Eevee is currently not working.
Oh well, i guess that explains it.
I had some thoughts about IOR/Specularity.
The Specular input of the Principled BSDF has been removed from the Principled BSDF node in favor of using the IOR since the “Specularity” of a material is supposed to be calculated off of the IOR.
This is good in terms of “realism”, but it’s not particularly artist friendly. Many people don’t know that IOR controls the Specular component. And on occasions a “Specular map” is provided with some textures. Without a Specular input in the node, most users will have no idea how to use the Specular map, and the people that do know how to use it will have to use a collection of math nodes to convert Specularity to IOR.
The second thing about using IOR as a control for Specularity is that it’s quite easy for someone to be turning down the IOR to reduce the Specularity and accidentally end up with a IOR less than 1. Which isn’t possible in the real world?
An IOR less than 1 can be useful for certain use cases, so I don’t think clamping the IOR between 1 and something greater than 1 is a good idea, but I don’t think the way the system works now is “perfect” yet.
These are just my thoughts. Feel free to do whatever you want with them.
1 Like
Which is not the same kind of specular, those maps belongs to a different PBR workflow, and in Blender we normally invert the map and connect it to roughness.
I don’t think the legacy specular input was artist friendly either, Blender Guru’s donut tutorial needed to tell people to adjust roughness when they want to adjust the highlight, and the specular is best to be left at default if you don’t know what you are doing. You can see how confusing it was for the users. Removing it was a good move.
I would assume the people adjusting IOR are the folks googling different IOR values for different materials? Otherwise I see people adjusting roughness most likely.
3 Likes
I have seen at least one texture pack that contains a “specular map” that appears to be designed for something like Blender’s Principled BSDF. However I haven’t seen maps like these that often, and you’re probably right, most of the time the specular map is probably for the specular PBR workflow.
On some occasions, particularly when working with a dirt material, I increase the roughness and turn down the specular value. And with some other materials I’ll do the same thing to get the look I want.
And in all those cases I based my specular value off of “personal preference”, not IOR. Although, I may be an exception to the “normal Blender user” in this case and I might be the “vocal minority” when it comes to wanting the specular parameter to remain.
2 Likes
I can’t speak for most Blender users either, but my impression is that many users just ignore the existence of the specular setting because it was confusing (different from the behavior of another specular setting from some other software, I would guess those software uses the specular PBR workflow), furthermore the beginners donut tutorial told everyone to leave it at 0.5.
Unless I am misunderstanding something fundamentally here I think specular Value should remain. There are cases where it’s needed. Even if it’s kept at default or automatic in 95% of the cases - in these remaining cases it’s important. It doesn’t hurt to keep it.
Also I don’t know how much of the shader is also going to be suitable for Eevee and realtime but for a specular PBR texture workflow it’s very much a good idea to keep specular value and have an additional color input.
It hurts for people to need to search for the arbitrary formula for IOR to Specular and pull up a calculator each time they want to have accurate IOR of the material. If you want to have the previous behavior I am not sure what’s keeping you from adjusting the new unified IOR input.
I think what we can do is to set a limit to the slider, and you can still enter those values with your keyboard. In that case, people who knows what they are doing will copy and paste an IOR from the material list website, and those just randomly sliding would not have values lower than 1.
Again the legacy specular parameter is not for the specular PBR texture workflow, it was a conversion from IOR via an arbitrary formula that you need to pull up the Blender menual to find it, and pull up the calculator to see what the end value is. For the specular maps you should invert it and connect to roughness. See how confusing it is?
2 Likes
actually, i’ve seen some assets that are using the metalness workflow and still control the specular parameter
1 Like
It could, but it would still be confusing as to figure out “which kind of specular map this is?” when you try to use it.
The legacy specular is really just another IOR input with a certain formula, unifying IOR setting is a good thing in that people don’t see the IOR as something else (like the other PBR workflow)
1 Like
There is also a report in the bug tracker demonstrating how the same texture used as bump or a displacement (bump only) look very different. In one release the bump node looked best and in a later release the displacement node looked best when I think they should look identical. No?
What if the devs provide a nice “specular map to IOR” node out of the box?
1 Like
The thing is the entire existence of “specular” that is actually IOR is confusing. Having such node would still make it confusing with the Specular PBR workflow.
1 Like
I have some thoughts on the name @lukasstockner97 . “Principled v2” IMO is wrong. The term “principled” is a reference to the Disney paper https://media.disneyanimation.com/uploads/production/publication_asset/48/asset/s2012_pbs_disney_brdf_notes_v3.pdf and the principles laid out there (section 5). It seems the design is a bit of departure from those, fine or at least a hybrid.
Other ideas for names:
Blender Standard (somewhat a play on Arnold Standard Surface)
Blender Surface
BSurface
Blender Material
13 Likes
Best not to hardcode the name to “surface” since the plan is to add volume support:
the idea here is to implement basic volume controls (homogenous absorption, maybe also scattering) in the main Principled BSDF.
1 Like
I think the name is still fine, having users learn a different name and changing docs and tutorials over minor stuff like that is not worth it.
It’s still a principled model over physically based parameters. The addition of refraction IOR in both the updated Disney model and our current implementation already deviated from some of the strict rules, Also volume controls doesn’t mean it’s no longer a surface shader, many of the others like standard surface offer similar controls, and the subsurface scattering is already a volumetric effect.
6 Likes
Also, “Principled v2” should not need to be a name anywhere in the UI. This should just be an update to the Principled BSDF shader node, no new name needed.
13 Likes
Just tried build out now. Had some issues.
If both GPU and CPU are ticked in preferences, then GPU rendering in Cycles is greyed out.
If just GPU is ticked in prefs, Blender crashes immediately upon hitting render image. Tried in a few different scenes and all did the same.
Mac used = M1 Max 32 cores, 64gb ram.
@lukasstockner97:
Do you plan to add possibility to load complex fresnel nk data? Or maybe it’s already possible somehow?
Will clear coat have separate IOR control?
Does current “tint” option for clear coat work as “absorption” or absorption will be available with thin-film option?
Hello, I have loved this development, thank you very much I am happy.
I haven’t tried much the new system yet, but the first thing I noticed is:
In the link I shared there are 3 initially important nodes (Specular, Base color and Roughness) to add these 3 important textures, How could I use these textures (Specular) in the new system? Do I have to use a Mix of the Roughness and Specular Textures to connect them to the new Roughness?. I have that concern.
I’m going to try more on other engines to make comparisons. Thanks for: Energy conservation, Sheen, Sheen roughness, Metallic Edge, Metallic Falloff… They are amazing
Edit byThomas Dinges: Removed screenshot, see Copyright guidelines for devtalk
