f13.net

...The press has no clue what they're talking about (http://www.bit-tech.net/news/2005/07/01/pgr3_screenshots).

Though, with this, we run into another problem. One which we're fast approaching (I give it one more generation, maybe two). If you remember a few weeks back to the article about electric town, Japan (http://forums.f13.net/index.php?topic=3591.0) being too geeky and scaring society, imagine what will happen when virtual worlds and games really do look better than real life.

Two things will probably occur:
1. The schism between reality and ...virtual reality will become much bigger and geeks will really remove themselves from society. Did you see the girl in the red dress?
2. Things will get creepy for your average person - fake people that look real freak most normal non-gaming types out (and even most gamers).

I'm interested in seeing how they avoid this problem since most designers don't have the balls to be abstract and really just want to make the MOST HUMAN THING POSSIBLE.

Wow, what a knob-slobbering article. Not that PGR3 doesn't look sweet, but the building detail? How often do you notice that while zipping by at 200 MPH?

Except for the growth in cel shaded games.  Or those with cartoonish graphics a'la WoW, or about half the GC titles.  Depends on what kind of game you're making, although games with relism are in more demand.  Then again, how is this different from other medium?  Take movies for instance.  People in suits or spaceships on wires don't cut it anymore.

I did say most designers. There's always exceptions and there's a lot of them. But look at most of the shit on the shelves. Even if the number of games that aren't realistic are higher in number, the shit that is fills up the shelves. EA stuff, Midnight Club, Gran Turismo, etc.

Everything you listed are sports titles (EA is hard on for sports stuff), which makes it diffuclt to buy off on a remake of Enduro.  Why, artistically, would you want a cel shaded sports game?  XIII and Killer7 do it as a nod to comic books.  Wind Waker does it to appeal to a younger audiance.  Mario Power Tenis is obviously looking to do the same, so again it isn't about having balls or not, it's about audiance.  If you hate realism, Mario is waiting.  With the average age of a console user being well over grade school level however, the tendancy for the market is to get a more realistic title.  It's hard to pull off a WoW-ish title to a mature audiance. 

What you get with realistic graphics is immersion.  So if you are trying to break immersion, you have to have a fairly compelling reason as to why.

Realistic graphics as they stand right now are immersive. Once they get too close to looking like humans themselves, it will break immersion. It will become something else entirely. That's the problem I'm talking about. Right now, I've no problem with it. This coming generation will probably be ok as well.

But one more generation? I think we're going to get into the realm of creepy weird.

I've heard that claim before and don't buy it.  On the one hand, 8-bit graphics are less immersive than what we have now.  On the other hand, perfect photorealism would be less immersive than what we have now, or so the argument goes.  Clearly, then, there is some sort of "immersion curve" relative to graphic realism, with its peak somewhere in between. 

Doesn't that seem slightly ludicrous?  Would movies shot with real actors be more "immersive" if they were digitally animated, but using something like the Source engine to make sure they weren't too realistic?  If this immersion curve existed, it'd stand to reason that we could improve on our current "realistic" filmmaking techniques by decreasing the level of realism.

Which is why I prefer my graphics abstract (Frequency/Amplitude, Rez, Darwinia) or functional (WoW, Kirby Canvas Curse). Realism in graphics is for people who can't admit they are playing a game.

So far, only two well-known movies have been shot with all sets computer generated: Sky Captain and the World of Tomorrow, and Sin City. Neither of these went for a realistic look, but purely stylistic/artistic.

I guess the jury's still out.

The problem with going for photorealistic graphics is that there sill always be some glitch - clipping, animation, collision detection, what have you - that eventually will manifest and break the realism. Not everyone notices such things, but I would wager gamers are more likely to than "normals".

I await the day I can upload my conscious to the 'net and live forever.  Until then 'realistic graphics' mean nothing to me.

Also, those buildings are pretty damn cool but I bet the flesh still looks plastic.

Everything Pixar?

Time for the requesite Uncanny Valley link: http://www.absoluteastronomy.com/encyclopedia/U/Un/Uncanny_Valley.htm

I'm gonna stop by and talk to one of my professors at UMD that discussed the actual term for the too realistic conundrum. But for now: HYPER-REALISM.


