Eevee Realistic Lighting And Rendering

1. Introduction: In blend five with Vulcan, IV looks insane. This can produce stunningly realistic visuals right up there with cycles and other serious path tracing engines, but often at a fraction of the rando time. Is it as good as cycles in every scenario? Of course not. But with the right artistic and technical know how, the results can be pretty darn close. In fact, in areas where cycles traditionally struggles like interior bound slight or fast volume rendering, a well tuned IV setup can actually outperform pure path tracing. Now, let's be clear. No engine is best at everything. I'm not saying IV beats cycles everywhere, far from that. But complex nature shots with millions of polygons, huge sprawling trains, complex light physics, that's still cycles home turf. So yeah, cycles is great. We all know that. Over the years, though, IV has changed. Part of it was the new Vulcan rendering back end. Part was new features and bug fixes from Blenders developers. And part was this community inventing new ways of doing things. Ivie has evolved. Sorry about that. And you missed it. It became a realistic rendering beast you probably never thought it could be. Long gone are the days when you needed third party add ons like SSGI to fake global illumination when the whole thing felt unfinished and glitchy. Now it can do things we only dreamed about a few years ago. This thing is furious now. Please welcome IV realistic lighting and rendering, a complete course produced and narrated by me, Gleb Alexandrov with the help of 80 Burrows. What will it do for you? This course is built on one idea, Launch blender, switch to rendered mode, and work with feedback that already looks finished. It's about creating stunning environments, characters, archives, and animations without the render weight. Inside, you learn it all. Dramatic environment lighting, AGI lighting that looks nearly identical to Shikos. Versatile lighting techniques and real time global illumination. Navigating EV's GI settings like a pro, lighting amazing character portraits from simple to full cinematic, realistic ambient eclusionRflections, believable glass and water. Baking light with light probes without the usual hassle. And much more. And a bit more. In Extras, we have a growing library of short videos from our YouTube channel all bundled on organized foo. This course isn't just about which buttons to press. It is about training your eye to see why one image feels alive and another feels dead. We break down the why behind every choice. The core principles of light and composition that pretty much work in every software in any style because tools change and these fundamentals, they stay with you forever, and even AI won't take that from you. We started with one idea realism rendered in fractions of a second. That's really the heart of it. I'll see you inside. Y. 2. Elevator Shaft Initial Scene Setup: Starting with level one, which will be the bare bones, yet super cool looking lighting scheme realized step by step to make this elevator shaft into a beautiful place. This lighting scheme was inspired by half life and many other classic games. So rise and shine, Mr. Freeman. Rise and shine. So imagine you want to design the lighting for this environment, which by the way, you can download from the project files. Where do you even begin? It can feel pretty intimidating. It's quite a typical environment, though, for movies, games, and more. Learn to light one environment like this and you can apply this knowledge to the rest of your scenes. And that's what we are going to do now. So here's the scene that we'll be setting up lighting for. Before we start, let's take a moment to check the blenders preferences, the system tab, the visual back end should be Vulcan if that option is available. Vulcan is so much faster than open gel, specifically for EV. It can process a heavy geometry much faster, literally millions of polygons, process GPU subdivision much faster, and even load textures and stuff much faster in most scenes anyway. So it's a no brainer. Obviously, we're going to use IV to render this scene. The material mode is useful for checking shaders and textures, but we'll jump straight into the rendered mode and start designing the cinematic lighting for the scene. What does cinematic mean? If you show these two images to any number of people, most will say that the right one is cinematic and the left one is not. Why? The left one is flat and the right one is not. The left one doesn't have any visual hierarchy or composition while the right one does. The left one doesn't invoke a mood while the right one is moody. And the left one isn't particularly realistic while the right one is. Let's try this again. How about something like that? Does this tick all the boxes? Nah, still flat. It breaks the shape rather than helping us to read it. I'm still not sure where to focus my attention. I don't feel any specific mood, and technically, it doesn't look too realistic to me, either. Now, with these guidelines in mind, let's start designing the cinematic lighting for this scene, the one that would make Roger Dickens nod approvingly. The camera serves the magical purpose of creating a frame for your composition. So to truly see and feel this frame better, I usually turn on Passeparto and max out this slider in the camera settings. So it really turns into a frame. Escape this view anytime to explore our scene, yet it's paramount to have this frame to see the final composition. As we are going to be using IV, I want to be able to debug the performance. The way I can do it is to hit play and watch the FPS counter. Now it's capped at the default 24 FPS, but I'll actually unblock it, switching it to 120 frames per second. So now we can see, Okay, so far so good. The viewport is still responsive, but things might change, so that's a way to measure it. All right, now on to lighting. We need to go from here to here. Not bad, right? Let's start with ambient light. In cinematography, they say, set up your room tone. In the world settings, the ambient light strength sets our room tone. And the way I want to handle it is to keep the exposure really low. You know, just a hint of ambience, so it's not completely black. Something like that. By default ambient light looks pretty flat and heavy, so I'll activate racing right away with the default settings to make it look slightly more three dimensional. Then for such darker tones, the view transform can actually make a big difference. For instance, filmic has a lifted base compared to AGx and we change how these darker tones are rendered. Or AGx has different looks or contrast presets to choose from. Low contrast will lift the floor a bit as well. Or and that's what I'll do, we can activate curves. And set the floor of the room tone manually. There's the black level here, and we can gently wash it out a bit. That will also make it feel slightly more vintage or even cinematic. So I will say to two -0.001. And that immediately gives us more leeway in the shadows. Here's our room tone and the room tone with a slightly lifted base to make sure there are no absolute blacks. This helps everything blend together nicer. Now, this is meant to be daylight. So should the main light come in from the camera direction like this, that would be the worst case scenario, in fact, as you make the foreground lid like a Christmas tree. That's just poor lighting design, even with softer lighting and all that. Lighting from the camera side is usually very complex to execute. It will be far better to light it from these windows, so we actually get a dark backdrop and the main light flow coming towards the camera, not from the camera. That will solve so many composition issues and will inherently sort the values by depth, lighter background, darker foreground, almost like sorting it by Z depth, indeed. Right, window lights. In cinema, every artificial light needs a motivation, a source that viewer can believe. So first of all, I will create the actual windows. These will be just planes, which will be roughly shaped like the window openings and will imitate the daylight, which is sipping in from these windows, supposedly. I'm roughly placing it near the window openings. It doesn't have to be exact. The material is going to be just a standard emission shader with a white color and fairly low strength. We can calibrate it anytime anyway. Now I'll copy this window light object to the rest of the openings. Some of them are going to be blocked, by the way, just to give it some interest and break up the rhythm. **** to D, duplicate. Now I will select all of them, press M, and move them to a new collection. Light windows temporary as I'm going to show you something and delete it later on. Mm hmm. The ate raised bound slide is a bit grainy and that's because we need to crank up the end the samples for the viewport. 32 or 32 multiplied by two. You can input math. Oh, 64 is better. Alright. Out of curiosity, how would cycles appear? Well, a bit more boundlide and a bit more photographic too, as well, but we'll get there. Now I'd like to mention it again that this scene has blockers covering some of the windows simple planes or something. It's basically for being able to manipulate the lighting and the composition. For instance, you can block some of the windows to avoid the uniform fill and instead go light, dark, light, dark, light, then a lot of dark. For moody shots like this, it's all about creating contrasts. These windows that I've just added, technically, these can be just emissive planes. But when you download the scene, you'll actually find the light windows collection there, which features slightly more interesting windows with actual textures on them. You know, just a photo from Pixie Bay or something, also using an emission shader. But if we have a look at the setup in the shader editor, I mean, you'll see that it has a texture within, indeed, used as a color of that emission shader. The curves there to tweak the color, and the strength regulates how bright it appears. See nothing complicated, really. As for the brightness, I don't want these objects to pump out a lot of light into the scene. I want them to serve as the motivation for the rest of the lighting setup that we'll add shortly. So I'll really dumping it down to the lowest value possible, honestly. So these windows still seem bright enough, but don't emit much light at the same time. Something like that. Light motivation is a critical concept in lighting. Technically, we could have used this window material to illuminate the entire scene like that. With ray tracing turned on, it would actually work in Ivy as this emissive material will actually create indirect illumination. So that will be alright. But the windows themselves will become completely blown out, like, totally clipped. And what I want for them instead is just to be bright enough so we can believe there's daylight outside. Then once we establish that belief, the magic happens, we can add artificial lights here or there or anywhere, and it will all look believable. Even the stuff like this, you know, the sunlight of a different color. Will all look convincing if it's roughly within the ballpark of the daylight theme we've set by using these windows. That's the power of light motivation. So now on to aerial lights. These emissive planes, the windows are just set up as a motivation. So now let's actually add the light. I mean, the aerial lights. The idea is to make them window shaped and basically line them up with the window frames. It also doesn't have to be too precise. Just roughly match the window shape and bring the light up really close. In the light settings, we can make it a lot brighter. Um, why doesn't it get brighter, though? When set to ten K Watts? Ah, that's because these emissive planes actually block all the light. Whoa, crazy. So with this emissive plane selected, I'll go over to the objects properties, visibility, and we'll switch off the shadow visibility. So now that should be alright. To avoid manually copying it to the rest of the windows, I'll select the entire collection from the right click menu. Make sure this first window is the active selection. Then right click the shadow. Copy to select it. Back to the area light settings. Let's fine tune its power. I'll start low and we'll slowly ramp it up. It's hard to judge exactly how punch it needs to be. I feel it's better to keep it fairly subdued to retain the mystery and the shadows and not get out of sync with the light motivation. I don't know. So maybe something like that. But the way this normalized checkbox just means that no matter how you resize the lights, the output will be consistent, which is handy. See, the power doesn't change at all on changing the size. All right. Then skipping the color and temperature because I want to keep the color neutral for now, I'll go straight to the custom distance. When you turn it on, it lets you calibrate how far the light from that source actually reaches. See this line emanating from the light source. It's the maximum distance the light from that source is allowed to travel closer to the end of that line, it will gradually fade to black. Now, of course, it can be used to tweak it to your artistic vision. But if you want to make it look like cycles, I mean, physically accurate, then this value should be really large, like 400 or whatever to cover the entire length of the scene. If you do it like this, your area lights in EV will pretty much look like area lights in cycles, AKA, no difference in follo. Actually, this is the physically accurate falloff. But there are times when you may want to over exaggerate the fall off like that, and this is where custom distance may come in handy. What we are doing here is the stylized photorealism. So if you want to do it like that, by exaggerating the fall off, nobody will blame you really. Actually, I will somewhat over exaggerate it, so it barely touches that right wall, a little bit of extra fall off. And then right away, I will duplicate this light a few times, actually, and position it near every window or at least near the windows that aren't blocked. Shift D is the shortcut for duplication. Once more? All right. So I've covered the rest of the windows. It probably doesn't matter if you omit one or two for creating your own rhythm. Like here, I've lit the elevator through that last window. So do you want this, or do you want to make it dimmer or leave it in the dark? The beauty of motivated lighting is that as we've mentioned, you can arrange the artificial lights however you want as an artist to the extent of your creative imagination, really. That means you can add all sorts of lights if you want, but we'll get there. I think the window material seems a bit too dull, now, so I'll give it a bit more strength. Its motivation started to falter. Alright, we've pretty much nailed the main lighting idea for these environment, and it doesn't look bad. There are hints of depth, shape, contrast, and atmosphere appearing. But let's ask ourselves, is it as non flat and compositional interesting and moody and realistic as it can be? It's worth stopping every now and then and asking yourself these questions. I think we can score more points on all these things by enhancing the tunnel separation between foreground and background like this. So the depth is easier to read. This will make composition kind of framed by these foreground elements, and this will make the viewers eye have nowhere to go except the shaft. Hey hey. And as a bonus, the whole thing will feel dustier and more abandoned. Basically scoring more points for all these factors. The foreground the foreground elements like this and this and this, they have sloppy texturing, but it doesn't matter because they are here to create a nice framing. Maybe if we reduce the background light in a bit, maybe even to zero. I don't like how other places become completely black, though. So for now, I'm going to be a bit less radical and will reduce the ambient light just a little bit. But the core idea is we need a better separation between the foreground and the background. The way we can do it is by adding some fog to separate that distant part from the foreground. So let's do the fog. Before and after adding new layers and elements, I'm checking the FPS if it's still healthy. That still good. All right. For this level one of difficulty, the easiest way to add fog is to simply go to the world shader in the shady editor. I then simply add the principled volume shader there. Shift A, shade principled volume, connected to the volume input, of course, with just a bit of density. We don't need it to be really thick. And maybe hin isotropy, as well. Something like that. The density, though. I really want it to be just a hint of a haze or I don't know, dustiness. Look at that. It also helped with a clue in the foreground for some reason, but okay, pretty interesting. So that would be the easiest way to create that framing and that separation that we need. It's not the most versatile approach, though, so I'll disconnect that for now, and we'll show you a bit more versatile technique in the next section. That is, if you decide to up the difficulty level with us. Alright. Awesome. Here's TLDR for level one lighting. We started with flat, then turned on tracing and set the room tone really low. Balanced by lifting the floor of the view transform a bit to give it a slightly washed out look. Next, we added the missive planes for the windows. This would become a motivation for our key or main lights later on, the area lights in this case, which created the actual illumination, c the main source of exposure. Finally, adding some fog helped to separate the composition planes and contributed to a thicker mood. Even though this setup is really simple, it does its job really well, filling this location that you can explore in real time, by the way, with a pretty thick mood. Right. Where am I? 3. Elevator Shaft Developing the Mood: Welcome to Level two of the elevator shaft environment lighting and rendering tutorial. In the previous part, we started with the flat default lighting of IV and gradually worked our way towards the moody daylight scheme. We established the main window lights and prepared the canvas for further creative exploration, and for whatever lighting decisions we can pursue next. And the next has come. In this segment, we'll learn how to set up a beautiful sunlight in IV and sneak a peek into the crucial techniques that make environment lighting visually appealing and realistic. Alright, we established a few goals for our environment lighting to make it look non flat, compositionally sound, moody and realistic. While in the first part of this tutorial, we moved in that direction. There are still creative techniques we can utilize to score even higher like adding sunlight. So let's see what we can do to push this scene into that half life to esque territory. First, let's take a look at how cycles would handle this. In cycles, what I would do is simply add one more area light. Point it at one of the windows at a certain angle if I wanted to create a slash of sunlight. Then in the light settings, I would make it seriously powerful ten K watts or something, like a really powerful spotlight with a slightly warmer color, perhaps. And then I would go straight to the beam spread and reduce it way, way down to really, really focus that light beam and turn it into this rectangle, basically. And that would be as perfect an imitation of sunlight as we could get while keeping it perfectly controllable at the same time. Easy to point, easy to move around, and easy to copy where we need those light strokes compositionally. I mean, these sledges are literally how we paint with lights, and the beam shape in cycles make it so simple to paint. Like, Hey, everyone, look exactly here. If we switch back to IV, the aerial lights don't have this beam spread, so it's a bit tricky here. Of course, there are different methods of achieving the same result in IV, and we cover all of them in the light leak chapter of this course. It's crucial to understand, though, what makes sunlight unique. Because the sun is so far away, technically the light rays from the sun should all be parallel, not converging or diverging, like in spotlights, but parallel. Technically, only the sun lamp satisfies this requirement in IV. We have the sun lamp hidden in this temporary folder. So let me unpack it. Its light type is set to sun, obviously, nothing special. So by default, it shines from an infinite distance and covers the entire world. So anything in the blender scene will be lit by that sun. So if you want to use the sun lamp, you need blockers. These could be planes or anything else, like this flat box at the top of the scene that blocks the part of the light that would otherwise come in from the top and turn the foreground into a Christmas tree. Logic is simple. You can move the blockers around to try to nail what part of the light you want to let through. Then some part of the sunlight enters the shaft through the windows, so you'll need some blockers there as well to manipulate where you want the light leaks to appear. This is an okay, totally workable solution, and the sunlight has the best shadow quality of lights in IV as we remember, especially when light is supposed to be really far away. So there's a decent way of doing it. Then you can make it really strong, so it actually gives off a beautiful global illumination or bounced lighting. Needless to say, real time bound lighting look absolutely gorgeous in IV. If you haven't tried that, just turn on retracing in ranger settings and make the light really powerful. We discussed RaltGI a lot in other chapters from this course, but here you can see how global illumination and IV looks and feels with just one light. Actually, the window lights are also important as they support this effect and make it believable. If we remove that visual grounding, it will start to look really uncanny. Actually, that's one of the things that cycles and other path tracing ran engines struggle with calculating the bounced lighting from tiny leaks in the interiors, and IV just traces it in real time, giving cycles a run for its money. So it's very cool, indeed, and sun lamp can definitely work fine for that. Getting back to our elevator shaft, this sun lamp could work really well. It's just placing the blockers that could feel a bit fiddly. So let me show you the alternative. Spotlights. Using spotlights would also work reasonably well despite the cone shape of the beam. In fact, it would be as simple as pointing it towards the windows, given it enough power. The exposure slider works pretty well for that as well. Then probably lower temperature, like 35 or 4,000 kelvins, imitating the warm evening sunlight. That would be it. It would be just a matter of pointing that flood light where we want it, making it as powerful as we want it. And that could also work pretty well. Especially given that for spotlights, beam shape can be tweaked and ivy. So that plus moving the light reasonably far away to make its rays reasonably parallel, imitating sunlight could work really well. Finally, the radius would control the blurriness of these light leaks. So it's a decent alternative to using the sun lap, as you can see, even though the shadow resolution won't be as good. T LDR, if I were you, I would just use one of these two methods, either a sunlight with blockers or a spotlight. It's just for a few flashes of light. Anyway, it will look great. Well, depending on how close the spotlight is to the scene, it may reveal its conical shape of the beam when shining through the fog, for instance, so it's something to be aware of. But again, I wouldn't worry about it too much, especially if you can move it sufficiently far away and aren't too concerned about the shadow resolution. Pick any of these methods and roll with it. I myself would my guy wear a nice little tube with an area light inside. That would go swooshy, swooshy and do the same thing and be perfect in that. So it doesn't really matter what option you choose for the sunlight. The point is to get the spots and the bounced lighting going. Now the quality of the bound sliding is to be defined in the rate racing settings. A, one of the most important panels in IV. It's here in the rand tab. I will tweak just a few things in there because the defaults don't give us the best quality. The first thing, I'll go to the fast GI approximation tab and do something about the steps. The steps define both the quality and the reach of the bound sliding of that fast GI pass. So increasing this value should make the GI bounces a bit stronger and of higher quality as well. At some cost to performance. Going 64 or maximum isn't probably necessary even 16 will already offer noticeably higher quality. It's pretty subtle here, so let me isolate just the sun. On step, two steps, four, eight, and then 16, a bit better. And then at 64, you hit diminishing returns. So 16 is good enough. For now, though, let's focus on conceptual things rather than technical. For instance, global illumination light fall off or how far the reflected light can travel. This is one of the things that can be modified rather easily in IV unlike cycles, and it can be used to thicken the mood and kind of limit where the light can reach. Notice those darker shadows in the IV example. If you switch to cycles right now, you'll notice that the bounced light actually completely fills in this area, for instance, flattening it out considerably, is because not only the directional light can reach there more easily as its flow hasn't been modified, but also the bounced light from this spot, from this one as well, can reach it more easily, and it just flattens it out. Even though, we can do a pretty interesting thing. As you remember, we've already limited the directional light from these area lights or the window lights. We've modified the custom distance so the light doesn't fully reach this wall, turning it into a Christmas tree. So we've already limited the distance in the previous part of this video, but only for a direct light. But we also have indirect light to take care of because it can reach places and flatten things out, as well. The bounced light generated by ray tracing, basically. It also has a distance parameter that we can tweak to limit the spread of the bounced light itself, which is pretty cool. So we can, for instance, make it barely reach that t wall. Cool, right? But let me reset it back to zero for a moment to show you something else. As you know for sure, there's another issue with unlimited bounds in IV, and that is screen space global illumination, meaning that the bounce light disappears when the camera turns away from the light. Because IV works in screen space. So that leads to that jarring popping effect. Here it's pretty mild, but something like that could be a better example. See it going on and off. Like, Whoop. Gone. That is a limitation of the screen space rate racing or screen tracing. Well, there are hacks to get around this issue, but we won't go there in this particular video. What we need to know right now is that this limitation exists by default. But fortunately, in the rate racing settings, there are things that might help us with that. So what happens when we override that distance and limit the indirect bounces a bit is that we also greatly reduce this popping effect. Let's have a look at it. Now, there's less popping because the bounced light doesn't reach that far wall anyway. But mostly it's all about controlling the light spilt and not let it spread too far. Actually, the windows emission material also generates bounced lighting that potentially can hit that wall and other stuff. Yeah. So if we don't limit it, guess what? It will also illuminate this wall, and it will all add up eventually. So I'll set the fast GI distance to something like 13 meters to limit it a bit. This also shows that this thing has probably incorrect dimensions, but it doesn't matter. Let's pretend that it doesn't matter. Finally, when all the light sources are back on, you might discover that there's something else that spills a bit too much. It can be indirect or direct light like this area light. The point is you can always fine tune these things and artistically direct them. And that is really a huge pro of IV. The controls for modifying the light fall off so easily. The rest of the GI settings should be fine. The bias in this scene works pretty interestingly, allowing to fine tune the direction of the light bounds. It's pretty cool for fine tuning the GI look, though, in this scene, at least. There is no such thing in cycles for sure, these visual sliders, but tweaking GI like that. The resolution is simply the quality of the fast GI approximation pass. Naturally, the lowest quality is the fastest. Then the default one to two is the nice balanced setting. Then one to one drops the FPS a bit while being the nicest. I think 12 has the best balance to be fair. Though it totally depends on your GPU. And finally, I'll switch the denoising off completely. Not that it would matter much. It's kind of subtle, and it mostly shows in glossy reflections. Usually, I really don't know if you can see it. All right, here's a better example. Reflective PBR materials look smudgy in Ivy by default. That's because denoising with its various filters and effects is turned on by default. And once you switch these guys off, which is the same as just switching the whole thing off, the reflection becomes much crispier. So that's the thing about denoising. Sometimes it's much better to just switch it off. Actually, if you want to deeper dive into the retracing parameters like these, we cover it in other chapters of the course, mainly in the HGRI and Global Illumination chapters. Watch those to learn more. And we are moving on to fog again. Well, haven't we done that already in level one? Yeah, we added some fog to create separation between the foreground and the background and make the whole thing more mysterious. Yeah, in level one of the tutorial, we did that via the World Shader simply by adding the principal volume to it. And actually, it looked pretty decent right away, surprisingly good, actually. But anyway, it's level two. So let's try something a bit more unnecessarily complex. Just kidding. I'm going to create a new collection for the fog object, and the container is going to be a simple cube, which I will scale up so it engulfs the shaft like that. And actually, I will stretch it down, so it spreads out in the shaft. Imagine filling this shaft with gas or something. Mm hmm. Right now, obviously, this object is solid. It needs a new volumetric material instead. But first, I want to make it invisible in the solid shading view. So I will switch the display mode to bounds. So it's nice and transparent there. Now in the shade editor, we'll need the object tab to be opened to delete this default one and replace this shader by the principled volume. True. I'll reduce the density right away. Why does it, yeah. It's temporal reprojection. Okay, what I want to do is give it a really low density. Right. As far as anisotropy goes, here, higher anisotropy makes the sunshaft kind of less prominent. So I don't know. Yeah, higher anisotropy like 0.6 makes the sunshafts less visible. They almost disappear in fact. But also, higher anisotropy kind of reveals the issue with the area lights in the windows. You see those harsh rays spreading out from the windows. Actually, it looks like that when jitter is turned off in the light settings. Upon switching it on, look what happens, soft scattering. And still, I will leave it off for now because it won't be that visible once we reduce the anisotropy. Okay, that's honestly pretty darn fantastic. The sunbeams look so great. Now the container fog is much more versatile than the world fog. Why so? Because you can move the container and thus direct the fogs placement and fall off and all these things. Or maybe you want multiple layers. No problem. You can just duplicate these containers and move them independently or change any parameters like color. That's pretty cool. Maybe you won't need it. Like, I don't think I'll need it in this scene, but when it comes to three yard and, well, stock market, having options is good. This is it for level two of the elevator shaft environment lighting and rendering tutorial. As a quick recap, in level one, we established the base layer of lighting. The soft overcast window light. It being simple doesn't mean it's not effective. It has a solemn overcast vibe, and if we needed that sort of mood, that could have done it. I Level two, we added sunlight on top of it. We haven't touched on the doughnuts anomaly. It's just there. Every blended tutorial is better with doughnuts. Anyway, apparently. Yeah, we went to be deeper in Level two with the light source types, screen space settings, global illumination, and custom light fall off. The artistic tools for tailoring IV lighting to your needs. It still renders fast as hell. And next we're going to move to Level three of environment lighting in IV. 4. Elevator Shaft Advanced Scene Details: Welcome to Level three of the elevator shaft lighting and rendering tutorial. In the previous episodes, we built a real time lighting setup for this scene and Navy and honestly, it already looks pretty cool. Most of the