Massive Content Generation Techniques In A Valley Without Wind

[BobTheJanitor]

I'll try to resist the urge to feel guilty when he glares at me then. 

[TechSY730]

I think that with a lot of companies, if some aspect of gameplay impacts the pretty-factor, they change the gameplay instead.  That's where we're backwards from most.

That is the reason many gamers are getting fed up with "big name companies" and coming to indie guys like you who are willing to do things the "better" way.  

[x4000]

I think there's room in this world for many various approaches.  I just know what my niche is.  I'll never out-pretty a big-budget company, or even close, but I can out-gameplay them simply by virtue of having creative freedom.

[BobTheJanitor]

"Pretty" is transitory. Quality gameplay lasts forever. Look at chess. The lack of pretty graphics didn't impede it from lasting over a thousand years. People can even play it without the pieces or the board. That's about as limited as you can get graphically.

[x4000]

That's my thought, too -- though this idea that prettiness is transitory is a strange and modern concept.  It used to be that the great works were timeless and outlived their period of creation.  Basically look at anything in a traditional art museum.  But now that we have technology for creating artwork... we are in this odd adolescent stage where the technology is growing so rapidly that our idea of beauty is constantly shifting.

Say what you will, but I still think that the original Half Life 2 looks great.  Yes, it's dated and there is a lot of other stuff that is more realistic these days.  But the graphics in HL2 still catch my eye in a pleasing way, and I still remember how amazing the graphics seemed when I first played it.  Actually... the same is true of Quake 2, come to that.  Those were some fan-freaking-tastic graphics in the day, and they still do such a good job of setting the mood for that game.

Some of the other ones, like Chrono Trigger or even Super Mario World have a rather timeless look to them as far as I'm concerned.  It's like arguing if Picasso or Van Gogh or Michelangelo was the better artist.  I know which pieces I prefer out of their collected works, but each one stands alone and apart as an artistic accomplishment.  But when it comes to ultra-"realistic" game graphics, there's this constant sense of one-upmanship that has everything to do with techno-culture and nothing to do with art.

Well, whatever -- I just have no desire to get into that sort of one-legged footrace.

[BobTheJanitor]

But when it comes to ultra-"realistic" game graphics, there's this constant sense of one-upmanship that has everything to do with techno-culture and nothing to do with art.

Absolutely true, and it's been like that for too long. The glimmer of hope is that we're starting to come out from under the shadow of that foolishness and realize that there's nothing that makes the most realistic graphics the best graphics. It was just an accepted delusion that the games industry was infected with. Recent successes like minecraft and in fact most of the indie scene have shown that there's no reason that games have to be reality simulators. Gamers are quite willing to accept anything along the line from hyper-realism to complete abstraction as long as the other elements are enjoyable. Good gameplay, good story, an enjoyable aesthetic quality are all more important than how many polygons you're packing in per model. Of course whatever art you go with has to make every effort to be good within its own bounds. Even if your character is a circle and your enemies are squares, players will grumble if the circle is off-round. Unless you've got an abstract scribbled look, in which case they'll complain if the circle is perfectly round. Internal consistency is important.

Even big studio titles like WoW or TF2 obviously didn't go for perfect realism in their looks. And they've ended up better for it. I'm reminded of Scott McCloud's book Understanding Comics (amazing book) where he devotes a whole chapter to various art styles and makes a well-reasoned argument that the more simple your characters look, the better the reader can identify with them. A perfectly realistic, lifelike drawing can really only represent one person (well, sans-twins, but work with me) whereas a completely simplified drawing with two dot eyes and a line for a mouth represents (almost) every person on the planet. And of course between those two there's a whole scale of styles from the near lifelike down to the very cartoony looks. Readers can more easily see themselves in the simple representation, and thus more easily be drawn in to the world of the story. So I'm glad when I see games that don't tread the worn path towards the most realistic graphics they can find. The game's environment is a completely invented world, after all. Why should it look like the one we've already got? I've been there. I go to games to take me somewhere else!

Edit: Found a panel from the part of that book showing exactly what I was referencing here:

[x4000]

All very true!

