20 replies to this topic

[alphster]

Hey guys,

I was curious as to why I see so many indie games being made using old-school graphics (like SNES graphics). Especially in 2d games..

My game programming is limited to lightly goofing around with XNA - but it seems to me that if you are capable of create good 2d graphics, it's super easy to import beautiful images that retain their level of detail even when rendered in the game.

So this leave me to believe one of these:
- game developers are purposely using SNES-like images to give their game a "nostalgic" feel.
- game developers are pushing really hard to make their game as little kbytes as possible by using low resolution images
- something else?

Can someone shine some light on this for me? It's just kept me very curious...

[Erik Rufelt]

Sometimes a bit for the nostalgic feel, sometimes because artists like to draw those images. The real important reason is that it's a lot harder to draw nice high-resolution graphics. There are challenges to both aspects of course, but when drawing a 128x128 smooth 32-bit color sprite, in order for the player to identify with it, you need to make it much more realistic than a 16x16 or 32x32 sprite. At those sizes you see the eyes, the facial expression, etc. It requires a lot more time and skill to draw such images and make them believable. Low-res graphics are a lot more forgiving.

For the player to properly identify with your sprite, you need to make it appealing. At 16x16 most things are up to the imagination, and giving a few pixels of hair and 2 red pixels for the mouth makes people see it as a pretty girl. If you want to do that at 128x128 you must be extremely good at drawing, and apart from that, you need to make the face you draw actually pretty. This is far from easy even for experienced artists.

[alphster]

thanks for the info, i now understand why graphics designers choose to create pixel-like graphics because it's easier to give "life" to it.

however, i'm still a little confused on the technical part of it...

lets say i was one of these game developers and creating a game in XNA (with SNES-like graphics) .
Do you just create a low resolution image in a photoshop program, and then save it as a .tga (for example), and then just import it? It seems weird to me that we would have these low resolution sprites which are built up of pixels, but then we import them into a game window where it really has nothing to do with pixels at all. In the end, it's just a "blocky" sprite where one pixel on the sprite has nothing to do with one pixel in the game window.

i thought there was something more "technical" about chosing to make these kinds of graphics... any more input on this thought? thanks!

[SuperVGA]

thanks for the info, i now understand why graphics designers choose to create pixel-like graphics because it's easier to give "life" to it.

however, i'm still a little confused on the technical part of it...

lets say i was one of these game developers and creating a game in XNA (with SNES-like graphics) .
Do you just create a low resolution image in a photoshop program, and then save it as a .tga (for example), and then just import it? It seems weird to me that we would have these low resolution sprites which are built up of pixels, but then we import them into a game window where it really has nothing to do with pixels at all. In the end, it's just a "blocky" sprite where one pixel on the sprite has nothing to do with one pixel in the game window.

i thought there was something more "technical" about chosing to make these kinds of graphics... any more input on this thought? thanks!

Well, drawing sprites is, although working with graphics and artwork also technical.
But even prerendered 3D models in low resolutions have the forgiving feature of low detail. The 3D modeler can omit many details.
Sure, using photoshop makes it much easier to acheive the nostalgic feel.
Stating that "a game window which has nothing to do with pixels" is untrue. Of course it has something to do with pixels! Everything in fact; it's nothing but pixels.
Yep, that's how i'd do it. Or cell-shading some low-polycount 3d models, rendering them in low res and applying a border. Will make excellent SNES-looking meshes, that are much easier to animate...

[Ravyne]

Its just a different aesthetic -- like choosing 3D, hand-drawn animation, or applying a cartoon shader (or even choosing something as simple as the "id" asthetic -- little light, lots of browns and greys). Sometimes its chosen to evoke nostalgia, sometimes its chosen because a simpler aesthetic just "fits" with simpler gameplay, sometimes its chosen just because that's what the artist wanted. Its also easier to produce lower-resolution graphics -- even in higher-resolution 2D graphics, every pixel is scrutinized, so doubling the pixels in both directions means 4x the work.

Ultimately you just have to take it how its served to you. Questioning this is like questioning a painter why she chose a particular brush or color -- there could be a "good" reason, or it could just have been a whim. Certain approaches may be more common, but that doesn't invalidate every other option, especially when you're just doing your own thing. An artist's reasons are their own.