core atmosphere is already here. You know, when we just started, the difference between this and this was pretty tremendous. Level one scored many points on our cinematic principles. Then Level two extended it, making it even more three dimensional, compositionally sound, moody and realistic, and it provided a bit of a vibe shift at the same time. Now, level three will be the mastering phase over the lighting scheme that we have already constructed, giving it a layer of professional treatment. Walking this last mile is always a very thankless work. The changes are sort of cosmetic and minimal. Working on tiny lighting details that nobody will ever notice unless you mess up, then everyone notices. And yeah, running the risk of ruining the work. Yet, the reward is the extra sauce that makes people go, Yeah, right, this lighting is so good. Actually, this is independent of the render engine. Alright, so what follows are some tips and tricks that go beyond that specific lighting setup. I don't want this to do this than this, that kind of thing. What I really want is for you to understand the principles behind it. Yeah, let's hand hold. In this one more about showing the lighting principles that actually make a difference. The first one, fake GI. Independent of the render engine, real light sources can be supported by the fake ones, channeling the artistic truth and just painting with light. So it works like this. You pick a light like this slash, for example, will switch off the fog for a moment. And now we want to augment it with the point light or any type of light for that matter. Its brightness and color shouldn't be wildly out of touch with that real light you want to supplement. So if it's orange, then this extra light should also be orange. And then you just literally paint with light, adjust its color, and try to make it believable. It can be any type of light. For instance, area. I will behave a little bit differently, yet it's just as good for imitating the directional light bounds, like the sunlight is bouncing off something. The size sets the softness of that bound. Then you can adjust the proportions of that area light and whatnot. The point is to augment that main light with the fake light. Tip two folders. Now, really, over time, lighting setup gets more and more elaborate. At first, it's not a problem, but once you get to the point where when you're adding details, once you get to level three, you will need some way to keep it organized. Otherwise, it will turn into a mess and folders or collections, if you wish, are a time proven way of keeping things tidy. One folder per light entity, the window lights, the sun, whatever. I actually checked the final version of the scene features, 32 lights. If not folders, it would be a total chaos, I tell you. Compound lights. In addition to the fake GI effects that we've covered before, we can do something more radical. Instead of using single lights, like a single light for each sun's streak, we can use multiple lights for each sun's streak, a hard light and a soft light, for instance. You can call it compound lights. It's an important technique to wrap your head around when it's time to add nuance to your lighting and lay the finishing strokes. So, for instance, the sun, it has a bright core. That's the sun lamp and blender, for instance. It may also have a wider glow around it. So these are technically two lights in one, the bright core and a more diffuse halo, making it a compound light source. I use this trick all the time. This is just how the sun lamp looks. Pretty cool, but see how complementing this harsh sunlight with a bigger softer light source will give it so much more richness. Pretty cool, right? It can even be moved around independently if you want more control over its shape. How do you turn a sun lamp into a compound light source? Just add a softer area light pointing roughly in the same direction. It should have much lower power, but here's the key. It should be roughly of the same color as the main source. Otherwise, the trick won't work. It should also be bigger and softer than the main light. There should be some difference in size between those elements to make it visually interesting. Otherwise, it will look like two separate sources and not the compound one, right? Now, suddenly the next step is going to be technical fixing the shadow pool issue. When you have many lights in your scene, you will eventually run into this problem. You'll open your scene and instead of proper lighting that you lovingly crafted, you'll see that most of the shadows are gone, like totally missing. And on top of that, you'll see this rather dreadful error saying shadow buf or full may result in missing shadows and lower performance and showing that you're of some budget. The way you can fix it is that you go to random properties performance, and there you'll see the shadow pool size, which is currently set to just 16 megs. And as it says, a bigger pool size allows for more shadows. So if you have enough GPU memory for that, you can just increase it to whatever higher value you can. Look how shadows are improving as we go higher. So basically, you can just go 1 gigabyte, and that will most likely cover it. This is sort of a nuclear option, though. You might still add more lights later on and surpass this memory budget. Then what? So let me switch it back to 256 megs and show you something else. Basically, we're way out of budget still. Damn it, the shadows look so glitchy. You know, the weird light leaks and stuff. Now, we're two times over budget for shadow maps. So what we can do is go to shadows in the end properties and decrease the global shadow resolution a bit. Like, Whoa, yeah, at 0.8, we just fit into the budget and the shadows are back to normal. Alright, let me switch it back to one for a moment. Actually, you can tweak the shadow resolution per light. It sounds like a lot of work, but it's not that bad. Basically every light, out of many we have has its own shadow resolution limit. By default it's set very low, meaning great quality. But if we crank it up a bit, see what happens. To. Boom, 300 megabytes shaved off. Now we can do it for every other light. Like, literally, for every other aerial light. We can just do it manually. And then for the secondary sun lights as well, Um, and then after you tweaked it for the individual lights, you can reduce the global resolution on top of it to just 0.9 this time. And this would fix this issue completely and bring back our shadow quality and performance. So that is how you fix this shadow pool memory issue in Ivy. This is a little TLDR for you. And where is my coffee? Another tip for adding a professional polish to any lighting scheme once the base has been set is to divide it into zones and treat each one as a separate scene if you wish. Taking a step back, that's something that we've already done to some extent in level one. We clearly separated the shaft and the dark brooding foreground where the character or the camera stands. These can be thought as separate stages within the main scene, and it is actually a very nice way of adding visual interest to any render. Actually, in Level two, we introduced yet another stage. Remember what it was? Sunlight. Part of the reason why it works so well is that it creates a whole new stage or zone, one can say, with a separate mini light in set. And then we kind of get yet another lighting zone here. The whole bottom part of the tunnel is also like a separate scene. It's way darker and all that. That's a really powerful concept to muster. It's level three for a reason, because that's not necessarily the kind of thing you think right of the start, although I imagine that professionals do. But you can definitely think of these zones at the later stages. Like, literally, you can think of it after the fact, after you have set up your lighting scheme and all that, and then you kind of start zoning it. So how does it look in practice? Okay, you just go in there, pick the part you want to zone, create a folder for it or collection. Then it could be something simple like a point light. You can even give it a different color, maybe something a bit more saturated. I don't know, a Sci fi blue or something that leans into that half life vibe. Maybe it could also flicker later on, you know, like a combined force field buzzing on this tech floor the player is standing on. Or maybe it's red. The point is to zone it to give that spot its own little pocket of lighting. It could be an emergency light, perhaps. Mm hmm. Just don't make it too bright or something, as it will probably break our setup. It still needs to sit naturally in the world. As for what lights to use and so on, feel free to treat it as an exercise. Maybe it's a bunch of area lights. That could definitely work. Or maybe something entirely different. I'm sure you got the points, experiment with it. I ended up adding these two can lights here pretty subtle intensity, just to kind of set this part apart. Now, as for other spots that could be zoned, I'll leave that up to you, as well. That's your chance to tell your own story. Maybe something's happening on this floor. Like, what's actually casting that con light there? Mmm. Or maybe it's a creepy red light. It's really a chance for some ambient storytelling, if you want to light up a certain area and give it its own little mini story. I'll delete this for now, though, and move on to the next tip. Practical lights. In cinema, practical lights are just practical. So those lights that actually exist in the shot, they're part of the world itself. It looks cool and it allows you to sneak in some colorful lighting where you normally couldn't. And it's a fun way to explore the world a bit more to think about what kind of lights would actually exist inside it. So in this location, I don't know. An industrial emergency lamp wouldn't feel out of place. You don't need to follow along with me here, but I animated the emission strength of this lamp using this little expression. If you're copying it, just remember to put a hash tag in front. Then the numbers control the strength and timing and all that. Even makes it super easy to preview it in real time, as well. But I digress. Practicals like these are great for your lighting refinement pass. What's really powerful about that is that it is a legitimate way to sneak in some really colorful lights if you need to, because if it exists in the world, you've got a perfect excuse to use any wild colour gel you want. Again, it mostly comes down to motivation. The tricks like these let you do crazy stuff composition wise without it feeling totally random. I mean, how else would you add a red light there without a practical? Hmm. And also, if you don't want to model any extra objects like this or anything, just add a hint of practical lights. For example, just a red omni over there as if there's some kind of emergency light nearby or just a tiny touch of red somewhere else. That's a nice little trick. Just don't make your practicals too bright, if it's supposed to be daylight. Well, by itself, I mean, if you remove the sun and all of that, these guys would look pretty bright, okay? Like proper emergency lights if it were nighttime. The practicals or just these omnolit simulating practicals would look really strong and powerful and Kudos to them. But because it's daylight, the stuff like that should be totally overpowered by the sun, so you should resist the temptation to make them strong and powerful, and that brings us to an important tip. Keep your light ratios and dynamic range believable. That might sound complicated, but it really isn't. When adding lights in general, but especially those little tiny details, it's tempting to make all of them super powerful. Like, Hey, I added this elevator light. Let's make it juicy in Hollywood. Sure, it looks cool when all lights are bright. There's that visceral response to it like kids get in a candy shop, and then it spirals. Maybe these lights can also be bright and punchy and so on and so on. As soon as you do that, you're committing a cardinal sin of breaking the light ratios. And once you do that, everything goes flat again. Remember our little checklist of what makes lighting and images look good. It becomes the opposite of that. So it's flat. I puzzles the eye. Maybe it's moody, but I don't think so. And most importantly, in this case, it breaks the realism. In the real world, lights have certain ratios attached to them. Let's say the sun is 100, just some abstract units, then the halo around it or maybe it's bounds will be 20. But then the skylight, ten, much weaker, and that's still more powerful than any artificial lights. Like elevator light will be five. It's practically nothing compared to sunlight. And then this light will be one, and then the room tone will be ten times weaker than that. I may very well be that the ratio between the room tone and the sunlight can easily be in thousands, probably way more than that. That's sometimes called dynamic range, and that is something level three explorers need to be aware of. It might not look as flashy right away, so it's very easy to fall into that trap, but this is the way. You can absolutely bend it to your artistic will, but in my experience, it's always good to keep the baseline realistic. This brings us to the next tip the glare. You might be thinking, What is it? So level three about a simple glare effect? You just open compositor, add a glare node, and the bright parts go bloom, right? Oh, and by the way, fog glow is slow. Bloom is way faster in Navy. So what's the big deal about glare? The thing about glare or bloom is that it works best and more realistic when your light ratios are set right. In other words, when you're scene actually has a wide range of intensities. Actually, you might not even realize how hot certain areas are until you add the glare effect. Like, take this sun right here. If you just render a frame, and then right click this bright part to sample its values, you'll see something crazy. Notice how the values in this part are way above one, and sure that's all relative, but here's why it matters. Watch this. If you go there and sample a darker value in the shadows, you'll get something like 0.001. So that's 1,000 times weaker. You can say here. Okay, this is almost black. No wonder. Fair enough, let's sample this gray area, 0.1, that's still like 50 times weaker light. Or this spot looks pretty bright, right? Also 50 times weaker than that sunny patch. Oh my God. Even this bit right next to the sunlight, 20 times weaker than than the sunny patch itself. And that's the point. The sun is really overpowering, and it will trigger a lot of glare, and it will look realistic. Like, Oh, now I see and subconsciously understand that some things were really bright, seeing how much glare they trigger. And that's the reason why it will look believable. We are good at noticing these contrasts because the real world has crazy contrasts. Then, of course, you can tweak how strong you want it to be the thresholds and all that stuff to include more areas into glare or reduce it depending on how stylized you want the look to be. And from here, you just fine tune it to taste. If that's not crazy enough for you, level three has more surprises. We didn't talk about this in Levels one or two of this EV environment lighting tutorial, 'cause that's a bit mind boggling, but now we can. It's level three. After all, it's supposed to be a bit mind boggling. The tonal ratios via color management. In addition to the physical light ratios, the power of your lamps and sunlight and all of that, there's another set of ratios, the tonal ratios, which are controlled from the color management tab. By tweaking these settings, you can totally reshape the relationship between shadows, mid tones, and highlights, which helps you emphasize certain lighting qualities that you have already set up. Let me unpack that for you. First of all, I made the render gray scale so we can clearly see the tonal relationships. Okay, so in the render tab, you have this color management panel. Every twig you make there has a huge impact on how your tunnel values are arranged. Exposure, for instance, not only it lets more or less light hit the virtual camera sensor, but it also changes how the tunnel ranges relate to each other. For example, with a higher exposure, the mid tones and highlights get really, really close together while shadows get pushed away. With lower exposure, the mid tones and shadows come closer while the highlights suddenly stand out like crazy. Do you see that? That's a totally different feel. Both versions are valid, but they emphasize different lighting qualities like that sunlight slash, it really pops in lower exposure. You can also tweak other things like the contrast levels, and that will also affect the tonal relationship, especially together with exposure. Follow the thought. Now we are basically left with just shadows and highlights, midtones, almost gone, and that sun patch becomes the main point of attraction. Now watch this, a lower contrast plus lower exposure will create a very different, very soothing balance, emphasizing not the sunlight but the room tone and all the ambient qualities of our lighting scheme. So this tab, my friends, is your secret tool for rearranging light values or even changing the entire look of your lighting scheme. And I'm not even talking about different color arrangements that come from different view transforms. This is way beyond the scope of this video. Tip number ten is light linking. Let's say you've added a few hero assets into your scene, and now you want to add a light that would affect just those assets to give them a little extra punch. Dish. Good news. It's super easy in IV. Yeah, but first of all, it's so awesome that lighting in IV is fully dynamic. I mean, the real time ray tracing. So when you drop new objects into a elite environment, they automatically get all the right direct and indirect light. And usually that looks great right out of the box, but maybe you want to go beyond realism and add that extra rim light just for these assets. Of course, you can just add new lights, normally, and illuminate these objects. But then it will spill everywhere, lighting up walls and other stuff you don't want. And that's pretty bad as it can ruin the mood. And such a spill can look really, really bad. So what you can do instead is create a light source and area light, for instance, outside the building. It doesn't matter where exactly. Just give it a lot of power, so it illuminates the whole thing like that. Now with that light selected, go to object properties and click New light Linking. That will set up the light linking group for that light. Now just drag the collection with your acts into that group. Swoosh. Pretty cool. Now this extra light should illuminate just these objects, but it still shines through the windows only, and I wanted to ignore the environment completely as if these walls don't exist. So I'll go back to this panel and create a new shadow linking group and again, drag the objects collection into it. And now this light will only take these objects into account when casting shadows, ignoring the rest of the scene, except the fog for some reason. So with this light still selected, I'll go to the light setting stub and kill the volume scatter influence there all the way to zero. Now this light should truly interact just with these objects, which is really useful because you can position it however you like. This way, you can add a really nice back light or some extra fill or whatever you want, really. And it will affect just these objects. That is so useful if you want to create some extra emphasis for it. Oh, and by the way, if your vi viewport starts shimmering with noise while playing animation, turn on temporal reprojection in the endo settings. That usually fixes it. Tip number 11, dust particles. A very, very level three thing to do is to add some floating dust modes. Like from our dust particles, Dondes acid, for example. The free version would do. Dust particles are the ultimate high frequency light in detail. You just need to tame their Bronian motion so it's not that crazy and make them much smaller, barely visible, just tiny modes, catching bits of light. And that's it. So that was level three, the polishing pass for the elevator shaft environment. Is that quiet invisible layer of care that takes us from looks nice to looks professional. Even though these changes are subtle, they teach you something crucial about lighting. It's not about the tools or even about IV. It's mainly about how you think about light. Okay, it's also about the tools and IV. And we've covered the main ones that will help you with environment lighting. If you've followed through all three levels, you've built a full cinematic lighting setup from the ground up. So congratulations. You've officially mastered the elevator shaft lighting setup. Now take what you've learned here and apply it anywhere because these principles will follow you far beyond this one seen. 5. HDRI Essentials: Be in nary let's start by comparing how HDRIs render in cycles versus IV straight out of the box in blender. This should come with no surprise, but cycles offers an extremely realistic simulation of the light from the 360 degrees HDRI panorama plugged into the environment slot of the background Sheda. You can really see how the light from the virtual backdrop wraps around the objects in a very realistic way. Switching over to IV by comparison, looks a bit underwhelming, if not straight out disappointing. Yeah, it runs in real time, but oh, boy, look at these harsh shadows and pretty flat and not super realistic lighting. What can we do to fix that? First, let's take a look at the environment settings, namely environment shadow. I will jump into the world tab, expand the options, and here we can find the sun threshold and the sun angle and the rest of the environment shadow controls. The Sun threshold tells IV how much of the bright areas in the HDRI should be extracted and treated as a sun lamp, since HDRIs in IV alone can reproduce that directional light alone. So using the threshold, we decide how much light from the strong sources within HDRI gets converted into the sunlight. For instance, I'll set it to ten here to get that main shadow mass falling in the camera's direction. This doesn't look very realistic straightaway because this shadow is raised or sharp, but we can fix it with the angle slider. As soon as we raise the value, the shadow and even the reflections become much softer. That's actually how HDRIs behave in cycles. They naturally give soft wrapping light, and that's one of the trademarks. So by pushing the envelope, we can get that same kind of softness. It's a little tricky to notice right now, though, since the ambient light is overpowering the shadow, but we'll take care of that in a moment. By rotating the HD Rye, you can see the shadows are definitely softer now, though still pretty faint. Alright, let's compare it to cycles. Oof. Not quite there yet. It's an improvement, but the result still feels a bit flat. In the world settings, there's also the jitter check box that would make it look slightly crisper, yet I'll leave this section untouched for now. The real secret to making the lighting look crisp in EVs rate tracing, it's settings. So next, I'll head over to the endo tab and expand this and enable the ate tracing checkbox. Right way things should look more plausible with the screen space, contact shadows and stuff. Nice. That is much better. Of course, that's as long as we don't put it side by side with cycles, which I'm about to do anyway, just to spoil the moment. Yeah, it's definitely closing now, but still not quite the same. Well, look at that. Disabling the fast GI approximation and denoising actually should bring the look much closer to what we expect from our beloved path tracing engine cycles. Well, obviously, that still doesn't look exactly like cycles, but it definitely gets into the same ballpark. Why did it work like that? Let's break it down. What does this have to do with us just nuking these two settings in IV? Alright, denoising is sort of self explanatory. When it's on, the IV's GI rays gets smoothed out and denoised using various methods that you can find in this tub. Spatial reuse, for instance, it kind of averages the neighbor's pixel rays at the cost of some accuracy. That kind of helps to fight the noise, but it also makes everything a bit smodgy like it's low rays or something. Now, add something called temporal accumulation on top, and it would make it even more smodgy like it has been smeared with vasolin. Without getting into the weeds, these settings greatly reduce the fidelity or precision of EV's global illumination. So they have to go if you want to make your render look more like cycles. Alright? Then what is fast GI approximation? Well, if you want a brief explanation, it's a way to make IV's GI look smoother at the cost of ray tracing precision. So it also has to go if we're trying to hit that realistic look. A longer explanation will sound like this. The way the raw rate tracing works in IV without any extra optimizations and with this checkbox turned off, I mean, is that IV just throws rays onto the scene using the screen tracing algorithm. This algorithm is pretty simple. It has just a few settings. It's just casting rays that hit objects and registers this noise that you see. You scroll to the top and set it samples to basically, if you give it just one sample, you'll literally see these rays and where they have hit it's literally these bright dots that's all there's to it pretty much. You throw more rays and the image gets cleaner. And when you give it enough rays, it eventually becomes really, really clean. It works a bit like cycles from that perspective, you know? So Ib casts these rays or samples, and there's also the resolution that affects the final number of rays. It's going to be one rape per pixel or one ray two pixels per four pixels, so on and so forth. And this is pretty much it. These are all the controls we've got over the quality of ray raising in IB in its raw form. Now comes the crucial point how well the raw ray tracing works depends on the materials, namely the roughness. If that sphere were made of solid metal and had a perfectly glossy surface, the roughness, in other words, then Ivy would have a very easy time tracing rays for that sphere, as they will just bounce off like they do on a mirror in a perfectly deterministic direction. The same for this floor. If it was made of a metallic material, which had zero roughness, then IV would have no issue tracing that single bounce, even just having one sample to work with or one ray per pixel. It would just know where the ray would go for every point of the surface. That's like the best case scenario for ray tracing. But as soon as IV has to deal with a higher roughness, it's no longer deterministic. The rough surfaces can bounce the rays in unpredictable directions, and some of these rays will hit and some rays will miss the camera. That's where the noise comes from. The same for the floor. Make it more rough and you'll just get more noise. That's how ray tracing works. Then, as we've mentioned, the blenders developers added denoising to help with that, but sometimes it's just not effective enough with this much noise. So the deep also added this fast GI approximation thing that is a completely different rendering technique meant specifically for high roughness materials. It has a threshold that you can control, and you can think of it as a mix factor between the raw ray tracing pipeline and this optimized technique. The biggest drawback of fast GI is that it is pretty imprecise. Kind of averages the light values and can't really reproduce sharper reflections and shading in crevices, that sort of realistic stuff. It's an approximation, really. This slider is actually pretty smart, though, as it allows us to mix the fast Gi and the real ray tracing based on materials roughness. So, for instance, if some materials were really shiny and metallic like this sphere and some other highly diffuse like this floor, then this threshold will make it possible to use real ray tracing for the metallic sphere and the fast method for the diffuse floor material. So that is pretty clever and sometimes it's pretty useful. Unfortunately, though, when working with HDRIs, it just over optimizes it a bit too much, making it stray away from the cycles ground truth a bit too far. That is easy to prove. And actually, without it, it looks much, much closer to cycles. So that's the entire reason we have switched these guys off. Sure, it makes Iv a little grainy, but honestly, that graininess feels pretty path tracy, too. First of all, let's get rid of that graininess. All we need to do is raise the viewport samples to something like 32 or 64 or even higher, like 256. In fact, if you set it to zero, it means infinite or automatic sampling. That clears things up nicely, but I also noticed an odd little outline. I wonder if you can spot it on your end. It's that thin, almost a pixel white bright edge. The culprit here is the resolution parameter in the ray tracing tab. It is caused by not enough resolution. I mean, this edge, and it gets a lot worse if you drop it down to one eighth of the full rays. Now, you could hide it with the fast GI approximation and denoising or even just denoising, but that's not really the right fix. The straightforward solution would be to simply set the resolution to one to one. And with that, it should look just perfect. Actually, with these settings, it doesn't look too bad, even with denoising enabled. Mm hmm. Especially if temporal accumulation is off. Since that tends to reduce both brightness and amount of detail, spatial reuse should be fine. Still, I'm going to switch it off for now. Now the screen tracing parameters become pretty important, especially since we're using pure ray tracing without any extra optimizations to hide imperfections. Take precision, for example. It has a noticeable effect now on the contact shadows. I mean, those tiny shadows where objects are close together. These contact shadows are, in fact, part of what makes cycles renders feel so grounded. And in EV adjusting precision lets us fine tune that look. The thickness parameter, let us fine tune the contact shadows even further. If it's near zero, you get almost now shadows, and it ends up looking pretty weird. Then thickening it out leads to bigger, thicker shadow in effect. It's a really interesting slider to experiment with. I think something like 0.3 should make things feel more grounded. In some cases, you might even push it higher like 0.7 or something. Why so? Look at that. See how that keeps the screen space shadows visible even from different angles. You know, because the lower values like 0.2, the contact shadows can disappear once the intersection between the objects moves out of the view. I could look awkward. That's why you might want to adjust the thickness. It just helps the shadows persist a bit longer. It's just one of those real time screen space trade offs we have to balance. Anyway, this is looking pretty good, and now it's pretty damn close to what cycles is producing. Honestly, that's already pretty mind blowing. It's impressive to see that level of realism in real time in blender. At this point, it's just looking great. If you want to stop here, you've already got the basics of the technique down, but we've got more to cover, so feel free to grab a quick coffee, and we'll continue to make it look even better. Now that we've nailed the overall look, there are a few more tweaks we can make to reduce graininess and fin tune the look of the scene. The fast GI approximation is one of them. We can turn it back on, which will undo the realistic look a bit, anyway. Remember, there's the threshold that we can set much higher to reveal more of that pure ray tracing effect while still reaping some benefits of that fast GI approximation smoothing. It could help to mellow things out and make the lighting feel a bit smoother, even though it will also wash out the shadows, but honestly, that can still look pretty nice. So this is one of those sliders you can experiment with to get the exact look you want. Next, the GI resolution doesn't have much effect. Since we are not relying heavily on this feature, the steps though can make a big difference. It can either make it much crispier. Or on the other hand wash it out completely at the lower values. See how the conduct shadows disappeared. So if the checkbox is enabled, it's usually a good idea to increase the steps from the default eight all the way to 24 or more. In my experience, 24 or 32 tends to be a sweet spot. It affects the performance though, so it's good to be balanced. Precision doesn't have much effect here in this tub. Distance can be used to guide the spread or the fall off. For now, I'm just going to leave it at the default zero. Thickness won't do much here, and the bias can be useful for tightening up the shadow a bit in theory, at least. So now I'll leave it at the default value since I think it already looks good. Just as a quick reminder, this is how it looked in cycles. And as you can see, it's pretty similar to what we've set up in IV. It's up to you to leave that fast GI thing on for a slightly smoother or arguably more polished effect or go for a more aggressive unfiltered rate raising effect. From here, we can revisit shadows to see what can be cranked up to 11 the render properties, shadows. There's a global parameter of steps there, making it higher makes the shadows a bit more defined, which is usually a good thing. From here, I'll jump to the world settings and enable the shadow jitter there. That should make the contact shadows even tighter in the render. And if we tick that checkbox, then in the viewport as well, although it would also cause this flickering effect when updating the view or moving the objects with a key or anything of that sort, this jittering effect can be a bit annoying to some people. Actually, now the trailing effect, too. That's also low key annoying. You can fix it by turning temporary reprojection off that will add some noise and grain, but at least the trails will be gone. So maybe now that would work better for you. So it's really pick your poison situation, choose the lesser of two evils. Now, ambient occlusion is one more card we haven't played yet. That is, in addition to jittered shadows and other such enhancements. So to activate it, I'll go to the view layer stub, check the ambient clusion pass there, and then turn on the real time compositor right away. Then we can just go to Compositor click CE Nodes. You'll now see the AO or ambient occlusion pass available there among the render layers, the one that we've just activated. From here, we can add the color mix, mix node. And plug AO into the second socket. Then we can blend between the beauty pass and the AO pass. That's how it looks at full mix, by the way. The key and really the only parameter to tweak here is the occlusion distance. It controls how far the self shadowing spreads. The value of 0.2 is a good starting point. The follow off can be somewhat adjusted with the help of the trusty and totally overused color ramp node. Coffee. We can squash the contrast levels like that, for instance, to tighten the transition if needed. Now we just need to change the blending mode to multiply. And Valla, you've got that extra self shadowing all still in real time. Again, the distance can be fine tuned here. It's better not to push this effect too much as it can make your three D scene look a bit dirty, you know? A resolved approach works best if you're going to use it. Something around 0.4 usually works good. Yeah. In some scenes, this could be the finishing touch that really helps to ground things if used in moderation and not in extremo, you know? It's. Anyway, that's the core of this technique. After following these steps, your HDRI setups should look just as good or almost as good as in cycles. Actually, if you don't like the noise in animation, you can bring back temporal reprojection to make it a bit smoother. I wonder if you agree that looks pretty damn realistic. Okay, as an option, you can also enable denoising for extra smoothness. That said, the temporal accumulation adds its own blotchiness, which I don't particularly like. Also, feel free to fine tune the fast GI approximation threshold if you want a cleaner or more stylized look if that's something that you want. And that is the whole technique. I hope you have a blast experimenting with GRI lighting in IV without compromising realism of that type of lighting and rendering. As a post scryptum cycles HGRI renderings, you may notice this double shadow or more accurately multiple shadow thing sometimes with subtle color variations. That's part of the signature look of the real world lighting data. And this aspect is missing in IV, but you can emulate it by supplementing the HGRI with a few smaller area lights placed at different angles like that. This helps to recreate that intersecting shadow's effect, and these lights can also be given some warm versus cold color variation. That should make it resemble that look. The difference will be subtle, but it should register as a bit more photorealistic. Feel free to download this scene from the project files. I can't wait to see realistic HGRI renders made in IV, renders that look just as convincing and strikingly photo real as those from Cycles or Octane or any other physically accurate render engine. 