Is it simply not worth it for developers today because only computers would be able to run those games and most of the market is on 6-7 year old consoles? Might it happen in 1 or 2 console generations?
Would it be possible to film actual objects from many angles and create games out of that?
Bear in mind, game designers have to contend with the Uncanny Valley too. There is a program that I have recently (within the past year) seen that lets you insert an image into a 3D environment that is quite realistic. I just gotta find its name again. When I find the name I will be back with information about it.
Of course, but we’re not there yet, and we won’t be for a while.
The most important parts we need to improve are the very complex interactions with light we take for granted and therefore become essential in order to sell the scene as realistic. The nuance in texture detail from one object to the next, and the complexity of the 3D mesh, both in detail, and in it’s ability to be modifed in real-time (desctruction).
Which bring up another important fact - animations. Animations for the ambient 3D world/scenery, the characters and the little details that come up when the two interact. Nothing breaks immersion more than swords sticking through walls, feet not angled against the terrain properly, etc.
Oh, we can do better than that… assuming we’re not talking about rendering the image in real time… actually at least 30 times per second.
That’s a big limitation of games, they need to render everything, as well as update all of the engine logic (collision detection, pathfinding, animation, physics, AI, etc) multiple times per second.
We can create graphics that are a heck of a lot better than whats on games, but each frame can take hours for an entire server farm to render.
Yeah, I mean once all computing is on the cloud, and by that I mean there’s an atom-sized 3.5 YottaHertz processor attached to literally every water molecule in every cloud, we can have Crysis 927 which realistically simulates billions of photons every millisecond, but even the best radiosity matrices are going to look slightly off. One of the main problems with simplification of lighting is that unless you use tricks and depth limiting (which undermine quality even further), you get infinite recursion problems with reflective surfaces pointing at each other. And we’re not talking about that two opposing mirror awesomeness (well, that too, but we can actually handle that pretty well nowadays), I mean how basically everything is a little reflective.
Yeah, I’m exaggerating a bit, I’m sure we’ll find some trick eventually that’s basically close enough that the difference is imperceptible to any human eye, but I don’t think we’ll have real time photo-realistic lighting in an arbitrary scene for a long time. (Also, Teeth have a long way to go, I’ve never seen video game, hell, CG movie teeth that didn’t distract me).
I doubt it, for one simple reason: reality has clutter. Lots and lots and *lots *of clutter. Also uniqueness, for lack of a better word - look at, I don’t know, a tiled bathroom closely and you’ll see that no two of these tiles are exactly alike. But in a video game, they will be because it saves computing power. And unless they pay level designers to go over every inch of a level to add bird shit, and dead ants, and a discarded shoe, random cigarette butts, make unique textures for every square inch etc… no computer-generated environment is ever coming close to the clutter and “asymmetry” of a real one.
The Crysis still shot is a very apt demonstration of this: look at that shack. Look how it’s all straight lines, clean & crisp textures, edges and flat surfaces when the real thing is all beat to shit, covered in rust and mold and water marks and whatever that white gunk is on the chicken wire fence.
And that’s just the man-made stuff. I have yet to see even a single believable tree in a video game. Not one.
Look at the grass in the Crysis shot - see how every tuft is exactly the same as the one next to it ? They’re forced to do this (or something close to it - Skyrim for example has multiple variants of each bit of primary flora, but not enough that you won’t notice you’re always seeing the same 3 bushes all the time) because otherwise the engine would have to load inordinate amount of resources just to display the fucken ground, even if artists could find the time and motivation to model hundreds of different tufts of grass. And even rendering, lighting, shading that one piece of *Crysis *grass in real time over and over puts most computers on their knees past a certain point.
I won’t say what we’ve got now is the best we’ll ever have, because we’re still seeing tremendous progress in all fields. But I honestly doubt we’ll have even a close approximation of reality - something a casual observer might mistake for a movie - any time soon.