[RooksBailey]

This is a really interesting discussion and I find myself agreeing with the sentiments that, generally speaking, less is more.  Gaming is a lot like reading:  the best games leave room for the imagination to fill in all the little details.  Going for realism is a trap:  once you commit to a "realistic" game, then you better well deliver it - something that is usually unattainable because of the limitations of modern hardware (as evidenced by this discussion).  So, what you usually wind up with is something that has pretensions of realism, but never fully delivers on the initial promise, leaving a bunch of disgruntled gamers to gripe about realism shortfalls.  

When you go for an abstracted, artsy look, you avoid that hopeless struggle entirely, as well as gaining far more artistic freedom (would it be wrong to say "realistic graphics" are the laziest graphics?).  There were so many great games back on the ol' 8-bit and 16-bit machines that managed to achieve so much with so little (MULE comes to mind, along with some of the early Ultimas).

But this discussion about realistic versus artistic also applies to gameplay.  Somebody mentioned chess - my favorite game of all time (sorry AI War    But after 1300 years of development, who knows?  ).  Chess succeeds where so many "realistic" wargames fail precisely because it is abstracted.  With realism comes complexity and with complexity comes a gameplay burden.  Chess, as one of the all time great works of human genius (and I suspect there was some divine assistance as well   succeeds precisely because it abstracts the minutia of war while managing to successfully capture the strategic and tactical essence of it.  Oh sure, there's no "realism chrome" to chess - no terrain, no logistics, no combat modifiers - but chess proves that you don't need such distractions as long as the gameplay cuts to the heart of the matter.  The human mind is more than capable of filling in all the details.

  

[stblr]

Has the team given any thought to using skeletal animation techniques instead of (from what I gather are) static animations generated from Poser? Using skeletons would allow you one-time generation of all animations for a given skeleton type, as well as greater flexibility in generating various body parts and mixing them for a huge variety of NPCs.

With the addition of editable color channels--through crafting, for instance--a single pass through the art pipeline could produce a shirt that could be worn by any NPC, and could be dyed any color. Skeletal animation would also allow for stackable items on characters e.g. armors, hats, even craftable weapons, leading to a nearly infinite combination of NPCs (and therefore players) with just a few static art pieces, since all you're doing is basically sticking layers onto a skeleton.

[x4000]

Skeletal animation techniques are great for pure 3D, especially when you have the exact same body type for every character or whatever, but it doesn't work in 2D that I've ever seen.  The only way would be if you rendered each arm, leg, body, and possibly head and hair separately.  That would be a really massive amount of work, and with the variety of characters we'd like to support I think it would really cut into our ability to make them distinctive.  I thought a lot about that sort of thing, but didn't see any way that would really make it work.

Granted, in the past I've used Charas for 2D character sprites, so maybe you mean something like that, but it only works with pixelart of very defined sizes.

[stblr]

Skeletal animation techniques are great for pure 3D, especially when you have the exact same body type for every character or whatever, but it doesn't work in 2D that I've ever seen.  The only way would be if you rendered each arm, leg, body, and possibly head and hair separately.  That would be a really massive amount of work, and with the variety of characters we'd like to support I think it would really cut into our ability to make them distinctive.  I thought a lot about that sort of thing, but didn't see any way that would really make it work.

Granted, in the past I've used Charas for 2D character sprites, so maybe you mean something like that, but it only works with pixelart of very defined sizes.

A skeletal system like Ragdoll Kung Fu uses is what I had in mind, but granted the application is way different from what you're going for in AVWW. But that game also uses high resolution 2D art similar to AVWW.

[x4000]

I gotcha -- had not seen that game before, that is pretty interesting.  It's definitely tempting, but I think a lot gets lost there when you do that, in terms of the sort of game I'm trying to make.  It also makes every onscreen character... 14x more expensive on the GPU?  I think that's right.  3 segments for each limb, and then one for head and one for body.  That's before you get to any clothing, so you can then up it even further.

With just a few characters that's no big deal, but if we got to a lot of characters and enemies that would quickly get pretty massive.

[PSimon23]

I decided to give a hand and sent a post into social bookmarks. I hope the popularity will rise in.