Fresnel IOR

[Byriel]

Hi

I've been testing around with the VrayMtl shader, and there is 1 thing i cant really get my head around..
In short: How am i supposed to determine fresnel IOR values, for any given material?

When using the PBR workflow (thus using metalness) Fresnel IOR values tends to be quite low in value (here referring to Vlado's IOR values, posted in Chris Nichols "Understanding Metalness" article) ranging from 1.002 for Aluminium, to 1.35ish for Gold...

However when i dont use a PBR workflow (so keeping metalness at 0) i need to use the Fresnel IOR at MUCH higher values to get conductive-like results - and nothing really seem to correspond "correctly" with any IOR lists or calculators i can find..
So, how do i actually determine which IOR value to use for any given material, in the different workflows?

Thanks

[Vizioen]

I use complexfresnel: https://www.sigershop.eu/free-3ds-ma...omplexfresnel/ and than tweak till it looks good You could do this by hand with curves like Grant Warwick used to explain but that's not very flexibel and a lot of work. So this is quite handy.

[Byriel]

Vizioen Wouldnt the problem be the same though ? having to do 2 different kinds of IOR, because of the use of Metalness in one kind of workflow, and not using metalness in the other? ("native workflow" vs PBR workflow)

[Ihno]

They say the values from the metalness blog post are correct (There might be a big variation from alloy to alloy, but we have a ballpark).
The old cheat to put an ior of 5+ is unrealistic and I don't think there is a reason for going down that route anymore.
The complexfresnel map won't work on gpu and is slower than the native metalness.
I don't see a reason to use something else than the native metalness.
Reflective IOR lists on the internet seem to be always wrong for metals, don't trust them.

[Ihno]

When you use metalness think of it like this:
The texture you put in the diffuse will be the color for looking at it at an angle of 0° and towards 90° it blends to the color you put into the reflection amount (white).
With the IOR you tell vray how the curve between 0-90 should look.

[Byriel]

Ihno Thanks - Yes i know how the IOR itself works - its just confusing when you get lists from Vlado that are corrected when using Metalness of 1, which in itself dosn't match anything from any other lists or calculators. And then you have the issue when not using metalness, that you need 5+ IOR values (as you call it, cheat) but - obviously this dosnt match anything either, not even the metalness workflow. So im just wondering how im supposed to figure out which IOR values to use when i setup a material using metalness and NOT using metalness (if i dont want to use a PBR workflow) - there must be some logic to the values that the VrayMtl uses surely?

[dgruwier]

Why not just use metalness, PBR or not? What in particular prevents you from using it like it's described in the blog any time you need a metal? You're simply not going to find IOR values for metals that work accurately in non-metalness materials, that's the whole reason metalness is there in the first place.

[Byriel]

dgruwier Hey dude - Well sure its possible - however, the list created by Vlado in Christopher Nichols "Understanding Metalness" article just dosnt add up when using metalness, in fact, alot of the values seem "flipped" when looking at IOR values from calculators, lists etc. example: Chromium IOR of 2,97 ish vs Vlado's Chromium IOR of 1.03 - Gold IOR of 0,470 vs. Vlado's 1.35 (IOR Values taken from pixelandpoly and cross checked with refractiveindex.info to make sure they are atleast close to something physical correct ) - It might just be getting way too technical with it, but it annoys me that everyone talks about using IOR lists and calculators, to figure out your values, when none of the values, seem to be whats supposed to be used or looks correct?

Fun fact: if you multiply the real world IOR values (ex. Chromium 2.97) with 6.5 you get 19.305 which is fairly close to what the IOR would usually be when not using metalness.. (so the "old way" i guess?)

At the end of the day, im fine with the propper workflow just being "guess your IOR" but im trying to create some material for our students, so they have some kind of logic they can use, when they need to address their IOR values, so im basically just searching for some sensible way to figure out what your IOR should roughly be, when you want to create either dielectric or conductive materials (Dielectric might not be such an issue, since IOR values seem to be ok then)

[Ihno]

I don't know If this will help, but after posing in this thread I wanted to test it and maybe its useful for somebody:
https://youtu.be/B_-MqgCAUnE
I used a Vray mtl with metalness at 1, a pure red diffuse color and a pure green reflection color.

