Author Topic: Ior vs reflection colour  (Read 9355 times)

2015-11-29, 12:50:31

Rhodesy

  • Active Users
  • **
  • Posts: 553
    • View Profile
I'm trying to work out the correct way to create darker metals. From how I understand it I can create a metal by just using the reflective channel (c4d version so diffuse is left unchecked = black). I can then adjust the Ior upwards to around 20 to create a reflective metal. The brightness of the material increases as the Ior increases and I as I understand it metals are generally about 20. So if I want a dark metal like a gun metal should I control the brightness of the material by darkening the reflection colour value, or, by reducing the Ior back down, or both? Then adjust the glossiness to suit.

Also for other Ior values, coming from vRay I would use a value of about 3 for the majority of materials like concrete/brick/wood and maybe 5 for plastics. Is this the same in corona? I see the default range in the slider on c4d just goes up to 3 which makes me think 3 might be a bit high.
Thanks for any help on this.

2015-11-29, 17:38:25
Reply #1

Nejc Kilar

  • Corona Team
  • Active Users
  • ****
  • Posts: 1245
    • View Profile
    • My personal website
Oh boy, a good one!

In general for pure metals you would put diffuse to zero and just use a high IOR reflection. But things aren't that simple as metals are usually oxidized, coroded and or painted which means you will probably need to focus on recreating the last layer that was applied (think painted metal here).

IOR by itself defines under which angle a material is reflective. So, if you put low IOR values in then an object will only be reflective at grazing angles (the parts that are facing you the least) and vice versa.

Reflection color on the other hand defines the strength of the reflection. Say you have a plastic material and it is a little worn out, you might want to put an IOR for the plastic in and tone down the overall reflection brightness to get the desired look and feel of the material.

From my understanding most dieletrics (non metals) fall into the default 1-2 IOR range. There are lots of lists online that have specific IORs for specific materials, a quick google for an IOR list might help you out best here ;)

I would suggest reading up on the Allergorithmic PBR guide (https://www.allegorithmic.com/pbr-guide) which is great for people who have a clue as to what does what but want to expand on that knowledge with a little theory.

Hopefully that made sense, and hopefully everything is correct. If I've gotten anything wrong someone please correct me ASAP! :)
Nejc Kilar | chaos-corona.com
Educational Content Creator | contact us

2015-11-29, 17:39:35
Reply #2

Juraj

  • Active Users
  • **
  • Posts: 4743
    • View Profile
    • studio website
There are two options, both are sound:

A )One more "artistic" (it's still physically correct):

Keep using the simplified IOR somewhere between 10-20, or just keep 20. Then adjust specular color to the average of perceived luminance + color. Basically what you see. Aluminium is bright gray, Chrome is medium gray, Iron is almost black, colored metal is color of that colorant,etc.. The rather flat curve of high IOR might not correctly reflect that given metal exactly, but as long as the diffuse is pure black (=0 ) it will look metallic.

B) Using correct 'Complex' Fresnel. Metals use additional parameter called 'extinction modifier' which you need to account for to represent correct reflective curve.
   Best way to do this is by using free SigerShaders map plugin you can download here:  http://www.sigerstudio.eu/shop/plugins-free/sigertexmap-complexfresnel/
   Everything else on that is on that page ;- )


By the time I wrote this (heh, minutes), nKilar wrote good advice concerning layering. Indeed, a oxidized metal is no longer metal (like Corten Steel for example), but just rust, and painted metal, is only paint. Combined materials should therefore be ideally created using masks.
Please follow my new Instagram for latest projects, tips&tricks, short video tutorials and free models
Behance  Probably best updated portfolio of my work
lysfaere.com Please check the new stuff!

2015-11-29, 21:49:54
Reply #3

Rhodesy

  • Active Users
  • **
  • Posts: 553
    • View Profile
Thanks a lot to you both for your knowledge. Alas I am a C4D user and as such dont have access to the Siger shaders which is frustrating as I would buy them all without hesitation. Some real beauties in there. Thanks also for the link to the PBR pdf, I will sign up for that.

I have looked into IOR lists in the past and just done it again now and to me they dont line up as Id expect. Perhaps because they might take the metal in its raw rough state? For example you will struggle to see a material over IOR 3 and most in the 1.5 - 2 range. So its a bit of a finger in the air job I find. But the beauty of corona is its simplicity and its fun and quick to tweak so I dont mind a bit of tinkering to get the desired look.

