Game Development Update – Animating From Scratch

It’s been several months since I wrote a proper update about my game development projects. That’s really what the purpose of this site is for, after all. So here’s an update.

I’ve been working for months on a new animation system for my next game. Unfortunately, not everything is working out smoothly…

My new 2D character in 3D... may not represent final gameplay.

My new 2D character in 3D… may not represent final gameplay.

Yes, I still work long hours at my day job and don’t spend as much time as I would like on my games. But a few hours each week for the last several months I’ve been slowly drawing my character as it would be animated in engine.

I’m still using Unity3D, and I’m still using the “2D-in-3D” methods I’ve developed. But this is a new innovation: before, I had a single perspective plane for the entire character. I did this on purpose, knowing from the start this would be the easiest way for an artist to create characters and see them exactly as they were drawn. But in an attempt to save time, hardware memory, and to create a better 3D effect, I’ve been trying to modularize the character’s body more. Now I have individual frames for the feet, lower legs, upper legs, waist, torso, and so on. These are attached to a 3D skeleton, and the 3D skeleton takes care of most of the animation.

Most 2D cartoons (especially on television) today use “skeletal animation,” where parts of the body are attached to a 2D skeleton to animate the character quickly like a puppet. This is basically that, except with a 3D skeleton. I’m pretty sure this hasn’t been done before (Live2D Euclid’s demo videos and my early research videos both touched on this, but neither really moved forward until now).

2D puppet animation often looks cheap and crude. Again, I knew from the start this wasn’t an ideal method, but I’m experimenting. I will probably make articles about this in further detail later this year. But for now, my early results –

my character looks weird.

Not in the usual “your drawing sucks” weird. Well, the design could be improved a bit. But mainly, the individual parts don’t always line up correctly, giving the appearance of disjointed, floating hands and arms in certain perspectives. Ubisoft, if you want help enhancing your next Rayman game, call me.

This month is the first time I was able to see the character come together in the game engine. Before, I had a good idea what it would look like (exactly like the single sprite I drew), but here I didn’t know for sure until I saw it in the 3D game engine, when all the body parts were connected. And I’ll have to make some tweaks.

A friend mentioned it might be the 3D skeleton, which I made myself without consulting typical body proportions or walking poses. I may try to modify that. But most of my time will be spent with the following development cycle: 1) see character in engine, 2) pause when ill-body pose appears, 3) see which modulized sprite is visible, open in Photoshop and redraw, refresh game, 4) repeat steps 1-3.

Once the proportions look ok, I can make sure the character design is acceptable using this test character as a template. I’ve already made a drawing of what I intended the character to look like based on the actual result, which I can use to help me redraw.

This looks a little better.

This looks a little better.

As to how much detail I give to the character, I’m undecided. I will give him eyes, and maybe a cape/cloak (2D or 3D?), but the shading effects are an extra that needs more time. I agonize too much over little details like this, I really can’t afford the time to think about that.

Concept art in stages for my new character.

Concept art in stages for my new character.

I think these concept images would be an improvement over what I currently have.

I still have hope for what the final game animation looks like.

I still have hope for what the final game animation looks like.

These concept images were traced over what I currently have, so I am certain they are possible. I can make my in-game character on left look EXACTLY like the character on the right in the image above. This is the power of my research. And this very much is research, experimental and still in the early stages.

I haven’t even talked about the animation yet…

Concept animation I made in a few hours.

Concept animation I made in a few hours.

The above image is a very quick sample animation, based off of recurring images similar to this I’ve had in my head for years. Yes, it looks similar to the promo images to the recent Legend of Zelda games… the artists behind those concept images do a great job, and while the final game never looks the same, their teams work hard to bring a painterly style to the game world. I look forward to the next one. Anyway…

While I hope my final game looks better than the gif above, I am confident I can make a game that looks exactly like that if I choose. In 2D or 3D. Which inspires me a little more to move forward and try.

This will take time. At this point I have no idea when my next game will be released. My development has always started with the character, making sure it looked good in the game, and building the world and gameplay after that was finalized. The world has enough games that rely on simplified pixel-art or detailed-3D models, there’s no reason for me to work so hard to make another one of those. I will definitely focus more on gameplay and controls this time around, but I may not start implementing that until late this year. By then I hope to show more of the game as it develops.

I feel lucky that I can smile whether or not I get fired from my full-time job, knowing I have this to focus on. But in the meantime, it’s hard to satisfy both.

Oh, and E3 live presentations start tomorrow. I can’t watch every press conference live this year, but I’ll watch them when they are uploaded afterwards. The gaming industry is a tough one to succeed in, either as indie or trying to join established companies. But in the last few years, it’s been fantastic to gamers. I look forward for what comes next.