6. Table Closeup Setting Up Product Lighting: First of all, we'll start with the cycles reference. Okay, take a mental snapshot. Here's what we're aiming for. Realistic GRI environment lighting, and we are going to achieve this level of quality in IV. The way I prefer working with cinematic shots like this is first of all, by turning on Passport two in the camera settings and setting it to maximum. That creates a picture frame. So now that is our canvas. It's too bad O Canvas right now is extremely flat and uninspiring, but good lighting will give it depth and dimension. And the first thing to check in IV to set it on this path is the right tracing settings. Right now, the scene is lit by the default wal background only, which has this solid gray color. It doesn't look super exciting, right? To make it a lot prettier, you can go to the render settings and turn on ray tracing that will already look better and definitely more grounded due to the extra shadowing underneath the objects and in the crevices and where the surfaces touch each other. It's also called ambient occlusion. Now let's set up an actual HGRI environment. It's usually done via the Shader Editors Worlds dab. By default, it's just a solid gray world light light in this scene. It's like a complete overcast sky. And not that it's inherently bad. The solid color world lighting can work really well for certain scenes, mainly carried by the red racing, generating all the indirect shadow information, which could make it look really good. Yet in this scene, I want this to be an HDR image. So I'm going to press Control T to add the texture nodes. And now we just need to load in an HGRI texture. I have one from Plyhaven which I downloaded a few years ago, and I'm just too lazy to download it again. So it's an HDR format and not EXR, I mean, which actually makes no difference here. Open image. So here we go. Now our scene should be lit by this 360 degrees spherical panorama. Right, from the start, though, it still looks a bit flat. So let's work on it. To do to environment shadow. If you rotate this edge DRI in the mapping node, you'll notice it actually casts a really sharp shadow. What is going on? If you check the world setting, you'll see it's the sun shadow. Basically an IV optimization, converting the brightest parts of the edge DI into a sun lamp, casting direct shadows. Again, what happens is that this feature analyzes the gGI image for brighter spots like these windows and converts them into sunlight. The threshold defines the contrast of that light. And the angle sets the softness. So a higher angle can be used to soften up this light. It's all nice and fancy, but by default, the softer shadows in Ivy aren't too precise. If you zoom in, you'll notice the lack of proper shading or shadowing, especially where the objects meet the floor. If you want it to look more photo real and grounded, there's a nuclear option called jitter, which also needs to be activated in the global settings as well if you wanted to work in the viewpod. It'll be much slower to render, and it causes this annoying flicker when moving the camera or just navigating in the viewpoard. I know for a fact that this flicker annoys some people. But it makes the shadows cast by the edge of the ye so much more solid and of a much higher precision. With Jitter, it should be possible to push the angle much higher to 50 and above and still retain high quality shadows even at that amount of softness. And that actually brings us way closer to the cycles look already. If it hurts the performance too much, it can be turned off for the viewpoart. But the jitter shadows will still show up in the final render. You'll still get them. So that could be a workable compromise if your viewpoard gets sluggish and all that. Next, light blocking. In cycles, objects placed behind the camera can actually aclude the incoming light. It makes it possible to cut the front light spill, that is making these render so flat and maybe cut the top spill as well to make it look more three dimensional. People also call it cinematic. But none of that works in the IV because IV uses screen space ray tracing, meaning it only sees and calculates lighting for the stuff that is in the camera frustum. So there is no behind the camera concept for IV when it comes to real time global illumination yet. So the way to block some part of the light in IV is to actually come to the world shading and manipulate the A. First of all, I'll change the strength from five to a lower number because it was a little bit too bright. And next, I'll add the mixed note from the Shift A menu, and we'll give it a purple debug kind of color. So we can clearly see what's going on. So currently, I'm just blending this new color with our HGRI. I'm going to add the gradient texture as well and plug it into the mix factor. I want to color ramp this gradient to fine tune its contrast or to be able to fine tune it. But actually, where is purple now? The purple now appears on that side of the scene. Basically, I'll press in and change the focal length of the viewboard so we can see the entire thing. So basically, the entire rear side of the world has turned purple due to that gradient. And this color ramp is how we can fine tune the transition. Okay. I'll press Control T now to add the mapping nodes. And basically, what I'm going to do now is set the X scale to a negative value to move this gradient to the other side to this side. So now it's here. Um, and watch this. I'm going to flip the color ramp. And tweak it so it actually becomes a window. See that? Now it's a window. Now it's just a matter of replacing that purple with black. Now it should completely block the frontal light from that gDI because the black part of the gGRI won't emit any light. I'll press M to toggle before and after so you can see what's going on here. And here is how it looks in comparison. Combined with ditter shadows, it should create a very nice three dimensional look with shadows falling in our direction, you know, on all objects. That's like blocking in a nutshell. If that's too much contrast, then the threshold can be set a bit higher. TLDR, blocking the frontal spill coming from an HDRI image is a way of making it more three dimensional, more cinematic in a way. So that's a really cool technique for that. And still, when zoomed in, the lack of contact shadows is a bit of a bummer. Contact shadows help anchor the objects and are one of the hallmarks of photorealism. To make it better in the rate racing tab, we can turn off these two defaults, denoising and fast GI, and it will actually improve that. For an in depth look at the settings, check the HGI essentials video from this course. In which we walk through these parameters in detail, based on a simpler example to make it easier to digest. All right, precision, all the way to one. Thickness also to one. Higher precision and thickness usually correspond more closely to the cycles ground truth if we take cycles as a reference. The resolution? Mm, let me show you something. I'll switch over to the diffuse light mode for a moment. So resolution is responsible for the indirect light bounces quality. Basically, anything that isn't the direct light isn't the sun, isn't area light and so on. So you can crank it up if you want to have a better quality light bounces, and it should also positively affect contact shadows. So I recommend to max it out if your GPU allows it. Okay, let me go back to the combined view in the viewpard. Okay, so maximum resolution. And these are all the settings at a glance. For now, anyway, it's fairly minimal, as you can see, if you know where to look. And guess what? At this point, it should be looking so close to cycles. You know, it might be even challenging to spot the difference. Maybe if you zoom in and check for contact shadows, cycles will still be a bit more convincing. But at this point, it does take some effort to spot the difference, right? It's getting pretty easy to mistake one for the other. If the occluded parts end up too dark, try fast GI approximation, but then increase the steps to at least 24. The rest of it can be default, as it doesn't matter here, except the threshold. At one, it will mean it's just the pure screen tracing again as if no fast GI was used. So for the best results, I recommend setting it to about 0.9 or something, or even a bit higher than that. So there's a tiny bit of fast GI mixed in, so the contact shatters are still there, but there's a tiny little bit of extra fill. That's how you can add that extra bit of fill, right? Then fog. It almost never hurts to add some fog. In, it's a breeze, just adding a cube, scale it to cover the contents. Then I always make it display as bounds or wireframe in the viewport. Then it's a matter of creating a new material and giving it pretty much any volume shader in the shader Editor The principled volume, for instance, which should go into the volume input of the material. Be careful with that. It doesn't have to be super dense or anything like that. Something like 0.01 should be fine. That I often go for higher anisotropy right away. It is always hard to explain what it does. But basically, higher anisotropy pushes the fogginess towards the light source to create that gradient. Here I made the fog much denser so you can actually see the effect. But then I'll go for an even lower density because I want it to be a really subtle effect. Yeah. The fact that we can pull off amazing volume lights in real time doesn't mean we need to overuse it. Oh, we can, if you wish. It's really your choice. Who, frankly, Coffee. Extra key light. HDRIs can be augmented by regular lights Aerial lights, for instance. I'll add one, make it nice and bright. And basically, I'll try to line it up with where I think the windows are on that HDI, roughly speaking. What I want to do is emphasize that main light direction a bit. The custom distance should cover the full length of the scene. Okay? And if you want the maximum quality, Jeter should be turned on. I think it can be brighter as well. Just a bit of extra punch. Yeah, it looks great. Next, depth. The cycles reference render actually included a shallow depth of field effect and a crazy vintage lens. But anyway, here's what we can do. First of all, turn on depth of field. A lower F stop makes depth of field more raiser thin and a high ratio, add an anamorphic squeeze to it, and I'll go for it. There's a comparison between the default anamorphic ratio and a heavily squeezed one. And just for comparison, 0.2 should stretch it in other way. That is as close to the vintage lens look as we can get in IV with the help of this ratio. But anyway, that looks really convincing to me, especially if you throw in some film grain. Alright, now I'll just focus on this horn and open the aperture a bit wider for a more shallow depth of field. Okay, this is it. Oh, before I forget, it won't show in the viewport, but the jitter camera setting in the depth of field tab simulates a much more physically accurate depth of field. Look at that off and on. Off, again, and on. It's simply a more photorealistic depth of field. It has better transitions between in and out of focus elements and even some extra noise and much better preservation of detail. Just look at that. The quality jump is especially evident in animations with shifting focus. I don't know if you can spot the jarring border between the in focus and out of focus elements, but it's really bad. Now, compare that to a jitter depth of field. The transition is much smoother now. You can't really tell where the border is anymore. It flows seamlessly just like in real lenses. Needless to say cycles and octane and other path tracing render engines are still kings when it comes to life like HDRI lighting and rendering. It's really hard to beat that. Yet with the right workflow that we cove it in this video, if we can look sufficiently close, I'd say incredibly close to that. 7. Table Closeup Final Animation : Here is a full animation, by the way, that I was able to render pretty quickly. 8. Light Leaks: One of the most effective tools in the lighting artist's arsenal is the ability to direct a light beam exactly where you want it and let that beautiful bounced lighting handle the rest. It's often so effective and looks so good that it can actually carry entire scenes on its own, injecting a mysterious mood and a specific aesthetic. Seriously, this simple trick is regularly practiced by cinematographers to make the mundane interesting. Let's actually start with some negatives. It's a shame that one of the best tools that we have in the blender for focusing the light exactly where we want it is actually only found in cycles. It's called the beam spread. It allows you to focus a beam of light to the point of making it laser focused and laser sharp, if you wish. It is very convenient, and in cycles, this makes it a breeze to shape the light and move it around like a tiny window of illumination. It probably sounds funny like it's such a basic feature. And yeah, we don't have it in IV yet. But wry not? In this video, I'll show you how to achieve the exact same results using only the built in IV tools. From the lighting theory standpoint, what do all these renders have in common? These are some cycles and IV renders mixed. What do they have in common? See these light leaks. These pools of highly directed light is what makes it look good, this contrast of light versus dark. And that trick has been well known to photographers and cinematographers for ages, if not centuries, you know? This is what Patrick O'Sullivan from wondering DP YouTube channel calls a little bit of light and a lot of darkness. You might have also heard the term Kiarascuro, which is an Italian term that literally means light dark. Okay, let's do some Kiaroscuro or light leaks in Ivy. Imagine you have a model like this, nice little hard surface snippet and you want to turn it into a cinematic scene or just showcase it in a cinematic way. How will you go about it? This is, by the way, just a blockout. I created it by generating depth from a photo. It's not even a prop hot surface mesh, but the lighting we're gonna use will make it look good, believe it or not, because adding a strong light leak and submerging the rest into darkness is one of the fastest ways to make anything look cinematic. The easiest way to go really directional in IV is to use the sun lamp with a piece of cardboard with a hole in front of it, casting that little rectangle of light or any other shape we want to project. So you'll need a sun lamp. Shift A, light, sun. By default, it doesn't matter where it's placed. It only matters from where it shines. It's very directional, so it's great, and it produces nice crispy shadows by default. This is also something that is going to be useful for us. In fact, Sun lamp casts better shadows than other types of lights in IV. Wait, what? Why does the Sun lamp cast better shadows? What are you talking about? Did you know that IV's Sun lamp produces better quality shadow maps over long distances than other types of lights? Shadow maps cast by the Sun lamp are nice and have clean outlines while the shadows from a spotlight positioned at 38 meters away from the object are super jagged. To fix that, you can apply some filtering in the spotlight settings. And that will make it look a bit cleaner at some cost to precision. The spotlight at 85 meters away will produce shadow maps that are even worse than that, though. Filtering this mass will somewhat help to hide that, but this is just the limitation of IV's cascaded shadow maps. The further away is the light source, if it's not a sun lamp. The lower resolution of the shadow maps you get. That is the way IV works currently. The solution to fix that jaggedness, either turn on the shadow jitter and that will make it look much better. Or use the sun lamp for the light sources that are supposed to be far away. This way, you'll get the nicest, cleanest looking shadow maps compared to all other light sources. Alright, so we've added the sun lamp and it's nice and bright and it has great shadows, but it floods the entire scene. And we want to direct our light like that into this nice little stripe. How do you do that? Now you need a blocker. It can be a simple plane. I'll control C to copy the location and rotation from the sun to line it up like this, as to scale. So for a start, I'll make sure this object covers the entire scene. Mm hmm. It also blocks the camera view. That's unfortunate, but that's okay for now. As I'll immediately jump to the object tub and switch off the camera visibility so it doesn't block the view. And I will switch its viewport method to bound, so it doesn't block the viewport, as well. So this is our blocker. A noun it's time to punch a hole in it. No, seriously. The easiest way to do that is to press tab to enter the edit mode, eye for inset, and delete the central polygon. And look at that. There is some light. Now, I'll press dot to switch to the median point so I can scale it from the middle of that selection. I'll hit S to scale, then shift that twice that will allow me to scale it like this so it stays on the surface of the plane. And so you can change the size and the shape of that hole to direct the light. There are some useful shortcuts seen on screen. You can slide it around like that. You can rotate it. And most importantly, you can reshape it. Now I want to make it into this kind of a stripe that slashes right across the form of that door or hatch or whatever it is. To make it a bit easier to move this setup around, I'm going to parent it to the sun by a control p so we can just move the sun and the blocker will stick to it. After that, you just set the settings of that light. Usually, you'd want it to be pretty darn strong. So you increase the power like that. And usually you want it to be sharp. But if you make it soft by adjusting the angle, then don't forget to turn on Jitter here and in the global settings and crank up the viewpoint samples. Otherwise, it will look not good. In EV, jittering makes softer shadows much higher quality and precision closer to how cycles would render it at the cost of some annoying flickering effect in the viewport, where you update the lights or move the camera and also slowing down both the viewport and the render. But if you're going to use soft shadows anyway and want it to look slightly more precise, then use Jitter. For render efficiency, though, and also for aesthetic reasons, I'll keep the jittering off now, both here and in the light settings. And I'll keep the angle low. I want this light to be really sharp anyway. Next shadow linking. I'm sure some of you will stumble across this issue, say, you want to copy this setup to have two strips of light. But once you do that like that, you'll notice that something is wrong. Basically, these objects are blocking each other, and it's a mess. So we have two blockers. How to exclude them from casting shadows for other lights. Select the first sun lamp, go to object properties, and under shadow link and click New and drag that second light blocker into this field and uncheck this box to exclude it. Do the same for the second light and drag the first blocker there, and again, unchecked to exclude. And now it should all work fine. Now these blockers should only affect those lights to which they are attached to. Now you can play with as many light leaks as you want. Think of it like painting with light. These are your brush strokes, and the object is the canvas. That is a bit of a cliche, but here we are literally painting with light. I also love using three Dcursor for rotating these lights around. So shift right click to position the cursor, then set it as a pivot point here, and then double tap R to rotate in a trackball style. And so this is a really fun way to come up with different shapes for these lights. Well, you can also make certain shapes softer by playing with the angle parameter in the light source settings if that's something that you want to try. But again, I will stick to the party line, and I will keep it sharp. Now, instead of cutting these shapes by deleting polygons, it's possible to use textures to influence the transparency of the planes, using the transparent shader. Yet, we won't go there in this video, as I want to show how much you can get just by using the simple rectangular shapes. But on screen, you can see the setup for doing it with textures if that's something that you want to try. It's really simple. It's just a black and white mask that goes into the transparent BSDF and this is it. For this tutorial, though, I want to keep it nice and sharp and geometric, so I'm going to stick with a simple approach. Anyway, simple light shapes can look really strong compositionally. Those angular, bright, solid shapes draw attention. It's like a magnet for the viewer's eye, but also in a peculiar way, it tells its own story. It's almost like this shapes suggest there must be a very specific window or windows right behind the camera from which this light is shining supposedly. This is a frame from Romulus, obviously. So even without showing a different angle where we could actually see that window, the viewer will be able to infer some information about it about that part of the ship or space station or whatever, just from that light. So there's some storytelling about choosing the light shapes. Alright, so the core of this technique or light style is made of two parts, the direct light or that bright shape and the beautiful indirect light that bounces off that shape. That is the soul of this technique, bounced lighting. The gist of it is simple. Turn on this checkbox that says rate racing, and that should do it. Thank God, if it has a built in screen space rate racing support or screen tracing. So that should look pretty good right away. I'll zero out the background to make it a bit more Kiara scuro. And just look at that. The nerd in me rejoices seeing GI simulated in real time. I still remember radiosity in three days Max times and so on. Again, it's pretty fun to orbit the lights around the three Dcursor that's such a blended thing, and I love it. Now, if you want to really see the pure screen tracing, you can turn off the other GI optimizations, denoising and Fas GI. Let me reiterate a few key things about how ray tracing and indirect lighting works in IV. And why do I want to switch to its raw mode? In IV, by default, denoising and fa Gy are switched on. And in most cases, that provides a noise free and pretty decent indirect illumination pass or these light bounces that irradiate from the brightly lit areas in all directions. Use in these settings, these are noiseless but not very precise. It's like a cheap version of the original indirect rays that IVt traced. And if you want to see the raw layer of this rays that IVt raced, you can do it by switching off these two settings, Fas Gia and denoising. Then you literally see those rays that also register as noise at lower sample count. So this is like an unbiased version of IV, and I'm putting it in quotation marks because it's not really unbiased, but it's as physically correct as you can get in IV. It's the closest thing to cycles that you can get in blender, aside from cycles. And the cost of it is the noise. This is noisy because IV has to calculate thousands of rays or samples for indirect lighting. If you change the viewport samples to one, you'll see that it would get much worse than that. Even though the GI will still be kind of truthful, there aren't enough samples, so we can appreciate it. And cranking the sample count up is the only way to fill the picture. 16 samples is the default and it's not enough. Setting up to 64 kind of works. But realistically, if you were dead set on getting the highest quality global illumination and IV you'd have to set this to a really high number. 256 or so, and naturally that won't be super optimal, either. So I'll set to 64 for now, and later we will explore some solutions to optimize it a bit like turning fast Ja back on. But as a starting point, I always want to see the raw screen tracing. Then I usually set the precision really high, just maxing it out and cranking up the thickness to 1 meter here. In this particular scene, it doesn't make a tremendous difference. So let me reiterate that part from other videos. Higher precision makes indirect lighting slightly more detailed, and thickness makes the AO and bounced lighting kind of spread over bigger distances, which often looks more realistic depending on the scene. Here's the thickness comparison from another scene. So here, the high thickness makes the indirect rays fill more space and bounce a bit more. High thickness almost looks here like a second bounce, which can't be true because if we can only simulate one bounce, but it looks like that. Now, this raw ray tracing is nice and fancy and unbiased, but we need to decide if actually going from 64 to a higher number of samples makes sense in this scene. After all, it will not only increase the quality of the image, but it will also increase the rando time. So is it worth brute forcing it like that? This scene is fairly minimalistic, and it will render pretty fast anyway, so I'm not too concerned about performance, and I will set 256. Then to double down on that, I will max out the resolution, as well. I want the bounced lighting to be of the highest quality possible. So I will just set it to one to one. It's worth noting that the higher resolution will also help to reduce the noise. Think of it as a multiplier for the number of samples. So cranking it up should help with both the quality and the noise. And now it should look pretty gorgeous. Man, I love this effect. If you also happen to like this effect and want to make it stronger, you can just pump up the light strength. That's the most straightforward way. Or actually the strength of indirect light rays in pretty much any rain ranging depends on the materials albedo, or simply put how bright its color is. And here's where it gets really interesting. There's no default slide of tweaking the indirect light and intensity in IV, at least not yet unless you're watching this from the future. But you can make the indirect bound stronger by making the materials lighter. Take this comparison, for example, the direct light here has exactly the same power in both shots in absolute terms, I mean, because it's basically albedo multiplied by light power, yet the image on the right has so much more bounced light, simply because the surface color is lighter. White versus dark gray. So yeah, altering material to albedo is actually one of the easiest and most natural ways to boost bounced lighting in Navy. Next, let's take a look at some more racing options. For instance, FAS GI and denoising that we have disabled earlier. What I want to do is smooth out our indirect lighting a bit. So denoising is actually not an option because it will just break it by making it way too dark, basically treating all the bounced illumination as noise. Without temporal accumulation, it will be slightly better, but still a bit splotchy. So it's pickier poison situation, so I'll just disable it altogether. The fast GI approximation is more interesting, though. What it does is that it switches EV to the faster and slightly less precise pipeline for global illumination rays, which at the same time looks a bit smoother. I think our image can use a bit of smoothness. Right away, I will set the resolution to the highest possible level, which is one to one, and we'll max out the precision. Next, the steps is an important one. If you set it to one, you'll notice that GI becomes super weird looking and super weak. So steps is like the quality, but also the intensity of the light bounces. With more steps, the bounces start to spread out beautifully, getting both smoother and stronger. I can describe it like that. So cranking it up to 24 or even higher could make things look great. But also, it comes at a cost at a performance cost. So these settings in the fast I approximation tab can make IV really slow. You can see the irony. It's called fast GI, but it can make EVA really slow. So it's all about balance. I'll leave the rest of the settings in the Fast GI tab at defaults. Since this is a pretty lightweight scene, I'll go pretty extreme with the steps, all the way to 64, actually, which is the highest quality you can get. There are some diminishing returns at that point, but hey, I'll take it. And lastly, I'll raise the threshold a bit to let in more true screen tracing as opposed to the fast GI approximation. Without going into the weeds, bringing this slide higher basically reduces the strength of the fast pipeline, letting in more raw pipeline. The values around 0.8 or 0.85 usually have a very nice balance. Here you can see a comparison between the EV default and the high quality settings when it comes to ray tracing. I know it can be pretty hard to pick up the difference. Yet by looking at these images, basically, the difference reveals itself mostly in the areas that are just indirect illumination. The high quality settings have a slight edge in terms of realism and also make the bounce look a bit fuller and stronger like it has better chances to bounce around. So look at these images to see if that makes any difference to you. It certainly does for me, but I have a professional deformation. This is it for the render settings. Now let's talk about something else. I want to show you an alternative method for shaping that hole. The light shines through. A pretty easy way to do it is to use the boolean operations. Select the cube, then shift click the plane, and then press Control minus on the numpad to cut that hole using boolean operations. And then you can adjust the shape of that object and move it around and do whatever you want with it. You can shape it in any way you like. You can rotate it and even animate it. It's a non destructive way of cutting that hole in the blocker object that also gives you the option to animate it. That concludes the first part of this video. So even though one of the best tools that we have in blender for focusing light, the beam spread for Aera lights exists only in cycles with a bit of technical know how we were able to recreate it beautifully in Ivy. For that, we use the sun lamp and the block your object, and it looks fantastic. The recipe for this dramatic mood is simple. Blast a certain part of your scene with light and let bounced lighting do the rest. The beauty of learning lighting techniques like this is that, A, you can use them for the rest of your creative career, refine them, make them your own. And, B, you can apply them anywhere in different scenes, different range engines, even outside of three D, and they will still look amazing. Some scenes naturally lend themselves to this. They already have windows or openings that can direct light leaks just where you want them. This one, for example, actually, it was rendered in 2020 using the old SSGI Adon fa blender, and it still looked great. In the next part of evilighting and rendering, we'll upgrade this setup by adding light shafts. 