Is Photorealism In Games The Right Direction?

Miscellaneous Forums/General Discussion/Is Photorealism In Games The Right Direction?

Amazingly brilliant article everyone here should read too:

http://www.gamasutra.com/php-bin/news_index.php?story=11489


In the quest for this immersive photo-real environment, creativity has died on the operating table. Given all these fantastic tools, the best we can do is attempt to duplicate our own environment on the same regurgitated topics and themes. In an interview with Tomek Bagiñski on his recent CG film 'Fallen Art,' Tomek states that he doesn't see any reason to do photo-real 3d graphics unless you are working for the SFX film industry. His reasoning is that stylized art is far more interesting to look at than just another attempt at reality. The photo-real painting movement ended after a brief period in the late 1960's for exactly the same reason - it was boring to look at. Guess what? So are the majority of today's video games.



Yep i think attemping to be photorealistic ends and stops beeing an issue when the technology is actually there and available and easy to use for everyone. Then people will move away from it again in order to find their own artistic style. Some will do realistic stuff, some will do another style, some will do pixelstuff again and so on. Actually some developers do it already now - hey me too! :O)

Much like any artform there is no wrong or right direction. There's room for any number of different styles in games, the problem is that there is a strong bias towards realism, fantasy and cartoon style, leaving little choice to people who find other styles (more) interesting.

I think the lack of creativity on the gameplay side of things is more depressing.

I second the boringness of realism.. but then again I'm a retrogamer, grown tired of anything 3d/fps.

Do me a silly Sonic the Hedgehog, a simple Gradius fighter with jumping canons on the ground, a silly Boulderdash guy, any day.

* liked Monkey Island 123, but not 4 .. 4 was spoilt with 3d, all character was gone.

The only 3d'ish game I played extensively was Lands of Lore II, aparently the story/gfx/music/game outruled the fact that it was 3d.

Did anyone ever use the X-rotating feature in Warcraft 3 btw? I happen to know quite some trained Starcraft players (of which some of 'em ranked in the world top100 in their years) and they just played W3 with the classic top-view.
So, what have we gained with that tilt function? Nothing!

Photorealism is the way forward because it is boring and uncreative, just like the majority of people who buy games. In short, "Financial Success" is a euphemism for "giving people what they want" and people get what they deserve, so, yeah, Photorealism is the right direction.

Who cares about artistic pretentions? Games are not art; it just takes artistic ability to polish them.

..photorealism suck..thats why I have to pay more to play games pumped with graphics without gameplay..

Photo-realism belongs exclusively in still images given the technology which is available to the masses today.

The games industry (at any level) has absolutely no business trying to foist the idea of it on the game buying public unless it is used as still images.

The sort of realism that 3d is most seriously short of right now is stereoscopic vision, and I've yet to see any sensible solutions for that at a truly affordable price. Come to that even a sensible design for such hardware has so far not been seen.

Clearly none of that has anything very much to do with photo-realism. It can be seen as having a lot to do with gameplay though. It could have a huge bearing on that.

If full colour stereoscopic vision was to be properly developed in such a way as to become available at a truly reasonable cost, then it would enable 3d games (where it actually adds something to the experience) to become viable.

Using the technology that is currently offered is rather akin to the idea of listening to a classical music piece, played by a full orchestra, on an mp3 player that only accepts 32K sample rate tracks. IE it's a largely pointless exercise. (I'd argue that listening to it in full on audiophile glory is a bit oif a waste of time too, but that's strictly about me being a philistine where old music is concerned)

I've no idea why so many lump 3d, photo-real, and stereoscopic vision into the same category, yet they do not really belong in the same category at all for most of the time, in fact they are barely related at all in a significant number of cases.

I think the term is misplaced. I don't think what we're pushing towards is "realism" as much as it is "authentism" (if that's a word). More powerful video capabilities give game developers a chance to sell their game world to you in a more sincere way. You don't have to imagine how that alien world looks - you can see it. You can interact with it. And the more power the cards provide, the more you can experience in the game.

Realism is the easy track because we can see it all around us and it's easy to know when you've got it right. Does that shiny car look like the one in the driveway? Then you're done. But I think that same power affords other developers a lot more flexibility and I look forward to seeing the crazy stuff that's going to be created in the coming years.

I couldnt care about photorealism, I have a real life, dont need a computer game to look it. Authough, I wouldnt really want to play a game that looked like the original Doom.

authentism


You mean authenticity?

..what Leiden said..

realism in visuals just make everything else look less believable and disappointing. Like rubbish AI in 99% of games, poor collision, physics. When something is stylised and obviously not real you can live with and ignore the flaws.

But were so used to realism in every day life that the slightest thing will stand out even more is its not quite right.

