Slashdot Mirror
Videogame Graphic Advances - Not That Important?
Thanks to the IGDA for its 'Culture Clash' column discussing the recent advances in graphics quality for games, and why increased detail isn't always a good thing. The author, referencing a previously Slashdot-covered article about "unsettlingly funereal" hi-poly face models in games, points out: "Dependence on increasingly real visuals alone to generate emotion will inevitably hit a wall: at some point game graphics will look as good as real life. Developers have an arsenal of emotioneering tools at hand; to limit themselves to just one, however prominent, would be ill-advised", before further warning: "Overfocus on hyper-realistic graphics and modeling, while not a bad idea in a general sort of way, can also impede quality of gameplay."
Graphics might be good to look at but if there's no gameplay what's the point of putting down $50? If it's no fun, no matter how life like it looks I'm not going to spend my hard earned money on it.
"Armed forces abroad are of little value unless there is prudent counsel at home" - Cicero
To say that focussing on graphics at the expense of gameplay is a bad thing is obviously true. However, the reverse is equally true. These days, when I hand over the money for a game, I expect to get a product which not only plays well, but also looks decent.
This doesn't necessarily mean it has to have the latest 3d modelling techniques and uber-realistic lighting; you can achieve decent visuals as much through stylised 2D work as you can through the latest 3d engine. What it does mean is that I don't want to end up with a game which looks like it belongs in the late 90s.
It's true that the pace of advances in graphics has slowed as we get closer to the point of photo-realism. Doom was gob-smacking visually, Quake was very impressive, Quake 2 was fairly pretty and Quake 3 left me frankly unimpressed. However, it doesn't mean that developers should stop trying to push back the envelope. Far Cry gained a lot of atmosphere from its stunning visuals and I'd be surprised if Doom 3 didn't have a similar effect. On the console side of things, the visuals of the PS2 Final Fantasy and Gran Turismo games continue to be among the most powerful arguments against those who say the PS2 can't manage decent graphics.
Dependence on increasingly real visuals alone to generate emotion
Just look at South Park, for example.
The characters are full of emotion expression, even if the graphics are ridiculously simple.
Parent article is definitely true.
Compare the "legs," or longevity, of games like Angband and Nethack to those of Quake and the Diablo games. No contest.
This is because there are different production values: the roguelike games have a lasting cerebral appeal, while games that are built on eye candy concentrate elsewhere. This may have to do with the business models of modern game companies.
Take id software for instance. For gameplay internals, it doesn't get much simpler than id games. Doom was actually a playable game from the map screen if you turned on display of objects, and doing this shows how moronically simple Doom and Quake are. The appeal of the games, however, came from the presentation of the data, and the atmosphere produced by the amazing, moody artwork.
Mid-end graphics are comparatively simple to do, and using OpenGL actually makes it simpler, once you get over a certain learning curve. The models are the sticking point: you're not going to be doing amazing mo-capped human character models, but there's quite a database of MDL format models already out there, and there are other types of games, such as modern military RTS, that don't really require extremely detailed models-- a good example is the amazing TA, a game that has excellent longevity despite rather dated graphics.
TA is a game where the graphics are just good enough. At the time, there had to be a lot of trickery to render that many units at once, and the trickery in the TA engine involved giving the graphics a stylization that is still quite capable of bringing its gritty, desolate image home. TA is a sterling example of turning flaws into advantages.
Linux games should focus on extensibility, replay value, using randomness (cf. Roguelikes), and multiplayer, which gives games far more gameplay depth than the engine would seem to warrant (cf. Quake, Diablo II).
We could have a hundred original, interesting games on Linux. Instead we have 45,000 versions of Freecell and Tetris. In fact, Linux is the indisputed king of these types of games, because of the minimal thought required in their creation.
One idea for curing this might be to leverage the existing codebases of games like Angband and grafting semi-modern rendering engines onto them. Even turnbased play is wonderful with these games, and I think realtime play a la Diablo might not be very difficult to achieve.
One thing we DON'T need is more Tetris and Tuxracer clones.
Intolerance for ambiguity is the mark of the authoritarian personality.
At GDC 2003, Jason Rubin, head and founder of Naugthy Dog, a highly successful development studio for PS1 and PS2, delivered a speech (slides available here, audio and slides available on Gamasutra (free painless reg. req.)) on a closely related subject : improvements in graphics quality will not be sustained over the next few years, and relying on them to impress potential customers is a bad idea.
Moral : as long as gameplay, character development and story do not suck, nice graphics are of course an asset, but they're useless in case of an already shitty game...