9. Advanced Bias & Resolution Fixes: Alright, guys, welcome to the second part of the light leak tutorial. Previously, we've talked about the basics of the QR squur directional lighting technique, and it's two main pillars a strong leak of flight and a gorgeous diffuse bounce. And yet there's a third thing, something that let's be honest, people just love. I love it myself. Even though it can get a bit cliche if you go too far with it, it is a staple in Hollywood, a staple in games. I'm talking about light shafts. Let's start this one with a super tip. How do you create a light lake in a room with no windows in blend? Hmm. Use a sun lamp with a go boo to shape the light beam, then go to object properties, shading tab and enable shadow inking for that light. After that, just drag your room object from the outliner into the shadow Lincoln list and uncheck this checkbox to exclude it from casting shadows. Using this simple method, you can throw a beautiful beam of sunlight from any direction into any environment you want, practicing the artistic freedom you deserve. You can even make it volumetric. So in this video, we're going to be talking about volumes and volume lighting. Direct light is perfect for creating those beautiful rays of light slicing through the air once you add some fog or haze to your scene, but it also makes the shape of the light visible. And that's why the type of lights that we use for creating light leaks matters. See this? If your beam is conical, it doesn't look like sunlight, as the sun rays should be parallel coming from a practically infinite distance. Actually, that's one of the reasons why cycles beam spread in aerial lights is so useful. It lets you straighten those rays, making them behave more like real sunlight. Okay, so back to our Kiara Scura technique. In addition to a strong leak of light and a gorgeous indirect illumination, we can add a third element to enhance this combo volumetric haze. It doesn't have to be over the top. It can be a pretty subtle effect, but you can very well go full load of the rings lighting department with it. Like, you know, in this balance tomb scene, you can do that if you want. Okay, so here we've got an ancient temple environment, and it looks extremely flat under the ambient light conditions. By the way, feel free to download this project file. Once we add the light leak and then global illumination via ray racing and then a bit of fog or haze to show those light rays dancing in the dust, that combo instantly makes the scene come alive. The more direct and sharp light source are, the sharper light shafts will look naturally. Sometimes it's called good rays or crepuscular rays when light passing through a volumetric medium becomes visible, rendering that effect is one of IV's superpowers. We cycle struggles a lot with clearing it all up, IV just excels. And So back to this scene, I'll start from the blank screen. No lights at all. Then I'll add a spotlight with a blocker object with a square hole in it to blast light through it right onto the temple, similar to how we did it in the first part of this chapter. Yeah, to double down on that, the only difference is that in the Part one, we used the sun lamp with a plane with a hole in front of it. And here I used the spotlight with a plane with a hole in front of it. Absolutely the same purpose of focusing the light beam into that tiny little rectangle of light. If you want to see the step by step explanation of this light blocking technique, check the first part. Okay? So the spotlight this time. Wait, wait to white spotlight. Well, yes, as a quick reminder, the sun lamp in EV produces the highest quality cascaded shadow maps over long distances. For Adults, spotlight, including the further away you place them, the worse the shadow map resolution gets. Mm hmm. To the point of being super, super pixelated. Really bad, right? Fortunately, though, with shadow Jitter, it can actually look okay. Here's without and with, even with lights other than the sun lamp. The workflow is the same, which is direct that spot where we want it. And we use EV's ray racing to get that marvelous bounced light with most of the lighting actually coming from that bounce. Before moving on to the volume lights, let me quickly share the ray tracing settings for this scene with this stuff is covered in other chapters and earlier in this tutorial, too, but there are a few interesting things you might enjoy. Okay, so basically, aside from the direct light, the rest of the illumination comes from IV's ray tracing. Without it, it would be extremely unimpressive. So that's a major component of the look when it comes to this lighting style. Here you can see this tab basically at its default settings, so I can quickly show you kind of pro preset that we create. Not that it looks bad by default. Actually, here it looks pretty fine, but usually we can push it much further. First of all, I almost always switch off denoising as it tends to eat a lot of detail, especially in the darker areas. Then I temporarily switch off fast ja approximation to see a pure screen tracing, which has just these few parameters in EV, so it's really simple to tweak. The resolution is basically the main quality control for it. And yeah, the lower values don't look great compared to higher ones. One to two can be okay, but one to one is a no brainer when you scene heavily relies on indirect elimination, like this scene does. Then I usually take precision all the way to one, but I think I just skimmed on it this time. Now, thickness to me reads like how far the light bounds or amd occlusion can reach. Or you can think of it as having more light bounces in a way, even though I can only simulate one. This setting when cranked up really high definitely makes it look like there are more bounces happening. If we set it to an absurd hundred meters, even those buildings way at the back will receive some indirect illumination compared to, I don't know, 10 meters or just 1 meter. So it's really like how far the light bounds can reach. Thickness, one, ten, 100 Coffee. Ten, one. I think ten works great in this case, and it's mostly an artistic decision rather than a technical one. After that, I usually enable FAT GI, the alternative algorithm for global elimination, which looks way different and has its own pros and cons, not necessarily better or worse and not always faster either. Sometimes it's way slower. Yeah, really. But anyway, it looks way different from the pure screen tracing and the great thing that you can actually choose the mix between the two. So the choice between the fast GI and the raw screen tracing also is an artistic one. Even without diving into technical detail, you can tailor the look that you want with this slider. In fast GI, the distance controls how far the light can reach a literal limit, which you can tweak very visually like this. At zero, it's like infinites. Usually, though, I go straight to steps to crank that up. The default eight steps is way too low in almost all cases. Steps control both the quality and the reach of the GI. It's like two and one. So I usually set it to at least 24. That's the minimum. Naturally, the higher you take it, the slower we can get so be careful with it. But at least 24. Okay, then raise isn't that important, but resolution is. The default is fine. You can go higher, but I certainly wouldn't recommend going lower, since that makes GI look all blobby. One to one, the max setting makes sense when the scene look depends on global illumination, which it does in this scene. And finally, the threshold, as we have mentioned, is a mixed factor between the two flavors of global illumination. At one, it basically disables fast GI entirely. So it's really a creative choice what the balance should be. Usually, I go for about 0.8 or 0.85, as it gives a really nice balance. Here's a comparison. The one on the right is the one with FAS GI. It looks a bit punchier and a bit less noisy at the cost of some physical accuracy and performance. Even though it's called fast, it takes its toll. Never believe three D software. Another random setting worth mentioning is clamping, specifically indirect light clamping. But if it's set to ten. But if you want to reduce the GI intensity and kill some noise, you can set it way lower. That will trim down the indirect bounds. So it's handy for calibrating light ratios or reducing noise and flicker. Flicker, I mean, animation. I personally love the bounced light in this scene, so I don't want to trim it. I think it looks fantastic. So yeah, I wouldn't dare to kill it like that. The LDR, if you are a visual person, here are the retracing settings at a glance. By the way, I don't always drink coffee while recording these videos, but sometimes I do. And right now, I've got this pitch washed, slightly fermented, strawberry compost tasting espresso. And I've just pressed pause and down it in one go. Okay, back to work, folks. Then on to volumes. I'm going to create a cube, and scale it up so it covers the whole scene. Then in the object settings shadows off just in case. This play a bounds. Then we need a new shader and we can just get rid of this default one and add the principled volume, for example. To volume, it should go. And don't mess this up as I do it very often. It's volume, not surface. Density needs to be, much lower so the light can actually penetrate that haze. At zero, it still kind of lingers. That's weird. You can either shake the viewport a bit or actually it's the temporal reprojection thing. I'll just switch it off for now. That ghost in effect is playing tricks on us. Now it should be snappier. How thick do you want it to be? Ask yourself that I this what you want? A really thick shaft? Or maybe something a bit less dramatic. Alright, anyway, now our light source should have this Volmeric effect attached to it. The stronger the light, the more visible it will be. If your directional light isn't super bright, the shafts will be pretty subtle. And if there's a lot of ambient light in the scene, too, you'll get a softer haze instead of those distinct got rays. I mean, diffuse lights create softer volume scattering. Directional lights create light shafts or crepuscular rays. Just by using different lights, you can achieve a variety of interesting volumetric effects in blender. You can have stronger and more focused light casting really pronounced shafts like Sauron's eye, you know? And you can also have softer, more diffuse lights to sort of enhance the overall haziness. So you can really build the look you want. When it comes to volume rendering, fog and haze doesn't have to be uniform. Both cycles and AV can render variable density using three denoise textures. In a nutshell, you need the noise texture, and then plug it into the density of a volume shader. Then just tweak the contrast using a color ramp, for instance. This way you define how splotchy the volume should be. After that, you add the mapping nodes and slide the noise coordinates around. This way, you find the best looking seed for the noise or best looking placement. And suddenly your fog looks way more interesting. Let's just say that again. Eve is great at doing this stuff in real time. Now, positive anisotropy should look more like vapor, while negative, more like dust. But often it just comes down to whether you want to make those got rays look stronger or on the contrary, tone them down a bit, since anisotropy directly affects their strength. The bottom line, experiment with the anisotropy to see how negative and positive and neutral values affect the volume light shafts. That is a decent tool for Taylor in the contrast of this effect. Finally, we'll need to multiplier, so we can easily make the fog thinner or thicker. That's the math node with the multiply operation quite literally. And then you just tweak this volume, and that's it. Okay, volumes have their own resolution control in IV that you can find under the volume section. Well, first of all, setting the start and end ranges properly to match the actual scale of your scene already helps to increase the effective resolution. Then there's this resolution drop down, and most of the time I wouldn't recommend setting it up too high, as it can get super, super slow. This is one of these really dangerous settings of Ib, that if you said it wrong, the computer just explodes. I don't know. Usually, it looks quite fine right out of the box, once you said the start and end ranges according Dev, if you want to tweak the strength of those light shafts directly in the data panel, every light source has the influence sliders. And what you want to tweak specifically is the volume scatter value. When set to zero, that will remove that light from the volume rendering, meaning it won't interact with the fog anymore. It is super useful, for instance, when you just want to have a light that illuminates objects, but you don't want to have the trails of light shafts when it passes through the volume shader, then you can just zero out this slider and you're done. Volumetric effects can make or break the shot, but they can definitely be such a beautiful addition to the light leaks technique. They really tied it all together and give us another important tool for realistic lighting and rendering IV. TLDR, Kiascuro is a very effective and very versatile lighting technique. Blast some area in your scene with a directional light suggestive of some environment, potentially, and allow the bold shape and the realistic bounced lighting to the rest. Optionally, use volume shaded to reveal those light shafts. Keep in mind the shape of the beam. To imitate cycles, beam spread, use a sun lamp with a gobo. Or if you don't care about non parallel rays and shadow flickering and animation, then spotlights are also fine. Light linking will help you to use the sun, even in the enclosed environments. 10. Global Illumination: Global illumination or GI in short. We all know that when it comes to light transport and global illumination cycles is a beast. You click a button and get a jaw dropping glee photorealistic result. But what if I told you that with a few technical tricks, you can get pretty much the same level of realism in IV. This video is all about tuning the real time global illumination in IV for realism using g tracing, light probes, and a handful of other tricks. If you're new to IV and want to make realistic renders, simulations, and animations, this is the perfect starting point. The ultimate benchmark for GI in blend rays, of course, cycles. So the first thing I'll do is switch over to cycles. And just to show you how easy it is to set up realistic GI in cycles, what we can do, for instance, is use the environment light to illuminate the entire room through these windows. So under the world setting, simply increase the strength value, and that will just fill the room with ambient light by tracing thousands of rays. It's really that easy. Cycles handles all the light bounces. The result just looks real. Of course, it's a bit noisy due to the way path tracing works, but a simple deniser will clean that write up. Then if we wanted to add sunlight in cycles, that's also incredibly easy. Shift A, light sun. I'll press G to move it here for convenience, although it doesn't really matter, then I'll hit M to put it into its own collection, and finally R to rotate or double R to use the track ball and aim the light through the windows. And that's all there is to it, really, when you use cycles. It calculates all the global illumination on the fly. And all we do is adjust the light settings and whatever, do our artistic stuff. We can make our lights brighter, give them different colors to simulate a cozy evening scene or whatever. Next, for instance, we can make the ambient light, bluish color like that. That's the thing about cycles. You just do things and you don't have to set it up in any particular way. It just works right out of the box and looks incredibly realistic. The photons are bouncing around. Everything looks perfectly grounded, and the materials are beautiful. Unlike in IV. Wait, isn't that IV course, though? Yeah, right, of course, of course. Uh, so now let's take a look at how to set up realistic global illumination in IV. Step by step. I'll switch the render engine over to IV now. And what's the first thing you see? It looks like garbage. It's not realistic at all anymore, especially that environment light. So let's start with that. Right away, the first thing that comes to mind is that we can turn on real timord racing. This is going to make everything look 100 times more grounded, but not nearly as realistic as in cycles just yet. So let's start with the basics, volume light probes. Let's work on just the environment light first, turn off the sun and Android tracing, as well for the time being. I'll also go ahead and create a new collection here specifically for light probes. Okay, so I'll go Shift A and add a volume light probe. This tool would capture all the bounced lighting within a specific area. Okay, in the front orthographic view, I'll make it larger like that, why the S shortcut the S and X to stretch it along the X axis. So it covers the whole thing, but the points are still inside. Or better yet, so they are inside the thickness of the wall. Now in the side view, I'll just scale it up so it covers the whole room. It has a weird irregular shape, so we're out of luck positioning all these points within the walls. But anyway, let's see how it works. I'll come over to the data tab of this volume probe and bake the light cache. That's, that doesn't look good. The light leaks are everywhere. Well, we can try to scale this probe somehow. Making it bigger like that to cover it up. But ultimately, that just doesn't quite work. It looks awful. What we can do, how to unscrew this, so to speak. You know, the issue is that some of these tiny probes are located outside the room. They're capturing the environment brightness and liking it in. I'll switch it off for now because that isn't going to work. And I will show you a much better technique for illuminating interiors and IV. Let me start with the light path trick. IV needs a ton of these to make things look real. So in the world tab, we have our environment or ambient light, which, as you saw, creates a mess when used with light probes. So in the shader editor world tab, I'll zoom in a bit, and here we have it the background color and strength. The goal here is to keep this brightness of the window, but at the same time, stop the light from actually coming into the room. The way we're going to do that is by adding the mixed shader first from the Shift A menu. Mm hmm. Then I'll shift D to copy this background shader, and it goes right into the second socket. I'll set this one to zero. So now we can mix between the full light and no light whatsoever. But I want to mix it based on what the camera sees. So I'll shift A to add the input light path node to separate just the camera array. The E camera ray switch goes to the factor like that. So Damn. I want that to be the other way around. So let me just swap this. This one should be 20, and this one should be zero. Okay, now the environment that could have been an HRI or whatever, by the way, still has some brightness, but it's not actually illuminating anything in our room. Cool. I'll tidy this up a bit by hitting control He to collapse these unused outputs, I mean, again, this color could have been an HGRI image from Ply haven or whatever. It doesn't have to be a solid color. From there, I will add the window area light all by myself, you know, manually. Back to three D viewport, I'll go Shift A and add an area light G to move, R, to rotate, S to scale. These are the shortcuts every blender user eventually etches into their muscle memory. I'll make this light roughly the shape of the window, maybe slightly bigger and position it right outside. I don't see any effect just yet. And that's because we need to go to the light settings and take custom distance. And if we want this to correspond to cycles, this should be a very big number as light doesn't get dimer over distance. It's just a matter of few rays hitting the objects, but this is not the same. And then I will give it a bluish tint to match the environment color, and I will make it stronger. This is our first Aera light. Let's create a new collection for it. I'll hit Shift D to duplicate it and place it here outside the second window. Okay, then we'll need the third one, as well. And finally, the last one at the top. R to rotate. G to grab and move, R and Z to rotate it like that. It doesn't have to be exact, roughly the size of the window. In fact, it can be a bit larger. No problem whatsoever. Yeah, actually, I'll make all of them a bit larger, just to be absolutely sure they properly cover the openings. So these are our four window lights. Nothing too complicated so far. A bit of renaming naturally. In EV, there is a setting for jittered shadows. Every light has this checkbox called Jitter. Turning it on should make the shadow rendering for these chosen lights slightly more precise or actually a lot more precise, depending on the lights, softness and a few other things. And then if you want to see it in the viewpoard, feel free to turn it on here as well in the rendered settings. Keep in mind this will make the viewport a bit slower. If you don't tick it there, this better shadow wing will still be present when you render out the final image. Even if you don't enable it in the render settings. I mean, the shadows will still be jittered. It only matters what you said in the light settings, in other words, just as a bit of a foreshadowing forgive upon, I want to remind you again that this joke box only works in the viewport, and we're going to leave it unchecked for now for the viewboard. So there is no flickering or popping on moving the camera and stuff like that. You know, jitter shadows tend to flicker a lot, but when it's time to bring it back, if you see a missing contact shadows or whatever, we can bring it back anytime. And now we can get back to our volume light probe now that we have proper window lights. I'll actually enable the sun collection as well, because I want the light probe to capture the sunlight, too. Okay, now for the light probe itself, doesn't matter as much where these dots are anymore, but it's still good practice to keep some of them, if possible, inside the walls, right? The thing is, now these extra guys won't leak the environment light into our room because there isn't any environment light anymore. A, gotcha. We turned it off completely, remember? It's been replaced with our window lights. Now, let's see if it actually works. I'll leave all of these at the defaults and it bake. And after a split second, we get this, which is way nicer than before. It might not look unbelievably realistic on its own, but it's a huge step up from where we started. Actually, on its own, it looks kind of like that. There's a bit of light leaking here and there, but that's okay. We'll fix that in a moment. And it's not as important as before. In Light probe, we have a bunch of options we can play with. The biases. Usually, I just kind of play with these to see if they help reduce the light leakage or smooth it out somehow. It's usually about the normal and view biases, mostly. These have a stronger effect on the look of that baked GI. This is looking pretty nice, by the way. This isn't supposed to be a final look, by the way. Then, of course, intensity is an important one. And as we'll see later, it allows us to easily adjust the amount of bounced light, which is a bit easier to do than in cycles. Again, this isn't really supposed to be viewed on its own. So let's switch our main lights back on. The probe's intensity now controls how much of that light bounce we want to have. And you can set it to a value higher than one if needed. As for the resolution, I usually don't try to crank it up too much. Counterintuitively, sometimes it looks nicer and smoother at a lower s. Then these settings doesn't do much, to be honest. This surf of resolution, however, can safely be cranked up to 60 or more, as the default value, like Andrew Price has also mentioned in his EV tutorial is way too low. The values of 60 or 80 usually give a much better and more precise GI bake. Well, to be fair, in this example, the difference isn't tremendous. Maybe the corners received slightly more detailed GI bake and got shaded nicely. This is a fairly minor thing, though, at least in this simpler example. If the scene was slightly more involved, though, like this classroom with lots of chairs and a slightly more complex overall shape, then the default surfalRsolution of 20 will certainly be way too low. It might look okay on the surface. Like, there are no horrible stuff happening, leaks or anything like that. But secretly, the tiny little probes capturing the bounced light, you know, these guys, they are secretly underperforming. I'm sure they don't capture as much information as they can. So let's take a look at 80 now. I'll re bake it. I bet the GI bake will look much brighter after it finishes baking and much more detailed, as well. Yeah. So now these guys do a much better job capturing the new answers of the light rays bouncing around that environment. By the way, there's still no ray tracing here. SflRsolution 20 and 80. See the difference. If anything, this setting pretty much always makes it look better when you bring it up to 80. Alright, where are we at so far? We've added our fake environment lights. Then the baked GI, then window lights. Well, I got the sequence wrong, but anyway, and the sun. So far so good, it's really starting to look good, and we haven't even fired our biggest guns yet. 11. Baking Indirect Light Perfectly: Rate tracing. That's what makes Eve's global illumination look so much more real. For a slightly less cluttered view in the viewport, you can switch off the overlays. Preset and toggle it off from there. So in the render settings, I'll go ahead and set that real time red tracing on. This is a big one, and that changes everything. It just took it to a whole other level, right? Let's take a look at the fast GI approximation tab. I'll go right to the distance. We can use this to limit our rate tracing effect, so it's not everywhere, although there is nothing inherently wrong about it. But in a moment, I will show you why we might want to limit it like that. This reason has to do with consistency upon changing the camera angle. We'll see it in a moment. So let me show you something. Let me solo the sun lamp. Now watch what happens when we change the camera angle. As long as the light pools from the sun are within the frame, it is okay. In another direction, and the GI effect will disappear. This is because it's a screen space effect. If there are no lights on the screen for Rive to sample, it has no data to work with. That's the main drawback of screen space effects. They're dynamic and fast, and they look good, but they have this fundamental limitation. Let me switch the light probe back on. It will already help with that transition. I mean, when we look in a different direction, there will be at least something a bounced light captured by the probe, but there's still a bit of an abrupt transition or at least not as smooth as I would like it to be. That is why I will limit the distance of the screen space GI to something reasonable, like, I don't know, something like 7 meters in case of this scene. But it very much depends on your project. That's going to mellow out that transition. There's still a little bit of it, but it's much more bearable now. Let me switch the area lights back on. And now that should make it a really smooth transition. It's already starting to look pretty damn fantastic, wouldn't you say? There's more we can do to make it even better. Let's give it a quick check against the cycles ground truth. I'd say it's pretty comparable already. There's some sampling noise here and there in IV or this kind of splotchiness. Let's dive into the advanced rate raising settings and unpack that. I will start by turning off the fast GI approximation and denoising. That's going to lay Es tracing samples or noise completely bare. And, oh, gosh, it's noisy. It's almost like cycles. Among these noise lets, there are even brighter dots here and there. What is that? Basically, it's caused by the sun. If we turn it off, it will become much smoother. With fast GI and denoising, it will be almost perfectly smooth. The sun, however, would still look pretty bad, even with those settings turned on. To fix that, come over to clamping. You see, if the indirect light clamping were zero or none, it would make things even worse. The violent sun rays that are bouncing around this room will be even brighter. So what we can do is limit it maybe to something like five or even less. Two, for instance. The direct light is noisy, so we can skip that, and after that, it should look much better. Setting the resolution to one to one in fast GI could also help, but one to two is also fine. Really, it's going to be fine. The step slider is a really important one. It controls the amount of screen samples per GI array, and that affects two things the power of the global illumination. By bringing it up, you can make the GI brighter, but also it would throw more samples at it, effectively reducing the noise. At 64, it looks pretty impressive, but it comes at the cost of the performance. Now, let's compare it to cycles. Yeah, it doesn't bounce as much diffuse light around as cycles at the moment, but we are getting closer. Well, if the distance would set larger, it would fill the room even more effectively, something to keep in mind. As for the steps, 24 usually works best, in my opinion. Everything about that creates an extra burden on your frame rate and doesn't really contribute much more light. The bias could be good for making artistic tweaks to the lighting. Honestly, I can't explain exactly what it does. Look, with some of the Eve settings, you just use your intuitive judgment to see whether it makes things look better or worse. And that is okay. Increasing the thickness near and far can help create those contact shadows, making the intersections between the objects produce more produce a thicker shadow, let's put it like that. Look at this space where this torus meets the ground. The high thickness makes it less likely for the GIR to miss these crevices, so you get better, more detailed occlusion. I know it can be confusing, so let's have a look at another demo that focuses on thickness. It's very instructive to see what thickness does when fast GI approximation is turned off completely. With that off, thickness becomes pretty much the only thing we can tweak. That makes it easy to see what's going on. So basically, if it's at zero, all the occlusion shadows will be completely gone, and your objects will start to look like they're floating. Raising the thickness will bring back those contact shadows. That's pretty much the gist of it. And not only that, but the higher thickness will make these contact shadows stick for longer when changing the camera angle. Otherwise, if thickness is going to be low, the GI rays will start missing the occlusion from certain angles. Like shadow, shadow, shadow, boom. Gone. Shadow, gone. Shadow, gone. This could be an issue for animations, of course. So a higher thickness helps those shadows persist. Okay, back to fast Ji approximation, thickness far has a very similar effect, thickening the contact shadows, but it also has a pretty interesting effect on the bounced lighting. How much of the red color from that monkey is bouncing around, for instance. It can be pretty tricky to tell the difference between the two, to be honest, or even see what's going on. So I'm going to show it in a new demo scene so we can get a better look. So thickness far and near the FAS GI approximation is on the rest of the settings at their defaults. Well, first of all, as I said, these two have a very similar effect. Still, there's something to be said about them. So I'm going to zero them both out. Thickness far can't be less than one degree, but anyway, this is how it would look. Very little conduct shadows and almost no color transfer from one object to another. Great. Now, let's nudge the thickness near or geometric thickness up. This is really helpful for getting the contact shadows going, like we've mentioned before, it helps put the objects on the ground. You usually set this value to taste, and usually doesn't have to be gigantic. Now let me reset it back to zero and check the far or angular thickness. Nudging it higher should bring up the colour bleed from these objects onto the environment and between the objects themselves. And it's kind of what is actually happening, right? There is some red from the monkey reaching onto the spheres and spilling onto the ceiling and walls, yet it's pretty faint, to be honest. It's pretty weak, even at maximum. There is something going on. There are some rays bouncing off the monkey coloring these other objects. But the caveat is at the default settings, it's just way too weak. The thing is, there's the steps parameter that can be adjusted. Think if it is running more and more iterations of the global illumination or throwing more and more rays, although it's confusingly called steps. Think of it as an intensity or energy of the GI rays. At 64, the color bleed is so much stronger. It's pretty crazy how much of that red color is actually bouncing around, reaching every object in its vicinity. And now the far thickness will have a much stronger effect. Now it legitimately works almost like a boost slider for the light bounce intensity, you know. And I get the impression it's meant for that kind of far interaction, like Ray's traveling a pretty big distance to colour the entire room, you know. While thickness near is meant mostly for capturing the shadowing that should happen near the objects, it's not so much about that far reaching light transport. If that makes sense, it's mostly for nice little crevices here and there. While the far sister of this parameter is mostly about far reaching color and light bleed and that sort of stuff. And indeed, if I move this monkey head over here, then thickness near helps with the occlusion shadows, while far is mostly for fine tuning the light bounds. The step setting serving as an energy multiplier for all of that. Its default eight steps is way too low, so in most cases, 24 works much better, but it can be higher like 36 or whatever. But 24 is usually enough. The GI resolution is the quality of that pass. It can affect your FPS as well. So if things are getting a little laggy, you might want to fine tune it. Or on the contrary, you might want to go one to one for better quality. But usually, I just keep it at the default one to two as it's a good balance. Alright, what can we do when the bound sliding looks noisy, though? Like, look at this unpleasant grain on the right wall and pretty much everywhere. Well, we can try to increase the number of rays in the GI settings to help reduce the noise. There's a chance the video compression will eat this, but hopefully you can see that it gets much smoother with more rays thrown onto it. Actually, to make this example a bit better, I'll reset it back to the default two rays per pixel. Okay, now, if we scroll all the way to the top, here we have a global samples parameter for the viewport and for the render. Here we're tweaking the viewport one, of course. I'll give you just one sample now. Now we should literally be able to see those GI rays and where they hit, we can also see it became noisy as hell as there's clearly not enough rays as it's just one sample per pixel or two rays per pixel or just one. You got the point. The race control allows us to budget more rays for each pixel, 16 being the maximum. If your GI was noisy to begin with, adjusting this control could really help with that. And don't forget that this number gets multiplied by the number of global samples, the viewport or the random samples, I mean. So at some point, it can start getting a bit laggy in the viewport specifically. So usually I try to keep it moderate. Let me reset the samples back to 16. 