Drawing vs Rendering Pixel Art

[Frostorm]

So 1st and foremost, I am NOT an artist. But as a solo dev, I have to get my art done somehow. So this led me to decide between 2 options. The most obvious and straightforward choice would be to purchase/commission/draw pixel art directly. As I've mentioned, I'm not very good at drawing to begin with, so I've mostly just saved some of the free assets found on this forum as well buying some sprite assets (like the Tokiwa Monster Packs). The other option is one less traveled. Basically it involves rendering and animating something in 3D and then pixellating it afterward. Dead Cells is a good example of a game that utilizes this technique. This can save on many manhours if your game requires a lot of sprites/frames.

For instance, each character in my game will require 132 frames for all their actions, plus its 64x96 pixels per frame. Why so many? Well, I use LTBS (LeCode's Tactical Battle System) so combat takes place on the same map as exploration does, at least with the way I have it setup. In addition to the basic up, down, left, right walking animations, there are also attack, cast, hurt, and idle animations for each direction. Also, non-melee attack animations, such as for bows, also add to that high frame count. This is honestly so much more daunting to me than anything on the gameplay mechanics side of things, at least for an unartistic individual like me. I excel at number crunching, theorycrafting, balancing, combat mechanics, skill/spell creation, and basically all things database, so art is basically the opposite of my strengths.

So my question is this: Would rendering and animating units in 3D and then pixellating them be easier in the grand scheme of things? I understand that would involve its own workload, but it seems to at least be less daunting that drawing sprites frame by frame. I could also easily feature gear-dependent aesthetic changes in the game this way. Has anyone here tried this approach? If so, what are your thoughts/experience on the matter?

 

[Waifus69]

I tend to just edit them instead of redrawing frame by frame.
Make/get a base, then edit from there.

 

[Frostorm]

I tend to just edit them instead of redrawing frame by frame.
Make/get a base, then edit from there.

Yea, that's where I started and is how I made my Orc. But he's missing some animations since the sprite sheet he's based off on didn't feature Bow and Spear (thrust) attack animations. Oh and he's missing the getting hit/hurt animation. I also just plain suck at anything beyond selecting, moving, or copy-pasting...

Here's another article on the subject: Dead Cells

Edit: If we leave animation out of the equation, I've found several YouTube tutorials on how to pixelate 3D objects. So I guess that's another choice to make. Streamline a rendering + animation process or simply render a 3D object then pixelate a single frame at a time.

 

[Waifus69]

Yea, that's where I started and is how I made my Orc. But he's missing some animations since the sprite sheet he's based off on didn't feature Bow and Spear (thrust) attack animations. Oh and he's missing the getting hit/hurt animation. I also just plain suck at anything beyond selecting, moving, or copy-pasting...

Here's another article on the subject: Dead Cells

That sound fun, but it’ll get pretty irritating making all of those sprites after a while.

 

[Tamina]

Yea, that's where I started and is how I made my Orc. But he's missing some animations since the sprite sheet he's based off on didn't feature Bow and Spear (thrust) attack animations. Oh and he's missing the getting hit/hurt animation. I also just plain suck at anything beyond selecting, moving, or copy-pasting...

Here's another article on the subject: Dead Cells

Edit: If we leave animation out of the equation, I've found several YouTube tutorials on how to pixelate 3D objects. So I guess that's another choice to make. Streamline a rendering + animation process or simply render a 3D object then pixelate a single frame at a time.

First of all, if you are not an artist and don't have one on your team, no method will magically create awesome looking animations. But I assume you already have one if you are considering this method.

Afaik you can pixelate any type of image, not just 3D, but 2D also. If my goal is to create smooth 2D pixel art animation, I would consider creating 2D animation in 2D software like Adobe anime or harmony first, then pixelate the output images.

This method saves the headache of having to model the character and rig them, nor having to deal with rigging limitations. 3D needs very different workflow from 2D so that's a lot of work unless you have a rigged model ready.

I wouldn't do it frame by frame unless it's 8bit style or something. In fact I'd use pixelate programs for any pixel art that isn't 8 bit these days. It's much easier to create gradient in illustration software like Photoshop or illustrator first then pixelate them.

 

[hiddenone]

Moving to Resource Support.

General Resources is for posting resources you made, not for asking for help/suggestions.

 

[Frostorm]

Afaik you can pixelate any type of image, not just 3D, but 2D also. If my goal is to create smooth 2D pixel art animation, I would consider creating 2D animation in 2D software like Adobe anime or harmony first, then pixelate the output images.

Lol, idk why that didn't cross my mind! But wouldn't I need to draw the art in 2D 1st? That's kind of the hard part for me. I'm better at making simple low polygon shapes than I am at making anything 2D look halfway decent. Basically, I wanted to avoid having to "draw".

General Resources is for posting resources you made, not for asking for help/suggestions.

Ah, gotcha. Duly noted.

 

[Waifus69]

There’s alway the way of start off small, then get bolder and bolder till you improve.