Works with environments:
(http://www.sauer-thompson.com/junkforcode/archives/Photography1.jpg)

Not so much with people:
(http://www.tate.org.uk/collection/T/T07/T07445_9.jpg)

Creepy enough for you?

Basically here's the problem. People who create something too realistic will go out of their way to do that. In other words, things that shouldn't be exaggerated will be and things that should be won't be. I'm not talking about the ass and titties of team ninja games, but things like noses, hands, feet, eyes - will look TOO real. It's a pretty sound theory that I'm willing to accept as I know that everyone in the world wants to create or be the perfect specimen of a human being.

If you're including that, you've got to include Final Fantasy:The Spirits Within as well. And as close to real as they got, you could still tell those people were CGI and looked 'wrong.'

This has more to do with mixing the art and the story, IMO.  CGI people have been used in movies for years, albeit as background.  SW3 is loaded with it, and it looks fine.  In other movies, actors have been crafted entirely out of CG as the main - for example, The Matrix did so with Neo on several instances.  You can't say super-realism is flat wrong, because there are so many other instances where it works just fine.  More likely, people mistake the notion that an art form, such as super-realism, sells itself - they don't.  FF tried to sell itself on the grounds that the CG was groundbreaking.  And it was; but the story wasn't very compelling, and the CG wasn't expressive enough.  The Incredibles' style of art added to the story, whereas Polar Express just looked bad (IMO) and had an uninteresting story. 

As for painting, realism faded out of favor due in part to photography.  Why spend so much time trying to be photo-realistic when you could just take a photograph and save yourself the trouble?  Merely being painted doesn't do anything to inspire you, the viewer.  The method of painting doesn't in and of itself convey anything; it has to be in context.  Again, in large part as to why art has gone so far into the extreme of abstract.

I've mentioned him before, but in cases like Gerhard Richter, he achieves photorealism by simply doing things a camera can't do. Some people don't classify him as photorealism. But get close enough to any painting and it looks shitty.

(http://www.thecityreview.com/s01sconri.jpg)

A little, but I think that's only because the legs look plasticky - in other words, it's not that the picture is too realistic, it's that there are breaks in its realism (I think AOFanboi nailed that one).  Cut down the specular lighting in that picture and I wouldn't be creeped out in the slightest.

(Edit) I might also be an abberation, though.  Some people are creeped out by finely detailed statues of people (like Renaissance-era marble sculptures).  I've always found them very compelling and not creepy in the slightest.

But that's the problem. Artists don't want female legs that look shaved that morning or the day before. They want perfectly good plasticky smooth legs. Getting rid of flaws. It's one of the faults of man. The results won't be more immersion it will be "Doesn't that look a little TOO perfect?"

There are no limits to how much every artist wants to mimic a god in terms of creation.

The feet look good, the kneecaps not so much.

You aren't arguing against realism, then, you're arguing against airbrushing.  Which is not realistic.

I'm arguing against using the imagination to create the perfect human. Which is what will inevitably happen. Also, I'd imagine seeing a hyperrealistic painting of my face would freak the hell out of me.

Why worry about this when some claim that we'll be able to design are own babies?  I mean what will japanese people look like once they have the ability to actually design their kids to look like their favorite anime characters?

Reality is too scary for me to worry about video games...


*edit*\/\/\/\/\/\/\/\/

Megaman Soccer for the NES was THE MAD SHIT FIZZLE YO.

Fuck Madden. You can get away with more the less realistic it is.

(http://img87.imageshack.us/img87/575/mlf8kd.jpg) (http://www.imageshack.us)

That's entirely different from "realism", though, so your original statement about realism breaking immersion ("too close to looking like humans") isn't what you really meant.  What you meant was "too close to looking like mockeries of humanity grown in a secret Nazi facility".  I agree, using our improving graphics technology to create Nazi zombie clones would be a bit creepy.  I don't think that every step towards photorealism is a step towards Nazi zombie clones, though.


Does seeing a photograph of your face freak you out?  Or seeing yourself in a mirror?  How about (hypothetically) a wax sculpture?

Of all of those, I imagine the wax sculpture would be the freakiest to most people, and the mirror the least.  The mirror is the most realistic, the wax sculpture is the least realistic.  You can't pin the freakout factor on realism there.