16 by nine should look pretty good. Anyway, I'll stick with the default for now, which is just two rays per pixel. It doesn't seem like it's enough, but it will be enough once we get all the other tricks working. But anyway, it's good to know that you can play with a number of rays if the global illumination starts to get noisy. That's one of the first things you can try to tweak. 12. Eliminating Bake Glitches: Now let's dive even deeper into the EVs rate as settings. If you notice that reflections, especially on highly glossy materials look a bit muffled compared to cycles, actually, let's take a look how it looks in cycles. Um, yeah. In comparison, the IV's reflections seem a little blurry and smodgy and that's because of denoising. Switching off the bilateral filter or even that and temporal accumulation usually fixes that. It might be a bit hard to tell by just looking at the smaller window. So let me show you a Zoomed in demo. This is the default relatively smodgy look caused by denoising, especially by something called temporal accumulation. And if you just switch it off, the glass and other reflective material should get much sharper. The spatial reuse usually isn't that aggressive, but it can also cause reflections to appear slightly lowers. So switching these denoising options off can make a world of a difference. It may not seem like a big deal, but it's the things like these that make you render look real, you know, just like contact shadows. So that's an important setting to be aware of. Another great example of this is PBR shaders. If your scene had a reflective PBR shader on the floor like this worn tile floor from Plyhaven, you may notice that by default, the specular reflections aren't as sharp as you'd like. And that's because these denoising settings mostly temporal accumulation were holding it back. Special use too, for that matter. Also guilty. And after switching them off, it might become a bit noisier, right? But it will also make PBR materials, especially highly glossy stuff like this wet floor look so much nicer. A long story short. If you feel like Eve isn't looking as detailed as it could, these guys might be a part of the reason. Often, I just turn them off completely. Usually, just make things look better. Now, there's a way to make IV slightly more precise and cycles like in other aspects. Here's how it goes. Switch off the fast GI approximation to see the primordial noise. That is as pure ray tracing as possible for IV. After that, give it more samples, actually, much more like 256. Go to screen tracing and max out the precision. And give the thickness a boost. 0.2 meters isn't quite enough. It should be around 8 meters to match cycles in a close away in this scene. Anyway, here's a quick comparison. These are the default settings, low precision, low thickness, and this is high precision, high thickness. The key takeaway is that the default screen tracing settings create a light bounce that is a bit too weak. If we are going to make it look like cycles, we need to bounce much more light around, and these settings bring it a bit closer to that look. Alright, that looks pretty believable for screen space rate racing, at least. Now, this isn't the real thing. Obviously, it's not proper world based rate racing, like in cycles and other such engines. If it still takes into account only what the camera sees to bounce that light around, but this screen racing looks pretty decent now. I'm very happy about it. It looks decent because we are throwing many, many samples at it. 256, in fact. If it was just one sample, it would be noisy as hell. Alright. So this screen tracing method requires an unholy amount of samples to fully converge and clear up that noise. And here comes the catch. How is this any different from just using cycles? If we have to brute force the samples like that, jacking them up and up and up and then having to wait until the image clears up, eventually, that is pretty insane, right? We could have just done the same in cycles. And we decided not to use any denoising. Crazy. So here's the next step. I'll dial in back to 256 samples, and here's what we do next. And next after tuning our screen tracing pass, we are coming back to fast GI approximation. What is that again? Basically, this is a faster and smoother global illumination technique for higher roughness surfaces and only for them. Why do we need to optimize higher roughness materials? Basically, screen tracing works great for perfect reflections like these glossy spheres or these pretty shiny plastic material or the material on this black doughnut. The only surfaces that actually look grainy in the viewport are the higher roughness or diffuse surfaces like these walls and the floor. That's because pure screen tracing kind of has to use a lot of rays for them. These things are diffuse, so the rays gets scattered around on bouncing off these objects. It creates a lot of noise to be resolved. So Blenders developers added the fast GI approximation method to solve that and give us a better technique for approximating these higher roughness surfaces. After switching it back on, they diffuse white material on the walls and the floor is now rendered more optimally, even though it's slightly less precise compared to pure screen tracing. So why did we bother setting it up if we are going to replace it with this faster method? Well, the thing is, we don't have to fully replace it. Among the fast GI settings, there's a slider right at the top called threshold. It's the maximum roughness allowed for screen tracing, and everything above that value will use the fast GI approximation. If we set it to zero, every surface will use fast GI approximation. Like, even this glossy sphere will use fast GI and this monkey and everything. Look at how this sphere has lost the light bounces from the monkey. It lost like all the precision of screen tracing. Now, by increasing the threshold, more and more surfaces will actually start using proper screen tracing, you know, instead of the approximation. And at some point, it will become a blend between the two techniques, the precise and the fast one. You know, somewhere around 0.8 or 0.9 perhaps near the end of that slider. It becomes a blend, kind of the best of both worlds. So it's pretty smooth now, but we are still getting some nice bounced lighting, a nice light bleed from that red plastic onto the floor, for instance. We get the best of both worlds. It's no longer as precise as pure screen tracing, okay? But at the same time, it's not as noisy either. Mm hmm. By the way, taking this all the way to one is the same as not using FAS GI at all. So there's no point in dialing it up to the maximum But near the maximum is a great idea. The sweet spot in my tests is around 0.85 or something. That is, if you want to reap the benefits of both techniques, it should be around 0.85. It doesn't have to be exact, though. When the GI is set up like that, we no longer need a crazy sample count. So we can dial it back to 64 or whatever. Pretty cool, right? Or maybe 128, just to be sure. Alright, that's brilliant. Actually, thinking of it, I think we've lost a bit of contact chasing with this fast Ji pipeline. So it might be worth dialing up the thickness near and far in the fast Ji settings, too. That might be a good idea. I'll probably leave that up to you, though. I think it looks pretty solid as is. I don't want to tweak anything. So that is how we set up rate tracing for realism in E. It's time to bring back the rest of our lighting setup, the baked GI probe and the window lights. If UI distracts you from seeing the beauty, again, just turn it off. Now, at this point, it should look largely like cycles, except for a few minor things, and I'll bet you can easily tell me what they are. Take your time spotting the differences while I'm drinking coffee. Here's a comparison to make it easier for you. Personally, two things still bother me. The glass that looks a bit weird in Ivy, not the same as in cycles. And then the other thing that mildly bothers me is that the contact shadows beneath this torus object looks great in cycles, nice and thick and a bit flimsy in Ivy. Why is that? What have we missed? Why the torus is floating again? It should look like it should look like this, isn't it? Nice and thick. When stuff like that is happening and you're sure you've set up right racing correctly, then it could be the jittered shadows turned off in the viewport. Boom, somewhat helps, right? It boiled down to shadows not being precise enough, and jittering them helped with that. There is something else causing this effect, though, which is our light probe. On switching it off, you can see that the shadows becomes even thicker. So it was the bounced light captured by that probe that was making it float alongside the jitter shadows. But it doesn't bother me as much. The main thing was still the jitter. That was the main thing, indeed. One thing that I don't like about enabling jittered shadows in the viewport is that it makes it flicker and flicker a lot, and it's a crazy effect, very tiring on the eyes. So I'll keep this checkbox turned off for the viewport. Even if it means a few floating objects here and there, it will still be fine in render, remember? This checkbox is just for the viewport, okay? We're largely there. It's just the reflective, glossy and glossy elements that aren't quite behaving in a realistic manner just yet. They're pulling us down. The rest is quite good. Really good, I think. Nothing reflection probes can solve, though. Okay, the shaders like glass and other reflective stuff might not look right in IV right away. Unlike cycles. Yet there are ways to tackle that. Namely, what we can do is use a sphere light probe or reflection probe. So I'll put cursor to select it. So it's right in the middle of the sphere. Then shift A, light probe, sphere. We can adjust its scale with the S shortcut, and that will affect reflections in that area. The thing is, it makes reflections more plausible, more realistic. Yeah, it has to be placed and adjusted manually like that. So I can imagine it isn't super convenient for more dynamic scenarios that involve animation, yet it works pretty good in static shots. I'll press Shift D to copy it out and move it to that other sphere, or we can just shift click that other object and copy its location. As you can see, that makes a big difference in how that reflection and refraction are rendered. Without it, it's close to being black. So probes make a big difference here. Of course, they deserve their own collection, which will be aptly called light probes. I'm going to run a quick cycles check next. What I can tell. Inching closer and closer. Maybe Ivie already looks a bit more interesting, even, hmm? Here's a comparison. I don't know. Ivie is even better, I think. It's slightly more artistic, slightly more stylized in a good way while still being realistic. It has a bit more color as well. I don't know. Both of these are good looking renders. After all that is set up, you can always go back to that volume probe, the baked light bounce and see how much of it do you want to have in the scene, ultimately. That's pretty cool that it can be tweaked like that on the spot, alongside its more artistic sliders, like biases and other stuff. At this point, it really boils down to your preference and some intuitive judgment. There is no single inherently realistic set of settings that would work every time. Sometimes you just have to keep tweaking it a bit until it clicks. Eventually it will click, I promise. By the way, feel free to rewatch this video as many times as you need. Before we call today, I would love to mention something that works arguably more conveniently in EV right away compared to cycles. It's volume fog. I will create a new collection, fog. For that, I'll use a regular cube. It would encapsulate the entire scene like that. It will need a new material. With appropriate naming. Now I usually check the object properties, shadow ray visibility, and I just turn it off. It might get candled automatically once we add a volume shadow, but I like to make sure. And then I also disable transparent shadows here for pretty much every material that isn't supposed to be semi transparent. And finally, I set the viewpod display to bounds so it doesn't block the view. I think we must have done that 1,000 times already, yet I'll jump into the shadow editor to show it to those who might be unfamiliar with volume shaders. Um, so I'll add the principled volume shader here, and there it goes into the volume input. Mind the input. It's the second one that says volume. Okay, I just want it to be a thin haze with high nisotropy. The light shafts are still a little bit weak, you know? So I'll grab the sun lamp and I'll make sure it blasts enough exposure into the sea. Uh huh, now we are talking. IV is just marvelous when it comes to these kinds of volumetc effects. Compared to cycles, it renders these light shafts 100 times faster. While it looks, in many cases, not that different. It really is a superpower of IV. I very much agree with how Andrew Price formulated it in his IV tutorial. It's a really good tutorial, by the way, too. This is it. This is how to tune EV's global illumination and other settings for more realistic rendering. It looks a lot closer now to path traced light transport, and even though it's not entirely physically correct, it's definitely good enough many times out of ten. In fact, it might even be better in its own slightly stylized way. So I hope you enjoyed this video and learned the thing or two about IV lighting and rendering. 13. Character Lighting: Welcome to the character lighting guide for Blender Zeb. The latest versions of Blender are absolutely incredible when it comes to rendering characters. Thanks to Vulcan and other advancements, it became possible to render hyper realistic portraits in real time using IV. And honestly, the results look production ready and pretty stunning, making it a total joy to work with. So in this very exciting tutorial, I'll show you four cinematic character lighting setups representing four levels of complexity and how to set up and render each one in IV. No matter if you're already familiar with IV or have just started dipping your toes in with this vendor engine, you're definitely going to pick up some useful techniques. Level one, studio lighting. This entry level setup involves just one light source and no backdrop. It is the most minimal setup possible, which certainly doesn't mean it's ineffective. In fact, it's super effective and is a fundamental skill worth knowing. A single light is often all you need to create a striking portrait. And it ties with the essential properties of light. We're starting here in the random mode using IB. We'll spend some time in the rand settings later in this video. But first, when working in a studio, you don't typically need the default ambient light. So you can switch it off completely. The simplest thing we can do really is just press Shift A and create an area light, one area light, a soft box, basically, to make it behave in a physically accurate way, turn on the custom distance, and set it to a high enough value. We can also make it nice and bright by increasing its voltage or power. To make the shadows more precise and realistic turn on jitter in the light settings and also in the global render settings. Shadow jitter makes a big difference when you're working with soft area lights in IV. Look how pro pisado appears under the nose and the chin and the eyelids in the right image where jitter is active. For small and sharp light sources like the sun, Jitter is less critical, though. It's mostly the soft lights that benefit most from jittering. I'll switch it off for now, though, because it can sometimes flicker. Alright, now you need to decide where exactly to position this light and whether it should be larger or smaller because that will affect its shadow. It's really simple. Larger lights, cast softer shadows. Here is a bit of a comparison to help you with that concept. Now, not all light angles are created equal, except for perhaps the dead front angle, which flattens the image quite a lot. And usually it's not good. So what's a good angle for the main light? Essentially, anything other than frontal. Any direction other than frontal is better from the standpoint of creating or sculpting the form. Even if this angle is just slightly above the model, you know, it's already a massive improvement. So you can bring it slightly above like that, and it will immediately look pretty nice for a studio portrait. Incidentally, this is the exact lighting scheme that was used in the original render by Y. See, you don't need much. Next, how close the light is to the subject also matters. If the light is very close like this, there's a dramatic light fall off from here to here. It's almost like manually reducing the influence in the IV light settings. You know, when the light is closed, you get a very similar effect. When light is further away, this difference in levels disappears. I want you to really look at this image and see that gradient of light to dark, which is looking totally different based on how close or far away is the light. That is called light fall off, and usually having some fall off is better than having completely uniform brightness. Well, maybe not as dramatic as in this demo, but you get the principle. When lighting for portraits and characters, there's one thing that can make or break the realism of the entire rematch no matter where you place the lights, and that's subsurface scattering. This is the shady property that allows the light to travel under the skin and illuminate it from within. Sounds gross, but it's a very nice effect. I'm going to hop over to the shade editor to demonstrate a few things there. First of all, here, we can tweak the specularity or IOR level. With IR, zero is just going to be the base color. So some specularities must have for realistic lighting of characters, but then subsurface scattering is really, really important, too. With the default settings, the characters might look a bit like resin, so I'll show you the settings that I typically use. What I'm going to do first is reduce the red, green, and blue contributions. 0.020 0.01. And then we are left with a scale, which is essentially a multiplier for this effect, defines how far the light penetrates under the skin. But also, there's the thickness output often missed. I'm going to set it to a very low value like 0.0 or something to make it look more like actual skin. Alright. See how changing that value allows to fine tune this scattering. It's pretty cool that we can do that in real time. And then the scale that we can tweak to honing on the skin look, I just can't stop thinking about the fact that it all works in real time, Ivy. Somebody wake me up. So here are the subsurface scattering settings that I ended up using. Now, as we've mentioned already, when using a softer light, shadow jittering is also vital. See how without it, some areas don't receive a proper shading. And these microshadows are absolutely critical for realistic character portrait. Absolutely critical. Then turning on tracing should help to fill in the shadows a bit. It's a little bit hard to tell here. There should now be a touch of light fell in the darker shadowed areas now. Alright, that is the first level of character lighting. One light, no backdrop, the most minimal setup ever, which can still look super effective with the right model, that is. Now it's time to explore level two, nature or HDRI. The second level will involve a full 360 degrees HDRI panorama of some nice location in South Africa, for example. The goal is to place the character within this location lighting wise, and then peek into the EVs weight raising settings to get the soft and natural portrait lighting going. So it will be just a little bit more involved. This time, I'll go straight to the shade editor, the World tab. Now I'll click the background node and press Control T to set up the texture nodes automatically. Now you need to find that HDI that you have downloaded from Ply haven, for instance. I'll use the Forrest Burg Mountain Lookout too. Mm hmm. Okay, now the character should be in this environment. It doesn't look great so far, but we'll fix that. First of all, it's blown out, so the strength should be much lower. Then right away, real tembra tracing must be turned on. It will look better right away. Just look how it removes those weird light leaks behind the ears and stuff like that by simulating a proper occlusion. That's already a ton better. Next, you can spin that HGI panorama on the z axis to find the best looking angle. I mean, for lighting the face. And again, I don't want it to be too fronty. I want the light to come in at an angle to create some shape and definition. And usually, there are a lot of different ways to rotate the HGI to achieve that. For instance, something like that looks pretty three dimensional to me. Now I want to make just the background a bit less bright, more like the evening. When I simply reduce it from here, the background looks pretty nice, but the face is now just way too dark. So let me show you a different way. To be able to adjust these things separately, we need a mixed shader first to mix between two backgrounds. Okay. But then we'll need the light path node. Which, by the way, works in IV, at least partially. We need its I camera ray inputs to mix between these two backgrounds. Control H to minimize it like that. Now, the first one will be just for the character lighting. Mm hmm. So you can fine tune it however you like. So you can make it weaker or stronger, whatever you wish, really. It's just for the character. The second one will be just for the backdrop. So you can also fine tune that, however you like. For instance, to make it nice and muted to make the character pop. You can do really cool stuff with that. T LER, you can always tweak the levels of the character and the background independently. The left image is technically a physically correct result, but what matters most is the artistic or cinematic realism. So it's absolutely fine to cheat a little like the middle image. On the other hand, if you go too far with that, like in the right image, Mmm, it may start to look a bit artificial, or as cinematographers say, it looks lit. How realistic the EV is HGRI lighting will look heavily depends on tracing and its settings. For better results, you can get rid of denoising, and that will add a lot of crispiness. Pay attention to the fine detail. Fast GI approximation can also be switched off. This one will be harder to spot, but there is a difference in precision. Maybe if we zoom in right down to pixels, you'll see the better contact shadows. See that? Or zooming in on the detail, see how it makes it look slightly more realistic. After that, it's a error of maximizing the precision. Check the hair and the detail. And increasing the thickness. Notice the ecluded areas now receiving a more natural shadow. The resolution here is the quality or fidelity of this effect. It affects primarily the indirect light. So if you set the samples to one, it'll be immediately clear. So that is resolution, and it's better set higher for higher quality portraits, at least one to two or better if your GPU allows to its maximum quality. The iter shadows doesn't necessarily matter much in this HDRI setup, but they can matter under certain conditions. If in the sun tab of the world settings, the jitter is also turned on and the sun threshold is low enough to actually create direct shadows from the brightest parts of the HDRI, which will be somewhere here, by the way, then it could really matter. Notice how these shadows are actually flackering when I'm changing this value. It means they are jittered. But anyway, if the threshold is set high, there will be no direct shadows at all, and there will be nothing to jitter. Notice there is no flickering anymore. So that is something to keep in mind. Under certain conditions, jitter can still be pretty important even in this setup. This was the second level of character lighting in IV, natural HDRI lighting. Level three will be the double HDRI lighting or location lighting. 