 

[Frostorm]

Sure, I get that there's always room for growth and improvement - and that applies to all subjects. However, I made this thread because I wanted to get some experience from those who have tried this technique. Being told "don't do it" w/o any solid reasons or accounts of why it didn't work isn't really helpful tbh. So far all I'm hearing is "don't bother, you're better off drawing by hand even if you're bad at it"...

It's supposedly a technique that's gaining popularity among indie devs so I'm hoping someone here has some experience with it.

 

[Tamina]

Lol, idk why that didn't cross my mind! But wouldn't I need to draw the art in 2D 1st? That's kind of the hard part for me. I'm better at making simple low polygon shapes than I am at making anything 2D look halfway decent. Basically, I wanted to avoid having to "draw".

3D art is not any easier than 2D IMO it's much harder, since you need the design and animation skill required in 2D AND modelling+rigging skill.

Keep in mind that in studios, character designer is 1 position, animator is 1, 3D is at least 1-3 different positions (depending on whether they want to separate modeler/texture/rigging). That's because it needs entire different sets of skills.

Making simple low polygon shapes very, very different from making a humanoid model with good designs, rig them and make sure nothing penetrates, then animate them.

That being said, if you are interested in playing with 3D programs, feel free to try them. Personally, I wouldn't touch 3D just because how much extra works involved in entire asset creation process.

If I'm a game designer that can't draw, I'd just hire someone or use stock assets so I can focus on finishing the game. Learning entire set of new skill that takes TONS of time is the last thing I'd consider since Id prefer to focus on game design and finishing a project instead. If you are solo dev, you are probably just digging a hole for yourself with all that 3D workflow which makes the project so much harder to finish.

Sure, I get that there's always room for growth and improvement - and that applies to all subjects. However, I made this thread because I wanted to get some experience from those who have tried this technique. Being told "don't do it" w/o any solid reasons or accounts of why it didn't work isn't really helpful tbh...

I've worked in a mobile studio using 3D to 2D method before. Although it wasn't not pixel art the process was similar. I don't recommend this method to solo dev or small team in any way.

First a character is designed by a character designer that can draw, then model was created by a 3D artist who also can draw(since they need to draw the textures), finally the model was given to animators who has to rig and animate.

Each character from start to finish takes 10-20 working days or maybe more.

It wasn't easy to animate 3D character with 2D style in mind. 3D rigs has "rigging limitations" so it was tougher to stretch out arms and legs for dynamic poses. This is less of an issue in 3D game but in 2D game silhouette is so important. So we often had to do all sorts of modifications to rigs to make it work.

Basically, this entire process involves multiple people that can draw as well as people with a more technical skills like animation or rigging.

I wouldn't consider this method unless I have a much bigger team than just a few people. If your team have difficult to even produce simple drawings with a few frames, then using this method isn't making simpler, it's the opposite. I suggest you cut down the work required by reducing animations, use stock or hire someone, not the other way around.

 

[Frostorm]

Isn't the whole point of this process to minimize workload? From what I've read in the links and videos, you don't need to texture the model and there's no need to fix penetration issues or details like that. You just need to make something with the general shape. The rest is done via cell shading and the pixelation filters. With traditional 3D animation, yea it's gonna be a lot of work. But since our goal is pixel art, we can just focus on keyframes. This would reduce the workload considerably. Also, unlike 2D art, we can reuse these assets if something needs to be changed, instead of redrawing the frame - for fixing a pose for example. I just don't think this method can be likened to a traditional 3D workflow...it's very different. There's also a bunch of already made assets in 3D just like there is for 2D. If I hadn't already exhausted most of the available resources on this forum (and felt it was insufficient to complete my game with) then I wouldn't have bothered looking into this technique. I need to find some way to get the work done regardless. If drawing 2D sprites frame by frame is more efficient, then so be it, but I highly doubt that's the case.

Edit: I found another video on the subject:

 

[Tamina]

Isn't the whole point of this process to minimize workload? From what I've read in the links and videos, you don't need to texture the model and there's no need to fix penetration issues or details like that. You just need to make something with the general shape. The rest is done via cell shading and the pixelation filters. With traditional 3D animation, yea it's gonna be a lot of work. But since our goal is pixel art, we can just focus on keyframes. This would reduce the workload considerably. Also, unlike 2D art, we can reuse these assets if something needs to be changed, instead of redrawing the frame - for fixing a pose for example. I just don't think this method can be likened to a traditional 3D workflow...it's very different.

Even if you bypass the rigging and texturing part, you still need 2 other skillset that you absolutely can't skip: design and animation. Both skill takes much longer to develop than "drawing". If you have difficulty learning drawing skill, then you probably have never learned design and animation in the past since those skills are related to each other. Trying to learn design/animation well without drawing fundamental takes years.

Reuse the asset can be a good reason to use 3D if your game has 100 monster or something, otherwise I'd just do a hue change in 2d if I want a new monster with the same look.

I mean, you can try to play with this workflow. I'd suggest you keep the schedule in mind when you try though. Too many solo devs wanted to try new things and their project never gets done.

