What a busy week! I haven't had time to do anything else, but it's been really fun to work on this project again, so I don't mind. This week I managed to make 49 individual CGs, successfully completing the first scene and meeting my goal! "49 CGs? So 7 a day?" Actually, no.
If you'll recall, in my first post I explained that I'd messed up with the renders and had to start late. I ended up only making CGs on four days this week — so I actually made 12 CGs a day on average. This is super encouraging because it's 3x as many as I thought I could do per day. And I probably could do even more in a day if needed, provided the prep was done beforehand. The idea of making the entire game out of CGs doesn't seem too farfetched now!
So how is this possible? Well, even before this, I've had a very streamlined style that emphasizes speed, with flat, sharp shading, black line art (as opposed to colored lines), and a very light touch when it comes to post-processing and effects, and I've been using 3D backgrounds ever since my first game, which takes way less time than hand-painted backgrounds. This speedy style helped me make eight sprites in just a few days for Methods. I've always been able to create my artwork fairly quickly, at least compared to other artists. That being said, often the quality would suffer.
On Methods, I opted for much simpler shading for most of the sprites. I often had to redo CGs because the proportions of the character were out of whack — and some CGs to this day have those issues. So my speed had some big downsides. For Catalyst Wake, I knew I had to step up my game and try to eliminate those downsides. With so many CGs, and how much I wanted dynamic camera angles and poses, I knew doing it the exact same way I'd done Methods just wouldn't cut it.
Time to go more in-depth on the process I'm using to create the CGs.
Each character has a unique 3D model, which I place in the scene. Then I sketch over it, filling in detail and improving on the shapes, and finally bring it into Adobe Animate for the final line work, coloring and shading.
The whole process takes around 15 minutes on average. A Methods CG took me around 40 minutes.
Using 3D models as a guide has a huge number of pros:
Sometimes when I was drawing a character on top of a background, I didn't correctly position the camera to make room for the character. This happened numerous times on Methods, and sometimes I needed to re-render the background or do something more hacky.
Dynamic angles and poses
I want my CGs to have a lot of cool, dynamic angles. But drawing them without any guide is super time-consuming and often the result is bad. By using 3D models, I can draw the perspective very accurately, so the angle can be as wild as I want.
I can also pose the 3D characters in some truly strange ways without fear. You've probably noticed I draw the same hands over and over — it's definitely something I could fix by practicing drawing hands for a few days, and therefore adding few dozen more hand poses to my repertoire... or I could just trace the 3d models and have an infinite number of hand poses I can now draw from literally any angle.
I want my characters to feel truly grounded in their environment, but that's pretty difficult to pull off. The 3D models help me easily see how the character is situated — what elements of the character need to be masked out, etc. Something like a character sitting in a chair is way easier to draw, for instance.
As you can see in this CG from Methods, the characters aren't really in the right perspective. It's serviceable — you get the idea that they're supposed to be sitting in the chairs, but they look a bit floaty, not the right height, and just off-looking in general. This is after I completely redrew this CG again by the way. Using 3D models largely eliminates this problem.
You'll notice in the Methods CGs that characters look a bit inconsistent from CG to CG, their height and various proportions change. It's kind of charming in its own way, but I wanted to improve on that for Catalyst Wake, since it relies so heavily on CGs.
Having the models as reference lets me easily make the characters consistent across all illustrations. Since the model remains totally consistent, the drawing traced from the model is also consistent.
Colors and Shading
Because the 3D model is literally placed in the scene, I can do a simple Toon Shader pass to get shading information, which informs how I shade the character, improving shading as well! The 3D models also have all the colors of the character, which is affected by lighting in the scene, giving me accurate colors based on the lights in the environment.
Placing and posing the 3D models for every shot essentially creates a storyboard sequence. I'm able to easily see how the CGs flow between each other, and get a really good idea of how the final thing will look. I could even whip up a quick prototype of the game with the unfinished 3d model shots if necessary.
When I wasn't using 3D models, it would take a lot longer to figure out how to draw the pose for the character, where to place them, and I'd often require a few drafts to get something I was reasonably happy with. Using 3D models eliminates all of that. And because the line-work is traced pretty much exactly over the model, it's waaay faster to color the characters, because I just eyedrop the color right from the image, then fill it in.
One big pro is that... this approach is just ten times more fun! When I draw a character out of my head, I try to change it up with an interesting angle now and then, but I largely stay in my comfort zone because going out of it leads to bad results 90% of the time... but now I'm free to experiment! Having a solid foundation with the model, knowing that, even on one of my "bad drawing days" it will still be perfectly usable? It actually gets me excited about drawing, which doesn't happen very often.
That being said, there are a few cons from this approach.
Sometimes the 3D models can end up looking a bit stiff, even after sketching over them. This is mostly to do with how they're traced and can be combated somewhat by following the 3D models more loosely. More dynamic poses usually don't have this problem.
Fortunately, I've been drawing without 3D models for a long time, so I know how I usually draw things like hands, gloves, etc. That's helping me make decisions about how to reinterpret the guide so it looks less like a traced model and more like my style.
Sometimes tracing a model results in the drawing looking "thicker" — it's sort of true for any tracing you do. The solution is to be on the lookout for that issue, and redraw the underlying structures if need be.
More Prep Work
You can't leap straight into drawing, you need a 3D model first, and spend time making it as accurate as possible to the character. You also need to create props and textures. However, I have enough humanoid models at this point to "kitbash" a character together fairly quickly. It helps that these characters all wear masks, which are very easy to model. However, some people may not have a good humanoid model available, have any modeling ability, or have more complex designs, which would require much more work. There are some free models online you can download, but nothing will beat having custom models based on your own personal art style.
Some people may take offense at the idea of tracing 3D models, because tracing is a big taboo for some people. Toniko Pantoja had a good video on this, which I agree with. I think this easily falls in the "working smart" category, but I understand if some people don't see it that way.
I do think it's a good idea to try to make the traced model look as much a not-traced drawing as possible — and I'll keep working on that as I continue streamlining this process. I think when you see the CGs in Catalyst Wake, you won't be thinking "This looks like a stiff 3D model", you'll be thinking "This looks like any other LockedOn illustration, just with better proportions, consistency and perspective".
This post is very detailed about this one thing because I want to be able to cite it in future posts whenever I reference this tracing technique. Now back to the usual update post stuff.
Other things I did:
- Updated the dialogue box UI
- Composed the dynamic music for the scene
So what's this next week got in store? I'm taking a short two-day break to recharge and get some other stuff done, then it's back in the trenches with prep for my next batch of CGs. I need to prepare character models, backgrounds, and set up the shots. Hopefully by the end of next week, I'll have everything lined up so that the remaining two weeks of the month can be nonstop drawing/rendering for the remaining scenes in this first sequel story.
I'm so glad I was able to reach the first milestone successfully. Seeing a full scene done this way was really the only way to confirm that this "CG-only" style works — and it does! I'm excited for everyone to see it!
Until next time!