14. Refining Mood & Accent Lights: The third level of character lighting is the on location interior setup using WH DRI. It's more complex than the second one because the lighting will be completely decoupled from the backdrop in this technique, and it requires some fiddling to make it look right, but it's such an incredibly useful setup that will breeze right through it. We already know how to separate the actual environment lighting from background from the previous part, right? So it will be actually pretty straightforward. First, I'll set up the HDRI for the backdrop. Right now it's on a bright side, killing the mood, so I'll tone it down quite a bit. And then I'll start rotating it on the Z axis to see what kind of composition we can get Mm hmm. You need to be patient with this. Okay, these lights are quite interesting, but maybe something like that would work. I'll roll the intensity back a bit. Okay, that looks great. Like a warehouse or something. Imagine this as establishing the location. Next, I'll use one more HDRI for the actual lighting here. Do do do Control T to set up the nodes. Now I'll just plug in the new HDRI. The first glance, the question arises, Why airplanes? How is this supposed to work with the previous one? What matters to me is that it has an amazing door over here, which will work as a key light, you know, nice and big light. So that second GI will be just for the lighting on the character. Your task as an artist is to find a good looking angle for it. You do it by changing the rotation of the HGI. Remember, all angles are good except the frontal. Frontal angles often simply tend to be flat. I thought about some wordplay, like flattening, not flattering, you know. Anyway, another task is to set the right softness. Remember, in the world settings, there's the sun tab, the one with the shadow Jitter and sun threshold. I want it to be super soft in this case, so I'll just put 45 there, and that will surely almost entirely remove the direct light. Obviously, it's up to you whether you want the light from this HGRI be harsher or softer, but that's how you control that. Ultimately, after we get this double HGI setup running, it's a matter of moving the light around, adjusting the levels of the backdrop and the character, and perhaps moving the light again. It's all it takes going back and forth. A quick tip. Don't hesitate to use X and Y axis for rotating your HDRI environment. It can yield some pretty interesting results like helping new model catch some nice reflections like the one here or just an unexpected light direction. Like, look at that. It now turned into a really interesting parking lot lighting slightly from below, a complete vibe shift. So that was the third difficulty level of setting up lighting for characters, the HDI technique for location shots. Now, I hope you're well rested and hydrated, my friend, because there's the last fourth level upon us. The one that old timers call full cinematic. By the way, it's not better than the previous three or anything like that, more. Involved. Using HDRIs to set the stage and mood, then cranking it up to 11, using the studio lights on top of that, resulting in a cinematic character portrait. For setting the stage, I'll use the vintage measuring lab HDI, which has really nice shapes and colors to it and works great as a base motivation layer for the lighting setup I'm about to build. I'm going to use the double HDRI trick for tilting the backdrop like that, straying away from the default orientation of it and giving a dynamic angle. Technically, we could also tweak the power of the first character GRI separately from the backdrop if we wanted to, but I wouldn't bother. This time, I'm going to set up the studio lights to create a moody dramatic look, still motivated by the colors and values of this environment. First, I will create a sun lamp, a pretty harsh and strong one. I want it to be a back light to separate our character from the backdrop, a classical cinematic kicker, still motivated by the bright lights of the background plate. Next, I'll duplicate this light with Shift D and an increase its angle to make it way, way softer and more wrappi. You know, so it wraps around the head like that, maybe off to the side a bit. Alright. Now, this light helps to sculpt a bit of the facial features, but mostly just wraps around the face a little bit. It is paramount to match the colors of these extra lights you set up to the environment, at least roughly green environment, greenish lights or at least cold by temperature. Otherwise, it may come across as artificial. For all the lights. For instance, this one, it is looking okay, but I think it can still be matched a little closer. Maybe to that tone marked with a red circle here. It may not seem like a big deal, but it looks a bit more from the same world when your lights match the backdrop. Compared to the previous three levels, this one has glasses because glass is notoriously difficult and heavy. It's not even clear to begin with, how to make it refract and reflect at the same time, you know? We definitely have a whole chapter dedicated to glass in this photo area lighting and rendering course. And now, the inspiration for the vitals has absolutely nothing to do with my recent recycling bin. Now, to wrap that back light around the face a bit more, I'll add another aerial light and make it more of a side light. I want to sculpt to the face a bit more. Right away, as this light is pretty soft, I'm going to activate jitter and as always custom distance. I imagine that this is a bounce of that back light, and I'm bouncing back that back light from that direction. Again, the color can be matched to the environment colors. There's a lot of green there, so technically, it won't be out of place, even if you go really bonkers with green. But I want it to be slightly more neutral, so I'll roll it back to this pale green. Now, one more light from the other side, I think that will look great as if the back light bounces off some surface there and highlights that side of the face, maybe it could bounce off something green like crocodile over there. Paints the face green from that direction. As long as you base these movie lighting choices on the environment, which is dictated by our HDRI, it won't feel out of place. Here's the progress so far. Set, ambient movie light. Sometimes they even add the highlights in movies, a tiny detail that works wondrous for adding emotion. An area light would be ideal. You know, you just place it a little bit further away. Take custom distance just in case I'm gonna place it really far out and try to capture that glint. Now, the problem is now the glasses also capture that glint, which, by the way, can be of different shapes. So what we can do is with that light still selected, come over to the object property stub, shading, light linking. Click New to create a new light linking group and drag the eyes into it. That sounds gross. Now this light should only affect the eyeballs. See that tiny little glint? That's emotion. Here's before and after. Oh, by the way, if you don't want the whole eye to get brighter and just the glint, what you can do is kill the diffuse influence on that light. Or just leave it a little bit. Raw, with the glasses and the eyelight. How can we elevate this even further? If I wear you, I would add some haze. Maybe actually quite a lot of haze. My Ai is a cinematographer, and he's a very generous person when it comes to haze. But really, it works wonders for making the characters pop. As usual, we can just add a cube, set its visibility to bounce, do not interfere with the viewport, then add any volume shade like principal volume and greatly decrease the density. We just want it to be a subtle haze. It really does make characters stand out. If you want some extra visual punch, there are transforms in blender that have more pop in them than the default AGX like A is 2.0 or AGX high contrast with a slightly increased exposure. Personally, I think Cronos PBR doesn't work that well for portraits and for realistic or photographic style at all as it over saturates, the colors way too much. I'd rather choose a more classic aesthetic with slightly desaturated highlights like in these other vietransfms. On this last level, you can also use the post pro effects like real time bloom to simulate how a real camera lens or human eye, for that matter, reacts to intense light. And to polish it up, maybe the film grain or sense of noise, anything that can add a touch of analog authenticity. Obviously, alongside the chromatic aberration to break digital perfection. Depth of field in IV is amazing. Honestly, it's so good. You can experiment with a lower F stop for a shallower depth of field, and a higher or lower ratio of assimilating the vintage lens weirdness or an anamorphic squeeze, if you wish. This could be that final photographic touch to your IV portrait for that fourth and the last level of IV carcighting. To recap, in this tutorial, we explored four cinematic character lighting setups, representing four levels of difficulty, from a single light in a minimalist studio to a full blown cinematic rig, all rendered in Ivy. Now let's be clear, great character lighting doesn't have to be complicated. Each of these workflows can produce stunning results on its own. But if you're looking to level up your renders, these setups offer solid options for compelling character lighting in blender. And now, thanks to Vulcan and other advancements, it's finally possible to render ultra realistic portraits in real time. In Eve what a time to be alive. 15. Ambient Occlusion: In a real world, even if the weather is extremely overcast and there are no direct sunshadows, there are still minus self shadowing in crevices, and we can still read the shape. If you strive for photorealism, this occlusion is extremely important to simulate. L compare these two renders, which one looks more believable to you? Obviously, the one on the right. It's the one with ambient occlusion. Actually, in cycles, that darkening effect, frustratingly for IB fans comes for free. The way path tracing engines work is that they actually simulate physical light transport by throwing thousands of rays into the scene, and naturally, some of those rays just cannot reach certain places like here underneath that car. It's where Pennywise the clown will be lurking. Whereas IV's ambient light render may seem totally flat in comparison. Well, of course, there are other scenarios where IV seems to outperform cycles, like simulating the light from a smaller window in an interior. That's where cycles clearly struggles. But in this scenario, cycles look realistic right out of the box while IV is not. That being said, I bet you won't believe how realistic IV renders with ambient lighting can look by just simulating some ambient occlusion. And there are three distinct ways of doing that in IV. And in this very, very exciting tutorial, I'm going to show you all three. Ambient occlusion. But if at Ivy's self shadowing falls short of rendering the details cycles capable of, and you may think that IV is a lost cause. But we can fix that with the help of the ambient occlusion effect for better realism in your real time renders. Ambient occlusion or AO for short originated in the early 2000s in film and game rendering as a fast approximation of global illumination, a way to get soft realistic shading without full ray tracing. It was either painted or baked right into textures or later on applied as a post processing effect to help ground objects in space and make lighting feel more believable. Since then, the technology for AO has advanced in EV, so AO comes in three flavors, a post processing effect, like in early games, a light probe effect, and a ray tracing effect. Let's see how it works. Ambient occlusion using tracing. Alright, so here we are in IV, and I want to compare it with cycles. So this is our reference for realistic shading of this guy. Look at all these folds of the space suit. These are shaded really, really nicely. Cool. Now back to IV. So B default, it's fairly flat in comparison. So the first thing we can do is enable ray tracing. And right away, this should make it look a lot better, especially if you turn off these two settings and maximize the resolution alongside the precision. And maybe crank up the thickness. If you watched other videos from this course, you are familiar with these settings. These are the settings which make IV's rate racing produce higher quality, more precise self shadowing or amient occlusion, you know, simply due to increased precision of that entire rate tracing system with all the knobs turned to maximum and all the optimization turned off. Compared to default parameters, these settings blow the rate tracing default out of the water. These settings work great in the outdoor scenes, too, dominated by indirect light, and in fact, it brings it way closer to the realism of cycles. So now if we compare it to cycles, we won't see such a striking difference as before. Actually, the thickness slider in IV can be used to artistically fine tune the depth and the balance of the indirect hedoin. The second method of doing ambient eclusion is to use a render pass in real tempos prosen. I mean, like in computer games. Let me turn on fast GI, tone down the rate race in occlusion because I'm going to use the compositor to add some more. The A or render pass can be activated in this tab, ambient occlusion with the default distance being 0.2 meters. Actually, this pass is always rendered under the hood and you can turn it on here. I mean, turn the visibility of that pass with the occlusion distance set in this tab. The higher you go, the wider the effect spreads out. So it's pretty cool to be able to preview it in real time like that. But let me go back. To overlay it on top of our viewpoard we'll need a compositor. Um, where is it? Yeah, compositor, I said. And to see it in real time, the viewpoard compositor should be set to always or camera. And now we should be able to control Shift click through the nodes to preview them. That's our ambient occlusion pass. Pretty amazing. I'll set it nice and thick. 5 meters. Now we need to mix it with our base render. It can be done via the mixed node. I'll drag the AO into the second socket of it and set the mode to multiply. Okay, the problem is that our HDI backdrop is also affected by this. While it shouldn't, we need a way to apply it to the three D objects only. And for that to work, we'll need to activate the environment pass as well. It's also available there. Mm hmm. And then in the render settings, enable the transparent film. It will show us checkers, but it's actually transparent if you never used Photoshop. Then I'll add the Alpha ov node. Plug our environment there. And the objects, then it should all be alright. Basically, what we've just done is separated the objects from the environment. There are other ways to do it, but it's the easiest one in this case. And now this AO pass or rend layer, oh, by the way, look at these neat background and foreground labels in five point oh. So this pass will only be applied to three objects, and that will help to ground them. If for some reason you want the AO pass or any passing blender to be mixed stronger than one, just uncheck the clamp factor in the mixed node, I mean, and this slider still won't let you go over one, but you can override it by using the secret technique of clicking and typing any value. And, hey, now we went totally overboard, but that's alright. You can go as bold as you like with Ambeclusion pass. I would have chickened out, though. I'll just stick to enormous value of one. Before we go any further and get over excited, there is a big limitation associated with this method of using amio occlusion as a post processing effect in blender, at least, as it's applied after everything else. So using depth of field or any other depth effects will likely break the image, as AO will still be drawn on top, leading to this funky cell shading effect. Could be pretty cool, to be honest, not very realistic, I guess. There's an easy fix to that, though. Go to render properties, depth of field, and take the Jitter camera setting there. And then this issue should be fixed, but only in the final render, not in the viewport. With this simple fix, all render passes, including AO will be rendered with Jitter depth of field. Another critical way of adding ambient occlusion effect in IV is light probes. So far we explore ray tracing, random passes. And the third method is to use volume light probes to make the ambient occlusion. But wait, aren't light probes supposed to be used for adding the extra light to the scene, the bounced light? Well, yes, but they're also good for subtracting or occluding parts of the environment light to simulate self shadowing. That will work better in the static scenes where lighting and objects aren't moving too much. But it's pretty cool nevertheless. All we need really is a volume light probe, which encapsulates whatever we want to make the occlusion for the astronauts in this case, obviously. Something like that. And in its settings, I will increase the resolution greatly to make it really precise, and I will bake the light case. I just took a few seconds. And yeah, it does something. Actually, if you can't see the effect, it could mean that there's not enough memory for light probes. So let's check the performance memory tab and dedicate some more memory to light probes. Yep, see? That did the trick. So now we can turn it on and off to see the occlusion effect. There's not much bounced light, so it mostly creates the occlusion. Now, the surfel resolution can be brought much higher for better quality. And now, actually, before baking it out by pressing this button, what you want to do to be able to manipulate the intensity when working with the environment light is tick the world checkbox and then bake because otherwise, the intensity won't work. I mean, this one. Yeah, so that's the light probes AO method. The catch with light probes is that they don't work with animated measures. For that, you'll want to use a tracing method instead or cycles for that matter. But for mostly static scenes like this forest one we are about to dive into, they can do wonders when it comes to simulating ambient light and ambient occlusion. Right, so right out of the box, Cycles handles ambient lighting so much better while IV looks super unrealistic with the factory settings and needs more bells and whistles. And as we've pointed out already, the most tremendous bell and whistle for that realistic overcast weather lighting is ambient occlusion. And there are three methods of creating ambient eclusion effect and Ib we can choose from or using combination to get that realistic look. So by default, Ib's ambient lighting looks so flat and boring. Nothing has contact shadows just yet, not even these leaves. Let's reiterate that. The first thing that you can do is use ray tracing. And if you want to see just the shading, you can go viewport settings, and actually, it's not the ambient eclusion pass. It's the diffuse light that shows the ray traced AO. In racing, ambient eclusion is kind of built in. It's part of it, so it shows in the diffused light pass. Okay, there's definitely some contact shadows visible now. It's a lot better. Yet, what else we can do? You probably remember from other videos that these guys should go denoising and fast GI. Then these two, and the resolution are the only things left. Resolution is no brainer now. Low residue doesn't really allow for much intricacy in the contact shadows. It becomes pretty gross and pixelated, so it should be set to at least one to two or preferably higher if you want to nail these tiny little shadows. Precision and thickness can be safely cranked up too for a deep, far reaching occlusion effect that usually, in my experience looks way more convincing with the higher thickness, a more substantial, you know. Here you can see the comparison of different levels of thickness. Here's what denoising doge to the retraced lighting. Smushy, smushy. It's the temporal accumulation but also spatial reuse that are to blame. So it probably has to go if we're aiming at photorealism here. As a quick reminder, we are still in the diffused light preview mode. Combined is the default view if we want to get back and see with colors, shaders and stuff. It's looking great, by the way. Who would have thought that IV can produce such a picture? Now, really? Anyway, next, I'll plunge back into the diffused light preview and turn ray tracing off temporarily to add some light probes instead. Baking light cache makes most sense for the static scenes. But, hooray, this is apparently a static scene. So we'll have a field day with that. Shift A light probe volume. It needs to cover or encapsulate the part of the scene that you want to bake the illumination cache for. As for the resolution, I'm going to set it crazy. I know my computer gonna handle it, and I don't have a super powerful computer, so it should be fine. It won't be a good look if blender crashes now. That would suck. I want to check the performance menu to make sure there's enough memory for probes. No 16 megabytes, but 512 or something. Now, I'll just yeah, the surf resolution can be higher and the world contribution turned on. And now I'll just hit bike Now, look at that. Already looking pretty good with all the shading info appearing out of thin air. Some of these parameters like dilation threshold can be tweaked after the BC is complete. If you set it to one, then you'd be able to use radius to blur the whole thing if needed. Then you can tweak the biases for certain artistic effects. It's not that they are right or wrong values. You tweak it to either fix errors or make it look in a certain way. The bottom line, it's good that it's fairly tweakable. So far, we've set up two different flavors of amen occlusion for this scene, each with its own quirks and advantages. The good news is that these two can be used in combination, and it's where it starts to be really different from cycles. It is as simple as turning on right tracing on top of the light probe, which will create an interesting mix of these two methods and a pretty intricate indirect shading. Then in the light probe settings, what we can do is tweak the intensity of the probe to make the whole thing brighter, for instance, and that will create a slightly different look. It's where it gets really interesting. We can start playing with these settings to make the occlusion slightly more or slightly less prominent, you know. This is where you can nudge it wherever you want, really artistically. This case, I would actually skip on using the ambient occlusion rando pass in compositor, I mean. I think it will just make it a bit too strong, as we're already using two different and two pretty powerful AO methods, although it is completely up to you. You know, you can still overlay the ambient eclusion rando pass on top of everything else. Did you know using all three methods of ambient seclusion in IV at once earns you the fabled three way AO badge. True story. This is it from me. We started this video with marveling at how cycles is so much better at rendering ambient light in the IV right out of the box. But after learning about ambient seclusion and its different flavors available in IV, we actually managed to make IV look just as realistic, or one may say photo realistic or beautiful in its own way, anyway. So thank you so much for watching. 16. Realtime Reflections: Compared to cycles getting reflections right in IV can be a bit tricky. Screen space rate racing almost gets you there, but sometimes it still looks weird. IV reflections can sometimes just completely vanish when you move the camera to a different angle or they might look a bit softer and mushy instead of crystal clear to the point where you start doubting whether IV can do it at all. Here are a few tricks to get clean, stable and most importantly, realistic reflections in IV. As usual, let's first see how it works in cycles for reference. Right, here we are in cycles. Immediately, it's pretty noisy this particular scene. So I'll turn on denoising to just get a clearer look. Then I'll jump into light bounces and set all bounces to zero to add just one glossy bounce after that, a single reflection bounce. So that is pretty much how we expect reflections to look in IV. I mean, IV can only do one light bounce. That's just how the engine works. It can't do two or ten or reflections within reflections like cycles and other path tracing engines do. Multiple reflections are the domain of unbiased path tracing engines. They are way slower than IV by nature, but they are amazing at physically correct light behavior. So if you want that level of precision, cycles is your friend. What I can do is render a pretty convincing single reflection or just glossiness at real time speeds. So let's take a look at a few tips to make it look great. You'll find this scene in the project files, by the way, if you want to follow along. It should look like this, basically a strip down room. When you open the file, you'll discover that right now there's only one light source in here, this baked volume light probe. In other words, the environment light has been baked into light cache of this probe, but just the diffuse light because IV's light cache doesn't store reflections. Okay? Loaded the scene. IV has a few types of probes. The volume probe is just for global illumination, mostly diffuse. The others are actually meant to capture glossy reflections, and we'll get to those later in the video. Hey, ray tracing is the first thing you want to enable to get reflections going in IV. Swoosh. And instantly you get like 80% of the cycle's look. Thanks for watching. See you next time. Just kidding. But seriously, it's kind of amazing how good it looks right away. It's not perfect, though. Look how IV's reflections vanish in the certain places. Whenever something isn't visible to the camera like the back of the red monkey head. Ivie simply has no data to reflect there. We'll fix that in a moment, but this is already a strong start. Now if you expand the ate raising panel, you'll see it split into two sections. This one controls glossy reflections, think mirror like surfaces. And this one below, fast ja approximation controls diffuse reflections. Basically, everything that isn't shiny. And this red slider is the mixed factor between the two. Before tweaking glossy reflections, I'm actually going to limit the diffuse bound distance a bit. Otherwise, if we can accumulate diffuse light too strongly at certain angles, I think it might be a bug. Anyway, if you notice weird bright spill or over accumulated light, maybe in really bright environments or something like that, reducing fast GI distance fixes it. Right, that's just something I wanted to mention really quickly. Alright, glossy reflections quality is controlled right here. For instance, changing the resolution here affects mostly the glossy surfaces, or you can say mirror like areas. See what parts got pixelated. That's exactly where the glossy reflections happen. And which part count as glossy is determined by this threshold slider. At zero, only the sharpest reflections are considered glossy and passed to screen tracing. And at one, pretty much everything is treated as glossy. Fast GI even gets grade out because screen tracing takes over fully a more truthful but also noisier rendering method. So the purpose of this fast GI approximation is to actually limit that noise and limit real time screen tracing to only the sharpest reflections and use the faster GI approximation pipeline for everything else. The LDR, Screen trace excels at rendering glossy surfaces yet struggles with diffuse, and FAS GI is great for rendering diffuse surfaces and cannot do reflections at all. And a threshold is a mix factor between the two. FASGI also has its own resolution, but it doesn't affect reflections at all. It's mostly for AO and global illumination, so we can ignore it for now. While this resolution, that's the reflection resolution. If you ever see a weird border around the objects, like a few pixels wide outline or something, switch to one to one pixel resolution. Now it should be pixel perfect. Here's a comparison. Look at the weird outline around the edges at 18 resolution and the pixel perfect outlines at 121. Again, screen trace is essential for reflections. Don't ever switch it to the light probe mode, or your reflections will just vanish like my coffee. Apparently, I run out of coffee. Tragedy. Anyway, moving on. Comparing it to cycles again, you'll notice reflections in cycles feel fuller and more complete. Like, look at that monkey. While in IV, some parts of the reflection just aren't there. That is because IV is screen space, meaning if the camera can't see the object, IV can't reflect it. Change the camera angle like that, so these parts can now be seen. They will also appear in reflection. Mm hmm. Just look at these parts they disappear from the camera view. They also disappear in reflection. I mean, these parts right here, And now they are gone. Just a fundamental limitation of screen space effects. But IV gives us a special light probe designed to fix exactly this problem, plain probe to show what's going on with floor reflections, I unplug the environment light. Now, if you look around, you may notice that from some angles, it looks okay, like from this angle. But if you tilt the camera down like that, a black patch creeps into the reflection. Pretty strange, right? Especially if the floor were fully metallic. That would become painfully obvious and really, really bad. The unintuitive thing about screen space reflections is that when the ceiling is on screen, the floor reflects it. Okay? When the camera can't see the ceiling anymore, if it has nothing to reflect and defaults to the environment color, which is black right now. To fix that, I'll make a new collection. Name it reflection probe, plane. Then I'll press Shift A and go light probe plane. Now you can scale it so it covers the entire floor. Actually, let me extend it a bit. Mm hmm. Now press G and Z to move it up or down to position it correctly. Then in the probe settings, you can adjust the distance, so it only spans the floor's thickness. In wireframe mode, this distance or thickness will be represented by these stacked layers. I like how Andrew calls it pancakes. It definitely looks like pancakes. So we want these pancakes to just cover the floor, and that should fix the reflection. So yeah, this is our reflection probe, which just fixed the reflection for us. And if you click capture, you'll see what it's actually sampling. It could be useful for debugging. Now I'll just revert the material to non fully metallic one, just a little bit metallic to retain that clear reflection. So now it shouldn't matter where you look that probe actually captures the scene in a world space, not in screen space. Think of it as an invisible camera pointing up from that floor and capturing the entire scene from that perspective. You can toggle it on and off to compare how it looked before. We had that black void and after. Now, for instance, if you want to enhance the reflection in that wall as well, you can just add another probe. So again, shift A, light probe, plane. So simply rotate it into place and basically line it up with that wall, hide and on height to see the difference. So multiple light probes can work at once. Just remember there's a performance cost associated with it. So use them only where it matters. Pretty cool. See, this probe captures that view of the scene, you know? How am I supposed to work without coffee? Actually, that's also how you do mirrors in IV. Now with that probe in place, move a plane over there and assign a perfectly reflective shader to it, zero roughness 100% metallic. Just like mirrors. If you place it such that the probe covers it, you will get a mirror. There's one big nasty limitation, though, if you won't reflect other reflections, as we've mentioned before, compared to cycles. Again, cycles can handle true multiple glossy bounces. Actually, let me show it to you. Zero means it's black. One is a single bounce, and then cycles can do as many as you want for true reflection chains. So yeah, IV definitely lags behind cycles in that sense. It is fast, but it just isn't that sophisticated when it comes to reflection accuracy. Back to IV, everything in the mirror now becomes diffuse, which is a bit of a bummer. But it works nevertheless. Incidentally, that's also why games rarely have mirrors everywhere. I know you like comparisons, so here's another one for you. Multiple bounces cycles can do. We can dig one Besides plain probes, there's also sphere probes in IV, and like its name suggests, they are used for 360 degrees panoramic captures. Light probe sphere. Let me hide the mirror. Okay. And here's a reflective sphere. Notice the black void again, even though it should be reflected in this room, right? It's also a pretty easy fix. Select the sphere, Shift S, cursor to selected, so it's in the middle and add a sphere probe. Dster scale to about that. Actually, it's the same as tweaking its radius in the settings. It should roughly surround the object like that. So that should remove that black void. Now it should get you roughly comparable to cycles, not identical, but close enough for most use cases. You can copy probes to other objects, too. You know, maybe to this one instead as it's more reflective. G to move, S to scale, blender hot keys alive. And I think it is looking great. Definitely much more believable than before and much closer to cycles, for that matter. With all three probes combined, the gap between IV and cycles gets much, much smaller. Lastly, if you want to tweak the resolution of the spherical probes, you'll find it in the scene properties, light probes. And here we go. In my experience, you rarely need super high resolution. It usually doesn't make a visible difference. Well, maybe in lab conditions, it does, but practically no. Especially if the objects themselves aren't perfect spheres, and the reflection gets warped anyway. Or if the objects have roughness texture disturbing the reflection. We haven't talked about that a lot, but when the objects have roughness texture, it can make it really hard to tell if the reflection is right or not. Mm hmm. Even if the resolution is low or probe is slightly off, nobody could tell once materials get complex, or maybe it's better like that, you know? In a scene with some materials and stuff, the inaccuracies of IV are kind of masked. See what I mean? In a way, that makes life easier. There is an important tip for roughness maps. If a floor or other material looks mushy or smudged, that's likely not because of the probe resolution. Why such materials as smudge it comes down to two things. First, it's screen tracing resolution, not probes or anything. Naturally, the higher it is, the crisper going to be the reflections, and it should peak at one to one, but it doesn't, and here's why. It is denoising that kills the clarity. Specifically temporal accumulation but also spatial reuse. Turn those off and reflections become razor sharp. So if you need max clarity, disabled denoising. Just turn it off. That's the trick to the crispy specular reflections in navy. Alright, last tip before I go get some coffee. Light leaks. You know, when environment, color leaks into reflections like a parasite. The fix, use a light path node. So you want a bright outdoor environment, not black, so you plug the environment light back in, and you get what you want, but your reflections get polluted. To fix it, you can add mixed shader and another background shader now, which will have zero intensity. I'll also give it black color just to make it clearer. Anyway, so that is the second shader in that mixer, and now just add the light path node and plug the e camera ray thing into the mix factor. So now it's black for the camera, and it should be other way around. So I'll just switch these nodes around and reconnect them like this. It's true. So now for the camera, it's a bright day. And for the actual illumination, it's nothing. It's void. So there's no environment lighting now. So this technique solves the light leaks, and crucially it frees you up to choose any color or any HDR texture you want for the outside environment without actually messing with the lighting. TLDR. IV has different types of light probes. Volume is great for capturing the diffuse light, and the plane and sphere probes are great for glossy reflections. We cover volume probes elsewhere. For reflections, the main things are enable ray tracing, watch your resolution, and focus on this panel screen tracing settings. This is what gives IV its glossy reflections. Plain probes are great for flat surfaces to capture their reflection right. Sphere probes are great for everything else. Put all those tricks together, and the will be almost as correct as cycles. Now, it's not entirely correct, and, yeah, we're definitely carrying a bunch of limitations, but otherwise, it looks absolutely ace. 17. Shading EEVEE Glass: Ready to finally make glass in IV that doesn't look like cheap plastic. Here's your IV Glass guide. The truth is, IV can create a pretty convincing glass if you know which switches to flip and which has to use. So in this video, we'll tackle all the classic pain points, including the cursed glass inside glass problem. We'll break glass down into simple building blocks like reflection, refraction, roughness, and so on, actually using cycles as a reference, and then assemble five different shaders from clean, simple glass to full on the life orb complexity. Before doing EV glass, let's take a look at cycles as a reference. So switching render engine to cycles for testing it. I'll add a UV