Already can't wait for the official release of blender 4.0! In my whole career, I don't think I've ever seen complete tutorials like yours. I can finally see what's new, and completely understand it!
That sheen slider is suh-weet! It looks amazing even without a texture and just those solid colors. Makes me want to glide my hand back and forth on that cloth. This is something that would gone completely under my radar without your amazing videos!
Thank you! Thank you. Not just for this video, but for presenting the information in a way that's thorough, thoughtful, and well-paced. So many tutorials are made by people in such a hurry that it sounds like they have something far better to get to. In contrast your video sounds like you have something very interesting and potentially helpful to share with me. You are a teacher, and I am blessed to have you as my teacher. I'm only halfway through this video but I've already liked it, subscribed to you, and rung the bell! Well done! Please do not change your style at all. Thank you again!
Thank you so much for this video, Ive been looking for a Sheen explanation and couldnt find a detailed video and this video is what i exactly wanted! Even though i dont use Blender as my main renderer it helped me so much to understand how it works exactly.
I am learning the basics of Blender right now but I am really looking forward to all the new features of 4.0. This is a great video about the changes and new features.
Thanks for the detailed explanation Christopher, I am currently using that new sheen parameter to mimic dust for a sandstone material and it really gets the job done. Maybe it's worth mentioning that it might simulate dust in general even though it was not designed for it in the first place
Already very early on when designing the "roughness" parameter of this new sheen model we actually did have dust in mind as a potentially useful application. Glad to hear you find it useful! :)
On the back of this video I tried using sheen to suggest peach fuzz on the face of the character I'm working on and it worked like a charm! She suddenly felt alive!
Fantastic showcase, really gives a good sense of how the underlying systems work together, with gorgeous examples, and you seem to know your stuff very well. I would appreciate it if you could turn off overlays while showing off stuff in the rendered viewport, they can be a little distracting.
Hey Christopher! After watching several of your videos, you are the only person whose judgment I trust to answer this question: Can the potential of Cycles, when used perfectly with post-production, compete with the most sophisticated render engines? Or is it still a step below? Thank you very much for your work and effort.
I would say that Cycles still has some catching up to do. It's not on par with a dedicated renderer like Octane for instance. It really depends on what you need to do. Another area that I think Cycles still needs work on is better color output and color management. It's making improvements in that area, but my feeling is it's still not quite there yet.
I only tested a while back, and in the standalone sheen shader. The new microfiber model handled very well normal modifications sent into it (say a fixed slope offset from actual surface normal, based on UV space). Whereas the old model (Velvet) is now completely broken doing the same trick to simulate nap direction? Not sure what that is about. So anyone using the old Velvet shader this way should really make sure the old scene is still working, and set to microfiber mode if the shading edges turn to garbage.
I've uploaded this and placed a link in the about section of my latest video on updated info about SSS in 4.0. I created this scene to help me test out material creation.
You could use an RGB node and pass it to the Base Color, and then also pass it to the Radius, but since the radius is a physical distance value I'm not sure how the color would be interpreted by the Principled shader.
Put a RGB node into Surface Radius and it will be easier to handle :) Tbh I don't know why they have a vector input and not an RGB one since that's what you are multiplying the scale with.
I don't understand this claim that the subsurface is now somehow tied to the entire scene. I thought subsurface scattering was OBJECT based. Like... I'm creating a candle. ONE object. I want the wax to have some subsurface scattering so I go into the CANDLE'S material and just set how much I want and the color. I don't understand why this is better in 4.0. 3:20 ? Then why couldn't they have put the letters R, G and B beside those sliders? And we need a separate radius for each RGB channel now? I need to see what effect those have so maybe it is explained in this video. Ah, 5:26 so just using the base color DOESN"T work! Now, we can't just use the color wheel to change the ss color, we have to know which of the three color components in the RGB system to reduce in order to FIX the ss color. Wow. People generally aren't going to know that if they want to remove yellow, they need to remove both red and green in equal parts. That's why your blue object with a greener ss got more blue because R+G is basically yellow. Can some genius please explain how this is improved? 5:52 omg... that was the perfect example of why this sucks. guessing how much of r g and b to muck with because that's COUNTER intuitive. The color wheel is perfect and intuitive. Just pick a color. If it's too red, then push it some into a neighboring color or closer to green to cancel some red. Too saturated? Slide the saturation bar down. This is not improved. Exact distances aren't needed. Anyone can LOOK at their object and quickly see how much ss they need. 0 = none, 1 = max. :( I guess someone will make a new add-on that let's you use a simple color wheel instead of unlabeled sliders that secretly represent RGB components.