Who needs a
Sooner or later all these high end engines are going to raise the cost of games. Every time these guys go from 10,000 to 100,000 poly models, that's just that much more time spent in a studio trying to get killer models done, and then there's the additional texture work, sound, etc.
... or they're small pro shops looking to make a name for themselves.
The engine might come out of the box to run at these level, but the asset work still requires that much more work to complete.
I'm seeing this all over the place in the (unreal) mod community. People don't want small mods anymore, they want commercial quality games. The problem is - making that quality gets more time consuming, requires more organization, etc. Few mod teams have the steam and those that do can rarely get out innovative work (which IMO, is the job of mods)
I'm actually looking to make use of the 2D aspects of Unreal to lower asset costs.
But back on topic - the high end nature of the graphics keep increasing the production costs, which eventually are going to have to increase product cost.
I play Half-Life Counter-Strike. Every map in the game consist of crates/boxes which are square-shaped. These are the few places where people can hide.
Now, the new game-engines out there seem to sport many new elements like trees, vehicles, grass, bush, etc. Which by the way makes us have to look really carefully for an enemy when we play. This really removes the "action aspect" of the game.
Mind you that when people make small arenas in real-life games, they often remove these complex things that slow down the gameplay.
Well.. the real problem I see is that companies are adding realistic graphics without the mandatory elements that go along with them i.e. if you are going to render each finger individually, you better make them move like real fingers. If you are going to make the characters mouths move when they talk, you better make damn sure that the speech lines up with their motuhs, or else it will stick out like a big fking X painted across your face..
I've seen plenty of games that only used 2d sprites, cel shaded or low poly(relative) 3d graphics that had more expressive and deep characters than some if not many of todays games with lifelike chars
They look like dead people, because they're about the right shape and color, but lack all of that "whatever" that makes a person alive.
Hence, funereal, or having to do with a funeral.
It's got to do with our perception of artistic representation of faces. The phenomenon is known as the "Uncanny Valley." We cut things a lot of slack when they don't look realistic at all (mario, for example) but when they get really, really close to real-- the tiny bit of difference sticks out like a sore thumb. There's this nasty place just before 100% realistic where even the tiniest imperfection makes things look awful. And they generally end up looking like dead people.
Consider the faces in games containing models of real people-- The Matrix, or Alias. They look like walking mannequins, and can be kinda unnerving.
That said, as someone who uses game technology for uses other than playing games (ie machinima,) I can say that the real-time lighting effects in Doom 3 are a huge change, and a sort of breakthrough in terms of what's possible.
When making Machinima, we are able to come very close to the techniques of real film-making. But the lighting has always been a limitation. Film-making is all about light. So the fact that we can now position lights in-game in real-time and create shadows, means we are that much closer to real film-making techniques.
Of course, if the past is any indication, we won't actually start to use Doom 3 for Machinima until Doom 4 is released. ; )
The ILL Clan - Machinima Pioneers
I've been thinking about this for quite some time and I realize that I'm more comfortable around cartoon and animated graphics than I am with pseudo realistic graphics. I've seen some movies and games where they try to make the face of a character as realistic as possible. However, it just feels eery to me. The closer they get to reality the more eery it feels. There is always something that just doesn't fit. Lips don't move properly, the skin is too shiny, the face too perfect, or the features too symmetrical.
In fact, I just looked on Google to see if anyone else noticed this and found this article.
Having worked for a gaming company and in the game industry for over ten years, gameplay and graphics go hand in hand. Yes, good graphics will improve sales but it will not make the game. I think most of us are smart enough to know that while eye candy is dandy, being real is the deal. But there is a lot more to developing a decent product.
There are four important factors in a games success:
1: open sourced/editable for improvements and new version (i.e. battlefield1942 morphing into desertcombat or starwars's galactic conquest, nethack)
2: gameplay that can extend beyond the original campaign(dynamic campaigns, add-ons)
3: good customer interaction and support for the game community (ie.combat mission, halflife, quake)
4: product that does something new or is not scared of rattling the conservative right (Grand Theft Auto).
The fourth will garner attention as free marketing. Rockstar used it for GTA:Vice and it worked brilliant.
Put those into a game, you've got a home run every time.
Management is doing things right; leadership is doing the right things. - Peter F. Drucker
Saying 'CG isn't good, look at how many people fail to make it look right' is a little like saying 'The invention of PAINT has failed! Look at how many bad paintings there are!" The technology is there, and those gifted and skilled enough can create terrific things, but the technology in the hands of a hack will create abominations.
I think the poor use of graphics technology is just a symptom of a more general, and obvious, problem with the gaming industry (and other industries as well):It's hard to make something good. It's hard to do something creative and innovative and 'well-done', and if you can make money stringing crap together and selling it with a gimmick, people will do it.