sphere. Right click smooth. And of course, now we need a simple glass shader, which in cycles is really simple, indeed. It's just glass, literally. Done. In path tracing engines like cycles, that's essentially all you need. You might throw on a solidify modifier to simulate actual wall thickness, so it's not just a solid sphere of glass. This is pretty much the whole setup. You're done, soldier. And to add insult to injury for IV, I mean, Cycles has absolutely no problem rendering multiple layers of refraction. In IV, that stuff is tricky. So to succeed in our mission, we need to manually create a checklist of essential properties in IV transparency or how much light passes through, refraction, how light bends inside. Reflection, how light bounces off roughness, how light scatters, thickness, solid object or shell, thin film, all the chromatic effect, and texture. Fingerprints, dust, or subtle surface and perfections that the game marketing departments always point out in their tech demos. Oh, look, attention to detail. Breaking glass down this way, makes Shader creation much easier. So in this video, we'll set up five different shaders from simple to more complex. Some will use only a few of these elements while others will take almost every single checkbox on our list. Starting with level one glass shader the easiest one. Quick reminder, grab the project files if you want to practice with me. So a good practice is to delete that default shader and start by adding a glass BSDF. That seems pretty intuitive so far, right? So this is EV glass. Mm hmm. These different types won't really work in IV, so don't bother with that. And before we tweak anything, we have to enable a few key things in the IV render and material settings. So turn on race the render tab, and next in the materials panel settings, turn on ray trace transmission so light can actually travel through that object. But now we have some refractive action going. Look how refraction is turned upside down, just like it should be. In addition to that, you need to make sure that random method is set to dithered and not blended. Otherwise, refractions won't work properly. Another parameter you can tweak is the thickness mode, which can be set to either sphere or slab. Slab will assume it's a flat object, essentially. So now it'll look more like a lens than a sphere. Notice how the refraction isn't upside down anymore. It isn't better or worse. It's just a completely different rendering method. For this simple level one shader, let's just keep it simple with sphere. Over in retracing panel, resolution is important because it literally controls the resolution of the rays passing through the glass. So at lower values, you'll get a really pixelated glass look, and you definitely don't want that. So setting it to one or two or even maxing it out is no brainer here. As for denoising, you can keep it turned off as it can sometimes make a glass a bit murky. It's fairly hard to see here, but once you add textures like a roughness map, for instance, it will definitely look better without denoising. Now fast GI approximation, it is completely irrelevant for gloss and glassy stuff. Remember, in EV, the pipeline is split. Screen tracing is used for glossy reflections while fast GI is resolved for purely diffuse surfaces. We cover it in detail in the reflections chapter of this course. It's a long story short, we don't care about fast GI in this one. Also, just ensure the rendering method is not switched to light probes accidentally or it won't work properly. One more basic parameter you can use for this level one shader is thickness. So I'll just add the value node. And plug it into this parameter into thickness. You can use this to simulate how thick is this object, basically, for a variety of interesting effects. This is a wisdom right here. Not a lot of people know about this. I got a mixture between Deckard cane and Yoda here for some reason. Anyway, I'm going to skip it for now, so it uses the object's actual dimensions instead to keep it simple. Then color is how much light gets absorbed while passing through. Usually you want it very transparent, but you can also add a colorful tint. And roughness, this blurs both the refraction and reflection. Great for adding that frosted glass look. I will keep it crystal clear for now. It seems we've already covered the basic properties like transparency, refraction, reflection, roughness, and even a bit of thickness. These are really the basic building blocks of the glass shader and so far so good. Yet, there are some things that are slightly more advanced. Like the IOR or index of refraction. This allows you to simulate different substances like glass, water, alcohol, different minerals, and so on. At one, the light rays will just pass through like air, and any other value will make the rays bend in interesting ways. It's pretty influential. You can see how the material looks completely different depending on the IOR. So it's a pretty important thing for tailoring the look. You can find IOR values for common materials online to simulate whatever, fused quartz. Maybe coffee, my God, or wine and beer, or even stuff like polyethylene and sapphire. Back to our simple glass shader. The other thing that can massively improve the look of IV glass are light probes, sphere light probes, for instance, as to scale, so it should be lined up to the object like this. It doesn't have to be exact, but if you want to be precise, you can shift, click the sphere, and then Control C to copy its location. What this probe does is sample a spherical panorama, basically an HGI of the scene to help render a more correct worldpace reflections. The radius can also be tweaked here in the probe settings alongside fall off and some other settings that aren't that important like clipping. Now the reflection of the environment should be a bit closer to what you'd see in cycles, for instance. Seem much more precise. Alright, that's how it would look with the probe, but I'll delete it for now. You guys, why to keep it simple. Alright, that's the Level one glass shader for IV. For level two, I'll duplicate this shader that we just created on a separate object, I mean, and next we can experiment with the thickness mode. Let's switch it to slab to see how that would change the refraction. Slab naturally better suits flat or slab like surfaces, but you can use it for anything else for creative purposes. If you do it, make sure to control the thickness in the shader. This is often overlooked. Don't make this mistake. Lower values will make it look like a super thin cellophane or I don't know, a thin polymer film, and larger values will make it look more like a lens. It's pretty weird, but it can be useful, like, depending on the specific look you're going for. And you can still adjust the roughness, however you want, really, and the IOR. And again, the thickness. So it's mainly between these parameters where the main look death happens for now. So now it's become a thin sort of glass that bends the light mostly near the edges. So that is another totally different glass like material you can achieve in IV. You know, sometimes you might want to make something like epoxy resin or other types of similar plastics. So the slab method plus increased roughness can give you some really interesting results. The problem with IV glass is that while it may look pretty good when it's just refracting the environment like that, it falls short when it comes to refracting other glass objects. And that's especially painful to see compared to cycles, which naturally has no problem doing that. That is just a limitation of IV's rendering engine, but there are ways around that, and we'll look into those in just a moment. This is quite important because the scenes where you might use glass shaders often involve liquids within glass. Whiskey in a jar, you know, milk in a glass, et cetera. Level three of the dungeon, presumably. For this one, I'll copy our first shader, the one using this sphere thickness mode. Uh huh. Here we have this new object with a fresh material. A is good. Alright. A nice little trick that often makes glass look pretty cool is to use multiple glass shaders for that extra bit of visual complexity. So I'll duplicate this one and then add a mixed shader then we'll use the layer weight node for mixing between them. Specifically, it's facing output. So now you can give these two shaders different IORs. So this one can be 1.5 or something, and this one below one. This makes the center and the edges of the sphere actually behave like two different types of glass. You can do some weird stuff like that. TLDR, the layer weight node is a black and white mask based on the surface angle relative to the viewer. You use this mask to blend between two types of glass or two shaders, generally speaking, to get a more complex looking shader with different IRs, roughnesses or even colors, as the surface goes from facing the viewer to viewing it at a grazing angle. It is a surprisingly simple and surprisingly powerful method of achieving vastly different looks. We're not using it as bluntly here. We're not mixing crazy colors or roughnesses, but still it adds a nice nuance. I'll move it there, and now I'll hold Shift and right click slash this connection to reroute it a bit to add a color ram here. This lets us fine tune that mix factor that fall off. This is how it looks when visualized. I'll control shift, click it, priv it. Basically, it's an angle based mask. Black is the first shader and white reveals the second one. So we can fine tune the reflection and refraction properties exactly how we want. You have to try it to fill it. So this is level three, if you want this kind of extra control. See? We're slowly adding complexity by introducing new blocks or new ways to combine the blocks. In Level four, we're going to tackle the glass inside glass issue that literally haunts IV users. Basically how on earth to do it and where to even begin. We're already familiar with the concept of angle of reflection. Sometimes it's called frenll effect. This physical effect makes dielectric materials, including glass more reflective at grazing angles. Even in cycles, if the glassy object isn't fully solid and has walls, both the reflection and refraction will mostly appear near its edges at grazing angles, leaving its center which is kind of facing us, mostly transparent. And we can use that to our advantage to implement one of the craziest IV hacks ever. Okay, our next shader is also based on the glass BSDF and now I'll add the mixed shader again. And we'll mix it with it. Transparent shader. Mm hmm. The transparent shader just does that. It makes things transparent without bending light or anything. And Iva has no problem rendering multiple layers of transparency, by the way. So it goes like that. You just mix in some transparency to the shader. And then you guessed it. Use that layer weight node. Layer weight. And it's facing output to make the surfaces facing the viewer transparent. By the way, there's also a dedicated frenll output for that. Yet the face output does practically the same thing. The mask seems to be inverted now, masking the edges rather than the center. So you can just swap the inputs, I think. And it should be good. Now, it's just a matter of fine tuning the fall off of that transparency. Alternatively, let me swap it back. Alternatively, you can use the color ramp node with the flipped flags. This will also allow you to calibrate the transition between these two shaders. And if you want to tweak the amount of transparency, that white flags color is the maximum transparency, so you can just tweak it. You can tone it down, and it will create a pretty cool half transparency effect, like seeing multiple things at once through that glass, which I think is really, really, really cool thing as it adds layers. That would be 100% transparency in the center, and this is 70 for some extra layering. There is no right or wrong here, just some pretty cool options to choose from. Lastly, for this level, you can add a solidify modifier on top to generate some actual thickness for this sphere. That will create even more layering of refractions and hopefully make it look even more believable, simulating that important property of glass. What does it gives us, including the transparency component and all that? Well, it's actually tremendous. This is our essential workaround for that limitation of IV where you just can't render glass through glass natively. This is the best hack ever. In fact, IV will have no problem rendering multiple layers of glass once we give it the solution or hack, if you wish. It's worth noting that it won't be true refraction in the sense that it won't be physically correct, but it will work. As a nice bonus to using transparency, the shadows will also become transparent, which is honestly pretty cool feature. Too bad if we can't rent a colorful transparency yet. But maybe that's something that will be added in future versions of Blender. So far, we've tackled four levels of making glass shaders, introducing new things, and ramping up the complexity. Now we have to deal with a boss on Level five, with diable of glass shaders, where all these concepts will be brought together. We will create a glass shader that ticks all the boxes on our list of things that make glass glass. 18. Realistic Frosted & Colored Glass: Alright, level five, the Uber glass shader. Alright, folks, we created a lot of glass shaders already. Let's create one more. The ultimate fusion. I'm starting from scratch with a new object and a new material, and I'll delete this one right away and start fresh. This time, instead of glass or principle BSDF, I'll start with refraction, going straight down to basics. Swoosh. This is the basic building block of glass and isolated refraction. I'll duplicate it right away, 'cause we'll need two of these and we'll need them mixed together. Okay, now shift again and bring in the layer weights. That's going to be the mask between the two. It's facing output. Remember what it does? It's our angle based mask. We'll need a color ramp to control the fall off one of my favorite nodes. I'll invert it now. And actually, now we can set the index of refraction for each shader separately. So maybe. Something like that for the edges, 1.45 and this one actually below one to get this sort of diminished image in the center, like a subtle fish islands effect. I think that's interesting. Feel free to play with these values with IORs and with the mask. I mean, with the follow anyway, you don't have to follow along exactly, but I think something like that looks pretty cool already. I'll put these nodes into a frame with Shift P and open the toolbar and add a label, refraction, because this is our refraction. Just like in our reference, remember, one of the most basic building blocks. Next, I'm going to grab another mixed shader. And what's coming is tremendous. Mm hmm. I'm going to use that transparent shader in the second socket. That'll give us the transparency for the angles that face the camera, which is essential for that glass inside glass problem. Once again, it solves the glass inside glass problem. Alright, just in case. Now I'll just copy these two guys with Control C and paste them here with Control B. That'll be our mix factor again or mask. Coloram tunes the transition and how gradual that needs to be. This blend factor is the angle threshold. Use your artistic eye for getting the look you want. I think it looks pretty cool when it reflects a lot of stuff near the edges. Like, it has this thick layer of reflection. Actually, maybe we can reduce the maximum transparency to add some visual layering to the center. Let me enlarge that for you. See that doubling effect near the center? Yeah. That totally works for me as a spice, you know. So this was transparency, and we already have refraction. Now we just need reflection. That should be pretty easy. Let me create some room over here, then shift A and add another mixed shader. Then shift again. And we can choose glossy, but we also have specula, which has some pretty interesting features. So let's go for this one. Plug. For tweaking its look, I'll temporarily set mix to one. So we just see specula. It looks like a pearl now, which is cute, but I'll zero out its roughness and make its base color pure black. Actually, it specular all the way to white or 100%. I'll copy that trusty lay weight plus color am combo again over here and plug it into the mix to make that reflection or specularity appear only at the grazing angles. I think we need to reverse it. There we go. Now it only appears at the edges. I'm tweaking these parameters to taste, not following any physical rules or anything like that this time, at least. Okay? This is before and after. See this nice reflection running around the edge. That's what color ramp is for if you want to tighten it up or vice versa. Oh my God, I could spend hours tweaking this stuff, I tell you, probably going in circles, you know? This pecularity can have its own tint. If that's something you need, maybe I'll just give it a slightly. Off white, beige color. Oh, actually, it can also be brighter than white for non physically correct effects, you know, way exaggerated brightness and all that. But I'm going to stick to the energy conservation law. Yeah? Let's not get carried away. Another way to infuse it with color is to tweak the base color. It can give it some nice glow almost, some sort of I don't know, some sort of washed out quality to it, almost like imitating some chromatic light scattering. Not physically, I mean, but visually. Of course, I'm not talking about more physically based light dispersion, like what you find in the small shades ads for blender, like splitting the spectral frequencies and all that. But we can at least add a hint of that color shift near the edges by offsetting all colors ever so slightly. I mean, by making them off white rather than simply white. So this is our reflection, reflection block. And so far, we added three blocks, refraction, transparency, and now reflection. You know, it's kind of amazing. We can get a whole variety of looks with just these three blocks. Okay, which building block comes after reflection? Roughness, and also thickness. So let's sort thickness out first. For thickness, I've just added the solify modifier that gives it some physical wall thickness for that nice multi layered look. Now, for roughness, you can use any texture you like. It can be a random grunge image or whatever. It's best if it's seamless. And just plug it into the roughness input of any shader you want. For instance, both refraction shaders. And let's not forget about the specular shade roughness. Where is it? Okay, found it. I need to admit this Uber shader is getting pretty messy. I want to make this roughness effect more noticeable. But first of all, I'll hold Shift and we click slash over these three lines because they will give me OCD. That will add a re rote. Now, I want to boost the contrast, so I will add the color mp. I'll preview it first, and then crunch it down very hard. So we get that super high contrast. Yep next up, I'll find the main shader and control shift click it again. Actually, I think I ended up fine tuning this contrast by flipping these guys eventually. It just looked more wholesome that way for some reason. Maybe it gave the texture more satisfying feel or something. Anyway, it should look a bit more like that now, like a frosted glass. Usually, you don't want it to be that obvious, the roughness variation effect, but here it works pretty well. As a finishing touch, the refraction colors here and here can also be made off white. That will add that chromatic flavor to it. I mean, like chromatic aberration. To make it slightly more interesting, I will add our trusty lay awight note again. To make a color gradient from it, we just need a color ramp now and the facing output. What I want to do is basically colorize it using this color ramp. This black flag will become, I don't know, bluish gray. Then I'll hold control and start clicking like mad to add a bunch of stops and assign a different color to each symbolizing different wavelengths of light. This will be our Pooh man's chromatic aberration. Finally, I'll plug this resulting color. Oh, it's getting messy into both color inputs of our refraction shaders. You know, this effect can be reminiscent of the thin film effect as well. You know, this rainbow like shimmer you see on things like soap bubbles, oil slicks on water, camera lenses and so on. So maybe it can be this. This gradient can look anywhere you want and go anywhere you want. You can absolutely turn it into a gasoline on water type of effect if that's what you want. Or you can keep it like I'll do slightly off white, you know, just a little bit of color as if the glass is a bit old, and that's it. So this is white, and this is slightly off white. See the difference? H. I called this effect thin film, but you can call it dispersion or whatever you like, really. But really, it's just a little huck. Anyway, there's only one thing left to deal with, and that's thickness. Wait, didn't we already tackle that with the solidifying bodifre? Well, yes, but here's the catch. If you drop another object inside this sphere, say, to make it look like a diablo style life or things can start to look a bit off. Something about the thickness feels weird. And remember, materials have their own thickness input, too. Easy to forget, right? Just give it a value. And you can tweak it until it feels right, like so many other things in Ivy. Here's what zero looks like. No thickness, but the inner object sits perfectly inside. And actually, that could be pretty useful for multiple Matrohka style nested objects. Or, honestly, just play around with this value until you find what works best. With IV, it's always that back and forth game tweak something here and watch something break over there. But hey, at least you're not descending into the catacombs unto the abandoned cathedral to take on the butcher or whatever. Right, that's all for me for this video, hopefully takes some of the paint out of making decent looking, believable glass in IV and helps you navigate those headaches like glass within glass and similar IV quirks. Here's the big picture. Cycles give us perfect glass right out of the box, but in IV, we have to build the illusion ourselves. And the way to do it with confidence is to break glass down into its core ingredients, transparency, refraction, reflection, roughness, thickness, thin film effects, and just a touch of texture for realism. In this video, we took these ideas and turned them into five shaders, from simple to advanced. So now you know exactly how to tackle IV glass and blender instead of just guessing at settings. 19. Lighting & Rendering a Realistic Bottle Scene: In this tutorial, we are taking on this beautiful bottle scene. We'll break down the cinematic lighting setup and figure out how to make it look both real and seriously eye catching. First up, you might be asking, why not just use cycles for glass? That's a fair question. In fact, this render started a few years ago as a test of cycles glass rendering. And let's be honest, cycles as a path tracing engine is the obvious go to for complex refractions. It's pretty much built for this. The downside of cycles is that it can still be pretty slow or pretty noisy, especially on mid range or the GPUs. And if you rely on heavy denoising, things can get a bit muffled or a bit soft. That can make producing longer animations a real challenge. Meanwhile, IV breezes through this at interactive frame rates, making it easier to work with animation, letting you navigate your camera in real time to nail the fill of your shots is just a different workflow, even if it's not 100% accurate. IV is all about smoke and mirrors. In IV, convincing glass boils down to three pillars. Your shaders, your render settings, and most importantly, your lighting. In the previous video, we learned to create glass shaders, and in this one, we'll focus on render settings, but mostly on lighting because lighting is absolutely crucial for anything reflective or refractive. What it doesn't boil down to in IV is physical accuracy. Mm hmm. As always, this scene is available in the project resources. When you open it, you'll see the main group of bottles here and some standalone objects over here. The render engine is naturally eavy. For previewing glass, sometimes it's better not to go straight to rendered view because with a default gray backdrop, the glass has nothing to reflect. So if you want to just see something really quickly, just go to material preview. At least it has some basic environment built in the default one or any other from that selection that you can spin around to get a quick feel of how your glass will react. That is even before plunging into the rendered mode. Now, these objects over here represent a bunch of different glass shaders. Let's open the shader editor in the top window. So each of these objects holds a separate glass shade. Glass Shader one, which is a simple glass BSDF then, for instance, glass two, which is the same thing, but with a thickness and the thickness mode set to slab instead of a sphere in the material settings. I mean, these are essentially the glass materials we built in the previous video. It's literally the same guys. I just reused them here. So if you want to see how they were created, watch the first part. This third one is a mixture of two glass shaders. The point is, all of these are a little bit different. Then the fourth one introduces some transparency, so on and so forth. We won't go into that right now. We also have a bunch of liquids there, which are also glass shaders, essentially. The important thing is I've arranged these bottles and randomly assigned the different shaders to each one. Then if you dig into these collections, you'll also find a few cameras. You can switch between them from the scene properties panel. You can select camera one, for instance, and then press zero on the on pad to get into the camera view like that. Or you can switch it to that second one which has a bit wider lens for a wider field of view. And the third one is actually animated. If you select it and hit play, you'll see it move. This animation will be super useful later. Motion reveals the qualities of glass better than a static shot, as the refractions tend to shift and change from every angle and all that. That will be very, very useful. For the time being, I'll switch either to this first camera or to the wider one, perhaps. Yeah. All right. The first step is to go to the world Shader and add the environment texture there. So I'll switch to the rendered mode now. Now we are officially rendering, and I'll actually turn off ray tracing just for a second. Without it, it will look completely flat, as ray tracing gives it something to reflect and refract, but that's okay. I'll click the background Shader now and Top Control T to conjure up these three magic nodes. For the environment texture, I'll select the circus backstage HGR file from Ply haven that I downloaded like 50 years ago when GR format was still cool. And instantly, we can see light bending action. Is it reflection, refraction? Hard to tell. By adding this 360 degrees environment, we gave our glass something to interact with and suddenly starts to look a bit more like real glass. Next, I usually rotate it on the z axis to try to find the most flattering angle, whatever it means. Um, maybe from the camera view, it will make more sense because it's view dependent. Of course, we still have ray trace in set off, so the refraction isn't perfect, but we still kind of see its approximation. I often switch all the UI off in the view portvoe to remove the distractions. It's so much cleaner without it. Now, I don't want this HDI to be so powerful, so I'll reduce its strength down to, like, 0.06 perhaps. So it's there, but it's not overpowering. So it becomes sort of a base layer for refraction, a room tone, one can say. Now with ray tracing turned on, things should get much more grounded and much more realistic, too, for that matter. We'll tweak these settings more in a bit, but for now, I'll leave them. Yeah, it looks more believable, not very cinematic. As in what? It's pretty flat. If we have a recurring theme in this course, it's exactly that. Flat is bad. Now, just by looking at the comparison, how to make it non flat? That's right. Lighting. Specifically, a reverse key light coming from behind the subject. It's probably the easiest way. Let's switch back to the three D view board to add a proper backdrop to begin with. Not that I dislike the HDRI or whatever. I just want to control it separately. So I'll create a new collection for the background plane, then shift A, and add a plane and immediately scale it up and basically nudge it to the back of the scene. Then I'll use a mortal combat style shortcut, a XX, 1.5 to give it a wider dimension. This is our backdrop. I'll give it a new shader via the materials panel. Through background. It'll be a simple emission shader. I'll turn on the material privy in the top window to fully see it here. But the color, I'll use an image texture of some kind. Open and select the photo that you like. I'll go with this public restroom photo from on splash. It doesn't have to be super high risk as there's some depth of field in our camera settings anyway, you know. Okay, this is our backdrop. I'll scale it up a bit more to fully cover the shot. Okay. Now back to the shader editor. Object. The emission strength can be way more substantial, so this can pump the exposure back into the scene through these bottles and stuff. That will look gorgeous, I tell you. Then right now the color of the backdrop is way colder or cooler than our HDRI. So I'm going to add the mixed note to bring it over to the warmer side with the help of the color blending mode and the warmer color in it. Mm hmm. And just to bring it over to this side of the spectrum as far as the temperature goes, I don't want to have this burnt sort of gamma, though, so I'll dw the mix factor back to about 0.2 or 0.3. I don't know. Uh huh. That looks good. By the way, the multiply mode would also work. These blending modes are very similar. We're just trying to unify the color palette and colour temperature a little so everything works together. Look how this brighter backdrop changed the entire scene and the glass. It looks so much better now, right? In fact, even if we completely kill the exposure of the HGRI environment here in the background shader with to mute it down, this background plate alone provides a really nice reflection. The direction is just perfect. All the light comes from behind towards the camera. This is just brilliant. It's still reverse key light, so it's not flat, at least. Maybe it looks a bit silhouetted now, but that's a different story. Uh huh. I'll reenable the GRI now. In addition to that backdrop, we can enhance things with an area light from roughly the same direction. This one purely for reflections. But first, let me select the background plane, go to object properties, expand the visibility tab there, and uncheck the shadow visibility check box. We don't want this object to block our new light. Okay. Now I'll create a new collection for it. Shift A, light area. I will make it roughly the size and shape of the background plane. It doesn't have to be exact, of course, but just roughly like that. Mm hmm. Then in the light settings, I'll make it much more powerful. And right away, notice the harsh refraction in the bottles. Yeah, that doesn't look great. We'll fix that in a moment. If I push the light back or change its angle ever so slightly, that bad refraction disappears. Also, I'll set the custom distance to 400 meters just to be sure it covers the scene. It's probably an overkill, but yeah, anyway, it works. And I'll pump up the output of the slide a bit more. Using the exposure slider is easier as it's more sensitive. Alright, maybe to about three for now. The role of this light is to draw those beautiful highlighted edges on the bottles, making everything look instantly more three dimensional. Okay, I moved the light, and that nasty, sharp refraction is back. It looks terrible. There's a way we can get rid of it, though. Under the lights influence settings, we can completely eliminate transmission annihilate it, even. I'm running out of synonyms. Okay, taking it to zero solves this. Now, no matter where we place our light, we won't run into this issue again. It will only give us those sweet, sweet reflections, which is exactly what we want. This light doesn't have to be too strong, so I'll reduce the intensity of this light a bit, um maybe to something like 130. Yeah, that's already way nicer, and we haven't even touched the shaders. This is all lighting. On our list of things that matter for IV glass, we also have render settings. So let's go straight to ray tracing, as it's the only render panel that really counts in this case. The resolution is super important. Setting it low will blur all the refractions like everything's made of frosted glass or something. Not a good look. For maximum clarity, crank this up, period. Precision. It's hard to pinpoint what exactly is happening in this case, but I think high precision helps small details catch glints better. Turning the noise and off