Realism has its place in simulations, and you can still have stylized environments and worlds with realistic surface materials. The problem is that realism is kind of boring, especially for the artists that have to photograph everything, or grab it off of google and basicly copy whats already there.


realism in visuals just make everything else look less believable and disappointing. Like rubbish AI in 99% of games, poor collision, physics. When something is stylised and obviously not real you can live with and ignore the flaws.



Bingo! Hence Super Mario being an attractive game :)

What was that famous line about Pacman again? Collecting pills in dark rooms, on repetitive music.. I'd say that's fairly realistic already. :D

To simulate physical reality is an extremely complicated endeavour and the closer people try to get to it the more complex the underlying code/systems/resources will have to be. There is on easy simple way to simulate its intricacies using a computer. There will still be companies that will keep expanding it until you have all realtime super-high-definition ultra-smooth billions of polygons raytraced HDR lit fractal bla bla bla trying to make it `immersive`.

I think people are bored with realism because they are trying to escape into fantasy. That's what a game is all about - an escape into illusions and fantasy. An escape from what? Reality. So at some point, it's actually counterproductive to be `too real` because then people start realizing that the realism blows the illusion. People want something that is `unreal` in enough ways as possible, which doesn't mean mirroring physical reality with precicely accurate physics and graphics. Doing so is more `science` than fantasy. Who would care if you had a perfectly real looking environment on the computer screen, the very kinds of behaviors that the `real` physics produces would not be interesting or exciting. People want things to be blown out of proportion, to be exaggerated, to be idealized, to be fantastic, to be larger than life, to present challenges that are impossible in physical reality, to pursue dreams that cannot exist as a physical reality, to get lost in an illusory daydream world that allows them to be what reality doesn't let them be. Who wants to be a totally photorealistic football player when you can be a totally unrealistic but thoroughly escapist wierd cartoon character. People don't want to buy reality, they want to buy dreams.

True, but there is a difference between looking real and being realistic. People want things to look real - even while being unrealistic.

Depends on the game. A First person perspective game/simulation (doesn't have to be a shooter) then yes. Basically the holodeck in Star Trek is the ultimate here.

But there is also a pick-up and play, more cartoony style of game that also appeals. Not all games work as well in 3d anyway.

Is Photorealism In Games The Right Direction?


Hell yes. Its pushing and improving the technology forward, which is a good thing. The fact that some games now lack any creativity is not due to photo realism, its due to the lack of creativity from the game designers themselves. Who says we can't have the best of both worlds?

Right now I'm having a blast with Orcs & Elves the cell phone game. It's cool to be impressed with things such as scaling, rotation, and alpha blending on a phone. The game definitely has more character than modern PC and console games because so much is left to the imagination. Photorealism has its place but it's not really for imagination fueld gamey games.

Photo-real painting movement ended when they invented PHOTOS! There was no need for photorealistic paintings anymore, too costy.

What we see right now, is the breaktrough of a new medium, using 4 dimensions. First there was the photo, with two dimensions. Then time was added as the third dimesion (movies), and now we got 3D plus time. Basicly we got virtual reality, that can be used as a medium, to entertain in many ways. It may also be boring. Especially when they don't have anything to tell at all. I mean, who want's to play yet trough another "good marine special ops guys are hunting for evil camel jockeys" game? I don't.

Take a good book. It's made of characters, like a bad book too. It's the story that makes it a good book. And that's exactly what the game industry has to learn. Artists need some freedom to make use of this new medium, and soon we'll have great artpieces that are more than only a lot of work.

And besides, there are many diffrent kinds of games. There will always be space for simple 2D games. But they dont represent the breaktrough of a new medium, as photorealistic games do. Additionally: there are many unique photo styles, so it's not the way that all photos are the same.

>Who cares about artistic pretentions? Games are not art;
>it just takes artistic ability to polish them.

A turd covered in goldleaf is still a turd.


Andy

Sorry, this is a bit long, I tried it short and it just did not fly. Actually even this is short, it's a wide reaching set of thoughts to do justice to.

I'm absolutely not bothered what the thing is called. I'd admit to just looking for the presentation to be "Adequately Convincing". I don't suppose one could get a snappy marketing slogan/spin out of that description so I can't seeing it getting coined as an expression.

However, it's spurred me to make content for flight Sims in a not so usual way. I'm not after making an average mesh and draping loads of paint over it to hide the deficiencies in that mesh, nor am I trying to get the absolute detail in the mesh. Getting a pretty good feeling of parts being where they should be and painting them with very simple and minimal colour patches is what I try to do wherever that is practical. Only resorting to more complex paint where it's needed or perhaps just a little more efficient/realistic as an approach.

What I am struggling with in flight is that visual cue of distance between objects in the 3d world.

No amount of photo-real or anything else like it will solve that, only true stereo vision will, or at least that's what I suspect. I'm not presently aware of any other way to get that visual cueing.