Let's all keep in mind the fact that EQ2 was touted as using the most polygons to achieve the most realistic environments and people for the absolute tops in immersion. That was the stated design ethic behind EQ2's entire art design.

How's that doing again?

The error there is that assuming more polys = more realism.  You also need models, textures, and animations that resemble the objects you're trying to represent.  From what I've seen of EQ2, its textures were pretty bland, and that on its own is enough to throw realism in the crapper.

Her kneecaps look friggin' enormous, and her legs are far too long.  Her face looks like it was severely beaten by the ugly stick.

 But wouldn't you be less likely to accuse the homely shy girl to be a robot, than the hot ninja chick in the post apocalyptic world we'll be living in when the machines take over?

Heh.  Post that on fark and you'd still get the requisite amount of morons posting 'I'd hit it' pictures.

That's realism.

I believe this is the error that is driving each generation of console hardware creation, as well as both Nvidia and ATI's current marketing strategies for the PC video card market. And all of them essentially forget that the number of polys don't matter if the artist can't use them well.

There are several other reasons to want high-poly cards.  For one, it allows you to throw more objects onto a screen.  Instead of three or four baddies, you can have an army.  For another, it kicks framerate up on lower poly loads, which makes the game play smoother.  Consoles are just tools to help make art, and it's again confusing the method of art with art itself.  Canvas and paint don't make art, context makes art.  Realistic graphics can be great for games where they want you to feel like you're "really there", because as the viewer you'll spend time marveling at your environment.  It's also one of the easiest angles to pitch

Is Nvidia supposed to care, or have something to do with that?  They're just making tools.

Yes, they are. I suspect it's not getting any easier to push a card to any sort of limit. If it were, no programmers would bitch and moan about the proprietary shit on consoles.

Edit: And yes, I understand it has more to do with the environment than just the graphics card. But then, I wouldn't be surprised if they aren't going out of their way to make things easier.

You damn right they should. They should do everything at the hardware and driver level to make it easy on the artists and programmers to make shit look good, without necessarily having to resort to adding more polys or things that choke the graphics processor.

Tha's what immediately leaped out at me.  Exaggerating leg length is a tactic used by porn artists to make images of children press our "old enough" button.  Polar Express definitely was about the creepiest thing I'd ever seen and does a fantastic job recapitulating the Uncanny Valley threshold.  The concept was stripmined in A.I. as well.  I thought STNG's backstory on Data/Lor was an interesting take.

But that was the same series that gave us a holodeck and humanity went on to evolve past thinking it was nothing more than a cheap way to go to Disneyland.  I'm of the common opinion a holodeck of that sort of realism would be humanity's last invention...

There it is.

Uncanny Valley at Wikipedia (http://en.wikipedia.org/wiki/Uncanny_Valley) - Mark my words, it will happen in video games.

Edit: I don't know how I missed Bunk's post earlier. Apologies. I'm a douchetard.

Can I quote you on that in the future?  :P

I may be off on a tangent here, but this sounds like something I read by comic book guru Scott McCloud.

Basically, he was looking at the idea of icons, and why do iconic characters (like Mickey Mouse or Homer Simpson or Tintin or Monkey D. Luffy) carry as much or more emotional involvement than more realistically drawn characters (like Bruce Wayne or Peter Parker).  

The idea he puts forth is that when we're interacting with someone else, you see their face in (basically) perfect realism, but you see your own face only in vague, indistinct form.  You maintain a rough mental picture of your own face (am I smiling, am I frowning, do I look like I'm reacting inappropriately to what this person is saying) in order to plan your responses and gauge how effective they are, but you don't really keep the complete details of how you look in your mind.

The upshot of this is that when you look at a super-realistic drawing, you see it the way you see someone else, but when you look at a more vague, iconic drawing, you see it the way you see yourself.

Applying this to video games, most main characters are somewhat "generic," their features fairly nondescript, and average.  The NPCs, on the other hand, are often much more visually distinct (Jade Empire is a game that comes to mind repeatedly with this example).  Adding realism to a game would help with making visually distinct characters, but the more detail you add to something, the less "average" it becomes, so it becomes progressively more difficult to make an identifiable protagonist.  This would be an especially big deal in an MMORPG like EQ2, where almost every character (with the exception of the monsters) will be a protagonist.

