[bjorn.syse]

.MXS file is here: http://dl.getdropbox.com/u/114792/testscen%20kuddar.mxs

[kami]

wow. the material looks really great now!

[Bubbaloo]

The mesh has plenty of triangles, but some are larger than others. I think the larger triangles (where the dark stripes appear) require higher precision. Maybe your precision of 16 is just not enough. Have you tested it with 32 or 64? It's such a fine displacement map that it needs a higher precision I think.

Anyway, it looks good as is!

[bjorn.syse]

Hi, yep, the blue and walnut version are with 32 in precision. However, the effect was larger when the mesh had a lot fewer polygons too...

Yes it does look good! I'll add them to the MXM-library!

-björn

[bjorn.syse]

Kami, thanks! is does. I mean, it still doesn't have all of that custom falloff i was looking for, but I think it resembles it enough at leasts..

[bjorn.syse]

I've added a few variants (colors) of the Plush-kinda-cloth to the MXM-gallery. They'll probably turn up soon.

Crushed upholstery fabric RUST, WALNUT and ENSIGN (navy)

Best regards,

- Björn

[tokiop]

very nice results, I think the dark stripe looks like "reverse brushed" fabric, which makes it look more realistic. Maybe this could be controled with an anisotropic texture..

[bjorn.syse]

I agree,

could I use an anisotropic texture in the angle map slot, without actually using any anisotropy?

- björn

[polynurb]

hi bjorn,

i had a similar issue with displacement the other day.
I also render from rhino, and had a grass displacement material which always seemed to reverse shadows in certain areas.

i used the rhino labs displacement plug in to create a first layer of displacement and then used the grass on top... so i figured out (cause i saw the mesh not the nurbs) that the direction of the triangulation affected the rendering of the displacement.
so i went back to nurbs, rebuilt as patch and forced 3x3 cm quad triangulation.. which solved the issue.
try to _extractrendermesh from your geometry and see if there is any significant change in direction of triangles where you see the black stripes.

see here:



:p

[bjorn.syse]

Hi Polynurb!

That's very interesting, thanks for sharing! When looking at the render mesh I can clearly see the areas where the displacement map seems to be directed at a slightly different angle. Btw, which direction is it that you are talking about, and how can I analyze the mesh?

What is the reason for this making the displaced geometry look different. Does polygon smoothing have any effect at all when using displacement?



btw, what a lovely looking render!

[polynurb]

..i think what we see in your render is the same effect. I was actually refering to the "direction" of the triangulation ("left-bottom / top-right or right-bottom to top/left") not the normals.. but in the end it's the same thing, something is behaving strangely at the edge of different mesh densities
looks like the rhino mesher and maxwell displacement don't always agree.

it becomes more visible with sunlight or other hard light sources.
the only way which seems to solve the problem is using fix sized quads instead of triangles, but then you have to set the max edge length to the smallest detail you want to achieve; hence you get large meshes .. its ok at rendertime since displacement works better with high subdivisions, but it is not good while working in rhino....

your mesh is very dense allready, is there any apparant reason for the density to be so high where the stripe shows, like surface knots.. or is this due to t-splines meshing?

try to adjust object mesh properties to min and max to the same value ~3 mm, uncheck/set to 0 everything else (not pack textures!) in the advanced mesh options. i think it should fix it.

btw. your material loos great .. super photorealism. forgot to mention

* i highly doubt that bump can get you the same "microshadow" qualities like the material you created there, but finding the "sweet spot" for the displacement setting should make it render at ok speed..