[MeshGearFox]

Most personal computers are *perfectly* capable of displaying 3D models which have already been created. This is because, one a model has been created and released, it has also been finalized. Finalizing a 3D model is a form of compression that causes it to need vastly reduced amounts of memory to display.

Unfortunately creating 3D models is a very different story. There are, at present, very few, if any, home computers that are capable of handling this very system intensive process. If creating a 3D model by yourself were a feasible process, I think most indie game devs would go that way, but it's not. It's just not possible.

[Erik Rufelt]

Unfortunately creating 3D models is a very different story. There are, at present, very few, if any, home computers that are capable of handling this very system intensive process. If creating a 3D model by yourself were a feasible process, I think most indie game devs would go that way, but it's not. It's just not possible.

.......... is that a joke?

[Servant of the Lord]

Unfortunately creating 3D models is a very different story. There are, at present, very few, if any, home computers that are capable of handling this very system intensive process. If creating a 3D model by yourself were a feasible process, I think most indie game devs would go that way, but it's not. It's just not possible.

My computer (4 years old) is perfectly capable of creating 3D models... My previous computer (which would at this time be 14 years old) was also capable of making 3D models. The reason why I don't make 3D models, is because I haven't yet learned how, not because personal computers can't create them.

[LorenzoGatti]

My game programming is limited to lightly goofing around with XNA - but it seems to me that if you are capable of create good 2d graphics, it's super easy to import beautiful images that retain their level of detail even when rendered in the game.

For most people's taste "good 2D graphics" are exactly "SNES-like" pixel art pieces: they are "beautiful images" because they are drawn, pixel by pixel, and they "retain their level of detail" because they are rendered faithfully as sprites rather than deformed as textures on a 3D model.

There is no "modern", different way to do 2D graphics: as already noted, sprites want to be small and technology improvement stops at faithfully rendering large numbers of large alpha-blended true-colour sprites rather than faithfully rendering few, small, palette-constrained and alpha-keyed sprites.

[Tom KQT]

Sometimes a bit for the nostalgic feel, sometimes because artists like to draw those images. The real important reason is that it's a lot harder to draw nice high-resolution graphics. There are challenges to both aspects of course, but when drawing a 128x128 smooth 32-bit color sprite, in order for the player to identify with it, you need to make it much more realistic than a 16x16 or 32x32 sprite. At those sizes you see the eyes, the facial expression, etc. It requires a lot more time and skill to draw such images and make them believable. Low-res graphics are a lot more forgiving.

For the player to properly identify with your sprite, you need to make it appealing. At 16x16 most things are up to the imagination, and giving a few pixels of hair and 2 red pixels for the mouth makes people see it as a pretty girl. If you want to do that at 128x128 you must be extremely good at drawing, and apart from that, you need to make the face you draw actually pretty. This is far from easy even for experienced artists.

Well it's not that straightforward IMHO. Sometimes it's harder to make a good 16x16 pixels large 16-colored character than to make him hi-res. There's incredibly little space in those 16x16 pixels and as you say 2 red pixels for mouth etc - just try it, try to draw a character on a very low resolution so that it will look like a girl or Indiana Jones or whatever. It's sometimes hard to describe something properly with such huge limitations. And it's even worse - your character must have walking, jumping and shooting animations and maybe must even have different representations for different views (depends on the game of course) and during all this, he still must look like a girl or Indiana Jones. In 16x16 pixels

You cannot generalise it and say that it's easier and more forgiving to make low-res graphics. I take the artists from the old days as geniuses Nowadays they make a 3D model and (provided we're talking about a 2D sprite-based game) render it as many times as they want, in different poses etc. It's also hard, no doubt, but in a quite different meaning...

Unfortunately creating 3D models is a very different story. There are, at present, very few, if any, home computers that are capable of handling this very system intensive process. If creating a 3D model by yourself were a feasible process, I think most indie game devs would go that way, but it's not. It's just not possible.

