forgot to init on the mesh and probably should have read the docs first (or rather again...)
sry

I just replaced all my Kristinka nodes with the latest version to make sure I was using the most recent. After looking at RRay's scene, I find that he is getting the taper from the shader transparency. I am currently using a shader that does not have transparency, so I have to rely on the fCurve. Using the standard fCurve and ShapeSizeAbsolute will keep me working for now.

Still loving these tools. Can't work without them now.

Yes I forgot to remove the shader transparency option - also I finally see what you mean now - I've never noticed before the thickness never goes below ~1 Pixels when you use Anto's Strand Size in SI Units compound.
I just checked this again and I think this is what the "Min Screen" parameter in that compound is for, it's like a minimal tapering width. I just tested it, set it to 0.1, removed the transparency - and it seems to work like it should now!

Yes, Perfect! And I agree, the rendering results do look better with the Screen Size method. The taper looks much more accurate along stand length. Many thanks.

well, nice if it works! As RRay said, 'min screen size' is for keeping the strand size at some 'screen' maximum.

BTW, there is a nice post about this stuff, by Moritz Moeller ('Mauritius', author of Affogato translator for XSI). It's mainly for 3delight, don't know does it all applies to MR too. Anyway, he also talking about 'screen size', in some very advanced way.

Hi Mathaeus

I'm having a problem when I use the "init meshuv emmiter" compound and assign the uvs, as you can see in the image http://duxen.wordpress.com/testing/ seems that the seams are still connected but as you can see in the uv of the texture editor they are not, what could be causing this?, thanks!

duxen wrote:Hi Mathaeus

I'm having a problem when I use the "init meshuv emmiter" compound and assign the uvs, as you can see in the image http://duxen.wordpress.com/testing/ seems that the seams are still connected but as you can see in the uv of the texture editor they are not, what could be causing this?, thanks!

hi,

a 'real' seam on mesh should be disconnected too.
Btw, in upcoming v 3.0 (hope it will be out this or next weekend), there won't be this procedure anymore, as it seems this produces more problems than benefits (benefit is only one emitting mesh) . I'm planning to replace it by tangent map and copy of mesh for deformation (so tangent map could be frozen).

cheers