There's also a bunch of already made assets in 3D just like there is for 2D. If I hadn't already exhausted most of the available resources on this forum (and felt it was insufficient to complete my game with) then I wouldn't have bothered looking into this technique. I need to find some way to get the work done regardless. If drawing 2D sprites frame by frame is more efficient, then so be it, but I highly doubt that's the case.

It's probably not. But reading what you say here I think the issue of your project is that you are inexperienced in art asset production, so you underestimate the resources required to create a game with custom art.

If this is the weakness of your team then digging more holes for this project doesn't help. I'd cut down the animation required in the game rather than looking into an entire new workflow that needs time to experiment and learn. Unless your goal is learning not finishing the game that is.

Do not underestimate the "art skill" required to design and anime, it's not an easy solution for people that can't draw. It's an entire different beast on its own.

 

[Finnuval]

I think in the end how efficiënt any of these to techniques is going to be comes down to skill...

I would be much more efficiënt drawing 2D then I would everr be in 3D - thats because in 2D I got skill and in 3D I do not.

So the question here is going to be do you have the skill in 3D to be noticably more efficiënt then in 2D and is that the case is it still more efficiënt then hiring someone else to do it in 2D?

If both 2D and 3D skills are similar though I think 2D is the more efficiënt method as the skills required are basically the same as for 3D with the exclusion of some and anyone can learn to draw Just like anyone can learn to model

 

[Frostorm]

Yea, character (or monster) design is probably going to be my biggest obstacle, whether 2D or 3D. Luckily for animation, I only need to make a few simple ones. Basically swinging a sword, thrusting a spear, shooting a bow, and waving hands around (casting). It looks like walking and/or running is already provided in several 3D programs, so I'm thankful for that at least.

I do have a bit more experience working with 3D models than I do drawing anything 2D. But they were objects (mostly swords and other weapons) and not humanoids/people. Still, it's not enough experience to be called "good", not by a long shot.

I guess in the end, there really aren't any RPG Maker users using this kind of method/technique... (Please chime in if you have)

 

[MadRamDesigns]

Is it an option? Sure. If you have the time and commitment to learn some new tools.
Due to computer limitations back in the 90s, it wasn't uncommon for developers to use 3d-to-2d for making sprites.
There's certainly are a lot more resources (forums, tutorials, assets, etc) for 3d programs like Blender than even a few years ago.
I use it for taking the guesswork out of certain complex shapes in topdown and isometric views. Nothing animated (yet).

"Easier" depends on the person, some will have a quicker aptitude for it than others.
It also depends on chosen style and complexity of the sprites.
If they're all chibis, then it probably easier to stick to learning 2d; if they're more proportionate to reality, then 3d-pixel has an advantage.
As for animation, an option depending on the style you're aiming for, you can probably find motion capture packages for some of those actions; or look into DIY mocap.
Though the others are right, you'd still need to have some design sense to keep enemies and characters visually cohesive, color choice, etc.

 

[Finnuval]

you'd still need to have some design sense to keep enemies and characters visually cohesive, color choice, etc.

Which is why concept-artist and design-lead are very important industry Jobs though often overlooked (well concept artist mostly lol) xD

 

[gstv87]

rendering and animating something in 3D and then pixellating it afterward

why don't you just render at low resolution?

 

[Frostorm]

why don't you just render at low resolution?

If that can achieve the look I'm after, then I'm not opposed to it.

As for animation, an option depending on the style you're aiming for, you can probably find motion capture packages for some of those actions; or look into DIY mocap.

I was actually considering this as well. The technology for it has become a lot more mainstream & available lately.

 

[gstv87]

If that can achieve the look I'm after, then I'm not opposed to it.

medium resolution 3D model, painted with medium resolution textures improved by high quality shaders, rendered "as is", no compression.
you can't really see the pixels from a distance, but if you go deeper, they clearly show up.
granted, the art style is bordering realism, closer to that of Diablo or Baldur's Gate (32-bit era)
it doesn't really benefit from the pixel look, which is more from the 16-bit era.

the key there being "the look *you* are after", which is going to have to be a compromise between what you can model, what you can paint, and how big the final sprite will end up being, so as to make the individual pixels noticeable.

my sprite? the pixels are barely noticeable at the size I'll be working with, and the grading of the colors blends them even more.
the thing I focused on there, was workflow: build the model, paint it with standarized textures, use the same shader for all characters, pose it, and send it to a preset renderer.
3Dstudio then does all the work of putting it together and saving it.

and yes, I'm using a tactical system as well, so you bet I know what it's like to render a 1000x800 sheet and how you have to shave colors off the palette to keep it lightweight.

 

[Frostorm]

@gstv87 That's very nice work you got there! My player sprites are going to be 64x96, though some of that is empty space so it actually looks like 48x72. This is what my Orc looks like atm, which was made by basically mix-mashing different bases together.

But he's missing a couple essential animations, as I've mentioned. So far he's really the only character I've made a sprite sheet for. Everyone else is just using placeholders. Therefore, I'm in a good position to decide on a method/process for making their sprites, which will undoutedly affect their final look/art style.