IOR for stone?

[tricky]

I am starting to play around with IORs and fresnel reflections - never really messed around with them before and have always used falloff maps in the reflection slots of reflective materials - glass/ceramics/metals etc.

However, I thought I'd use the table of IORs at vray.info. Lots of info in that table, but what would I use for stone - a nice, limestone, like 'freestone' or ashlar I guess.

[zoranm]

I don't think you'll find a set value anywhere. Value depends on how much is a stone polished, therefore reflective. Lighter stones reflect less than darker ones. Honed stones got a strong glossy component.
In other words, what looks right is right.
I had scenes where the same stone texture had IOR=1,3 and in the other camera motion IOR=2 and it looked OK.

Hope this helps

Zoran

[womble]

Just a quick question as I have trouble understanding the math involved here but:

Am I right in thinking that a higher incidence of reflection = more reflective. As in a stone material with IOR = 1 will be less reflective at glancing angles than one with an IOR of 10?

[vlado]

You are correct; in fact, IOR=1 will not produce any reflection. Larger values will produce stronger reflection.

Best regards,
Vlado

[andrewjohn81]

doesn't Fresnel actually make the object Zero reflectivity at a directly perpendicular angle? When would you really want to use that?
We've got some fancy color scanners here, some of which do calculate reflective values at different angles. First off, those are Always different, and I have also never seen a material that has no reflectivity at 90 degrees when it has any reflectivity somewhere else.

I am confused at our Fresnel computer model. I would think the curve would be relative to the reflectivity, but it doesn't seem to be.

Does anyone know if this is actually correct? And care to explain just a little?

[vlado]

doesn't Fresnel actually make the object Zero reflectivity at a directly perpendicular angle?

Nope. The Fresnel function is never really zero for IOR > 1.0. The IOR only changes precisely the reflectivity at directly perpendicular angles.

Best regards,
Vlado

[andrewjohn81]

well maybe it just doesn't look right to me. If you put in the IOR for plastic, wouldn't that be perfectly smooth plastic? If it's perfectly smooth plastic is fairly reflective no matter the angle. It's definitely more reflective at more extreme angles, but I would think that I should see reflections perpendicular as well.
I have seen materials that only change the IOR and keep the reflectivity at white. Even those materials seem to loose nearly all of the reflection at the perpendicular angle, so where it may not be zero I see it visually as too extreme. And isn't what we do only visual anyway. If it looks right it is right. Right?
It would be nice to just plug in the "correct" values for things like IOR and reflectivity at least as a starting point though.

[vlado]

Many materials do not follow the Fresnel law when reflecting; it is relevant mostly for dielectric materials like glass, water and metals; this means that it may be impossible to get the desired look by adjusting the Fresnel IOR only. For absolutely correct result, ideally you'd measure the BRDF reflectance of the particular material and use this data to drive a shader. Of course, in that case you'd loose all possibility to adjust the shader itself beyond its scanned appearance...

Best regards,
Vlado