Variable Roughness, mixing with roughness map for more realism.

[@MrWoundedalien]

I was today years old when I learned the difference between mix and multiply! lol. Great video! While I'm watching your videos, I often think, "How does he know this?" haha!

[@LacklusterFilms]

Genuinely astounded that your videos don't get more views. I have no doubt that if more people used your tips, blender art on the internet as a whole would improve.

[@dunravin]

Shhh... It's our little secret 🤫😃

[@Fleig.]

Well he's not a clickbaity comedy channel. Education is serious.

[@BaronFactotum]

Great video, Chris. I've been developing some techniques for variable roughness myself. I've produced two videos on the subject. I've standardized some of the language describing the phenomenon, I believe 'roughness truncation' is a more precise term than variable roughness. My last video is a tutorial video that goes over the node setup I use in Blender, Unreal, Unity, and the shader code for Godot. You should take a look at my methods, the map range node allows more control and cross program compatibility which is of course important for many workflows. Feel free to make a video on the alternate setup if you'd like!

[@christopher3d475]

Thanks for the tip, I'll check it out.

[@SFE-Viz]

Was wondering on how to implement roughness maps too, thanks for making this follow up. This is going to enhance a lot of my renders.

[@王诩-k7v]

This is really important for improving the realism of Blender materials! Thank you so much for sharing! I just finished watching the previous video, and this couldn't have come at a better time!

[@FlickerJab]

We asked and you delivered! Thanks again, very helpful series

[@anjanadilshan7057]

These Shader technicalities is very important. Keep doing more please ❤❤

[@artffan5413]

You make me happy whenever you upload

[@Jonah_Anthony]

Something I didn't even realize I needed. Excellent video as always Christopher

[@BahMah-z9b]

Best blender channel ❤

[@kevinmanley5043]

Agree. The best! Thanks again for sharing your knowledge Christopher.

[@Mranshumansinghr]

Always something new to learn in your videos. Thank you.

[@nikko3d]

excellent tip!

[@laanrohs]

and he provided!!!!

[@dunravin]

Thanks Christopher 👍

[@asabovesobeloweternal]

great tutorials to get that realism.

[@john-ants]

Fantastic, thank you!

[@swannlecocq6083]

Exaclty what I wanted to ask you about after your last video ! Thanks

[@Floatharr]

Very minor improvement I thought of for the videos: setting viewport denoising to always on allows the video encoder to use the available bitrate much more efficiently, making the video less blurry than if it had to try to compress the random noise from rendering. I tend to use Openimagedenoise with Accurate pre filtering and Fast quality with GPU turned on, with viewport scale turned down to 2x instead of Automatic in the render Performance section.

[@Dhruv1223]

Thank you ❤

[@hardwire666too]

So after the first video I've started playing with the Layer Weight node more. I have also found it handy for water. The IOR of water doesn't change, however turbulent water does have a lensing effect. All of which can change heavily depending on your position relative to the surface. Using the same idea But or other properties yields some great results. Esp i you use something like Shaders Plus ake Caustics. My water went from looking like chrome to actually looking like water.

[@affanity.x9268]

I think id like another demonstration with a slightly darker surface, additionally real life setup of the same setup in blender, so we can see both clearly.

[@MrHugoHD]

I thought the Principal BSDF already had the Fresnel effect. Crazy that we still have to go through this process in 2025.

[@christopher3d475]

It does support the fresnel effect. In fact Blender 4.0 saw internal improvements to the fresnel code. What Blender isn't doing is this more sophisticated process that can happen with apparent roughness changing with visual angle. I explain this in the first video I reference (the video right before this one).

[@metajerk]

I find this fascinating, that said, I’d be interested to see how you feel about this in Octane.

[@christopher3d475]

I don't have a lot of experience with Octane so I couldn't comment on its abilities in this area to be honest.

[@metajerk]

@ that’s why I’d be curious to see your approach. Im a new blender user from the modo massacre… I have found your videos very insightful in my learning blender. At the same time, to truly understand issues I’m always interested in the approach from multiple perspectives. Many thanks for your work.

[@christopher3d475]

@@metajerk Yeah, I also used to use Modo. I was kind of happy to move on from the often unintuitive shader tree.

[@TheThebeep]

Octane doesn't have variable roughness too.

[@JorgeMarquezIO]

Wouldn't a Color Ramp or RGB curve be easier to control and explain than the first multiply node and achieve the exact same result? That result being controling the lower AND upper range of the texture.. or is there something I am missing?

[@christopher3d475]

I tried a color ramp and didn't get results that I felt were usable, But an RGB Curve could also work. I simply showed one general way to approach it.