Good to know that metals will always have a diffuse colour of 0 and everything is done in the reflection channel, makes it easier to focus on just one set of settings.

2015-11-29, 22:22:04
Reply #4

Juraj

  • Active Users
  • **
  • Posts: 4743
    • View Profile
    • studio website
you will struggle to see a material over IOR 3 and most in the 1.5 - 2 range.

Which is correct, because for most, all the non-metals do in fact fall into this range, which is n=1.333 (water) -- 2 (some shiny plastics or woods), but you can keep the default 1.52 in 99perc. of (non-metal ) mats without issue.

Metals do in fact also have this IOR so low, but also the extinction parameter (which is =1 for non-metals, thus we don't account for it), which together create a curve that can be sort of similar like high IOR number.
Thus, using something like 20 in simplified IOR model, or correct numbers for both n and K in complex model, can look 95perc. visually identical to our eye. It depends on individual cases, it's mostly the lower reflective metals that can benefit from complex formula (or rich metals used for jewelry where exact look is needed).
Please follow my new Instagram for latest projects, tips&tricks, short video tutorials and free models
Behance  Probably best updated portfolio of my work
lysfaere.com Please check the new stuff!

2015-11-30, 00:08:33
Reply #5

Rhodesy

  • Active Users
  • **
  • Posts: 553
    • View Profile
Thats interesting to know about the extinction parameter and its effect that can be simulated by high IOR as things like steel were only coming in at IOR 2.5. Good to know that just using the basic system we can get visually close enough withought going too complex. Thanks again, Rob.

2015-11-30, 00:22:47
Reply #6

Juraj

  • Active Users
  • **
  • Posts: 4743
    • View Profile
    • studio website
The 'PBR' system doesn't even use fresnel controls in this way because visual difference is fairly minor with exceptions (like precision for jewelry rendering, but those aren't done in UnrealEngine :- )  )
I mean, our eyes can't really distinguish between 60 or 65perc. reflectance which will occur at certain angle, which is what the complex fresnel describes. So for absolute physical corrected, and specific needs it's good to use,
but by all means, not at all necessity.
I create Chrome completely different for every application :- ) Sometimes I use intentionally super low reflective chrome to get contrast in room where there is none. It looks physically correct because it's still metal, it just isn't real chrome.
But who cares ?

The "inverted" way in PBR comes from Disney's paper where they defined certain range for non-metals (1.3-2.4 IOR ) and interpolated them to [0 - 1] values for artistic control, which default 0,5 being something like 1.52 IOR (default in Corona).
For metals, they interpolated something like 8-40 IOR into [0 - 1 ] again, with default somewhere in middle.

It's pretty smart solution because you will always work with physically correct reflectance, and everybody can tell if something is metal or not, or most of the time :- ) And default values reflect that. Going more complex than that is rarely benefitial. Good textures are far more important.

Please follow my new Instagram for latest projects, tips&tricks, short video tutorials and free models
Behance  Probably best updated portfolio of my work
lysfaere.com Please check the new stuff!

2021-01-06, 02:02:20
Reply #7

n2graf

  • Active Users
  • **
  • Posts: 55
    • View Profile
Hey

What is the physical reason why there cannot be more IOR than 3 for dielectrics? The angle of refraction varies according to the composition of the material, the denser materials at high pressure could create higher angles, right? Is there any limit?

2021-01-06, 11:22:57
Reply #8

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12708
  • Marcin
    • View Profile
What is the physical reason why there cannot be more IOR than 3 for dielectrics? The angle of refraction varies according to the composition of the material, the denser materials at high pressure could create higher angles, right? Is there any limit?

The short answer is that there is no dielectric material in real world that would have IOR higher than 3 for visible light. If you believe that's not the case, and having the possibility to set IOR higher than 3, could you provide an example where this could be used? I would then find out more.

Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2021-01-07, 14:23:43
Reply #9

n2graf

  • Active Users
  • **
  • Posts: 55
    • View Profile
No, its only a phisical question. Im wondering wy is no materials higher than 3. I want the long answer jajajaj

2021-01-08, 11:09:57
Reply #10

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12708
  • Marcin
    • View Profile
There is simply no real life dielectric material that would have an IOR higher than 3. You can confirm this by checking the IOR lists:
https://pixelandpoly.com/ior.html
There are elements like Germanium with IOR over 4, but this dude is really special and it is even called a "metalloid".

Here is a more professional website: https://refractiveindex.info/?shelf=main&book=Ge&page=Aspnes
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us