spice

Thanks. Here is a version including the shadow call in the illuminance loop.... -Arno arno_hair_far_field_v_1_9_shadow.zip

I will check that tomorrow... -Arno

Thanks for testing ! I would really appreciate to see some results obtained with the shader. :-) -Arno

Hi All, Here is a modified version of my shader (for hair with a circular cross section) which hopefully fixes the "bright spot" issue. It would be great if you could perform some tests to see if it works properly. Ciao, Arno [EDIT: Please reload this version !] arno_hair_far_field_v_1_9.zip

Thanks. No, this is not an "x/0" issue. :-) -Arno

Hi All, it sems that those bright spots are a consequence of the limited floating point accuracy of Mantra. However, I'll try to fix this issue. Hence, it would be great to have a simple test scene (with artifacts). Ciao, Arno

So then.... if I find the time there will be a new shader... -Arno

Marschners approach works well, but is quite slow. For a faster accurate solution have a look at this paper: http://cg.cs.uni-bonn.de/project-pages/hai...tion-sicg06.pdf (Note, that my paper was first ! :-D) Moreover, I developed some simpler physically based approximations which produce plausible results, but which are MUCH faster (There will be a definite paper about efficient GI techniques for fibers/hair next year). Cheers ! Arno

Hi, since my shader is quite a complicated thing, I want to give you some additional hints about the usage... The shader consists of four components: R,TT,TRT,diffuse,ambient, where R means a direct (white) surface reflection (first highlight), TT the forward transimission (especially importantant if lit from back) and TRT a strong secondary highlight. These three modes are strongly realted to the model of Marschner et al. The diffuse and ambient components are similar to the usual Kajiya&Kay model and ar NOT physically correct. The total outgoing intensity is calculated as follows: Out=Light*scale_total*(scale_R*R + scale_TT*TT + scale_TRT*TRT + scale_diffuse * diffuse + scale_ambient*ambient) With the scale parameters each component can be scaled individually. Note, that for physically based hair both diffuse and ambient components should not contribute too much intensity. The most important parameters for changing the hair color are the absorption coefficients. They give you the right colors for both the TRT and the TT components. The diffuse and ambient colors should be set according to this absorption (see the tutorial). Beside the color one might want to have more (less) radiant hair. This can be done by changing the scale of the highlights (scale_R,TT,TRT) and their widths (width_R,TT,TRT; the bigger the numbers the smaller the width !!!). For realistically looking hair/fur one needs furthermore variation of the shading parameters. To get more "texture" try to vary: -index of refraction (slightly !!!) -absorption coefficients -the diffuse/ambient colors Cheers! Arno

All right, thanks a lot .....

I mean without the additional coloring, but with the cam and the lights... Thanks Arno

Hi DaJuice, is it possible to get a "clean furball scene", with the ball and the hair only ?! Arno

As always... one version is not enough Here comes the supa dupa release with some nice additional features. The most prominent one is the grouping of the parameters with respect to their meaning and an additional factor for scaling the total intensity. This scale_total factor is especially usefull to adapt the shader to your lighting (it basically scales all incoming light) while preserving the rest. Cheers ! Arno arno_hair_far_field_circular_v_1_6.zip Juice_furball_default_v_1_6.zip

To play around with the shader I attached DaJuice's fur ball hip (with some point lights and the new shader) as a test scene... Arno Juice_furball_default.zip

arno_hair_far_field_circular.zip Hmm... you are right, the vfl is not included, but the shader should work anyway ?! If the file is broken try this one: