Тёмный

The Furnace Test, does Blender 4.0 pass? 

Christopher 3D
Подписаться 18 тыс.
Просмотров 24 тыс.
50% 1

In this video, I show some subtle but quite significant improvements to the foundation of Cycles for Blender 4.0, dealing with materials and shading, by running what's called The Furnace Test. This is a test to determine if a renderer is properly handling energy conservation and preservation. Passing the Furnace Test is a really good thing. As part of this I discuss the difference between the shading algorithm GGX and Multiscatter GGX, and improvement to it in 4.0.
00:00 Introduction
01:10 Furnace Test, setup in 3.6
03:45 Furnace Test in Blender 3.6, specular+diffuse
05:25 Adding Clearcoat
05:52 Adding Sheen
06:58 Energy preservation, GGX/Multiscatter GGX
10:00 Energy preservation and transparency
12:13 Furnace Test in Blender 4.0
16:25 Comparison Example #1
17:05 Comparison Example #2
17:43 Comparison Example #3
19:00 Comparison Example #4
You can see all of my videos about changes to Cycles and Blender 4.0 here:
• Blender 4.0 Feature Vi...
#ggx #multiscatterggx #blender4 #blendertutorial #energyconservation

Кино

Опубликовано:

 

4 июл 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 150   
@smeece85
@smeece85 9 месяцев назад
These 4.0 videos you’ve been making are incredibly helpful. The content is concise yet in-depth, and your straightforward presentation brings refreshing clarity. Thank you so much for taking the time to share.
@darrylros681
@darrylros681 9 месяцев назад
So this is the reason why my renders are a bit off sometimes because those edges seems to be brighter that the environment which can subconciously be picked up by our brain that that is not suppose to happen. Thanks for this video!
@nimerbushnaq7850
@nimerbushnaq7850 9 месяцев назад
Professional explanation, a breath of fresh air from all the noisy videos out there, thank you !
@digital0785
@digital0785 9 месяцев назад
i feel like i'm as excited for 4/4.1 as I was for 2.8 theres so much HUGE stuff going on with the next few versions
@christopher3d475
@christopher3d475 9 месяцев назад
Yeah, 4.0 is doing some important things for the future.
@4.0.4
@4.0.4 9 месяцев назад
Yeah, I remember the hype for 2.8 and I'm impressed they pulled it off again. Blender is amazing.
@Mranshumansinghr
@Mranshumansinghr 9 месяцев назад
Blender has more features every month. This is more than 10 times what I was used to when using Autodesk 3dsMax and Chaos Group Vray Render. So happy I shifted to Blender. Thank you for your detailed explanation.
@madjidbouharis729
@madjidbouharis729 8 месяцев назад
It's been years since I am noticing this issue generally when comparing cycles to Arnold and I just couldn't understand the WHY and HOW thank you for this concise education
@WonkyGrub
@WonkyGrub 8 месяцев назад
I would love to see 3.6 filmic vs 4.0 multiscatter and agx to show the how much the two new things that might seem subtle actually help realism so much!
@antialfrednt
@antialfrednt 8 месяцев назад
It's nice to see a return to educational form in the Blender space. Hoping to see more channels follow suit!
@AhmedNassef
@AhmedNassef 9 месяцев назад
Your videos are so perfect in the sense that though they're about high concepts yet you manage to explain them in depth and a polished presentation that I always reach the end of the video with greater comprehension than I might do other Blender videos. Moreover your channel was mentioned in last week's Blender Today on the official RU-vid Blender channel as a great source to learn about the new updates in the Principled BSDF. Looking forward to your future stuff..
@FutharkStudios
@FutharkStudios 8 месяцев назад
These are some amazing high-level stuff explained so easily! Being a Blender only studio, it makes me very happy seeing these improvements coming our way!
@christopher3d475
@christopher3d475 8 месяцев назад
Yeah, energy conservation in 4.0 is a big deal and it's what I'm most excited about.
@UCD33
@UCD33 8 месяцев назад
I'm still fairly new to this and have dredged through way too many videos trying to get it straight. This finally helped make it "click"! Thank you for an excellent presentation.
@timothycook
@timothycook 9 месяцев назад
Great explanation and helps me wrap my head around the new settings. I also noticed that 4.0 converts older more complex node setups to align with 4.0’s inputs I.e. I use a separate image file sometimes for subsurface color. 4.0 automagically combines with a mix node instead of throwing the image node out.
@christopher3d475
@christopher3d475 9 месяцев назад
Yes, I've noticed some of these conversions too that I've been meaning to explore in more detail to see how they're handling different cases.
@ShiroiAkumaSama
@ShiroiAkumaSama 9 месяцев назад
Thanks for making us aware of all those changes! I really digging those videos!
@superkaboose1066
@superkaboose1066 9 месяцев назад
These have been so helpful, best in depth shader explaining I've ever seen, thanks!
@Carkoon
@Carkoon 9 месяцев назад
Thanks for another great video! And also for including the side by side comparisons.
@leestripp7387
@leestripp7387 9 месяцев назад
Thank you for all your hard work making these, looking forward to watching more.
@xDaShaanx
@xDaShaanx 9 месяцев назад
One of the best channels when it comes to detailed in depth blender tutorials. You explain the how and why as well which is very helpful. :)
@JoshuaMerrill042
@JoshuaMerrill042 9 месяцев назад
Excellent demonstration of these new features and wonderfully detailed explanations!
@gabrielmoro3d
@gabrielmoro3d 8 месяцев назад
This release is just brilliant. Thanks for this class!
@josuam5502
@josuam5502 9 месяцев назад
I really really like this kind of content. Please do more of it. It's not dry at all. You definitely deserved more subscribers.
@godsofentropy
@godsofentropy 9 месяцев назад
Amazing video! Thanks for sharing your deep knowledge about rendering, lighting and all engine quirks behind it!
@NightVisionOfficial
@NightVisionOfficial 9 месяцев назад
This video is FIRE! keep up the good work man! Love this deep look into the material :)
@muscle__rr
@muscle__rr 9 месяцев назад
Wow this kind of deep dive and knowledge sharing will be very useful for the community. Thank you
@4.0.4
@4.0.4 9 месяцев назад
Thanks, in the past I had seen energy conservation discussions and didn't understand why it was such a big deal, thinking it was just a technicality.
@christopher3d475
@christopher3d475 9 месяцев назад
Yep. If only one changed had happened in Cycles for 4.0, this is the one I would have gone to battle to get in.
@andrekling
@andrekling 9 месяцев назад
Wow amazing explanation, keep up the amazing work maestro
@nobleengine
@nobleengine 9 месяцев назад
This is an incredible video. Thank you so much.
@guilhermecampos8313
@guilhermecampos8313 9 месяцев назад
Thank you very mush for the work doing this video. Here we can see how cycles will be more precise now.
@sbrokmann
@sbrokmann 7 месяцев назад
this is just the video i needed to see, thanks!
@EmiMartinezAdrian
@EmiMartinezAdrian 9 месяцев назад
Thank you for all these explanation!! great work!
@MarcosBaldin
@MarcosBaldin 8 месяцев назад
Wow! What an explanation! Thank you man.
@mariuspuiu9555
@mariuspuiu9555 9 месяцев назад
Thanks for the comparisons and explanations.
@AirmanCS
@AirmanCS 9 месяцев назад
You are the most knowledgeable youtuber I know on color, rendering, and light physics on blender. Thanks for your videos I could say im an "advanced" user and this is fantastic even if I dont ever do realistic renders I find this VERY useful
@Jason_Silent
@Jason_Silent 9 месяцев назад
Your videos is such interesting, I always liked that under the hood stuff, really helps achieve better results with this lots of important details
@Atimo133
@Atimo133 9 месяцев назад
Thank you so much Christopher for everything you've done!!!
@MitchTube
@MitchTube 9 месяцев назад
I have learned so much from these fundamentals based videos. Everything I’ve been confused about is starting to make sense!
@andrea_ciani
@andrea_ciani 8 месяцев назад
Extremely useful video, thank you Christopher 🙂
@TiredCommentary99
@TiredCommentary99 9 месяцев назад
So much quality content, thank you very much.
@nightlyknight7970
@nightlyknight7970 9 месяцев назад
Thanks for the video, very informative!
@MrBrax
@MrBrax 9 месяцев назад
Very cool, makes a lot of sense with the sheen around it
@fumier21
@fumier21 9 месяцев назад
Great video as always, kinda takes me back on memory lane back in the days when Brazil RS was a thing and Neil Blevins used to make tutorials about photorealism
@christopher3d475
@christopher3d475 9 месяцев назад
I remember seeing that years ago. Those poor guys were trying to pull off sophisticated algorithms on much less powerful hardware.
@rahulujjal8245
@rahulujjal8245 9 месяцев назад
Loving these insights, It's not limited to blender now this information is helping me rethink my custom shaders in Unity as well.
@GaryParris
@GaryParris 9 месяцев назад
grear video, nicely demonstrated with examples :O)
@jessebrady2614
@jessebrady2614 9 месяцев назад
Hey! Chris Tyler, haha! Our studio used Strata 3D like a decade ago, and I recognized you from the Strata University videos. Glad you made the switch! I feel like a lot of the hardcore Strata guys are pretty stubborn in spite of a complete lack of support, haha! Good stuff! Me and the other animator in the studio had your "Art and Science of Strata 3D" back in the day...blast from the past 😂
@peacfulmusicwaves
@peacfulmusicwaves 9 месяцев назад
I'd really like to thank you for your effort to bring closer to non expert users the huge changes that take place towards the 4.0 release. Keep it up. These videos are something that was missing for a long time. A breath of fresh air. I'd like to see more about the changes of Cycles and path guiding in the future, if you please. Thanks again!
@christopher3d475
@christopher3d475 9 месяцев назад
Did you catch my video on updated render features in 4.0 where I talk about the improvements to path guiding?
@peacfulmusicwaves
@peacfulmusicwaves 9 месяцев назад
@@christopher3d475 I'll check it out. Thanks!
@vuk8550
@vuk8550 9 месяцев назад
Killer Information mate. Subed
@evanlane1690
@evanlane1690 9 месяцев назад
Excellent content. Dry by its nature, but you present it very well. Thanks!
@HUE3D
@HUE3D 9 месяцев назад
Great stuff!
@GiPa98
@GiPa98 8 месяцев назад
super interesting video! Looking forward to Blender 4.0 then!
@frittenpeter
@frittenpeter 9 месяцев назад
these are the real tutorials, please more
@AtaurRaziq
@AtaurRaziq 9 месяцев назад
Loving how much better 4.0 will look without any effort from the user!
@blikkcg
@blikkcg 7 месяцев назад
bro i want u to give a talk about anything in blender con next year, u'r soo damn good
@euclois
@euclois 9 месяцев назад
the tutorials we need
@cg1582
@cg1582 9 месяцев назад
Incredible videos. Nuff said.
@shlokbhakta2893
@shlokbhakta2893 9 месяцев назад
So excited for 4.0 stable to be out!
@gottagowork
@gottagowork 9 месяцев назад
Some comments. 1. It's not "Blender Diffuse" that is causing the issue, but Principled which uses Disney Diffuse. The latter accounts for "Exit IOR" which is known to cause energy loss. 2. Perfect energy conservation is only expected with regular Diffuse (not Oren Nayar or Velvet like effects) in combination with fresnel mixing in Glossy. 3. "Glowy" usually refers to "glowy edges" caused when mixing Diffuse with *ROUGH* glossy using a standard fresnel that doesn't account for glossy roughness. 4. Sheen 4 are unnatural and should be avoided because they cause problems. The "legal" vanilla range is 0-1. The "Sheen" in Principled (Disney) isn't about adding natural Velvet like effect like in upcoming Principled, but to account for the energy loss caused by Disney's "Exit IOR". (Edited, as topcoat now works as you explained - I checked - was it always like this though? Feel like I must have tested this way back when). 5. It's worth mentioning that testing out with the furnace test is best done using Filmic if going by visual clues. Using sRGB you'd have to measure the >white values. 6. The "simply added" clearcoat isn't usually a problem for what it's designed to do, as albedo values are supposed to be realistically low anyway, and user is supposed to lower the specular to account for the topcoat. Even the fixed Principled v2 can't properly simulate a complex topcoat layer, which should automatically lower specular on lower substrate and can even enhance saturation in the diffuse and cause complex total internal reflections that picks up base color per bounce until it leaves (can be somewhat handled with old specular tint). Look at what happens to wood when adding some types of lacquer, polyurethane topcoats, and shellacs to it. Trying to mimic these, although you can sometimes "get away with it" using a simple topcoat layer, usually requires more complex manually setup materials. Note that we're not able to simulate Disney Diffuse outside of Principled, as the Diffuse shader can only be Lambertian or Oren Nayar (using the diffuse roughness value). 7. Old "Sheen" was completely broken with simplified AO bounces, or at least used to be, haven't used it in ages because of it. 8. Many real life materials does exhibit energy loss. This is accounted for in the diffuse part simply by having a lower albedo, but the specular will reflect according to fresnel unless specular is lowered. All porous materials (unrelated to roughness), and all materials with microtubular molecular structures; light hitting one of these "pores or tubes" will "get trapped" and never be allowed to bounce back out. This is why specular value is called an "artistic choice", since it's not part of IOR value given for the material. 9. So basically, there is more to shading in Cycles than always relying on Principled; Principled can't handle complex topcoats, Principled can't make materials designed for "thin geometry" yet, and Cycles (at all for now) can't do anisotropic reflections without roughness (I think only Filament renderer can do this) with Eevee not supporting anisotropic at all (without weird hacks). I haven't tested if v4 allows subsurface scattering to transport light at all, but I kinda doubt it. 10. Other than Mitsuba, *ALL* renderers will suffer some of the above problems to one extent or another, with associated workarounds. There is no perfect renderer.
@IridiumZero
@IridiumZero 9 месяцев назад
OK, this guy knows some stuff for sure. Have you written anything elsewhere? I haven't seen anyone go into this much detail before.
@gottagowork
@gottagowork 9 месяцев назад
​@@IridiumZero Only youtube comments and Blender forum. A thing I missed is that new Principled have abandoned the GTR distribution for the topcoat (which looks a lot nicer with its longer tail) as it was "completely broken" (changed from sharp to GTR above a certain threshold, making low values hard to map) in Cycles and didn't really catch on with the industry for some reason (VRAY/Houdini has tail falloff control though), and replaced it with standard GGX (generalized GTR without tail control).
@christopher3d475
@christopher3d475 9 месяцев назад
1. Whatever its genesis, the diffuse in 3.6 wasn't ideal, the diffuse that's there now in 4.0 is working better. Because this is a video about Blender/Cycles and the diffuse algorithm it has been using, it's what I referenced, not Disney. But you could split hairs about that. Regardless they're moving away from Disney's standard towards MaterialX and OpenPBR. 3. This doesn't make any sense. I used the term 'glowy' to describe the visual most prominently visible at glancing angles on a surface with glossy reflections that isn't properly energy conserving. Even so F0 will still exceed proper conservation by 4% using a standard IOR of 1.5. but that's no longer the case in 4.0. 4. I used Sheen 4 to amplify the effect, I didn't say it was a correct value. I used 4 for Clearcoat to give the common IOR 1.5 reflectance function. 5. I used Filmic in the video for this reason, in both 3.6 and 4.0. 6. Whatever you're trying to communicate here, Coat works much better in 4.0 7. 4.0 Uses a brand new algorithm, for the reason you outline. 8. Perhaps, but the point here is that Cycles was always exhibiting energy conservation problems. Using proper conservation is a much better starting point for common materials. 10. All renderers are indeed using approximations of physical reality. Even spectral renderers are using 'hero' wavelength values to sidestep the problem of having to render massive numbers of samples to represent an entire visible spectrum.
@gottagowork
@gottagowork 9 месяцев назад
​@@christopher3d475 1 & 8. It's not about diffuse or Cycles, it's only about Principled. Which I believe was even called Disney early on. And yes, Disney always had shortcomings which was explained in their paper. 3. I mean "glowy" the way it has been described on the forum since long before Principled even arrived. 6. I haven't tested updated Coat much (I'm still on 3.6 for work), but since it's now GGX at least the horrible interpolation problem should be gone. But I have to say, the GTR generalization "looks" far better than GGX. Glad they fixed the energy conservation thing, but for me it wasn't what made or broke my renders. If I needed more accurate, I could make it manually myself. 10. Indeed. In real life there is no such thing as "diffuse", only volume absorption and scattering. But I wouldn't have the patience to wait for everything doing volumetric rendering 😁
@sleepCircle
@sleepCircle 9 месяцев назад
thank you for finally making me get what the energy conservation problem is hahaha.
@peihengzhang5647
@peihengzhang5647 8 месяцев назад
This is huge🎉
@itellyouforfree7238
@itellyouforfree7238 9 месяцев назад
This is HUGE! I've been complaining about these issues, starting from incorrect Fresnel, for years.
@christopher3d475
@christopher3d475 9 месяцев назад
My understanding is that the fresnel code was improved also for greater physical accuracy which is a general theme for Cycles in 4.0.
@slimeball3209
@slimeball3209 9 месяцев назад
I feel soo pleasing every time i see energy conserved image after old version, i know why i have so much trouble with color correction in blender. im just instinctively tried to replicate energy conserved look. When i frist time used lux render i was catch on this look.
@andreykovalyov3131
@andreykovalyov3131 9 месяцев назад
Thanks for so awesome explanation. Are there a sweet changes for glass rendering in 4.0?
@christopher3d475
@christopher3d475 9 месяцев назад
Not for the initial 4.0 release, but there is a change that's on the developer list called nested dialectrics which is quite important. I don't know the timeline for it though, but that'll be great when it comes.
@shishi-cc
@shishi-cc 9 месяцев назад
holy shit im looking forward to 4.0
@danieleColomboChnl
@danieleColomboChnl 9 месяцев назад
I started the video not being excited "about it", I ended the video being excited "about it"
@ShrikeGFX
@ShrikeGFX 7 месяцев назад
After reading the paper from Disney - Physically Based Shading at Disney, they describe the issue of the current diffuse models vs real measurements and they describe exactly this edge darkening and highlighting on roughness which is not covered by Lambert or Oren Nayar, so this must emulate this or tries to approximate a sort of Disney Diffuse I strongly assume (They say, the measured data shows a highlight on rough materials but a darkening on gloss)
@mdoerkse
@mdoerkse 8 месяцев назад
These kinds of changes are going to make Blender much more attractive for use in VFX in movies and such.
@christopher3d475
@christopher3d475 8 месяцев назад
Yes agreed. It also lays the foundation for Cycles to gain newer functionality because it's working more correctly at this low level.
@clonkex
@clonkex 9 месяцев назад
I owned a 2009 Audi A3 cabriolet for several years (2017-2023) and I'm intimately familiar with its panels, lines and trim pieces (after having to spend so much time repairing it lol). SO MUCH of that car model looks exactly like my A3 that it makes me wonder if you (or whoever made it) based the model on the A3.
@christopher3d475
@christopher3d475 9 месяцев назад
It was a model from a project a few years back. We bought it as a stock model for the project. It has some topology issues which kind of bug me, but it otherwise worked well for the tutorial.
@dagobject
@dagobject 9 месяцев назад
Man, your videos are amazing! I'm in CG for 15 years now and yet you manage to tell me something new =) Now I know why I don't like diffuse roughness slider in Arnold though oren-nayar is kinda more accurate than lambert. It makes surfaces too dark.
@christopher3d475
@christopher3d475 9 месяцев назад
Yes, there are different diffuse algorithms that behave in certain ways. TheaRender has one that simulates scattered diffuse light such that it behaves like the way the moon scatters sun light. Cool stuff.
@dagobject
@dagobject 9 месяцев назад
@@christopher3d475 I believe Oren-Nayar was developed to do exactly that. To mimic how light scatters on the Moon compared to Lambertian shading.
@christopher3d475
@christopher3d475 9 месяцев назад
Ah, that very well could be then. I don't think they used the name specifically. I think they used a variable called Sigma to invoke it.@@dagobject
@rodrigolucena1698
@rodrigolucena1698 9 месяцев назад
What a wonderful video! Thanks! Is energy conservation a property of the render engines or it is in the material? I wonder if other render softwares that uses v-ray, corona or lumion also could pass in the Furnace Test?
@christopher3d475
@christopher3d475 9 месяцев назад
Energy conservation/preservation is a feature of high quality raytracers. I would assume Vray, Corona and others are energy conserving.
@rodrigolucena1698
@rodrigolucena1698 9 месяцев назад
@@christopher3d475 Thank you. Around the Blender 2.8 release, I used to work with 3Ds Max which used V-ray render. At that moment I moved to Blender (2.8) and felt (I had no idea why), V-ray deliver more realistic renders than Cycles...
@BubbleVolcano
@BubbleVolcano 9 месяцев назад
thinking with my monkey brain, it's a huge improvement with edges details of rendering. thanks for the explained that is clear & simple! I pefer Arnold render because the edges reflection is more real than other render engine. Because object's edges shows darker reflection in the real world, blender render also is ok, but lighter and blur with reflection on the edges of object, that's make me feel fake. 4.0 is better and more real, good job!
@christopher3d475
@christopher3d475 9 месяцев назад
You can affect edge reflections as far as reflection sharpness. I did a video on that. Reflections often become sharper with increasing incidence, and you can simulate this easily enough with two nodes (layer weight and RGB curves nodes).
@Ntom24
@Ntom24 8 месяцев назад
I have a wierd issue with my scenes lately, materials seem to be kind of emmisive, like a regular pbr material seems to be light up and have saturated shadows, I cannot for my life find what the problem is. Is this a common issue? (Default cycles GPU render settings, AgX color space base contrast). Your videos are great, very to the point and clear, thank you for the content it really helps.
@LeOtter
@LeOtter 9 месяцев назад
Great and informative Video and even understandable for a newbie :) Quick Question about Example #4, when switching from 3.6 to 4.0 a shadow on the wodden cutting board seems to vanish or more likely the light which hits the french press gets "correctly" refracted. Is this also a side effect of multiscatter GGX or something else in 4.0 ? ^^
@christopher3d475
@christopher3d475 9 месяцев назад
I was wondering if anyone would notice that. 4.0 has an updated version of a technology called Path Guiding, and I assume it's a result of this improvement, since it deals with light through refractive objects.
@LeOtter
@LeOtter 9 месяцев назад
Ah I see, thank you for the enlightenment :) @@christopher3d475
@josebarria3233
@josebarria3233 9 месяцев назад
I guess a fraction of the incident light will convert into heat in 4.0 MGGX in the case of translucent materials and thats it looks slightly darker
@IridiumZero
@IridiumZero 9 месяцев назад
The new energy conservation, combined with AGX, will go a long way toward fixing the 'washed-out' look that plagued Cycles before.
@christopher3d475
@christopher3d475 9 месяцев назад
Yep, that's exactly what the car example shows..
@xdanic3
@xdanic3 9 месяцев назад
And at some point spectral rendering, which might be more minor but is another step into realism!
@christopher3d475
@christopher3d475 9 месяцев назад
Yes, there's already been work done on a spectral version of Cycles. It'll be awesome.@@xdanic3
@goldenmikytlgp3484
@goldenmikytlgp3484 8 месяцев назад
So im guessing, for people that use 3.6, until 4.0 is released it is best to use multiscatter GGX, even if the 3,6 version isn’t perfect Thanks for pointing this out. Since i want to do realism Im learning as much as i can
@christopher3d475
@christopher3d475 8 месяцев назад
You know for a lot of general materials you may not notice a huge difference, but it doesn't hurt to test switching it over to multiscatter. however, once you do migrate to 4.0, make sure and switch it over since older materials will continue using what is specified in the material which is GGX.
@dancaldwell910
@dancaldwell910 9 месяцев назад
Nice video! Is there a link to a paper or technical information on the Furnace Test? I want to look deeper into this but can't find anything
@christopher3d475
@christopher3d475 9 месяцев назад
Not that I'm aware of, it's just an informal test people have used.
@gerdsfargen6687
@gerdsfargen6687 7 месяцев назад
Hi Christopher, I love the ball stand type model. Ive seen it in a few videos. Is there a project or model download for it or is it your own model?
@christopher3d475
@christopher3d475 7 месяцев назад
I developed that for testing purposes. I put a link to it in the latest video I just posted.
@gerdsfargen6687
@gerdsfargen6687 7 месяцев назад
​@christopher3d475 really? Awesome! Thank you!
@Robjohn37
@Robjohn37 9 месяцев назад
I wonder how many people noticed the word Blender misspelt in the description
@digitalartworm
@digitalartworm 6 месяцев назад
Looks better but what impact if any in render times between the various identical scenes? Or is there no real difference (GTX 1080ti here)
@christopher3d475
@christopher3d475 6 месяцев назад
Negligible to the raytracing performance.
@VFXCentral
@VFXCentral 9 месяцев назад
Blender should pay you. Lol. Those are fantastic videos. Cant wait to try out 4.0
@jakobholland
@jakobholland 9 месяцев назад
What is the model that you use for the examples? the model in the introduction section
@christopher3d475
@christopher3d475 9 месяцев назад
It's a model I created for testing materials. It's available in the About section of my video of SSS & Sheen.
@jakobholland
@jakobholland 9 месяцев назад
Ok thanks!@@christopher3d475
@rbettsx
@rbettsx 9 месяцев назад
AAA content, as always. Just to put a spanner in the works, I guess a full energy model would have to be spectral? It would have to take into account light absorbed from the visible spectrum and re-emmitted as invisible IR heat, and vice versa: UV fluorescence.
