Hello guys, Nauval here! And in this video, I’d like to talk about my workflow and experience in making my 3D characters. So, first of all, creating 3D characters is one of my favorite parts in the 3D world, well it's because I just like characters from movies, games, and I was just thinking of creating my own original character.
So, of course, the first thing you want to have, before modeling your character in 3D is concept art, absolutely important. Concept art gives you so much information about how your 3D will look like in the final result. So, basically, your final 3D will depend on your concept art, I’d say.
So, you want to spend your time creating your character proportion, clothes, hair, and any other accessories as your concept art. In my case, I draw my concept art using software called Krita and Clip Studio Paint. And maybe this first step can be tough to do, but here is the thing, you always can look for references online to enrich your concept art. I’d say, using references is not cheating as long as you don’t blatantly copy other people's works. I mean like it's just like you covering a song but of course, you add your personal touch on it.
I myself have been using references a lot, I mean I like always looking for references in my drawing process. And if you ask where, well it's everywhere, it can be from Pinterest, Pixiv, awesome people on Twitter, and even Google Images of course.
Now, what you want to draw in your concept art is not the stylized poses illustration, instead you want to draw your character from front view, back view, side view, and ¾ view if it necessary. But for me, having only the front view of the character concept art is pretty much sufficient.
Okay next step in the workflow is the modelling process itself. Now, I bet you guys already know that there are at least two types of modelling techniques. The first one is the traditional geometry modelling, and the other one is sculpting technique. For myself, I have experiences in using both techniques and both are just useful.
But, in my case, because I model for game engines, my model should be low poly. So, in my opinion geometry modelling is just perfect for this case, because you won’t be bothered to do retopology for your model which means you don’t work twice. Maybe there is an addon for automatic retopology for Blender but I am quite sure that it won’t perfectly work as you would expect. But, if you use ZBrush I think it's not the case, ZBrush automatic retopology is very neat and I am very impressed with this feature, but hey, we talk about Blender so yeah, geometry modelling is the winner in my opinion.
Also, if you have a neat concept art for your model, you will be able to model without having a problem. This is why concept art is very important, because it gives you information about what you should model and what is not, how much poly do you need and etc. And with that information alongside you, believe me you will also use Ctrl-Z lesser compared to without using any concept art.
So, while modelling, I usually UV unwrapped my model but it's just a simple UV unwrapped not very complicated. I just use this UV to give a glimpse texture on the model by using my concept art as the texture of course. And this is one of the benefits of using concept art. By using this technique, I can tell if my model is correctly shaped or not.
Talk about modifiers, I only use mirror modifiers and that's it. Well sometimes I also use solidify modifiers to emphasize the shape of the model especially when I create the character clothes. But, I will remove this modifier when the model is finished because you know, using a solidify modifier will add up a new poly count.
Some of you guys asked me how to model hair, well it's pretty much basic. If you watch my other videos, the way I model hair is just to follow up with the concept art and it's pretty much about moving vertices around to match the concept art. And I don’t really think much about how actually. Well, there is another way tho, you can use a bezier curve and you know add a sample geometry on it and move the bezier curve the way that you want. But, in my opinion, using a bezier curve can be very messy sometimes because you need to apply the geometry and then tidy up the geometry.
Okay, moving up to the next step, which is the rigging process. Well, I watched several 3D modelers out there, some of them put this step after the texturing process. But, it's different for my case, I put this process before the texturing. Maybe you are wondering why? Well it’s because in the middle of the rigging process, you might want to change the poly of your model a bit, adding new edges, fixing up vertices, in order to deform your mesh correctly. So, this is why texturing should come after the rigging, because if you change your mesh after you textured it, of course it’s most likely possible that your UV map will mess up. So in conclusion, if you are not confident with your mesh topology then do the rigging first, and then texture it later.
Well, the way to rig in Blender and I believe in most of 3D softwares, I think there are at least two techniques, the first one is to assign weight to vertex manually, or using the second one which is to assign weight by using weight painting, which is both of them are manual actually. In Blender there is a way to assign weight automatically which is when you assign your model to an armature by pressing Ctrl-P and then choose automatic weight, but of course it wont gives you the correct weight influence.
In my case tho, I use to assign weight to vertex manually and then fix it in weight painting later on if necessary. Either way should work but you need to know how you should assign weight to your mesh. The rule of thumb in my opinion is to give a gap between joints as you can see here. This is a post I found on Twitter and this is super legit guys, this is just the perfect way I can find to rig a low poly model.
Some troubles on modeling arms, legs etc is the deform arc 🤔...... look this #tip for better deformations on #3D #modeliing , #tutorial #Maya #Blender #cinema4D
— Quetzal (@AnimQuetzal) May 31, 2019
🎨 @Wunkolo pic.twitter.com/kNgXwPP7ou
For the facial expressions, in Blender you can create shape keys to deform your mesh. Based on my experience, shape keys can be exported to most game engines, including Unity 3D. So for facial expressions, I recommend using shape keys instead of assigning vertex weight manually to bones. And then, the way how you control the shape keys value is by adding a driver so you can control your shape keys by moving your rig.
Okay, next step is texturing. And of course as you know, you need to UV unwrap your model first. I usually use a checker texture to help me to know better whether my UV is correct or not. For texturing, I usually use different layers to avoid mistakes and you know it can be used for back ups purposes too.
Unfortunately, Blender doesn’t have a layer organizer like in Photoshop, so you should create on your own. And the way you do this, is to create appropriate material so it can manipulate such features. So to set up this material, you will need your shader, a mixRGB shader, and your textures. So as you can see here, we use the alpha of the second texture as the mixRGB factor. This way the first texture will be rendered on the second texture alpha. And there you go, you just manipulate the layer feature that Blender misses. Well, there are addons to do the same exact thing tho but I found it quite complicated and not very practical for me, so I use the traditional nodes instead. But, I do really hope that Blender will add layers feature in the future tho.
Simple tips, before you start to draw fine details on your hand drawn texture, first of all, fill up your texture with gradient colors. In my opinion, it gives a slight shade so if you are going to export your model and consider applying an unlit shader, your model will just look fine even without any shade. So, this kind of technique is just what I found on game models out there, especially in mobile games. If you look closely when you play games, you’ll notice that some of them do not cast any shadow. This is one of the methods to improve game performance so it can run on stable FPS even on low spec devices.
So, I usually just use the Draw brush in Blender and I would say the brush is just okay. I mean like Blender brush doesn’t really represent real world brush in my opinion, because when you start making strokes, the colors do not blend. Maybe I am the one that doesn't know that a blending brush exists in Blender, but I can't find this feature anywhere in Blender. So, if you guys know about this stuff, let me know in the comment.
Oh yeah, when you draw on texture and you don't want to affect the alpha channel of your texture, uncheck the “affect alpha” properties. This way, whenever you draw, your strokes will clip only to your texture alpha, which is very very handy if you are working on many layers.
And yeah, I believe that's pretty much it about my workflow. I hope you guys find this video helpful. And be sure to subscribe for more content like this, and with that being said. I will see you next time. Bye!
No comments:
Post a Comment