 So if you're a 3D artist looking to train your own AI on your own character, basically the easiest way to start is to create a small folder of images showing various poses and angles of your character. Now I'll be going exactly what those poses are over the next few days, but for now let's just talk about simple ways for you to prepare an environment for this process. Now I think it is very worth your time to spend a few hours to create your own environment first. And you can do whatever you want, but just make sure that it has the ability to have a green screen as the background. The reason you want the green screen background is because it can easily be identified by the AI. And I have been told that green works best because during the caption process you can easily tell the AI, Hey, everything that is this ugly green color is part of the background and you do not want to train it. Now, if your character is green, you might want to use a different color like white or blue. But from what I've seen green works extremely well. Now something you will also want in your green screen environment is a flexible lighting system. I recommend the traditional three point light one because it looks good and it's really easy to adjust everything. But whatever lighting system you decide to use just make sure it just has to be good enough to get some solid lighting for your character while it's doing different poses. Now you can use whatever software you want for this process. You can use Blender, Maya, Max, ZBrush, Houdini, Marmoset, it really doesn't matter. I've personally decided to make my environment in Unreal and the reason for that is because I love Unreal's ability to get photorealistic renders in real time while being able to control and move the character or camera around either manually or with programming and blueprints. So I've basically made a virtual photo booth that allows me to swap the mannequin with whatever 3D character I'm trying to train on. As long as it's rigged with the Unreal 5 skeleton, I can cycle through and drag and drop any animation inside of Unreal's default engine or the marketplace. And this allows me to quickly add any animation I want to my characters, then pause the game, rotate the character, detach the camera, and screenshot whatever I need without worrying about how long it takes to render. I've also created a simple material instance with parameters for the skybox that let me control the background exposure, rotation, or just swap out the skybox texture with any color that I want. It's really easy to swap out backgrounds. And I've also put in some basic 3D props like a trophy stand and some studio curtains, which are really useful whenever you just want to get a nice looking render of new characters. Now, if you are also an Unreal dev and you don't want to bother making your own environment, you can just pick mine up from ArtStation for 25 bucks. Or if you want to get your hands on it for free, you can just follow the links in the pinned comment to learn how to recreate this booth on your own from scratch. So as usual, the information to learn how to do this stuff is always free. The paid version is really just for those of you who don't want to spend your own time making everything yourself. The personal license is everything you need, including the props for the background, blueprints to rotate and animate your character, multiple skybox backgrounds and textures. But the commercial license has a few extra things that probably make your life a lot easier, which are mainly 18 different poses that I thought were particularly helpful when training my AI. One of the most time consuming parts of training your own data set with 3D characters is going into the rig and creating new poses. But if you get the commercial version, there's 18 poses ready to go right out of the box that you can drag and drop to any Unreal skeleton character. So you don't have to use any time at all. You can just change the skeletal mesh to your own character. And as long as it's using the default UE5 skeleton, you can use the poses and start screenshotting immediately. These were the same poses that I used to create my first data set for my own character. So you know they're good for training in general. But like I said, you don't need all the bells and whistles. All you really need is your 3D character, good lighting and a green screen. So if you do that and you join me next video, I'll share some of the techniques that I've found to be really useful when it comes to quickly generating new screenshots. Hope that helps and as always I'll be having a fantastic day and I'll see you around.