Maybe somebody smarter than me will have an answer for your questions regarding internet IOR lists vlado ^Lele^

[dgruwier]

dgruwier Hey dude - Well sure its possible - however, the list created by Vlado in Christopher Nichols "Understanding Metalness" article just dosnt add up when using metalness, in fact, alot of the values seem "flipped" when looking at IOR values from calculators, lists etc. example: Chromium IOR of 2,97 ish vs Vlado's Chromium IOR of 1.03 - Gold IOR of 0,470 vs. Vlado's 1.35 (IOR Values taken from pixelandpoly and cross checked with refractiveindex.info to make sure they are atleast close to something physical correct ) - It might just be getting way too technical with it, but it annoys me that everyone talks about using IOR lists and calculators, to figure out your values, when none of the values, seem to be whats supposed to be used or looks correct?

Fun fact: if you multiply the real world IOR values (ex. Chromium 2.97) with 6.5 you get 19.305 which is fairly close to what the IOR would usually be when not using metalness.. (so the "old way" i guess?)

At the end of the day, im fine with the propper workflow just being "guess your IOR" but im trying to create some material for our students, so they have some kind of logic they can use, when they need to address their IOR values, so im basically just searching for some sensible way to figure out what your IOR should roughly be, when you want to create either dielectric or conductive materials (Dielectric might not be such an issue, since IOR values seem to be ok then)

You need to think of the IOR for dielectric materials and IOR for conductive materials as two different things. Calculations based on dielectric materials (which I think is what you're finding online) will not translate into conductive materials, and definitely not into Vlados implementation. He wrote a script to convert the values from refractiveindex.info into something that metalness understands, and that's where the values you find in the blog post come from.

The problem is that there's much more going on with metals than regular materials, it's not just mapping the reflection strength in an even curve from 0-90 degrees. Metals have a more or less constant reflection strength that only curves upwards right towards the edge, and they have different reflection strengths at different wavelengths (which is where the color comes from), all converging to white near the edges.
I'm not sure exactly what happens behind the scenes in the metalness shader, but all you need to know to use it is that the diffuse color becomes the color AND reflection strength of the metal, and the IOR value only influences how it transitions to100% reflective white near the edges.

Bottom line, there doesn't have to be any guesswork involved: if your material is not a metal, look up the IOR online, done. If it IS metal, use metalness in VRay or another shader that mimics a true conductive material, then follow its particular instructions.

[Byriel]

You need to think of the IOR for dielectric materials and IOR for conductive materials as two different things. Calculations based on dielectric materials (which I think is what you're finding online) will not translate into conductive materials, and definitely not into Vlados implementation. He wrote a script to convert the values from refractiveindex.info into something that metalness understands, and that's where the values you find in the blog post come from.

The problem is that there's much more going on with metals than regular materials, it's not just mapping the reflection strength in an even curve from 0-90 degrees. Metals have a more or less constant reflection strength that only curves upwards right towards the edge, and they have different reflection strengths at different wavelengths (which is where the color comes from), all converging to white near the edges.
I'm not sure exactly what happens behind the scenes in the metalness shader, but all you need to know to use it is that the diffuse color becomes the color AND reflection strength of the metal, and the IOR value only influences how it transitions to100% reflective white near the edges.

Bottom line, there doesn't have to be any guesswork involved: if your material is not a metal, look up the IOR online, done. If it IS metal, use metalness in VRay or another shader that mimics a true conductive material, then follow its particular instructions.

That is an interresting way of seeing it... I dont think i've thought about it as the IOR lists being for dielectric only (even though it annoys me that the lists include conductive materials) but i guess thats where the script comes in handy, as long as metalness is used.. Which might be the correct and only path from here...