I think that we are going to hit the endgame for graphics advancement soon. That is to say that graphics will cease to have massive improvements in the way they have from say 2002 to today, but will instead be about shoving more particles and animations in the backgrounds. Those will make the games look better yes, but not always in an immediate way. The new Unreal demo is a good example, the graphics and environments in and of themselves are no more impressive than most of what’s available today, all of the visual improvements come from real time lighting and particles. My hope is that this actually helps to avoid photo realistic graphics, I would much rather see beautiful and visually interesting worlds than an even more realistic look at the boring dirt and sand that’s so common in games recently.
You are confusing technical graphics capability with artistic design.
You can have a beautiful ruined castle in the sky with color rainbows and vast, deep green forests which look extremely “realistic”, or in other words, where the graphics “sell” you on the scene.
That would be achived by a combination of the technical capabilities of an engine, and the artistic design behind it.
You seem to think that better graphics (more photo-realistic graphics) must equate to Call Of Duty clone, all grey/green world.
That’s not true.
The more tools the artists have the more they can do with them, and the more they can get done. Half of what better graphics bring to the table are purely to empower the artists.
I would argue that we’ve pretty much already hit this as far as the ‘common man’ is concerned. Yes, there are graphics enthusiasts who are still very impressed with the latest bump mapping advances, and things will still look visibly better in an “all settings turned on” vs “default” side by side comparison, but fundamentally, it’s just not making people stop and go “Whoa! Look at that!”
I took hope that sooner or later, we’ll get over our self destructive obsession with photorealism (Hint: Film is not “photorealistic” either - untouchedup/filtered/blahed stuff on film looks like crap. Photorealism is not the goal.) and get one with making art.
No, my point was that the increasingly “realistic” graphics that have become the standard will hopefully fade away once improvement rate of graphics slows. “Realistic” graphics are popular right now because they wow, but once the leaps and bounds stop they will cease to be impressive. You can see this to some extent now, with more and more gamers grumbling about the bland browns of gaming. My hope specifically is that soon once the tech on its own stops impressing more games will focus on more artistic worlds to wow in new ways.
How much of this alleged “graphics improvements slow down” will be attributed to some type of diminishing returns from hardware hitting the “it can’t get any smaller or faster” wall?
To tell you the truth, when I play, I don’t want “realistic” graphics. I don’t have time to do it anymore, but when I played WoW, I did it to get away from reality, not to immerse myself in it. If I wanted reality, I’d step outside. The cartoonish oversaturated WoW pseudo-reality was fine.
I don’t think it will be from a slowdown of tech advancement, rather I think it’ll be from diminishing returns from the improvements themselves. Once you’ve got a certain number of polygons and particles in a game adding a few more will only look so much better. Yes games will continue to become more impressive visually, but eventually we’ll get to a point where the average person probably won’t notice. At this point I think the major hurdles will be more dynamic lighting and water effects, and it looks like the next generation of engines will be working on those areas in particular.
I think there’s a place for both. I can appreciate and enjoy art styles like those from WoW or Team Fortress 2 but some games really benefit from an immersion standpoint. I enjoyed my random wanderings in Skyrim or Fallout 3 due in large part to the realistic looking world.
I would assert that both of these games would benefit from not being “realistic” so much as a sort of “superreal” that is actually LESS realistic. Gorgeous vistas, truly blighted landscapes, etc that capture the feel of realism without actually being there. And I daresay these games are actually doing it - probably the same way film is. Stepping the colors up or down, adding bloom or lens flare or ‘unnatural’ lighting effects, etc.
Games should aspire to film levels of realism, not to reality levels of realism. Of course, there’s still a lot of room for improvement, but mostly I think we need to focus on animation more than rendering right now - most game models actually kinda look like arse in motion right now, if you look at them in any sort of objective way. There are far more ‘realism’ gains to be gotten there than there is in “look how realistic and well rendered our mud is”.
