Anisotropic surface texture

Hi,
I am trying to apply an anisotropic surface to this filter but I really don't know what I'm doing. Can someone tell me if the surface of this filter appears anisotropic? And if it does not; what term is used to describe the surface as it appears.

Thanks,
Ron

CG education

try this its worth looking at the whole site would be of interest I hope for you

Wether you achieved it or not, depends on what level of anisotropy you are focusing on. If you are only talking about the texture surface of the orb, I guess it makes sense to speak of anisotropy ("property of being directionally dependent"), due to the surface light/shader model. However the whole orb and the way the texture wraps around, I think, doesn't appear anisotropic - you need the texture to wrap around the orb.. i.e. transform the heightmap using spheric mapping

Check out how the dimpels follows sphere curvature in this example: http://www.filterforge.com/filters/4633.html

Very fake looking but I think I can correct it with offset.

The difference is the noise is now preferring a direction whereas in the first two the noise was random. From what I gathered from the article anisotropy requires a predominant direction. I will post a better one shortly.

Thanks

Quote
Sphinxmorpher wrote: anisotropic lightning

Wait a minute; we are talking about surface characteristics not lighting, is that correct or am I missing something?

Quote
ronviers wrote: Wait a minute; we are talking about surface characteristics not lighting, is that correct or am I missing something?

The two things are connected more than you'd think - bidirectional reflectance distribution is something the shader produces based on surface properties/characteristics - so in order to make this work at all, you must have a shader that supports illumination of such properties - do a search on BRDF

Quote
Sphinxmorpher wrote: connected more than you'd think

Indeed.

Here is one where I fake a light source shaped like a fluorescent tube. Not pretty yet but I think I can get there. I will do another one with a light shaped like a ball.

anisotropic has quite a few technical definitions, depending on what field of science or endeavor you're talking about, but the simple definition is:

Showing differences of property or of effect in different directions.
www.mpoweruk.com/glossary.htm

Not possessing the same properties in all directions (the opposite of isotropic). Because of its fibrous structure wood is anisotropic.
www3.sympatico.ca/3jdw8/glossary.htm

and for a slightly more technical definition when applied to optics and light:

Materials that are anisotropic have non-uniform spatial distribution of optical properties (for example, refraction, transmission, reflection). ...
www.olympusmicro.com/primer/techniques/dic/dicglossary.html

and notice in that last one that it's the material that is anisotropic.

well, most of this is over my head at the moment, but what little i can glean makes me think you might want to take a look at the normal maps in FF.

Quote
Kraellin wrote: dispersion routines

Here is a simple way to fake it using motion blur.

Anitest.ffxml

cause here's what i did:

Anisotropic test 1.ffxml

the rainbow gradients look more like some lens issue IMO... I think for the effect to work visually you need some curvature and shading

I just submitted a custom shader filter - perhaps you can modify it somehow, you should check it out (though its at a prealpha stage)

http://www.filterforge.com/filters/4648.html

That looks interesting Sphinxmorpher but I'm not really ready for that. Besides, and this is nothing personal, I don't like looking inside other peoples filters. It looks like a really cool project and I will keep an eye on your progress. Good luck.

Quote
Kraellin wrote: the rainbow part isnt right

Yes, stainless steel is entirely the wrong material for rainbow refelctions. The reflections should still be offset and blurred according to viewing angle and surface characteristics but just not rainbows. That's why it would be nice to be able to take luminosity information from the HDRI environment to use to define the shading network - as it is, I think, we have to fake anisotropic reflections.
I like what you did with your filter using the threshold and alpha components but wouldn't you need to stretch or otherwise give the surface definition a preferred direction?

ron, glad you liked it, but i dont even remember what i did now i'm guessing there is a tighter definition for aniso than just 'irregular', which is pretty much what i based my version on. i just made 'irregular' lighting/reflections... sort of. i'm sure someone more versed in the 3d industry could do much better.

The colors are from the environment - not faked.

ok, those are interesting.

but correct me if i'm wrong here, but does that last one look like it's rising in the middle, like it's being pushed up from underneath?

oh, these look MUCH better within the filter rather than saved out as compressed .jpgs!

I realized I was multiplying across a quad the same as the single - I just used a forth as much - no problem. Thanks for spotting the problem for me!