Page 4 of 9
Re: Skin Rendering
Posted: Sun Sep 01, 2013 8:56 pm
by tonysculptor
Sorry for the double post, but here is a lighting test with only HDR. The one above was only spotlights. Which do you all prefer?

Re: Skin Rendering
Posted: Mon Sep 02, 2013 7:55 am
by ice13
WOW!
Amazing work Tony!
Re: Skin Rendering
Posted: Mon Sep 02, 2013 1:13 pm
by gordonrobb
Love them both. They're just conveying different things. I have to say, I'm really looking forward to understanding this shader, and applying it in LW.
Re: Skin Rendering
Posted: Mon Sep 02, 2013 5:25 pm
by tonysculptor
These are the maps I used
These are my material settings
This is the material network
Material breakdown in the next post (its going to be a bit of read).
Re: Skin Rendering
Posted: Mon Sep 02, 2013 6:04 pm
by gordonrobb
You have a gloss map there, does that mean Spec? or are you using it somehow differently?
Re: Skin Rendering
Posted: Mon Sep 02, 2013 6:10 pm
by tonysculptor
Material Theory
The idea for this material was based on reading many papers on the structure and optical properties of human skin. I decided that for a physically correct renderer, I should build a reasonably close to physically correct skin structure. Many papers have measured the reflectance of skin at the utmost oily layer (Sebum), the absorption of light at the Epidermal layer, and the absorption and scattering of light at the Dermis or Subdermal layer.
Basically, I embedded the reflectance component into all the specular materials, so I wouldn't have to blow out my reflectance values. The Epidermal layer only absorbs light, so there is no scattering int it. The dermal layer does almost all of the scattering of light with very low absorption. Although there are components of the dermis that actually scatter the light in a forward direction, based on research, I would say that the scattering direction (Phase) should be set to 0, there is plenty of scattering going in all directions to justify it.
The defaults I have put into the material reflect these findings for the most part. The only place I diverged from data was for the Backscattering effect, and the Diffuse component needed in some cases. I found that lower scale values in specular materials allowed light to pass wholly through the material, so I picked a scale value that allowed the light to pass through thin portions of my model. For best results, models should be built to real world scale (including thickness of parts like the ears). If done properly there is no need to mask the backscatter effect from any part of the model. I also added a diffuse component for times when there is not enough light in the scene to truly support the scattering of light through the material. I normally would keep the Diffuse power for this component at 0.
The material network is built as follows:
Mix Material (Backscatter)
Backscatter Power => Mix Material (Backscatter) Amount Pin
Mix Material (Epidermis) => Mix Material (Backscatter) Material2 Pin
Epidermis Power => Mix Material (Epidermis) Amount Pin
Epidermis Material – Specular Material => Mix Material (Epidermis) Material1 Pin
Specular/Gloss Map => Reflection pin
Color Map => Transmission pin
IOR ( 1.43) => Index pin
Bump (Turbulence) => Bump pin
Normal Map => Normal pin
Roughness Map => Roughness pin
Epidermis Scattering Medium => Medium pin (Absorption Medium)
Epidermis Absorption => Absorption pin of Medium
Phase:Scattering Direction should be set to 0
Scattering Scale => Scale pin of Medium
Fake_Shadows should be enabled
Smooth Normals =>Smooth Pin
Mix Material (Dermis) => Mix Material (Epidermis) Material2 Pin
Diffuse Power => Mix Material (Dermis) Amount Pin
DermisMaterial – Diffuser Material => Mix Material (Dermis) Material1 Pin
Color Map => Diffuse pin
Color Map => Transmission pin
Bump (Turbulence) => Bump pin
Normal Map => Normal pin
Dermis Scattering Medium => Medium pin (Scattering Medium)
Dermis Absorption => Absorption pin of Medium
Dermis Scattering => Scattering pin of Medium
Phase:Scattering Direction should be set to 0
Scattering Scale => Scale pin of Medium
Smooth Normals =>Smooth Pin
DermisMaterial – Specular Material => Mix Material (Dermis) Material2 Pin
Specular/Gloss Map => Reflection pin
Color Map => Transmission pin
IOR ( 1.43) => Index pin
Bump (Turbulence) => Bump pin
Normal Map => Normal pin
Roughness Map => Roughness pin
Dermis Scattering Medium => Medium pin (Scattering Medium)
Dermis Absorption => Absorption pin of Medium
Dermis Scattering => Scattering pin of Medium
Phase:Scattering Direction should be set to 0
Scattering Scale => Scale pin of Medium
Fake_Shadows should be enabled
Smooth Normals =>Smooth Pin
Backscatter Material – Specular Material => Mix Material (Backscatter) Material1 Pin
Specular/Gloss Map => Reflection pin
Color Map => Transmission pin
IOR ( 1.43) => Index pin
Bump (Turbulence) => Bump pin
Normal Map => Normal pin
Roughness Map => Roughness pin
Backscattering Medium => Medium pin (Scattering Medium)
Backscatter Absorption => Absorption pin of Medium
Backscatter Scattering => Scattering pin of Medium
Phase:Scattering Direction should be set to 0
Backscatter Scale => Scale pin of Medium
Fake_Shadows should be enabled
Smooth Normals =>Smooth Pin
Re: Skin Rendering
Posted: Mon Sep 02, 2013 8:10 pm
by gordonrobb
Wow. Need some time to absorb that

. Thanks.
Re: Skin Rendering
Posted: Mon Sep 02, 2013 8:57 pm
by tonysculptor
That's why I've been putting off doing the brain dump. Lots of layers of materials.
Btw, I added a macro to the LiveDB that reflects the updated material network I dumped here. Its tonysculptor Skin2. I also have a specific material for eyes and hair. I am going to start another portrait and probably tweak those materials a bit before adding them to the LiveDB.
Re: Skin Rendering
Posted: Mon Sep 02, 2013 10:08 pm
by immortalartscom
THIS IS SICK! grreeeaaat image man!
Re: Skin Rendering
Posted: Mon Sep 02, 2013 11:53 pm
by Kevin Sanderson
Excellent work! Thank you for sharing!