Help A Newb Out - Reflection Color

[torsu]

Watching tutorials, I see that people use anything from full white down to 180 or so in the reflection color slot. Megascans imports, on the other hand, come in at 50%. So, what is the correct value for reflection color? And, perhaps more importantly, why?

[dgruwier]

The correct value for reflection should in theory be 100% white unless you know what you're doing. The reflection strength of the material is controlled by the IOR value, which assumes the reflection color is white.
Reflection color maps (or specular amount maps) used used to be commonly used to emulate glossiness, but now that glossiness maps are common, it's less important.
Sometimes you do want to vary the reflection strength across a material if your mater contains multiple substances, like dirt, moss, glass. In theory you could map the IOR value, but that can be cumbersome, so you might just want to use the reflection color for that. Use wisely.

The other exceptions I can think of:
- Iridescence, if you're doing soap bubbles, oil spills, or thin film interference effects in general (Some ordinary materials like plastics can benefit from a very, very, very thin film interference effect)
- Metals, if you're not using the more correct metalness feature (Read up on dielectric and conductive materials and find the "Understanding metalness" blog post on the Chaos Group blog if you're curious)
- Metallic car paint, like with the stochastic flakes material.

[torsu]

Excellent, thank you for the thorough answer! This helped a lot.

[Karol.Osinski]

dgruwier Thanks for reminding me about the metalness implementation being more physically correct. I gotta go back to that blog entry (3rd time I think lol). I remember I literally checked it once while making brushed brass material and used it instead high IOR value. The results were surely different but I am not sure if nicer or it was placebo effect. Need to do some more testing on and show it to others in studio. Cheers!

[dgruwier]

Yeah, can’t guarantee that the difference in realism is going to be immediately obvious, but small inaccuracies accumulate. Might as well use the calibrated, physically correct values when they are right there

[torsu]

Sorry to revive an aging thread but I've going around thinking about what you, dgruwier , said about how plastics can benefit from very thin film interference effects. Could you expand on this? Plastics are something I've been struggling to reproduce - they never seem to look 'plastic-y' enough. I've had a look at the Siger thin film shader. Cool stuff, especially with a map thrown into the thickness slot. I see that there's a V-Ray page for a thin film OSL shader, not sure if it's any different or 'better'. Nonetheless, I'm not quite seeing how this will apply to plastics in general but this is probably due to my lack of knowledge in the material science bit.

[dgruwier]

It's very unlikely that it'll help you, it's something very subtle to add in a few cases when you already have a good looking material, but you can read some about it here:
https://blog.maxwellrender.com/tips/...-of-materials/

Don't bother with any of the thin film shaders IMO, too hard to control. Who knows what the right values are to use, so why bother using real world values anyway. Just use a fresnel node and remap it with a gradient and pick your own colors using reference images, so you have full control.

[torsu]

Thanks for the response! In that article about Maxwell, it looks way better and hardly subtle at all. We'll see how it works out in V-Ray. 90% of what I do is on a hobby basis, so I've got all the time in the world to mess about with "trivial" details.