actually add a lot of definition to those glands, too, especially if you compare before and after. And fast GI turned off, lets more light pass through the glass, making it slightly more truthful. D say physically correct. Then higher thickness value, arguably because we are getting subjective now, makes it look more three dimensional by introducing more ambient occlusion and stuff at the cost of some extra artifacts. It's totally fine to go really bold with that to seven and above, if you like that type of look. I think I'll settle around a slightly smaller value like 3.8. Okay, there's a lot more noise now since we switched off denoising and FAS GI, so I will compensate it by upping the viewpot samples. Straight to 32. That should help smooth things out. We're getting really close to that look from the inter shot. It's not completely there, but it's already pretty cool. Actually, feel free to bring denoising back if the grain bothers you, if that gets on your three D artist's nerves. That would also look pretty good. It's just that some details, especially the textural details will be lost or muffled. I prefer the look without denoising. It's just again, more truthful. It doesn't hold back. Like it's telling you exactly how drunk you were last night when you scratched the car. I like that honesty. And by the way, dear viewer, don't hold back, either. We love hearing from you, even if it's a critique. It's all useful. Feel free to reach out to us with any sort of feedback anytime on any platform. Okay, after sorting these settings out, there's nothing left to do here in this panel anymore, and we can safely move on to creating more lighting layers. This time, let's add some light probes, specifically sphere light probes. Here's the thing. In cycles, reflection complexity arises naturally from light bouncing around the glass and so on, especially with caustics turned on, it just makes it look more interesting and complex. In EB, we have to use all kinds of tricks to simulate that complexity. One trick is using probes to capture and manipulate light. Shift A, light probe, sphere. As you move this gizmo, you'll see the reflection start to change. This sphere probe samples the 360 degrees panorama like mini HDRI of the scene from its viewpoint, and you can make it larger or smaller to change its projection, and it allows you to manipulate reflections. You can duplicate it, by the way, and populate your scene with multiple such pbes. I'll create three in total to cover the entire area with bottles. I'm not saying this will necessarily make it look more realistic. It's more of an artistic tweak in this case, because it's hard to tell what's even going on. Just the reflections change ever so slightly. So it's mostly an aesthetic tweak here that allows you literally paint with reflections. Like, if you want to make this plastic bottle reflect some extra stuff or on the contrary, make this reflection darker, you can achieve that by moving that probe around. Okay, as I've mentioned, I ended up using three probes catered around semi randomly. So far, the effect is kind of negligible, but in the next step, it will make a lot of sense, as I'm about to add some spotlights to this scene, imitating caustics. You know, the wild bounces that happen within glass. And for that effect, light probes will be instrumental. So next up, spotlights. I'll create a new collection. Light, zero, one, zero, two, whatever. These naming conventions remind me of the Mm. Eh, close enough. Who cares anyway? I created the collection for it and now shift a light spot. My idea is to aim it at the bottles from the back of the scene. I want it to be a backlight that complements the rest of our set and I want it to be really, really punchy. But first of all, custom distance to 400 or whatever to cover the entire length of the scene and more. Now, let's make it punchier. How about 1,000 watts and exposure all the way to the right. So ten, basically. Now, it's probably too powerful, but the idea is to blast it with light like this, Imitating some wild caustic rays bouncing around This light needs to be harsh, as well. The smaller lights are harsher and this one is small, so that's okay. I mean, the radius, of course. Now, watch this. If you grab and move any of the probes, now, the reflection will change dramatically. It's because these probes have something really bright to sample and carry around. Now we're talking for the spotlight, I want it to be much more focused, so I will constrain its angle or its cone as if it's a focused sliver of light, you know? Some random rays refracted through something and bounce off like that. To avoid the crazy bright stuff, I'll suppress the transmission influence. I already used obliterate, I believe. So let's go with suppress. Again, even after setting up just this single light ray adjusting probes will have a huge visual impact. It will carry the bright spots around, so you can really paint with that. That's totally smoke and mirrors, which, of course, is one of the themes of this course. I tell you, this is a really, really fun process. It feels like painting with light, really. After setting up this first light leak, you'll want to copy it a few times to create more. Shift D or Control C, Control V, then just move it around. You can see where it's going. It's really nice random light scattering effect. I see it like that. You can shift click all of them, to select them all at once to move them around together. This part is very free form. Use your creativity to place them. Anyway, I played with these lights for 20 minutes straight before landing on this particular placement. And I also made these lights really, really powerful. The exposure brought all the way to ten. And with this addition, our glass setup looks way more alive and ironically, a bit closer to how path tracing might handle caustics. I'm stretching, but hear me out. Sometimes by going very stylized, we stumble into realism from an unexpected angle or even get past realism straight into hyperrealism. The entire history of modern art proves that anyway, not to get too philosophical. Um, let's get back into our scene. If it looks too noisy, feel free to enable denoising at the cost of some accuracy. The most aggressive element of denoising in IV is temporal accumulation, so you can single it out and mute it, and it will make the image a lot clearer. See, spatial denoising isn't that bad. That could be a nice compromise between precision and smoothness. Most of our glass shaders we created in the previous part are so fake. I mean, you often can't see one glass object through another with them with most of these shaders we created in the previous tutorial, I mean, except the few. But when seen together in context, it looks cool. And it runs in real time on my fairly old RTX 28 ETI. Alright, I really dig what I see. Now we can switch over to the animated camera to see how it feels in motion because glass materials are so dynamic by nature. When you hit Play, IV will drop to just one sample per frame, which isn't nearly enough to calculate the noise free screen traced or bounced lighting. Hence the extra noise. So if you want to see a better quality viewpoint render, first of all, boost the samples. You'll need it, and then run viewpod preview if you want a static shot, which you probably don't, for instance, for animation, you can go 50% of the resolution, then set the output folder. Then define a video format like MP four, for instance. It should render fast at 50% and give it a play blast. It should really be fast, like half a second or second frame. So let it cook. This should result in a nice preview animation. Just find that file and open it. And that's it for lighting and rendering the bottle seen in Ivy. A very quick side note. You can export these playblast as opening SAR sequences from blender and process them in DaVinci resolve, for instance. By default, it will display wrong. And I won't go deep here, but you can import opening SARs, then open the color tab and use the colspace transform node and convert it from the default reg seven oh nine color space of opening SARs and linear gamma to R white gamut four and Rox C four, for instance, and no tone mapping. So on top of that grayish log footage that we convert at HRs two, we just need to add one more node. So right click, one more Carrecor node. And then you can simply double click any of these lots after you install them. Blenders AGx or AgXKraken or any other Vetransfm you like, like filmic, or the creative ones like classic Cinema negative based on the 24 99 transform or any other I think for the final animation, I used 24 99 DRT print version one. Yeah, great. That's our look at lighting and rendering this bottle seen in IV. As you've seen, it's all smoke and mirrors, pretty much. But with the right mix of HDI, mec plates, Airy lights, probe, and a few spacious spotlight caustics, you can push IV far beyond what seems possible at first glance. 20. Compositing Realistic Hair & Fur : How to render realistic hair and fur in IV. Rendering realistic hair and fur is one of the biggest challenges in a real time engine like IV. Out of the box, it often looks unconvincing and lacks complexity and nuance. Well, we're solving that today. We're taking it from this to this. And by the end, you'll know exactly how to make the jump in quality and actually make a decent looking fur in IV. Here we have EV selected as our render engine, and right away, I'll turn on the viewport rendering. By default, to keep things simple, we just have a sun lamp here. Feel free to play with it. But no matter how you angle it, the fur rendering will look a bit weird and clumpy right out of the box. So before we do anything else, let's go straight to render curve settings and switch the shape from strand to strip. And just like that, we fix that rather horrific rendering. The difference is already night and day, so that's the very first thing you should always check. Let me turn on overlays here so you can see there's a base mesh underneath. The fur itself is made using a particle system, by the way. If you scroll down to the children tab, the display amount, there is how dense the fur looks in the viewport. Needless to say, if you set it too low, our bear will look shabby and bold. Density is a major factor of what makes hair look good and realistic. Usually hair looks better when it's dense. 25 here is the bare minimum for good look. See what I did there? For your final render, you can set it even higher. We keep it lower here just to not strain the viewport too much. After sorting out these fundamental particle properties, it's time to check ray tracing settings. Try playing with a sunlight by pressing R twice to rotate it like a trackball, and you'll notice a peculiar thing about IV's default lighting. By default, the direct light looks okay, if a bit harsh, but the areas in shadows like this arm look completely flat. Like, really, really flat. If in the random settings you turn on ray tracing, then suddenly those shadow areas receive a lot more definition. This is a massive difference. And it will look even crispier if in the tracing settings, you turn off fast Gi and turn off denoising. Switching off these two usual suspects has become a staple of this course. I recommend watching the earlier videos where we explain the reasoning behind this in more detail. The short version, if you want it to look better, just pull the plug on these two. Now, it's not that bad, but denoising does make it look a bit blurry. So it is just crispier without it. Like, look at that. And Fazis pretty inconsequential here. Well, maybe it adds to the performance cost. So out it goes. Next, when dealing with fur, setting precision to one can have a big effect. It's like a high pass filter for ray tracing. The high precision makes it tune into finer details, which is perfect for hair because hair is made of details of chinese strands, basically, you know? Thickness is another important setting. A low thickness makes rate racing work in sort of a higher frequency, ignoring bigger shape changes, deep shadows and all that, and focusing on the small shadows between the individual hair strands. And higher thickness makes this ambient occlusion effect spread out. It has diminishing returns. I'll just set it to 5 meters, which is anyway larger than the bear itself, just to make sure we have the deepest possible and direct shadows. Deepest shadows here really means the spread of the ambient eclusion effect, simulated by EV's real time rate racing. Check the ambient occlusion chapter if you want to learn more about that. I will also go with the highest resolution. I don't see a particularly big difference right now, to be honest, but trust me, there is one. Let's zoom in a bit to prove it. So that's one to one resolution. And now I'll switch it to one to two. See, there is a difference. Just look at that as I'm switching it. And this difference becomes much more obvious if we go even lower. It will lose definition really quickly and become pixelated. That matters when you're working with something as delicate and inherently detailed as hair. Naturally, I want to preserve all the detail I can, so I'm maxing it out. After this treatment, the lighting in the shadows should look much more grounded and much more real. And that's one of the biggest changes you can make to hair rendering in Ivy. Now that we've tweaked the render settings, and it looks much better than before. And from a rendering perspective, it might be okay. But from a lighting perspective, it's still a bit flat, and uninspiring. Let's just say that. We can do so much more with lighting and materials to make this hair and fur look absolute best. Let's turn it into this cutesy. Lighting. Now, the sun is gone, and we're starting from scratch to build a softer more beautiful lighting setup. Right now, it's just the environment light or the world background shader to be exact. That is illuminating this scene. Let's make it a bit stronger while we work. It only looks decent because of ray tracing. Ray tracing is what gives it the entire volume and dimension. Then again, denoising would pretty much ruin it by making it blurry. And FAS GI would also alter the look, arguably making it a bit more flat as well. Well, if steps were set higher and threshold also, it would be a bit crispier. So if you must use fast GI for some reason, push this factor towards around 0.9 or something. I'm just going to skip Fast GI altogether as it unifies the values a bit too much, and I'll use raw screen tracing instead. Raw meaning without optimizations. Okay, the environment light is working fine as a base. I'll set its level pretty low, so it becomes our ambient tone or room tone. So that is our first layer of lighting. Now, what kind of lighting setup makes hair and fur look great? Well, of course, a lot depends on your creative goals, but there are a few lighting principles that work well in almost any context, and they work fantastically for hair and fur. There are things that are almost universally good and things that are bad when it comes to lighting, be it furry bear or an elevator shaft, for instance. The same core mechanisms are at work. Good lighting is usually non flat, has some composition idea, creates a mood and feels realistic and not only, like, realistic as a realistic rendering, but also in its motivation. We talk about it in the elevator shaft chapter. Look, pretty much, these same guidelines apply here for fur and hair. The difference between the left and right image is that the right image is non flat, is compositionally sound, has a mood and looks more realistic. So how exactly do we achieve that? Yes, first of all, let's get rid of the sun here, as it's the first thing that makes it look flat. Instead, let's press Shift A and add an aerate. Make it pretty large, so it becomes a soft light and position it really, really close to our bear for even more light fall off, and I'll give it a large custom distance. So it covers the entire scene and much more power. Maybe 1,000 or so. So it becomes the main light coming from the back. Soft shadows can look a bit weird in IV right out of the gate. So to fix that, we need to turn on jitter and also jitter here in the render settings as well. This dramatically improves shadow quality. Now, thanks to this aerial light, our bear has some real dimension to it. So this is our area light, and you can say key light as it's currently meant to scopes the form in a very soft way. The position or angle of this light matters most. Frontal angles tend to be really, really flat. A perfectly frontal light is the most flattening. It's a total no go. If you want any sense of dimension, we covered this concept a lot throughout the course. It should be at least two thirds angle or a side light or even better, a reverse key light, which means placed slightly behind the subject. This creates a nice light rap effect. I'll press Shift D to duplicate this light, and this will be our second aerial light. Let's isolate it for a moment by switching off the first one. I want to make it a slightly more side light, if you know what I mean to wrap it around the bear a bit more. I will also make it much larger to make it much softer at the same time. Our first lite positioned more at the back, emphasizes the contour and the shape, while the second one is meant to mellow things out, smooth things out and provide an ultra soft fill. Then I'll select the second alte. Shift D again to make a third one and this one will be much smaller and placed even more at the back to act as an edge light or rim light. Its role is mostly to emphasize the silhouette of our bear and make the hairs at the edge glow. And now, actually, let's check how it all works together. The key, the fill, and the rim. A lights that loosely follow a three point lighting scheme. Well, kind of, it's not a classic three point lighting scheme, of course. It's more of a reverse key variation since the key light is positioned slightly behind. But still, the first one or key sculpts the main shape. The second one or fill provides a soft gradient, and the third one or edge makes the hair strands glow and helps the bear pop from the backdrop. That looks a bit like the three point lighting scheme, you can say. This reverse key lighting scheme works great for sculpting the shape because all the light pretty much comes from the back and wraps around the subject, not from here from the front, because frontal light would kill the volume and make it flat again. It's a no go. All the light should come from there, from the back. The lighting that we've built from the ground up makes our hair and fur looks much more three dimensional, and that is, of course, amazing. Yet by itself isn't enough to create a truly wow rendering Ivy. We need something more and that more is the shader. The last mile is often the shader. There's just no way around it. So that is our next goal. So far, we have this principled BSDF which I'll delete, so we can build a better one from scratch. Shift A, shader, principal BSDF. But this time from scratch, It's a nice yeti hair. Wait, I just realized that the particle system for the ears was turned off. Let me fix that real quick. LOL. Okay, back to the bear. Let's start working on that shader. For that first base layer, I'll turn off the IR or reflection. So we can focus on the diffuse component only for the time being. I'll make the base or diffuse color dark gray. Well, technically, it can be given any sort of color, but I want a fairly neutral tint for our bear. Fur and hair can benefit greatly from subsurface scattering in this tab, so I'll tackle that next. The SSS weight, I'll just set it to one, and then I will tweak the scale of the effect until we get that nice shine from light scattered under the surface. By default, it's reddish, yet I'll set its color to a neutral gray or an equal value in all these three sliders, red, green, and blue. 0.6 and all three makes it gray. You can use these colors to tweak the balance of the scattering effect. So, for instance, if you said blue to something like 0.5, it will be tinted slightly brown since red and green are dominant now. See, it has this slightly brownish color now. The thickness also plays an important role. I mean, this thickness. If you just give it a value, it will allow you to fine tune the look of this effect alongside this scale. So it's really a back and forth between setting the scale here and thickness there. So I'm setting the scale to taste really not paying attention to the physical values. So 0.21 meters. That's just how I see it. And the thickness, um well, hair should have a pretty low thickness, isn't it? Maybe I'll set it to a really small value this time, like 0.0 something because it should be really, really thin. Well, maybe in here as well. 0.07 or something. Aside from that, it's all about tweaking these individual red, green and blue sliders to colorize the effect. Personally, I don't want to make a peculiar tint like purple or so this time. So I will stick to this neutral brownish gray. So that is our subsurface scattering to fast track it forward a bit, we'll add some more subsurface scattering later on, but that is the first pass of subsurface scattering that we add. Here's a quick before and after. As far as this effect goes. At the moment, it's too much scattering, but that's just one of the layers, as I've mentioned, we'll be adding more layers in a moment, including reflections and all that stuff. Yet, so far we've added this effect of the light trapped under the hair strength, giving it an internal shine. Speaking of shine, there's also the shen parameter in the principle BSDF and it can be tweaked. Remember the velvet shader, Shin does something similar. It simulates tiny fibers on the surface. For cloth, these add a soft velvet like reflection. The weight is the strength of this effect. It's fairly self explanatory. I would set it pretty high to about 0.8 for now. So it has a lot of sheen. Roughness changes the reflectance angle, simulating different fabrics like silk or velvet or anything else. The tint lets you add color, and I will go for a darker, fairly saturated brown. Something like that. Yeah. And as for roughness, lower roughness looks pretty interesting, I think. I love the look it creates the patches of color here and there, you know? Actually, something a bit higher around 0.5 doesn't look too bad either. So shen adds a very important layer of color and dimension to this shader, if you look at it now. This is where those clothy microfiber or velvet qualities come from in our custom hair shader. And funny enough, most of the color comes from sheen and not from the diffuse shader. That's an interesting detail, I think. Usually hair has some reflectivity, as well. So shift A, shader. And typically, if you want to mix in other shader, you use mixed shader no brainer. But here, I'll actually use the add shader. It's practically the same with a minor minor different. A mixed shader blends in that other material, glossy in this case, replacing the original one. The add shader does something very similar, but it lets the original material to pick through a bit more in case you want to preserve the original colors or whatnot. I want to add one of the reflection shaders, like specula, for instance. Okay? I will change its base color to black fo to work on the specular reflection only. All hair should have at least some reflectivity, you know? I don't want too much of that, or it will look almost metallic or synthetic. So I will stick with the lower reflectivity, something like a middle of the road gray, which is still a bit too much, like it's made of nylon or something, but we'll fix that shortly. This is our base specular reflectants. This stuff can be confusing, this shader in particular, how the base color is different from specular reflectants and all that. So let me briefly explain how the specular input is different from the base color. Let me unplug it for a moment and add the standard glossy shader. So by default, the glossy shader adds a metallic reflection, meaning it reflects light equally from all angles. So if you reduce the amount of reflection, it would still act as a metallic reflection. So that's how it works by default. In a metallic mode. Okay? So let me unplug it and plug the specula back in. And zero out the specular color for a moment. So the base color in this shader is the exact same thing. In essence, it's a metallic reflection. The absolutely same thing as with glossy shader, no difference whatsoever. So what is specular for then? The specular is a dielectric reflection or a frenel reflection, the kind that materials like wood, glass, water, and plastics exhibit. The frenel node plugged into the glossy shader imitates this kind of reflection, basically a non metallic reflection. Nonmetallic reflections depend on the angle reflecting more at a glancing angle. Okay? It doesn't make sense so far? So that is incidentally exactly what the specular input does in that specular shade. It has a frenel index of refraction built in and it imitates dielectric materials. It's exactly the same as the setup. Now, whether wool or fur is metallic or dielectric is another question, and you get to decide. Yet having both of these options available makes it easier and more enjoyable to twig the shader. It makes this little heart beat excitedly. Okay, we've added some specular reflection to our hair shader. Next up, I will give you some metallic reflection as well. Hair often exhibits a metallic like reflection component, AC silky shine, often colored by melanine or other pigments. Actually, I'll give it this light purple tint. And then I'll give an opposite direction tint or greenish one to our specula. You know, maybe a bit mossy greenish, how to put it. Basically to tint, two reflectant is slightly different way. It should make it look slightly more three dimensional. I will crank up the roughness to about 0.8 or something to make it look like old shabby hair. Press to toggle and check the before and after. I don't know. Maybe making it slightly more silky would be more interesting. Yeah, I think I like that. It looks very nice at this stage with the extra bit of reflectivity that we've added. And now, it didn't shrink. It's just a different angle. Again, notice how the reflection is colored to match the overall hue of the bare fur in that ballpark. Transparency slider can add an ethereal quality to hair, making it look a bit airy. In the material settings, the random method and this transparent shadows checkbox would also change its look. Turning off transparent shadows would make it look a bit differently, especially the silhouette. Yet, I'll leave that on to get that cotton candy quality, and I'll set the transparency either to zero or really, really low number, hoping it adds some visual interest to our Shader. Hair is really fully opaque, but it's not very transparent, either. It can be semi translucent. That's a different thing. The transparency slide Shenanigans play into that. It's a very subtle effect, indeed, though, pretty hard to tell what's going on, but higher transparency can bring a glowing effect if you look very, very closely. Now, in addition to the reflectivity that we've already added, I want to add yet another layer of glossiness. This time, let's use a glossy PSDF to shake things up. I want this to be a frenll or nonmetallic reflection. I'll use the layer weight node. Again, just for a change. Specifically, it's frenll output. It should go into the color. So I'll set it around 0.2 or maybe 0.1 or something to get an ice reflection fall off. To control this better, color ramp node is always handy. The white flag is maximum reflection. So I'll pull it back just a little really gently. Mm hmm. I don't want this shader to be rough and shabby. On the contrary, I want a silky shine where a few hairs pick up specular glints. So zero roughness this time. It's perfectly glossy. This helps the hair look alive and healthy, just like they say in shampoo ads. So it actually has a bit of shine now. Now comes a bit of a strange turn. When I was recording this, after adding this extra glossy pass, I struggled to make it look healthy and shiny. What do I mean? Actually, I realized that in the color management setting, someone set the exposure to one. Or slightly increased it. Maybe it was me, but it doesn't matter. The point is that made the image brighter and that's okay, but somehow it also affected the specular reflections. It muffled them. When I reset it back to zero for a test, the glossy reflections popped out much more to my surprise. I'm shocked that this is something that would affect the look of the hair in Ivy. But you're a witness to it, it really does. To me, it looks like an IV bug, some sort of clipping, but maybe I'm wrong, and this is how things should be. Hey, there seems to be a link between exposure and how contrasted the glossy reflections look. I really think it must be a bug in IV, and we all missed this bug. At exposures other than zero, no matter negative or positive, IV seems to apply some sort of clamping to the highlights, muffling the contrast. I can be wrong, but I'm pretty sure it's not how it's supposed to be. No, I don't have clamping applied. And yes, I will file a bug report. If it's fixed by the time you were watching this, brilliant. Until then, it might even be worth using exposure in the compositor instead because it preserves the highlights much better. I honestly hope it's fixed by now. This effect was similar to using direct clamping and setting it to two or something like that, which would also clip the highlights. Or to one or something like that, leading to this slightly muted look. Actually, light clamping in v can be pretty useful sometimes. If you set it reasonably high, it can help reduce noise without crippling the highlights too much. So maybe even setting it to three or so could help. Otherwise, the amount of noise in these specular highlights can be pretty high, and it can lead to flickering and animation and other such unpleasant stuff. So there is a point in clamping or clipping, but only if we do it ourselves and intentionally, anyway, I think in the end, I succumbed to the bugs, said, Okay, fine, switched the exposure to one. After all, I thought, it doesn't look that bad. No, really? It made it look like really floffy like a ball of flof. So, yeah, that's my story. Okay, snapping back to our hair shader. Another component that I want to add is more subsurface scattering. Shift A, shader, add shader again. And right away. I'll also add the subsurface scattering shader, which will go into the free socket. Wow, I made it ultra shiny. And this isn't about the RGB coefficients. Actually, I'll set those to 0.4 or zero point. Less. 0.2, which is a neutral gray. Mm. And it's not about the scale, I think. This supernatural brightness is because of the color, which would be way weaker since we're using the add shader. So let's make it black and then raise it up carefully. Yeah, something like that is more like it. Now I just want to give it some colour because in real hair, various pigments colorize the passing light, so brownish it will be. The scale of SSS shader controls how far light travels under the surface. Think of it as the overall translucency. And you can definitely make it super translucent, but it's always better to keep it in the ballpark of physical plausibility. What looks real to you, basically. So maybe a darker SSS color would do. You know, so it's kind of there, and without it, it's not as cool, and it makes the hair slightly translucent and traps the light inside. But it does so in moderation. Mm hmm. That's the key. It doesn't have to be super strong effect. Well, of course, it's completely up to you, and maybe you want to make it, I don't know, orange. Like synthetic foam rubber or something. Silly, but fun. Actually, it looks pretty good. But I think I will reset it back to darker brown. Make the safest color for bears. Here's at a glance the translucent quality that subsurface scattering adds to a custom shader. And this is our resulting shader, if you want to pause and take a screenshot. Feel free, by the way. Otherwise, just check the project file. To recap our four blocks are diffuse plus sheen, specular reflection, extra glossiness, subsurface scattering. The rest is up to you. How do you want to go about fine tuning any parameter in each of these blocks? I'm reaching for my water, and you can take some more screenshots, if you like. And don't forget about lights. We've set up three lights in total, and this can be adjusted too. The look of the heavily depends on lighting. For instance, maybe you'd want more rim light. You can definitely crank up this light and see how it looks, how it emphasizes the hair strands glow at the edges. That could be a major part of the look, you know. Adding some color to the lights can also help tweak the mood and create a juicier final render. Maybe something more on the fantasy side, who knows? I like our original neutral vibe, so I'll do a final render with that. Thank you for watching. I hope this tutorial helps you create neat looking hair and fur in IV. It's not the thing that people usually associate with IV, a decent hair rendering. Yet, as we've seen with a bit of technical know how and an understanding of the core elements of a custom hair shader, it can definitely be done in IV.