Anyway, that's my take on it.

Laying the blame for bad art direction at the feet of the hardware manufacturer seems like a bit of a stretch, unless you're positing some sort of shadowy conspiracy between ATI and nVidia to make games look ugly.  The better graphics the cards produce, the better their cards will sell, so it's in their best interests to do anything they can to make their cards easier to produce pretty graphics on than their competition's.

A large part of the problem is that game programmers don't want to write hardware-specific graphics code, because then they either have to replicate that work for other hardware, or limit their target audience to owners of that particular hardware.  I can only think of one game off the top of my head that's that hardware-dependent, Bridge IT (http://www.chroniclogic.com/bridgeit.htm).  Everything else uses "neutral" APIs like OpenGL or DirectX, and those APIs don't necessarily provide direct access to everything a card can do (though I think the reason DirectX has been revving so quickly is that it's trying to catch up in that regard where OpenGL is lagging).

The original theory behind Cg (http://developer.nvidia.com/object/cg_toolkit.html) was to provide an API (actually a distinct language) that made it easier to directly access the hardware, and theoretically work with ATI cards as well as nVidia ones, but I don't know if ATI is actually getting on board that bandwagon, since it sort of gives nVidia a home court advantage.

I wasn't laying the blame for bad art direction at ATI or Nvidia's feet, I was just saying they need to make sure they don't make programming for the cards incredibly difficult.

I think this is a problem dealing with our generation versus the generation of kids growing up nowadays.  I mean we grew up with 8 bit Nintendo games that looked like crap, but that didn't matter, for the games were extremely fun to play.  Where are the "A Boy and His Blob" or "Bionic Commando" games nowadays, it's all the same shit, and when everything is the same the only defining aspect would be to make the game more realistic than the others out there.

We complain about innovation in games while it seems others (not just the kiddies) have grown accustomed to the same gameplay over and over, and even more have no idea that such innovative gameplay could exist.  All they want is the coolest looking game out there, and if the majority wants that they shall receive it. 

There also is that pesky little human trait of wanting to create and understand the world around us, it also takes much less creativity to create what you see rather than what you imagine.

Sorry if this doesn't make sense, work is over and I have to book, hence no reread.

That's pretty much the entire reason those companies have software divisions, so I think it's safe to say they're aware it's important.   :wink:

I don't know if this is covered by the Uncanny Valley theory, not having read the original research, but one of the main problems I have with "realistic" character modeling and animation is that as the models look more and more human-like, deficencies in animation and character physics become more distracting to me. In other words with hand drawn stuff I'm unconsciously cutting the animators a lot of slack but when the characters look realistic I expect them to animate realistically and the physics to be accurate and when they aren't it's very jarring to me.

I second that. The more perfect the graphics become the more nitpicky things I notice wrong. It seems to just be a part of human nature to only really be jarred by things that're "slightly off". Seeing something flat out unusual will draw raised eyebrows or stares, but when something nearly intangible is different enough to be noticed people seem to get MUCH more uncomfortable.

I don't think it's too good of an analogy, but imagine coming home and finding your place wrecked. You'd be upset right? Now imagine coming home and noticing that all your stuff has been moved very slightly. For some reason that would creep me out much more than finding my house ransacked.

One time my mom lowered a painting and moved a flowerpot. I thought they changed the entire room.

They got a new TV once and switched a whole mess of picture frames. I didn't notice for 3 months.

You don't normally program for the cards anyway, so not sure what the complaint is.

Theoretically you don't program for the cards, the reality tends to be different.

Not if you have any desire to use a card beyond just that one, it isn't.  Sometimes people try to bypass abstraction layers, but it often comes back to bite them.  Maybe you can get away with it if you're targeting a single console, and that one only, but beyond that you'd be asking for trouble.

Things may not be written card specific but they are certainly written feature specific and often features are not available on cards of even the same family much less between competing products.  This is one of the reasons many games require a driver update to function properly.  Non-standard calls to resources are done for various reasons (mostly speed optimizations though) the cutting edge games rarely stay entirely within the DirectX or OpenGL libraries.