Author Topic: ELI5 - How do i work with CIELAB and real world gloss levels in corona?  (Read 8600 times)

2019-05-29, 16:18:59

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
Im trying to replicate some specific RAL finishes and im not sure of the physically correct way of working with CIELAB values and gloss levels. Never had a client supply the data in this format before so im not sure if the conversion to rgb would be accurate given the smaller gamut of RGB within CIELAB.

RAL 7038   delta E: 4,49   Gloss level:13   L: 76,88; a: - 0,16; b: 1,68
RAL 5008   delta E: 2,23   Gloss level:15   L: 33,93; a: - 1,53; b: - 7,04

Do i simply convert the lab to RGB?
Does real world gloss translate directly into a glossiness value in corona?
e.g. gloss level 13 would that be 0.13 in corona?
This feels like something dubcat would have experimented with but he hasnt been around for a little while.

« Last Edit: 2019-05-30, 11:30:10 by jpjapers »

2019-08-04, 06:14:52
Reply #1

dubcat

  • Active Users
  • **
  • Posts: 425
  • ฅ^•ﻌ•^ฅ meow
    • View Profile
Hey!

I'm having a hard time sleeping and thought, why not do a random dubcat google search, and here I am.

RAL 7038 is pretty easy to convert to Albedo. RAL 5008 will get clamped because the LRV is 4%. I don't know if Autodesk is the culprit, but both Corona and Vray will clamp these values to 0 RGB.



edit:

I forgot to answer the gloss question. Real world materials usually don't go bellow 0.4 PBR glossiness. 13 gloss level does not equal to 0.13 pbr glossiness.

This is fstorms glossiness level mapped to PBR glossiness.



Current PBR glossiness does not exceed the material IOR. So if you do a cross polarized material scan of asphalt, and normalize the scan based on the F0 of the asphalt. And generate a glossiness/roughness map base on the asphalt F0. The gem stones in the asphalt will not have the correct IOR. Because the gemstones can not exceed 1.5 IOR. We need a new shader that read both glossiness/roughness and IOR/Specular at the same time. Modern render engines can not fake this stuff with geometry. Corona is stil using complicated IOR instead of real world easy math Specular.
« Last Edit: 2020-03-06, 04:10:08 by dubcat »
             ___
    _] [__|OO|
   (____|___|     https://www.twitch.tv/dubca7 / https://soundcloud.com/dubca7 ( ͡° ͜ʖ ͡°) choo choo

2019-08-04, 12:22:43
Reply #2

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
Thank you for finding this haha. Thats very very interesting.
I take it youre an fstorm user now?

2019-08-06, 23:07:48
Reply #3

dubcat

  • Active Users
  • **
  • Posts: 425
  • ฅ^•ﻌ•^ฅ meow
    • View Profile
I take it youre an fstorm user now?

Hey!

I still use Corona / 3dsMax professionally, but I always stress test other render engines. My current absences is not the result of an engine switch. I'm happy with the current Corona development, but there are currently two SUPER important features that are lacking in Corona. Fabric BRDF ( You have to copy paste the URL, the direct link URL will redirect you to the index) https://blog.selfshadow.com/publications/s2017-shading-course/imageworks/s2017_pbs_imageworks_sheen.pdf and Burley Diffuse Roughness https://disney-animation.s3.amazonaws.com/library/s2012_pbs_disney_brdf_notes_v2.pdf.
Both these features are included in MaterialX https://www.materialx.org/assets/MaterialX.v1.37.PBRSpec.pdf that even Substance Designer/Painter has implemented.
« Last Edit: 2020-03-06, 04:07:12 by dubcat »
             ___
    _] [__|OO|
   (____|___|     https://www.twitch.tv/dubca7 / https://soundcloud.com/dubca7 ( ͡° ͜ʖ ͡°) choo choo

2019-08-07, 12:25:23
Reply #4

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
Yeah im getting a little jaded of having to do workarounds for fabric if im honest. And brushed metal/microdetail. And displacement.