@christopher3d475
@christopher3d475 9 месяцев назад
Yes, spectral rendering will take this up a notch for sure.
@stubman5927
@stubman5927 9 месяцев назад
what's a full energy model? if you mean energy conserving/preserving, then you don't really need spectral or even RGB if you want to stretch the truth. energy conservation/preservation is only concerned by the quantity of light that is scattered from a surface is same as incoming... and you can define the light to be whatever you want in rendering, be it spectral, rgb or even hsv. spectral rendering would still be cool to see tho
@rbettsx
@rbettsx 9 месяцев назад
@@stubman5927 Hi! This was more a question than an assertion.. I was just thinking that the current model only does the book-keeping for visible light? That, by contrast, in the real world, the conversion of total energy in and out of visible, to and from invisible radiation, has a part to play ... what we wind up seeing is only part of the picture. By all means, shoot me down.. :)
@stubman5927
@stubman5927 9 месяцев назад
@@rbettsx if that's what you meant, then my answer remains unchanged. but i suppose for extra clarification, energy conservation/preservation in shading models are not necessarily concerned with quantity of light energy but rather electromagnetic radiation. visible light is one of many possible ranges of electromagnetic radiation that can exist. what this means is that you can do the typical spectral rendering of visible light by tracing rays with defined wavelength range in visible light range, or maybe even simulate old cameras' sensitivity to infrared by adding additional infrared range for tracing (blackbody will come in handy), or literally do microwave range by having the wavelength range within microwave range and the principled bsdf would do energy conservation/preservation just fine for any of those scenarios (though since microwave wavelength is quite large, the fact that it is modelled as a ray than a wave is a more concerning matter for realism...)
@edivigroom698
@edivigroom698 9 месяцев назад
Why transmission roughness got removed?
@christopher3d475
@christopher3d475 9 месяцев назад
I don't know the answer to that honestly.
@tenns
@tenns 9 месяцев назад
i didnt understand the difference between energy conservation and preservation, could you explain it ?
@christopher3d475
@christopher3d475 9 месяцев назад
Sure. Energy preservation problems happen with GGX because the light ray strike the microfacets and bounce just once. This causes a loss of energy because that single bounce can be occluded by another facet. That causes the darkening we see. The multibounce GGX models multiple bounces inthe microfacets and results in a more accurate preservation of reflected light off a surface. I have a more in depth video on the energy conservation problem here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-kgORQ5tMe2I.html
@tenns
@tenns 9 месяцев назад
@@christopher3d475 but that's not called preservation, that's conservation isnt it ?
@christopher3d475
@christopher3d475 9 месяцев назад
No, because we're preserving a more correct amount of information that's otherwise being terminated too early. Conservation refers to too much energy accumulating on a surface that has multiple reflectance layers.@@tenns
@tenns
@tenns 9 месяцев назад
@@christopher3d475 aaah ok! thanks for you answers! In physics "preservation" is not the used term, we use the violation of conservation of energy for both when there is too much or too little energy. I couldn't find much literature defining it either! fascinating video, and your whole library too! truly spectacular for someone coming from a physics background!
@christopher3d475
@christopher3d475 9 месяцев назад
Yeah, preservation is the specific term used in 3D development for this situation, although conservation really could apply to both in a sense, given that nature doesn't like to do things with any more energy than it needs to.@@tenns
9 месяцев назад
A super bright object that radiate light in any direction will be darker on the side, we can see this on our own sun, best is to look for images of ISS passing the sun where we often can see that the sun look brighter in the center and darker on the sides, the same goes for the moon even it is lit a bit from the side, so is that not how the light shall work in Blender too?
@christopher3d475
@christopher3d475 9 месяцев назад
All rendering systems approximate what's happening in reality. There are multiple scattering equations to simulate a diffuse surface each with somewhat different characteristics. For instance, the Diffuse BSDF node uses a basic lambertian surface, which is a kind of simplified diffusion, and transitions to the more sophisticated Oren-Nayar diffuse algorithm which models light more accurately, including what happens at glancing angles. For whatever reason they've chosen a model for the updated Principled BSDF to make the layers work more consistently from an energy conservation standpoint which on the whole creates a more realistic image.
