In this Essentials video, I'll go over how we can take high poly detail, and bake it onto low poly geometry, so we can get all the detail, the edges, and the chamfering from a high poly model and put it onto a game ready one. In this Essentials video, I want to show you a little bit about what the different types of smoothing are and what exactly baking does for us. So let's say, I have a couple of cubes here, let's make three cubes. First off, we have one cube, it's the default cube that generally comes out. Now, we don't know this, but it actually is hard shading on everything. That means every side at every corner, it changes directions of the verts. We can actually see this by looking at the verts, go into vert normals and displaying them. We can see that each normal has three verts coming off one in each direction. If I was to take this object, and soften all the edges and now display the verts, we can see they're all averaged out and each vert has one direction. Now in this case, it looks like one vert with three directions. Each vert can actually only have one direction. This is technically from the background as far as the 3D modeling programs are concerned, three verts. This can be a bit confusing because Maya right here, tells us there's only eight verts. But really if we open this up in something like Unity, Unity would actually tell us that there's 24 verts, because there's one for each corner here, as opposed to Unity would consider this eight verts. Because again, in order to control all the shading, we actually create some duplicate subverts that we use to control the edges here. So this is a soft or a smooth shading. Now, I'm going to be going through this a lot in the next couple of lessons when it gets into modelling but if you press three on your keyboard, we can see that it turns our cube into a sphere. Now, it hasn't actually changed anything, it's just a preview. If I press two, it will actually show me the preview cage with the final result. What it's trying to do is basically it takes this cube right here, and if I shift, right mouse and I go to smooth, what you can see it's done is it's taken every face and it's turned it into four phases. If I apply this twice, we see it's the exact same result as this, a smooth preview by default and Maya is showing us what two smoothing operations would look like. It's really handy to model things out and features out before we need to use them in the final station but like we can see what it's going to look like, as opposed to needing to set up all our edge loops and get all those things correct and then guess what this is going to look like. So for practical terms in modelling, we find this smoothing is used for two things. For one, sometimes I might need to create a complicated shape. Like the example I like to use, is maybe something like a funky door handle that take a shape like this, and let's delete the bottom here, and I can see from smoothing it out, I'm producing a fairly soft shape that would otherwise be very difficult for me to model out by hand. So maybe I'm trying to create like a door handle on like a sci-fi car or something like that. I can take this whole thing now and run a smooth on it once or depending on what I'm going to finally go for twice, and I get a good piece of low poly geometry that would've been really hard for me to model. Computers are really good at curves this way. The other thing I can use smoothing for is thinking about my edges when I'm trying to make something like a normal map. So I'm going to make a couple of versions of this cube here. I'm going to have the default cube. I'm going to take all the edges on this cube. I'm going to run a bevel with a chamfer on them, and just tighten up the fractions. You can see that by comparison, this cube when it's properly smooth, looks a lot more realistic than this cube right here. It's just got that little edge lighting, something we often don't notice, but it's got this little shining, white, bright edge on everything. We call this a Fresnel reflection, so it's often the idea is that the light is bouncing and some part of this edge is facing directly towards us, because as it smooths from one surface to another, it's much more likely one of these edges gets faced directly at us. This means, as the light bounces directly into our eyes, we're likely to see it as a brighter, wider edge. If you look around your room or on any object on your desk, you will see that every time you see what looks like a 90 degree edge, you'll get this fine little thin line of reflection running across it. Now, the problem with that is so in order to achieve this effect, we often need a really high piece of geometry. So for instance, to get a really fine reflection we might need 3,458 verts, whereas we only need like eight verts to get this cube to look correct. So the goal is, how can we make a cube that has this number of verts but it looks like this? This is where normals come into play. So I'm going to take both of these. I'm going to bring both of them into our chosen baking tool which is marmoset. Let's just name them really quick. I'm going to call this low and I'm going to call this high. Let's select these together and I'm going to do a quick export. So I'm going to bring both these objects into marmoset. Any baker will really do. I personally really like marmoset because of I find it really easy to use and it has an interactive baker, unlike a lot of other programs, I don't get to control the baking cage manually. But what I'm going to do is take this. I'm going to create a new bake group. It's this little low symbol. We can see that there's a little spot for our high and our low poly. So we take our subdivided model and we take our low poly model, drag it into low, and now the baker we can easily turn off the low to see just the high, or the high just to see low. I'm going to turn everything else off just normal is on for now, and resolution just 1024 by 1024 should be fine. What we can see is if I press this little preview bake button, and let's just dump it into a really quick spot, I'm going to call this bake demo, what we can see here if I turn on my wireframe really quick is that I have the same low poly eight vert geometry but it appears as if it has all of the resolution, those edge details, and it reflects light with that little beam, the same way that our high poly does. In this way, we can take a low poly piece of geometry something that could actually run inside of a game engine, virtual reality, or even just work better in our films for rendering, and actually give it the details and all the surface stuff that we're going to be putting into our high polygon model.
