You can find or cheaply buy premade models that are textured, rigged, animated, etc.

If you're avoiding normal maps, you're avoiding the wrong part of the art creation process. Professional artists don't usually paint those by hand; they're mostly all automatically generated by the artist's tools.

And of course your characters will be affected by the lighting, cel-shading or not (cel-shading is still *shading* and still requires light calculations... and the screenshots of Megaman Legends are _definitely_ lit and... definitely not cel-shaded, at least not on the main character).

Of course, you just drop all the basic art assets into a Unity or Unreal game and you get all the graphics math done for you and you can concentrate on the hard parts of making a game that you're overlooking (making a fully modeled, painted, rigged, and animated character; designing and implementing all the gameplay; actually finishing a project instead of getting bored and quitting two months into it; etc.)

I don't think you could normal map a face automatically with low res models in a way that looks good. But then again, I know jack all about it.

In Megaman Legends the textures/shadows on characters are pixel art, and drawn, which is what I'm referring to. I meant cell shaded as in the aesthetic style of 2-3 tone flat colors using pixel art or otherwise, as opposed to a shader that effects lighting. The environments are lit, but it's pretty basic, and just to show depth. MML2 appears do some lighting with the characters as some of the screenshots have colors with a slight gradient, but I can't say what it is specifically. I'd be aiming for flat colors anyway.

Unity certainly isn't going to do graphical work automatically in any way that's graphically acceptable, everything takes some degree of tweaking and work. And no, I'm not overlooking any of the other stuff you mentioned. That's just stuff that's necessary to development, and I'd be more interested in how to simplify those processes the same way I aim to simplify visuals.

Let me tell you, yes you can!

It depends on the tools you use (some are better at baking normal maps than others, 3D Coat seems to be much better at it than Blender for example), and your own skill at creating high poly and low poly meshes (if your low poly mesh is crap, baking will be a nightmare, but I guess than actually rigging and animating that mesh would be even worse of a nightmare).

Of course, creating detailed high poly assets is very time consuming and will require lots of skill... which is exactly why many low cost Indie games go with low poly assets without detailed normal maps.

That doesn't mean that every kind of normal map will cost you many hours of time to create. For technical assets that might be much less of a problem... actually, ANY asset that is not a realistic looking human is easier to create a high-poly asset for.

Why so dismissive of Unity? Its not the most advanced engine out there, and true, you still need to tweak and configure yourself. As long as you go with a simple 3D game and have the skills, using a framework or writing from scratch CAN be the better idea. But really, you don't seem to have this kind of skill, so Unity WILL be faster for you (unless you want to learn to write games from scratch of course).

Let me tell you, when you talk about "simple 3D art", and you actually mean it, Unity is the right engine for you. Name one thing that needs to be tweaked in Unity to make simple 3D graphics possible, actually. Should be pretty much out-of-the-box when you leave out normal mapping, postprocess effects, or any kind of complex lighting.

The shader discussion is definitely good for imrpoving the looking. As far as the assets themselves, honestly, Unity Asset Store and OpenGameArt.org

It does limit what you can do (unless you're going for generic fantasy / horror / scifi) but it can cut out on the amount of assets needed A LOT, especially if you spend some time researching and make some artistic compromises.

There's also a lot of ways you can creatively cheat and reuse assets. Here's an article from our own game: How we Recreated All Character Models from Scratch in Two Weeks

I mean, buying assets sort of goes beyond the boundary of working on a game alone. It goes without saying that working in a team or outsourcing stuff will be easier, however this is more about ways to simplify development as opposed to..avoiding it lol.

?

Are we now going into a philosophical rant about what is the correct way to 'lone wolf' game development? About what is correct game development and what is not? Really?

There was a good article on gamasutra about that by some Indie devs months back, about staying scrappy to survive... sadly I cannot find it, and it is actually about surviving as an Indie, which might not apply to your case.

Fact is, as an Indie, especially as a lone wolf, game development is all about making compromises (as you do in the opening posts by trying to go with simple 3D graphics)... for many that means that they come to the realization that they cannot draw or 3D model to save their life, and rather stay sharp in programming and/or game design, concentrate their precious little time on that, so they compromise by outsourcing art.

Which is an extremly clever thing to do. Creating art is, well, an art, and needs 1000's of hours of expierience to get to proficiency. Not many programmers can do that while keeping their programming skills fresh, AND then having to learn about game engines, 3D graphic techniques and game design.

While on the other hand a trained artist might crank out a simple asset in mere hours or less, thus quite cheaply, while the end product looking vastly superior than anything you could have done in days and weeks of work (especially if you don't work on it fulltime).