That's crazy. 3d studio started in DOS times and was capable of making 3D models. Do you remember the computers back then?
I don't see what's so system intensive on making 3D models. On the contrary - it requires almost nothing from the computer. The 3D modelling SWs use DirectX or OpenGL to display the model, exactly as a game would. And even if the model is made with a large polygon count (to create normal maps etc.), you are usually working just with that one model at a time so it's still probably MUCH less polygons than modern games display all the time.
Maybe you're confusing it with photorealistic rendering - that can be very CPU intensive, but still is absolutely doable on any current PC machine, you'll just wait longer for the result. But that's not related to making 3D models for games.
And also - there are tons of indie games with 3D models.

[frioux]

Why painting a portrait when you can take a photo. Thats the same. Its art.

[Dawoodoz]

Sometimes, you want to be able to destroy anything like in Worms and Liero but you don't want to spend years developing a voxel engine like in Outcast or get holes from rounding errors like in RedFaction.

[MarkS]

Most personal computers are *perfectly* capable of displaying 3D models which have already been created. This is because, one a model has been created and released, it has also been finalized. Finalizing a 3D model is a form of compression that causes it to need vastly reduced amounts of memory to display.

Unfortunately creating 3D models is a very different story. There are, at present, very few, if any, home computers that are capable of handling this very system intensive process. If creating a 3D model by yourself were a feasible process, I think most indie game devs would go that way, but it's not. It's just not possible.

As someone that has created some rather complicated and polygon intensive 3D models on my home PC, which was out of date when I built it four years ago, all I can think to say is: "WHAT?!"

"Finalized 3D model?" "Compressed?"

[NEXUSKill]

There are quite a few reasons to go pixelated, space is rarely the case, in indie projects the most important rule is to go with what you have, most people would go pixelated mainly because it comes easy to them, Indie teams if not one-men armies, tend to be very reduced and get together through circumstantial availability rather than to supply specific needs, like searching for a 3D modeler because you want your game to be 3d, you work with what you got with the people that wants to work, so if your artist is good with vectorial graphics, thats what you go with, if its pixelated, you go pixelated.

It is also a pretty effective style choice if you do it right, a lot of people like pixelated games out of nostalgia, or mere taste, that's why from time to time larger companies also choose to go pixelated.

State of the art high definition graphics DO NOT make a game good, that is a rule, I've seen some of the worst games in the industry have the most overproduced high tech graphics of their time, and they still suck.

With current download speeds and storage capabilities, and as far as an indie developer would be concerned and able to produce, striving to make a game super low weight is rarely a concern unless you go crazy with insane amounts of content, which as an indie project is bound to fail to publish due to eternal unfinishedness (yeap, made up word, but descriptive enough)

Indie developers tend to do their projects parallel to their day job, few indie developers are self sustaining and have the luxury of being full time indie developers, the focus of any indie project is to produce a fun, creative game in as little time as possible and withe the few resources they tend to have at hand reach.

[Enders]

My indie game is going NES like. 56 colors to choose from with 8x8 being my smallest sprites. I am doing 16x16 for tiles and sprites mostly.

Why am I making an 8-bit game when I can do it with better graphics?

1. I suck at making graphics. If I have a small palette of colors and a small canvas. I can make some good pieces of work.
2. I want to make a RPG that is 8Bit style.
3. There is strangely a market for it. Capitalize!

[MrDaaark]

1. I suck at making graphics. If I have a small palette of colors and a small canvas. I can make some good pieces of work.

#1 is actually a misconception too many people have. There is a big difference between shitty graphics and retro graphics. Retro graphics were still created with traditional art skills. No amount of color or resolution limiting is going to make bad art look good. A big part of making good low res graphics is using proper art knowledge and experience to overcome the limitations.

[jameswilddev]

I'm surprised how few games there are using flat shaded polygons to make 2D graphics. Resolution independence, yum....

[MeshGearFox]

Pixel art is harder to make than regular art. Have fun manually dithering your everything.

[Enders]

Pixel art is harder to make than regular art. Have fun manually dithering your everything.

I will take that challenge, I will be working on my art assets this weekend. I will show you what I will come up with for my map tile set next week.

Specs I will be using:

52 color Palette (same as NES)
16x16 tiles
64 tiles in total.

[pikzel]

I love the pixel style. Don't try to make 3D games if you're not a 10+ years experienced animator with all the best tools.