Game-makers, not having a sure-fire way to make good games, want a gimmick that will guarentee some sales. If they can throw in some cheesey visual effects, make a sequel to a tired series, or attach a game to a successful movie/TV franchise, they will, because these things are easy, whereas making the game good would be hard.
Games have increasingly getting more and more complex over the years, and thankfully so have the graphics associated with them. The higher production values are noticeable, and the entire affair is more engaging than it's ever been. While highly simplistic gameplay can be great, no one mentions that it's also damned hard to pull off so one doesn't get bored of it.
Does that mean that all old games are junk? Definitely not, but classic games that can hold one's attention indefinitely like Tetris are flukes. They aren't the norm, and on balance the majority of games being produced today are far better than their predecessors.
Not necessarily. You could just paint in the shadows into the background and when a character stands in the shadow you just use a darker palette to render the sprite, signalling to the player that you are now hidden. The advance in graphics technology just makes the effect better-looking.
There are many games that fall outside of the four factors.
Myst, Far Cry, Baldur's Gate, Splinter Cell. They don't have customization, open source, or anything breakthrough like you list; They just executed an existing genre really well.
Myst was just an adventure puzzle game, with [at the time] mindblowing graphics, and really well executed puzzles.
Far Cry is just an FPS with mindblowing graphics and really good physics. The story, the multiplayer, even the gameplay is pretty good, nothing terribly special.
Splinter Cell is a good stealth shooter with good gameplay but not spectacular, once again graphics and level design make the game.
Baldur's Gate series was just a great story, mediocre graphics, average gameplay
On the flip side there are games that have one or more of the four things you list, but aren't great (there are no sure homeruns) because they executed poorly. Tons of me-too RTS that have dynamic campaigns, lots of sucky editable FPS, and even more badly made "controversial" games.
The only one that I think does make for a successful game is number 3, but often that is more something that happens as a result of a good game. It doesn't matter how much a dev supports the gaming community if the game is bad.
I think the items you list can take a good game and make it a great game, ie Neverwinter Nights, or take a great game and make it genre defining. But none of those things really "make" the game either.
D6 63 0D 70 89 81 BB 8E 7B 7C 5F 5D 54 EA AB 73
I guess we have to do with someone very unimaginative, who is discarding lighting/gravity as a non-gameplay item.
Bontago (www.bontago.com) proves for once that gravity can be used for the gameplay, and not gimmicky-like as in , for example, Mac Payne 2.
For the lighting : I see plenty of options, since it's the cornerstone of most 3dengines (the shadows give something 'real ' depth) , ifnot our real lifes.
I can see how someone can make a stand that only graphics, or only a cool gravity engine, can't make a good game on itself : But there are plenty of examples that are/will be.
In my opinion, there are multiple uncanny valleys that gasmes can fall into. Graphics is only one of them and the most obvious of them. The other two things that have uncanny valleys are AI and physics.
One of the problems that people are having now is the ability to make characters in the game behave in a realistic fashion. In older games, you had things that behaved in such an artificial manner that it didn't jar our expectations. Now that we're trying to make games more realistic, creating characters that act like humans, we're going to find the ways they fall short of actual humans rather jarring, for the same reason that we find the zombies of modern games disturbing. We're wired to react to people socially. We can deal with artificial things easily enough, but someone that acts like a weird human will push mental buttons that clearly artificial things won't.
Likewise with physics. I think one of the reason a lot of very old games do very well in replayability is that they had totally unrealistic physics. Of course they had totally unrealistic worlds so we weren't jarred by the fact that things did not obey the normal laws of physics. Why did the things in Centipede or an early platformer act the way they did? That's just the way the world worked, and that was that.
Now we're trying to create games with realistic looking worlds. And people wonder why they can't pick up a rock and break open a window. Or move aside crates blocking a hallway. Games are getting more real, and that means we're sliding into the Uncanny Valley again as our expectations rise up to demand realism and what we are wired to expect.
Eventually things will get better, as we get good at creating synthetic digital actors who can express a range of emotions, and artificial personality programs that process player-NPC interaction and generate appropriate NPC reactions, and we have libraries that automatically model the physics and behavior of realistic objects.
Incidentally, even as the polygon count goes up, I don't expect the artistic cost to go up proportionally. I do expect the artistic tools to get better over time. An artist who wants a forest scene will just tell the computer to create a forest and he'll be able to tweak parameters and make a few manual adjustments over time. Just because an object has a zillion polygons doesn't mean an artist has to specify each one by hand. I do expect the demands on artists to level off.