[@JorgeMarquezIO]

@@christopher3d475 Interesting.... Is there any chance you know why? would it have a different interpolation?

[@asrafhadjirashid2162]

Is it okay to apply it in all materials ?

[@bclaus0]

Where did you get the information that Blender uses a uniform roughness across viewing angles? The Fresnel effect is built into every PBR standard, including Blender's Principled Shader.

[@christopher3d475]

Roughness is a separate thing from the Fresnel effect. Roughness affects Fresnel by attenuating it as roughness increases. I explain this in multiple videos I've done and it's demonstrable. The GGX microfacet roughness mechanism decreases the strength of fresnel at glancing angles. Watch my previous video on variable roughness.

[@bclaus0]

@@christopher3d475 Wow, I can't believe nobody ever taught this before. This is basically the fresnel effect but on a mesoscopic level, where the principled shader only accounts for it on a microscopic level. I think you're mixing up the words a bit, you can't really "affect the Fresnel'. But I get the gist and will definitely look into it more.

[@christopher3d475]

@@bclaus0 Yes, Fresnel is still being calculated at a microfacet level, so if you want to nit-pick, yes, fresnel is still fully operating. But we're really concerned with macroscopic results here obviously.

[@SeeYouInBluffington]

I tend to mix noise (or other roughness/color variation) into the b socket of the mix node and use the factor to set the strength of its influence. Is that wrong?

[@christopher3d475]

Use whatever mechanism works for you. Blender is pretty open how it comes to manipulating data through various nodes. I chose one approach, it's more the overall variable reflectance setup that I wanted to convey.

[@TheThebeep]

I wonder why it is not implemented in the Cycles by default? Does anyone know? This is not something new.

[@christopher3d475]

It's not implemented by default because it doesn't happen on every surface. The default behavior is logical, but it would be nice to have this variable control as a built in option for the principled bsdf.

[@fairplex3883]

I dont think this is physically correct but good to know for a more artistic approach

[@christopher3d475]

It is a physical property of some surface. It's all over the place when you know to look for it.

[@MaxKnol-uz4vl]

This is because of clearcoat. The floor has been sealed. "variable roughness" isnt a real thing. It doesn't exsist The effect you're showing is clearcoat with a lower IOR Get your regular rough concrete texture with it's dedicated roughness map and layer clearcoat on top with a lower IOR than the base layer (1.2 to 1.4) The roughness of the clearcoat will also obviously be lower than that of the concrete. You can put a blurred version of the concrete normalmap in the clearcoat normal and a shiny, scratched roughnessmap in the roughness slot Again, variable roughness based on angle doesn't exsist IRL What you're doing works for artistic purposes but isn't physically accurate as far as I know Edit: I apologised I was wrong

[@christopher3d475]

No, that's not correct. If you watch the video I reference, I show real world examples of this phenomena. Clearcoat is one way to go about adding a secondary roughness, but what I'm demonstrating is something that happens on a single BSDF reflectance layer.

[@AliasA1]

You can take sandpaper to a piece of wood and observe this effect in the rough surface that is created. You can also see it in freshly etched frosted glass, and on untreated plastic parts like textured Lego bricks. None of these surfaces have any kind of secondary coating of note, and the few research papers I could find on this all seem to try to integrate the effect into the base reflectance model, as is demonstrated in this video. Do you have any further explanation or source for saying that this effect only appears with a clear coat?

[@ZacDonald]

It's really noticeable on textured surfaces, head on you can see every cavity and it appears rougher, from glancing angles those cavities get occluded and the surface looks less rough. Valve has a paper where they have the normal map influence the roughness because at a distance the normal map will blur out with MIPs and the roughness gets glossier than it should be.

[@dunravin]

Did you watch the first video? Christopher shows diagrams explaining how it does happen on IRL surfaces.

[@MaxKnol-uz4vl]

Excuse my monday grumpiness, I did some more research and I was in fact wrong. Surfaces can be treated in such a way that the roughness apprears sharper at glancing angles. It appears if the top of a rough surface like concrete or wood has been sanded/honed. the top microfacetsare are flattened, but the lower laying microfacets are still oriented in more random directions. When you look at the surface from a glancing angle the lower laying, more random microfacets are occluded causing you to only see the flattened, smoother microfacets. That should put the record straight! regards, a random 3D artist working for a 3d studio in the netherlands PS: many 3D artist worth their salt already did something similar to this but just in a different realm: darkening velvet fabrics when viewed head on since there is AO between the hairs of velvet. Same is often done with bouclé. Though personally I use real displacement for bouclé