It is completly fine if you want to also do the art alone (I do so myself, but then I was always doing art as a hobby besides being a fulltime programmer, and I have to say my personal projects get slowed down considerably because I take the double role of programmer and artist)...

Just do it for the right reasons, and be aware that you throw out the best way of finishing your game/project in a sensible timeframe instead of wasting years even on a simple project. Which can be completly fine if you do this as a hobby, and you are interested in art.

If you CAN find an artist that is willing to work with you, and shares your vision while not costing you an arm and a leg, you SHOULD try to work with him definitily. It would make you more than double as efficient as you can now concentrate on one thing while the artist frees you of art workload. Of course finding such an artist might not be that easy...

I mean, buying assets sort of goes beyond the boundary of working on a game alone. It goes without saying that working in a team or outsourcing stuff will be easier, however this is more about ways to simplify development as opposed to..avoiding it lol.

?

Are we now going into a philosophical rant about what is the correct way to 'lone wolf' game development? About what is correct game development and what is not? Really?

Hear, hear. Kyrieru, your complication ("it's impure to buy assets"), flies in the face of your desire to simplify the process. To use readymade assets is to simplify the process greatly.

Both processes are valid lines of thought:

A) Where can I get assets from that I cannot realistically create enough of?

B) How can I change the nature of my game to not require workload-heavy assets?

Some games successfully pursue the latter method, and are arguably better off for it.

Ofcourse, both can be used together (i.e. outsource or use stock music, but simplify your art style, and/or use procedural generation).

I don't think you could normal map a face automatically with low res models in a way that looks good. But then again, I know jack all about it.

I promise you, it's done that way with modern tools. You don't author low-poly models. You author art to the highest quality you are capable of doing and able to afford. Then you use tools (with perhaps some light manual assistance) to simplify the polygonal model and automatically generate various texture maps from the high-detail version. Many art workflows now outright require this as the "best" tools for creating 3D models (i.e. ZBrush and the like) will result in millions and millions of polygons and outright require automatic simplification to be usable in a game.

The main problem I had the last time I messed around with it was jaggy shadows, and I thought normal maps were the only solution, but if there's a way to get around that without too much extra work then I'd have no problem with flat drawn textures + cell shaded characters.

If you look at the games that actually used a lot of low-poly models, shadows weren't exactly a big thing in those games. Normal maps don't make rough models have "smooth" lighting; they actually do the opposite by giving models more fine, realistic details to make lighting _less_ smooth.

You use normals (not normal maps... just normals) for smoothing lighting on models. The problem is that you can still be left with bad silhouettes on low-poly models, which is one reason we keep pushing for higher-detailed models, and now use techniques like parallax mapping or other displacement mapping (normal maps on steroids, sort of).

Unity certainly isn't going to do graphical work automatically in any way that's graphically acceptable, everything takes some degree of tweaking and work.

Of course. But to counter your second point... you can either spend your time working on that graphical tweaking or you can spend your time doing EVERYTHING else. Which is a truly stupendous amount. If you were just going to whip up a graphical demo, there's the core windowing, input, timing, core mesh rendering, model importing, skeletal animation, debug drawing, basic lighting, etc. If you want more than a basic demo then there's also the object system, serialization, audio, gameplay code, GUI integration, level editor, object editor, scripting language, etc. to do.

Starting with an engine means you can dedicate your time to the parts you actually care about. Plus, engines like Unreal or Unity have online stores, so (a) you might just be able to buy a premade set of shaders that already do exactly what you want, or (b) if you make your own you can sell them for a profit to other developers.

If you want to simplify 3d the fastest route is to use an engine.

You then have all the pbr rendering, model loading, animation, the full whack. You'll easily spend all your time creating these things yourself and if you can shortcut all that you massively simplify the entry path into 3d development.

I tried to create a game from scratch in 3d. I simplified by making all my models cubes (no it wasn't a minecraft clone) and still it took me months and I didn't have shadow maps and lighting working right.

I then started a much more complex game using an engine and the head start I had boosted the development by years easily.

Good luck!

Edit: I use ue4 and it does all those things listed out of the box so all you need to do is find or buy some good 3d art, but I'm sure unity can do the same this along with many other engines.

I promise you, it's done that way with modern tools. You don't author low-poly models. You author art to the highest quality you are capable of doing and able to afford. Then you use tools (with perhaps some light manual assistance) to simplify the polygonal model and automatically generate various texture maps from the high-detail version. Many art workflows now outright require this as the "best" tools for creating 3D models (i.e. ZBrush and the like) will result in millions and millions of polygons and outright require automatic simplification to be usable in a game.