@pikachufan25
@pikachufan25 9 месяцев назад
Learn something new woo... now i know why it exist that glow and how to Control it better.... Yay! xD.... Neat that we are Gonna have a Better one Soon! xD
@MrTomyCJ
@MrTomyCJ 9 месяцев назад
Clearcoat normally goes from 0 to 1, where does that 4 come from? Values higher than 1 are supposed to stop being energy conserving, I suppose.
@christopher3d475
@christopher3d475 9 месяцев назад
In 3.6, it functioned differently than it does now. Watch my video on Coat to understand the difference. 4% was the F0 reflection amount which matches a 1.5 IOR. You just needed to know that. It behaves more logically in 4.0.
@depechejung7917
@depechejung7917 9 месяцев назад
Why so far the metal parameter has multiple values?
@christopher3d475
@christopher3d475 9 месяцев назад
I'm not sure I understand what you mean.
@BoyceBailey
@BoyceBailey 9 месяцев назад
my eyes are now open (in 4.0) Where are you learning this from?
@christopher3d475
@christopher3d475 9 месяцев назад
Just studying other renderers over the years. For instance, TheaRender, which is still one of the best renders around, talked about this somewhere. I owned a copy and probably 10 years ago would do tests with it. You could output each of the layers (diffuse, spec, etc.) from one of its biased mode and actually see the conservation on the layers. So that was one tool. Also, the Blender developers discuss this need in a presentation at the last blender conference about a year ago. But the furnace test is really the gold standard in this regard.
@Aldebaran80
@Aldebaran80 9 месяцев назад
rendering is a fucking art... how many technical variables to know how the light works
@johannes-vollmer
@johannes-vollmer 9 месяцев назад
with any clearcoat factor above 1.0, why would you ever expect it to not glow? :P
@christopher3d475
@christopher3d475 9 месяцев назад
Well, clearcoat behaved differently in 3.6 than it does in 4.0. If you watch my video on the new Coat update in 4.0, you'll see how they've changed it.
@johannes-vollmer
@johannes-vollmer 9 месяцев назад
@@christopher3d475 I'll watch that, thanks for the hint :)
@souharmhammed766
@souharmhammed766 9 месяцев назад
❤❤❤❤❤❤👍👍👍👍👍👍👍🇲🇦🇲🇦🇲🇦🇲🇦🇲🇦🇲🇦💕
@ruslandad365
@ruslandad365 8 месяцев назад
This Light theme Burn my eyes!!!!
@Gaalets
@Gaalets 9 месяцев назад
Oh, yes! Blender renders have always been very annoying. Now I understand why.
@hameddesign70
@hameddesign70 9 месяцев назад
people are complaining about energy leak , me yelling about unnecessary memory leak over two years and no one cares ...
@christopher3d475
@christopher3d475 9 месяцев назад
Memory leaks can be tough to track down. I assume you've been filing bug reports. At least for the ones I've submitted they've been pretty good about following up with them.
@blckops712
@blckops712 9 месяцев назад
welcome to 5 years ago blender lol
@SlyNine
@SlyNine 9 месяцев назад
In real life if you use three polarized lenses each at a 45 degree angle from each other, it seems to violate energy conservation. Not really relevant, but interesting.
@rbettsx
@rbettsx 9 месяцев назад
Hehe..a lovely one.. no violation, only 25% gets through.. the weird bit is that it is not zero, because of the collapsed align/don't align 'decisions' taken by the photons on the way.
@robertYoutub
@robertYoutub 9 месяцев назад
While that is great, its not a revolution and not yet enough. The thing was in other render engines since almost 15 years. Slowly Blender gets up to date.
@christopher3d475
@christopher3d475 9 месяцев назад
Better late than never.
Далее
Subsurface Modes & Anisotropy in 4.0
13:32
Просмотров 25 тыс.
КТО ЭТО БЫЛ?
25:31
Просмотров 1 млн
MC TAXI: АК-47
35:14
Просмотров 420 тыс.
Improved Coatings in Blender 4.0
16:20
Просмотров 17 тыс.
Blender Path Deform
1:38
Просмотров 11 тыс.
Light Path Node Demystified and Visualized in Blender
23:26
Subsurface Scattering & Sheen in 4.0 are awesome!
13:33
LightWrangler Updated! Point and click lighting.
14:20
Просмотров 3,1 тыс.
AgX in Blender 4.0. It's coming, are you ready?
18:53
Просмотров 123 тыс.
Become a Topology PRO with these Five Tips
9:13
Просмотров 279 тыс.
#фильм #кино #фильмы
0:55
Просмотров 7 млн