I think you misunderstand what I said. SSS is now more RELATABLE to the scale of the scene because it's using the units system YOU specify for that scene, as opposed to an abstracted apparently abstract unitless value from 3.6. But if you watch my video carefully, you'll see that in fact when you bring in a 3.6 scene, that apparently unitless 'subsurface' value was tied into meters, regardless of your scene's units. So now, in 4.0, when you specify a value, you can related it to the actual physical size of your objects in a more obvious way. The units of your scene don't determine the scattering distance, you do, on a material by material, or object by object way. It's just now more obvious now.
@christopher3d475 thx for the reply. Ok, units are now exposed and tied to metric or English. I understand that. The rgb setting though needs to change, I think. Thx!
Any advice on how to use sheen for a dust effect? I’m looking for a solution to create an epoxy finished concrete (coat over concrete base) with dust imperfections over the epoxy reflections. It’s a floor material.
what if you want to apply subsurface without thinking of scale? any workaround? i never took scale into account, i make abstract compositions, but i like subsurface. if i don't want to look up the scale i'm working with, should i just calculate scale by eye? thanks for your videos, amazing info!
It's actually not behaving too differently than 3.6. For instance, if you were setting up SSS for your abstract materials in 3.6 using the default Scale value 0.5, you just configure it the same in 4.0 by setting your Scene Units to Meters and the Scale value is 0.5m, it's the same. So you shouldn't really have to do it too much differently.
Strange values for the Subsurface Radius parameter: instead of making a more logical color scale, the developers made it in the form of a 3D vector, and now we ourselves have to think about which parameter corresponds to the desired color. A very illogical approach.
In the current version, you can simply connect a color to the blue dot and it works as intended. I imagine in 4.0 it will work the same way.@@andre_julius
no, stupid approach. No one thinks in RGB. Like no one counts in hexadecimal. The SS color needs to be set using the COLOR WHEEL with all the settings like saturation, value, hue etc. I bet a LOT of people complain about this and maybe the idi... uh, person who thought this was a good idea will fix it and implement the old color wheel widget.
Maybe the initial plan was similar to what LuxCore renderer does now, which lets you set a different anisotropy (from -1 to 1) for each channel. So a material can backscatter a color and forward scatter a different one (like milk).@@TruthSurge
How would you plug in a skin subdermal texture into this new setup? Do we plug it into the radius and dial up and down the value of the texture to drive the SSS color?
Can you pass in a node of a texture map that describes the subsurface scale or depth at local positions? I don't know if I am using the right terminology, but maybe you understand. This seems like it would be necessary for mimicking skin, since skin is not a consistent depth and density depending on body location and what is under the skin.
Oh my god, the subsurface scattering is so much better now. Although, it's a bit confusing about the "Subsurface Radius" - if those 3 values represent RGB, then why no color picker?
Think of them as secondary scalar factors for each of the channel. If you set them to 1.0 they have the full effect of the scattering radius. It's pretty straightforward.
@@christopher3d475 my principle bsdf node don't even change, it's the same as in 3.6, I tested it on render and it still renders the same, so, waiting for initial release
Out of curiosity, Since you are more of a professional and might be able to give a definitive answer. You used a partial metallic value in the video. I have heard conflicting information on whether or not having metallic values outside of 0 and 1 makes physical sense. On one hand, I hear the explanation that something is either metallic or not because the value represents whether or not a material conducts electricity. However on the other hand, I hear the explanation that other values are fine because they represent more complex composite materials which have a more complex interaction with light that wouldn't otherwise be able to be simulated unless you actually modeled the quantum effects of the light. I am curious to get your opinion on this. Both on whether or not this models something physically plausible, And also what such a surface implies on a material science level.
A lot of metallic materials have very complex surface scattering properties.This can include a very rough component mixed with much more shinny or sharply glossy aspects. The way material systems like Blender's principled BSDF transition between diffuse and metallic is just one way of potentially simulating this. I actually don't often leave the slider below 1.0 and instead mix multiple metallic roughness values together, but at the end of the day, it's an artistic call. That's what I did in this video, I just made an artistic call as to what I wanted from a visual standpoint. And all 3D material systems are abstract representations of what happens in the real world.
@@christopher3d475 That makes sense. To be clear, When you refer to mixing multiple metallic roughness values, Are you referring to mixing multiple principled shaders together which have differing roughness values? Or to a different process? Additionally. Do you have any specific resources for research to do in order to build up a base intuition for replicating more complicated real world materials than could be replicated with the standard single principled shader 0/1 metallic workflow? Obviously when it comes down to it, The end result is artistic interpretation. But I would love to step up my material game beyond where I currently am at and your channel has already been such an invaluable resource for that.
Correct, mixing either multiple Principled shaders or with other BSDF nodes. One thing I wish Blender had is in Vray's implementation of the GGX microfacet function that includes a 'Tail' parameter. this simulates secondary roughnesses within a single BSDF layer. But we can do something similar by mixing two shaders together. As far as specific resource, there are databases of complex IOR values for specific materials, but Blender doesn't use those (IOR extinction coefficient) and it really just comes down to observing what's happening with real world surfaces and then trying to replicate them. I'm going to be doing another tutorial using the material tester I show in the video where I show some basic mixing for metals. Hopefully I can get to that soon.@@Dryym
@@christopher3d475 Yeah, I have noticed with my research that Blender's principled shader is missing a lot of features which are standard in other packages. One key example being the fact that the industry standard for SSS in skin shaders appears to be a triple layered system with epidermal, Dermal, And subdermal layers. And as someone who's still actively learning about these things, I don't actually know the most efficient way to replicate this. My gut tells me to mix three principled shaders together with different SSS settings, But that also feels like it'd be much slower to render than a more optimized and integrated version. However even still, It's really nice to see that it's getting some love to be more accurate and more flexible. Likewise, Your channel had been an absolute goldmine of information to help with more accurate materials.
Woah woah woah, Why would they remove subsurface colouring colour? If this has done what I think it has; you'd have a creature covered in a thin veil of blue paint; how are you going to get blood colours to show through the sss if your base colour is all blue?? I would have preferred it got the base colour by default, but also give me the option to plug in a specific map. I believe that's how it currently works. Can we roll this back?
In a sea of reaction content where people just read off patch notes without truly understanding, your channel is a diamond in the rough.
Thank you so much for your lovely helpful tutorial...❤❤❤
Sheen roughness at 1.0 is almost like suede leather. Very nice to have.
I am so happy that I decided to shift to Blender from 3dsMax. Great additions. Cant wait for the final release.
Congratulation, you are now a free man.
Wait till you realise that you can't actually produce quality work within time without 1000$ worth of add-ons. Blender is a trap. @@MarquisDeSang
Much appreciated that you screen record at a large scale which makes it easier for with smaller screens. 😁Amazing as always.
Already can't wait for the official release of blender 4.0! In my whole career, I don't think I've ever seen complete tutorials like yours. I can finally see what's new, and completely understand it!
That sheen slider is suh-weet! It looks amazing even without a texture and just those solid colors. Makes me want to glide my hand back and forth on that cloth. This is something that would gone completely under my radar without your amazing videos!
I've been using blender since 2019 and I wish I would have found your channel before.. This channel is pure gold quality knowledge. Thank you!
Thank you! Thank you. Not just for this video, but for presenting the information in a way that's thorough, thoughtful, and well-paced. So many tutorials are made by people in such a hurry that it sounds like they have something far better to get to. In contrast your video sounds like you have something very interesting and potentially helpful to share with me. You are a teacher, and I am blessed to have you as my teacher. I'm only halfway through this video but I've already liked it, subscribed to you, and rung the bell! Well done! Please do not change your style at all. Thank you again!
Thank you so much for this video, Ive been looking for a Sheen explanation and couldnt find a detailed video and this video is what i exactly wanted! Even though i dont use Blender as my main renderer it helped me so much to understand how it works exactly.
I am learning the basics of Blender right now but I am really looking forward to all the new features of 4.0. This is a great video about the changes and new features.
This channel has become the kind where I instantly hit like on a new video and then proceed to watch it 🔥
Thanks for the detailed explanation Christopher, I am currently using that new sheen parameter to mimic dust for a sandstone material and it really gets the job done. Maybe it's worth mentioning that it might simulate dust in general even though it was not designed for it in the first place
Yes, that's quite true. I talked about it being used for dust at one point but I may have edited that part out.
Already very early on when designing the "roughness" parameter of this new sheen model we actually did have dust in mind as a potentially useful application. Glad to hear you find it useful! :)
Now I finally know how to use the subsurface scattering !!
On the back of this video I tried using sheen to suggest peach fuzz on the face of the character I'm working on and it worked like a charm! She suddenly felt alive!
Just the explanation I was looking for. Thanks!
One of the best channel hands down
10:30 that was eye opening and informative and mind blowing way of explaining 🎉🎉🎉❤❤❤
This is absolutely fantastic! Thank you for your clear explanation.
so cool to see Blender catching up to Unity and Unreal engine!
Really well-made tutorial! It definitely has the production value of much bigger channels! Hope your subscriber count will match it soon 😊
Very, very good as always!
Although a nitpick I was using a lot the subsurface color different than the diffuser to get hued controlled caramel finish on my models
I'm going to do a quick video today where I show those materials transfer over to 4.0, so you can see how they're handling this situation.
Well done , Well explained and intuitive. I like it I do !❤
Can we have more stuff like this
You produce very nice Videos!
Fantastic showcase, really gives a good sense of how the underlying systems work together, with gorgeous examples, and you seem to know your stuff very well. I would appreciate it if you could turn off overlays while showing off stuff in the rendered viewport, they can be a little distracting.
That's a good point, I'll turn those layers off next time.
Hey Christopher! After watching several of your videos, you are the only person whose judgment I trust to answer this question: Can the potential of Cycles, when used perfectly with post-production, compete with the most sophisticated render engines? Or is it still a step below?
Thank you very much for your work and effort.
I would say that Cycles still has some catching up to do. It's not on par with a dedicated renderer like Octane for instance. It really depends on what you need to do. Another area that I think Cycles still needs work on is better color output and color management. It's making improvements in that area, but my feeling is it's still not quite there yet.
@@christopher3d475 Thank you again Christopher :)
I only tested a while back, and in the standalone sheen shader. The new microfiber model handled very well normal modifications sent into it (say a fixed slope offset from actual surface normal, based on UV space). Whereas the old model (Velvet) is now completely broken doing the same trick to simulate nap direction? Not sure what that is about. So anyone using the old Velvet shader this way should really make sure the old scene is still working, and set to microfiber mode if the shading edges turn to garbage.
Loved this, please do for everything lol
This is the Velvet Gold
Thank you
Hi! Thank you for your videos, where can I get this model from 1:31?
I've uploaded this and placed a link in the about section of my latest video on updated info about SSS in 4.0. I created this scene to help me test out material creation.
This the great news for skin shading...now it separate the scale setting, which is the most thing I was confused with.
nice breakdown, would there be any way/benefit to feeding the base colour into surface radius for ease then just adjusting the scale accordingly?
You could use an RGB node and pass it to the Base Color, and then also pass it to the Radius, but since the radius is a physical distance value I'm not sure how the color would be interpreted by the Principled shader.
Put a RGB node into Surface Radius and it will be easier to handle :) Tbh I don't know why they have a vector input and not an RGB one since that's what you are multiplying the scale with.
There's always room for improvement, 4.5, 5.0...
Maybe you can separate rgb and then load the vector. Of course, this is a bit cumbersome.
I don't understand this claim that the subsurface is now somehow tied to the entire scene. I thought subsurface scattering was OBJECT based. Like... I'm creating a candle. ONE object. I want the wax to have some subsurface scattering so I go into the CANDLE'S material and just set how much I want and the color. I don't understand why this is better in 4.0. 3:20 ? Then why couldn't they have put the letters R, G and B beside those sliders? And we need a separate radius for each RGB channel now? I need to see what effect those have so maybe it is explained in this video. Ah, 5:26 so just using the base color DOESN"T work! Now, we can't just use the color wheel to change the ss color, we have to know which of the three color components in the RGB system to reduce in order to FIX the ss color. Wow. People generally aren't going to know that if they want to remove yellow, they need to remove both red and green in equal parts. That's why your blue object with a greener ss got more blue because R+G is basically yellow. Can some genius please explain how this is improved? 5:52 omg... that was the perfect example of why this sucks. guessing how much of r g and b to muck with because that's COUNTER intuitive. The color wheel is perfect and intuitive. Just pick a color. If it's too red, then push it some into a neighboring color or closer to green to cancel some red. Too saturated? Slide the saturation bar down. This is not improved. Exact distances aren't needed. Anyone can LOOK at their object and quickly see how much ss they need. 0 = none, 1 = max. :( I guess someone will make a new add-on that let's you use a simple color wheel instead of unlabeled sliders that secretly represent RGB components.
I think you misunderstand what I said. SSS is now more RELATABLE to the scale of the scene because it's using the units system YOU specify for that scene, as opposed to an abstracted apparently abstract unitless value from 3.6. But if you watch my video carefully, you'll see that in fact when you bring in a 3.6 scene, that apparently unitless 'subsurface' value was tied into meters, regardless of your scene's units. So now, in 4.0, when you specify a value, you can related it to the actual physical size of your objects in a more obvious way. The units of your scene don't determine the scattering distance, you do, on a material by material, or object by object way. It's just now more obvious now.
@christopher3d475 thx for the reply. Ok, units are now exposed and tied to metric or English. I understand that. The rgb setting though needs to change, I think. Thx!
thank you❤
Any advice on how to use sheen for a dust effect? I’m looking for a solution to create an epoxy finished concrete (coat over concrete base) with dust imperfections over the epoxy reflections. It’s a floor material.
what if you want to apply subsurface without thinking of scale? any workaround? i never took scale into account, i make abstract compositions, but i like subsurface. if i don't want to look up the scale i'm working with, should i just calculate scale by eye?
thanks for your videos, amazing info!
It's actually not behaving too differently than 3.6. For instance, if you were setting up SSS for your abstract materials in 3.6 using the default Scale value 0.5, you just configure it the same in 4.0 by setting your Scene Units to Meters and the Scale value is 0.5m, it's the same. So you shouldn't really have to do it too much differently.
thanx
Strange values for the Subsurface Radius parameter: instead of making a more logical color scale, the developers made it in the form of a 3D vector, and now we ourselves have to think about which parameter corresponds to the desired color. A very illogical approach.
Yeah that's the only thing I could think of as well. If the radius is obviously influencing the color then... why not have it be a color input :D
In the current version, you can simply connect a color to the blue dot and it works as intended. I imagine in 4.0 it will work the same way.@@andre_julius
no, stupid approach. No one thinks in RGB. Like no one counts in hexadecimal. The SS color needs to be set using the COLOR WHEEL with all the settings like saturation, value, hue etc. I bet a LOT of people complain about this and maybe the idi... uh, person who thought this was a good idea will fix it and implement the old color wheel widget.
You talking about this update or when subsurface was added years ago?
Maybe the initial plan was similar to what LuxCore renderer does now, which lets you set a different anisotropy (from -1 to 1) for each channel. So a material can backscatter a color and forward scatter a different one (like milk).@@TruthSurge
I think the RGB in the subsurface should have been indicated for better understanding of what people are adjusting
Yes, I agree.
How would you plug in a skin subdermal texture into this new setup? Do we plug it into the radius and dial up and down the value of the texture to drive the SSS color?
Watch my followup video here. ua-cam.com/video/yn3v9xyRM4o/v-deo.html
Can you pass in a node of a texture map that describes the subsurface scale or depth at local positions? I don't know if I am using the right terminology, but maybe you understand. This seems like it would be necessary for mimicking skin, since skin is not a consistent depth and density depending on body location and what is under the skin.
Yes, you can use 3D procedural texture nodes in the material editor to affect the SSS. I just didn't show it.
Oh my god, the subsurface scattering is so much better now. Although, it's a bit confusing about the "Subsurface Radius" - if those 3 values represent RGB, then why no color picker?
Think of them as secondary scalar factors for each of the channel. If you set them to 1.0 they have the full effect of the scattering radius. It's pretty straightforward.
@@christopher3d475 I guess I can just connect the RGB node :)
Does this version fixes SSS glitch that highlights mesh wireframe?
I'm not sure to be honest. You could download the 4.0 builds and test it out.
