I am creating materials with node wrangler and notice it doesn't use the OCC textures ever, so I was wondering if you could point me to where to use it.
1 Answer
Assuming the OCC texture is ambient occlusion, there's no direct input for it on a Principled Shader.
It's also really important to note that it's not really a sensible texture to use with the Cycles renderer. See these other questions. Cycles already does, using ray tracing, what the ambient occlusion texture is trying to "fake".
If you're using Eevee, then there's an ambient occlusion option in the rendering tab, though it's only an approximately, and you might not like the results. For Eevee, a reasonable thing to do would be to use the ambient occlusion texture as a factor to mix the diffuse texture with black. If you use it, it's probably a good idea to make sure Eevee's own ambient occlusion feature is shut off.
Using the low-res versions of this crate texture, here's the node set-up I used to incorporate the diffuse, normal, ambient occlusion, and reflective occlusion maps:
Here's the result in Eevee with Eevee's own ambient occlusion turned off. The middle crate is shaded using the above node network. The left crate shows what happens if you leave the ambient occlusion texture out (just diffuse). The right crate, for reference, isn't shaded by Eevee -- it's just the exact texture (displayed using an emission node) from the reference photo for the texture.
-
$\begingroup$ A very enlightening answer, sir. One thought - is it wise to connect directly into the specular input, considering how Principled BSDF handles it? blendernation.com/2018/11/05/… $\endgroup$ Commented Nov 20, 2019 at 8:02
-
1$\begingroup$ Good question. I think it's probably okay. I believe the shader's specular input specifies the fraction of power reflected at a head-on angle (which for PBR materials should be given by the formula [(IOR-1)/(IOR+1)]^2). It's scaled so that 0.5 corresponds to the approximate 4% head-on reflection of a glass surface, which is a good value for most dielectrics. If the ROCC map represents percentage of full reflection (at angles close to normal) the multiplication probably make sense. $\endgroup$ Commented Nov 20, 2019 at 16:50