While that makes sense for modern poly-counts, I'm not so sure that it would be faster when you're creating models with extremely low poly-counts that rely more on textures. (Like these, for example)

http://i.imgur.com/noNscbp.png

https://s-media-cache-ak0.pinimg.com/736x/99/31/78/993178756108b65fc4d69e1dbdd0a04e.jpg

At least for me, it's much easier to build out a model from scratch, or from a model sheet, rather than with sculpting. But that's only because I focus on extremely low res

characters.

If you look at the games that actually used a lot of low-poly models, shadows weren't exactly a big thing in those games. Normal maps don't make rough models have "smooth" lighting; they actually do the opposite by giving models more fine, realistic details to make lighting _less_ smooth.

You use normals (not normal maps... just normals) for smoothing lighting on models. The problem is that you can still be left with bad silhouettes on low-poly models, which is one reason we keep pushing for higher-detailed models, and now use techniques like parallax mapping or other displacement mapping (normal maps on steroids, sort of).

I didn't say they made things smooth, I said that shadows can appear jaggy with low res models. However that probably isn't the right word for it either. What I meant was that there aren't enough edges/normals to get lighting to behave the way you like on say, a face. For example, you may get the silhouette you want, but you would have to add a bunch of edges just to get shadows to "frame" certain features properly. I was under the impression that this is where normal maps would come in (By smoothing the model, and using it as a normal map, or by starting with a high res model as you mentioned.)

Of course. But to counter your second point... you can either spend your time working on that graphical tweaking or you can spend your time doing EVERYTHING else. Which is a truly stupendous amount. If you were just going to whip up a graphical demo, there's the core windowing, input, timing, core mesh rendering, model importing, skeletal animation, debug drawing, basic lighting, etc. If you want more than a basic demo then there's also the object system, serialization, audio, gameplay code, GUI integration, level editor, object editor, scripting language, etc. to do.

Starting with an engine means you can dedicate your time to the parts you actually care about. Plus, engines like Unreal or Unity have online stores, so (a) you might just be able to buy a premade set of shaders that already do exactly what you want, or (b) if you make your own you can sell them for a profit to other developers.

What? Why do people keep thinking that I was saying not to use Unity or an engine?

Like I said in the response above. I was simply saying that even if you have something like lighting being calculated automatically by an engine, it doesn't mean that you don't have to set up lighting in your scenes/areas, which is still an extra step over relying more on textures. (Unless you focus very little on textures at all, at which point it might be the same amount of work to have good lighting.)

I'm not being dismissive of Unity. I would be using Unity. What I was saying is that Unity isn't going to create normal maps for you, or do any of other leg-work needed for lighting and shaders to actually look passable on your models. Not to mention, even if you had the best dynamic lighting and normal maps handed to you, it still means setting up lights in your scenes.

@Tom and Gian: Stop putting words in my mouth -__-

I didn't bash outsourcing assets in principle. I merely said that avoiding a process isn't the same as simplifying it, which is what the discussion is about. Sure, not everyone can make graphics, but that goes without saying, and so it's much more productive to talk about simplifying that process for those who will be involved in it. I was just trying to steer discussion away from outsourcing, not decry it.

That said, I understand that I didn't make it clear in the OP that I wanted to focus on individual processes, as opposed to simply "game development" as the act of releasing a game at all. I'll edit it to make it more clear.

Thanks for the responses so far.

Sean and Braindigitalis have said it very well already, just to strengthen their points:

No, the engine will NOT do the work creating assets for you. It will lighten the workload on creating the technical part of the game though, putting a ready made renderer, and a framework to run your game logic into your hands that you do not need to care about anymore, giving you MORE time to create your assets.

Of course there are people that like to write their games from scratch, and depending on your game and your expierience in programming, it might not be that much more work to do so (and might result in a leaner game with less overhead as the "engine" part of the game is more adapted to your game)... but using a readymade engine DOES save time, else the professional studios would not have switched to using third party engines or re-using their own engines with most of their AAA titles.

If the pros do it, it cannot be completly wrong.

Totally fine if you do not want to talk about outsourcing. Also totally fine if you not WANT to outsource for other reasons. Fact is, if you want max quality for the least amount of your own time involved, outsourcing to a professional artist is the best bet (might not be the cheapest moneywise, but that is a different topic altogether).

Of course, there are many valid points why you wouldn't want to outsource, and I respect that decision... maybe you should have made that clearer from the start in this topic (as you wrote above)...

I promise you, it's done that way with modern tools. You don't author low-poly models. You author art to the highest quality you are capable of doing and able to afford. Then you use tools (with perhaps some light manual assistance) to simplify the polygonal model and automatically generate various texture maps from the high-detail version. Many art workflows now outright require this as the "best" tools for creating 3D models (i.e. ZBrush and the like) will result in millions and millions of polygons and outright require automatic simplification to be usable in a game.