@@christopher3d475 my principle bsdf node don't even change, it's the same as in 3.6, I tested it on render and it still renders the same, so, waiting for initial release
I donwloaded 4.0 but didn't found this changes in shader
I would try the latest build.
what about characters with blue skin that you want red subsurface underneath?
Use the subsurface radius to alter the scattering color.
What's the difference between sheen and fresnal?
Sheen is a reflection layer that sits above fresnel reflectivity. I have a couple of videos about specularity and fresnel that describe it in detail.
I wonder how this will effect skin shaders
It's a good question. I suspect they'll be largely the same.
Out of curiosity, Since you are more of a professional and might be able to give a definitive answer. You used a partial metallic value in the video. I have heard conflicting information on whether or not having metallic values outside of 0 and 1 makes physical sense. On one hand, I hear the explanation that something is either metallic or not because the value represents whether or not a material conducts electricity. However on the other hand, I hear the explanation that other values are fine because they represent more complex composite materials which have a more complex interaction with light that wouldn't otherwise be able to be simulated unless you actually modeled the quantum effects of the light.
I am curious to get your opinion on this. Both on whether or not this models something physically plausible, And also what such a surface implies on a material science level.
A lot of metallic materials have very complex surface scattering properties.This can include a very rough component mixed with much more shinny or sharply glossy aspects. The way material systems like Blender's principled BSDF transition between diffuse and metallic is just one way of potentially simulating this. I actually don't often leave the slider below 1.0 and instead mix multiple metallic roughness values together, but at the end of the day, it's an artistic call. That's what I did in this video, I just made an artistic call as to what I wanted from a visual standpoint. And all 3D material systems are abstract representations of what happens in the real world.
@@christopher3d475 That makes sense. To be clear, When you refer to mixing multiple metallic roughness values, Are you referring to mixing multiple principled shaders together which have differing roughness values? Or to a different process?
Additionally. Do you have any specific resources for research to do in order to build up a base intuition for replicating more complicated real world materials than could be replicated with the standard single principled shader 0/1 metallic workflow? Obviously when it comes down to it, The end result is artistic interpretation. But I would love to step up my material game beyond where I currently am at and your channel has already been such an invaluable resource for that.
Correct, mixing either multiple Principled shaders or with other BSDF nodes. One thing I wish Blender had is in Vray's implementation of the GGX microfacet function that includes a 'Tail' parameter. this simulates secondary roughnesses within a single BSDF layer. But we can do something similar by mixing two shaders together.
As far as specific resource, there are databases of complex IOR values for specific materials, but Blender doesn't use those (IOR extinction coefficient) and it really just comes down to observing what's happening with real world surfaces and then trying to replicate them. I'm going to be doing another tutorial using the material tester I show in the video where I show some basic mixing for metals. Hopefully I can get to that soon.@@Dryym
@@christopher3d475 Yeah, I have noticed with my research that Blender's principled shader is missing a lot of features which are standard in other packages. One key example being the fact that the industry standard for SSS in skin shaders appears to be a triple layered system with epidermal, Dermal, And subdermal layers. And as someone who's still actively learning about these things, I don't actually know the most efficient way to replicate this.
My gut tells me to mix three principled shaders together with different SSS settings, But that also feels like it'd be much slower to render than a more optimized and integrated version. However even still, It's really nice to see that it's getting some love to be more accurate and more flexible. Likewise, Your channel had been an absolute goldmine of information to help with more accurate materials.
Why on earth didn't they just write RGB?
I don't know, many of us suggested they do that for clarity purposes.
"We can see that I have 2 centimeters. Let's make it more relatable by coming down to..." Yikes.
It's a boomer I can't subscribe twice
Thanks
Woah woah woah,
Why would they remove subsurface colouring colour?
If this has done what I think it has; you'd have a creature covered in a thin veil of blue paint; how are you going to get blood colours to show through the sss if your base colour is all blue??
I would have preferred it got the base colour by default, but also give me the option to plug in a specific map. I believe that's how it currently works. Can we roll this back?
I just recorded a video that shows how a 3.6 SSS containing material transfers over to 4.0. You can see if this addresses your concern.
💪💣💥🗯💭💨❤🕊 nailed....
4.0,...4.1.... amazing milestone ....