VRayDirt

[CCS]

I've been scratching my head all morning.

It seems based on the options in the vraydirt map that there would be some way to control where dirt is placed. What I am looking for is a way to have dirt have most effect adjacent to horizontal surfaces, and little effect adjacent to vertical surfaces. I've tried using a perp/parallel falloff in the radius, which seemed to make the most sense, but couldn't figure out what it was doing. I also tried to accomplish the same thing using the occluded color slot. I can't figure out how this map is supposed to work, beyond the basic parameters.

Anyone have any ideas?

A related question...
There are X, Y, and Z controls for bias. XYZ what?
World axis directions? Ray normals relative to the surface receiving the dirt?

Likewise, if I put a falloff in any of this map's slots and I specify World X or Camera Z or whatever, what does that direction refer to? The normal of the surface receiving the dirt? The rays cast to determine dirt placement?

[Dynedain]

Try a falloff map with a world Z axis in the dirt slot... but I think you'll probably have to tweak it for each object or close groupings of objects.

[joconnell]

Heya!

Dirt is based on the mesh itself and most of the time it's based on the surface normals - anywhere you have polygons meeting at a 90 degree angle or thereabouts you'll get dirt - If you have a look at your model and take a look for these areas, it's normally the places where dirt would collect in the real world, the parts of the model that would be sheltered from the elements.

Could you paint a small example of what it is you want and maybe we could offer some suggestions?

[CCS]

Try a falloff map with a world Z axis in the dirt slot... but I think you'll probably have to tweak it for each object or close groupings of objects.

That was the first thing I tried. I couldn't figure out the results though. Not what I expected.

Dirt is based on the mesh itself and most of the time it's based on the surface normals - anywhere you have polygons meeting at a 90 degree angle or thereabouts you'll get dirt - If you have a look at your model and take a look for these areas, it's normally the places where dirt would collect in the real world, the parts of the model that would be sheltered from the elements.

I know how dirt works. I understand the basic dirt parameters and have used it for some time. What I was experimenting with is trying to get more dirt on horizontal surfaces than on vertical surfaces. For instance, one way would be to have a radius of 4" adjacent to horizontal surfaces and 1" next to vertical surfaces. Or - another way would be to have a black occlusion color next to H and much lighter grey next to V surfs.

[joconnell]

Okay - so it's almost like having some kind of gravity type controls over the spread of the dirt for the bottom of posts in the ground so you'd have control over how far up on the pole the dirt would go versus the radius of its spread on the ground?

[joconnell]

Just did a quick mess in the demo version - if you increase the Z bias it'll do as you want - it seems to work in world coordinates - biggers numbers pull the dirt lower down the objects.

[CCS]

Not exactly, but close. It's actually simpler that what you are saying. It is closer to how dirt/weathering would actually occur. What we have now is like AO in that it identifies little nooks and crannies that don't receive as much GI. Actual dirt and weathering on an exterior surface is going to occur more on horizontal ledges (window sills, etc) that can gather grime, and even moreso on horizontal undersides which are protected from rain washing it away. Tight areas such as corner trim on a building might get dirt in something like a 0.5 radius, whereas protected areas under a window sill might have a 6.0 radius.

I was thinking I could accomplish this by using a falloff controlled by perp/parallel to world Z. But it doesn't respond as I would have expected it to. If it did, you could even nest an additional falloff in there to separate +Z from -Z. (Actually, a toward/away falloff might cover that.)

Does this make more sense?

[Neilg]

Have you read the spot 3d vray guide?

Those XYZ bias controls, you want those. I would assume the bias is the distribution of the dirt in any of the 3 directions.

[joconnell]

Yup, getcha. Maybe something like normal maps would be a good way togo on this since they'd be more sensitive again to direction. I'll have another fiddle and see can I get better results again.

[CCS]

Have you read the spot 3d vray guide?

Those XYZ bias controls, you want those. I would assume the bias is the distribution of the dirt in any of the 3 directions.

Yup. Sure have. Tried it too. Not what I'm talking about, although that was another question. XYZ meaning.... what? World XYZ? Relative to surface normal? Ray direction?

[Neilg]

I would assume world, but its only 3 variables for trial and error.

On another read I think I mis-read your original question.