While that makes sense for modern poly-counts, I'm not so sure that it would be faster when you're creating models with extremely low poly-counts that rely more on textures. (Like these, for example)

http://i.imgur.com/noNscbp.png

https://s-media-cache-ak0.pinimg.com/736x/99/31/78/993178756108b65fc4d69e1dbdd0a04e.jpg

At least for me, it's much easier to build out a model from scratch, or from a model sheet, rather than with sculpting. But that's only because I focus on extremely low res

characters.

If you look at the games that actually used a lot of low-poly models, shadows weren't exactly a big thing in those games. Normal maps don't make rough models have "smooth" lighting; they actually do the opposite by giving models more fine, realistic details to make lighting _less_ smooth.

You use normals (not normal maps... just normals) for smoothing lighting on models. The problem is that you can still be left with bad silhouettes on low-poly models, which is one reason we keep pushing for higher-detailed models, and now use techniques like parallax mapping or other displacement mapping (normal maps on steroids, sort of).

I didn't say they made things smooth, I said that shadows can appear jaggy with low res models. However that probably isn't the right word for it either. What I meant was that there aren't enough edges/normals to get lighting to behave the way you like on say, a face. For example, you may get the silhouette you want, but you would have to add a bunch of edges just to get shadows to "frame" certain features properly. I was under the impression that this is where normal maps would come in (By smoothing the model, and using it as a normal map, or by starting with a high res model as you mentioned.)

?While I am not sure why you are so concentrated on making "low-poly vs high-poly" look like a radical difference in process (it is not, at least as long as you use a good tool), there certainly is a tangible gain in speed by going more low-poly and not sculpting high-poly details that need baking (baking does NOT take much time, sculpting details does as said before).

That said, it depends on what art style you want to achieve. If your art style is built around the desire to go low-poly, or even better, if your art style just happens to look better with low poly modelling, your game will look good even without sculpting high-poly meshes and baking out normal maps...

BUT: you will most probably want to think twice about modelling faces then. Your faces will most probably not look to good with low poly sculpting, as long as you want to achieve a realistic result... hence most new low-poly art styles leaving faces unmodelled.

You will HAVE to adjust your art style to accomodate the lower poly count... textures can only do so much.

Of course you can experimentwith leaving away lighting, in which case the amount of polys might not matter so much anymore.... but if you start turning the camera, you might still start to see ugly signs of the low poly count.

You will HAVE to experiment a lot to find the art style that makes your low poly assets look good, you will HAVE to experiment with lighting or how make your asset look good without lighting, you will HAVE to experiment with your pipeline to get the speed benefits you hope for.

Maybe you should talk about what exactly you want to achieve, so people know what we are talking about. There are many different low poly art styles, and there are many things that are VERY hard to achieve with a low poly art style.

@Gian: Again, I never said anything against using an engine, you even quoted the part where I said so -___-

However it's worth mentioning as a general tip.

As for the rest of what you were saying, I pretty much agree.

Low poly and high poly aren't always that different, but it depends on the method. Personally, it's not so much that I prefer one over the other aesthetically, but more that I prefer the method of being deliberate with every vertices, and I find simple styles somehow appealing. That's just personal preference though.

In terms of what I want to achieve, I don't think it's too important. While I have an idea of what I'd do, my thoughts on making it practical only go as far as visuals, and in terms of making development in general practical I only have the 2d games I've released to go by. While I want to figure out what I should eventually work towards, I'm just as interested in what people think to do themselves, and how it might change what I plan on doing.

However, to at least give a visual example of the tests I did when I was think about styles.

https://i.gyazo.com/48f851cacec6c80a50c5f40dd4990036.mp4

https://i.gyazo.com/c2ced30e025bac9bcc69d4311e6623bb.png

https://i.gyazo.com/c71822c54a9feb3f2328acd162a3dca9.png

https://i.gyazo.com/c1071b458bf8f4a08929a914301de8b0.png

I wouldn't really say that those tests represent what I want to achieve, though. I'm still not 100% sure what kind of style I'd want to go for, other than the fact that I'd go low-poly. I'd also take the time to learn more about modeling than I know now before actually attempting anything.

Okay, just to be sure... is your last post a question of some sort? It sounds like it even though I am not sure what you are asking.

Is it that you are unsure where to go, which style to choose? If this is what you are looking for, you will find it hard to get good answers. There are a ton of lowpoly styles, each with its own pros and cons, and it depends completly on your own game and visual concept what fits or enhances the game or what is out of place.

I would first try to come up with a game idea, and then choose a visual style that is ideal for this game idea.

BTW, your test art is really good! I like what you did there. That kind of 3D art would go well with a JRPG or similar style of game, something where a japanese anime style would fit.