Getting a nice well controlled landing with a rotary winged craft is very tough in the last few inches of travel from air to ground, and that is when it's important to really see what the relationships of ground an craft are.

I've expressed this in terms of that which I know, but I bet there must be dozens of examples which others could describe where this sort of feedback is missing, and where it would improve dramatically if it were to be there.

It's worth noting that while most flight sims are no more than games, there is an element of great satisfaction in mastering "flying" in such an environment. It's clearly a different form of flight in a sim/game, but that just makes it different and not really any less valid than the conventional version of flight.

I got a free pair of 3d LCD "shutter" specs with a video card I bought some years back, it was absolutely great in terms of how the whole visual experience felt. Why not continue with it's use? Well the specs were not well designed, they worked, but kept falling off the whole time, and in between that were painfully uncomfortable. What is needed is some sort of gizmo that attaches to any pair of specs, specs designed by someone who knows how to design frames and lenses. It would then be simple to use, even on safety glasses that don't change what one sees, but do protect while also offering a suitable platform for such hardware.

Why add this to the debate? Well, it seems to me that it might be the case that video cards could exploit this style to great advantage without too much work. Video cards are good with throwing up lots of triangles, they are a bit slower doing triangles with paint (not by much, but it's still a potential area for savings in horsepower and there might be a potential gain in performance if it were not quite so complex a paint map). I'd imagine these concepts also apply for computers generating the underlying components for all this too, I can't see how it would not be the case in fact.

We are seeing a time where it's getting so complex that some sort of teaming is a requirement to even make a game these days. Looking into this area might help to remove the requirement, or at least slow it's spread.

I don't know all the answers of course, far from it, but it's got to be worth exploring all the areas that might help deliver an adequately convincing experience I'd have thought. The more people that bring their bit of knowledge to that table, the more likely we are to discover better ways to do it all, and that does not have to make it harder, but it does need a better understanding, the kind that diversity of know how can bring to it all.

I have trouble getting into car games for the same reason. In a game you have a fixed camera perspective and can't see past the next corner, neither can you guage speed well.

Then you have the whole feel thing, a decent car will comunicate how the road feels to the triver through the wheel, and their backside so you can feel the traction going a long time before you hear the noise of the tires giving out.

You also have gravity that is non existent in a game. It's bad enough that I can have more fun throwing my own car around street corners at 25-30 mph, heeltoing the throttle for rev matching down through the gears balancing the car before turning in and enjoying feeling the threshold of the tires as oversteer starts to kick in as you accelerate hard out.

If you could simulate all those feelings and the buzz when you get it right, you wouldn't need the graphics. There really isn't that much immersion in current realistic games, and the only thing you really do have is like someone else said. A story that negates the repetition and makes you want to find out what happens further along in the game.

what we need now is 'real-realism' - lol


What I am struggling with in flight is that visual cue of distance between objects in the 3d world.

No amount of photo-real or anything else like it will solve that, only true stereo vision will, or at least that's what I suspect.


People who only have one working eye don't constantly bump into things - depth perception doesn't stem from having two eyes, it comes from getting used to seeing with the number of eyes you do have.

... actually.. Depth-perception relies on 3 things. For-shortening, Stereo vision, and shading.... As long as one is apparent, you can gauge depth. the more that are present, the better the gauge.

If all 3 are nullified or played with... you can't tell depth. (you can fool the whole thing by faking 1 or 2 of the things, like in a lot of fun houses.)

DampeS8N:

Good call.

How (in practical terms) might you see it being made better beyond adding stereo?

Shading would seem do-able to some extent, but I'm not sure whether such benefit as one might get would be eaten by the horsepower required to get it that good. Might be OK though!

What of foreshortening then? Again, it sounds reasonable, but how would you propose to get it fairly cheaply in raw horsepower terms?

I guess I'm just not clear on "the how" with those other two yet. Must be a good area to investigate, I'd say the methodology is far from commonly understood, never mind documented. Looks a bit unexplored given a cursory glance.

Quite possibly worth a deeper delve into those.

I just want a room with wall to wall screens on all the walls/floors/ceoling and motion detectors and voice detectors so I can walk round the room and make actions to play RPGs etc. Like a primative holodeck.

Soul Calibur 3. Excellent graphics, and not exactly realistic looking graphics, but excellent artwork and very stylised. Great gameplay too, which ultimately, makes it work. Th efirst time you play a game, you might be taken in with the graphics, but as it wears on, it's how the game plays that hooks you in. The storyline, the characters, the animation, the controls, and i would say, the framerate.

Ever tried playing GT 1 on the PS 1 again? God, it's hard to play, because your used to the smoother frame rates available today. But mario? Zelda? on the SNES, still excellent.

well I don't see any games that try to mimic a photo . they have their own stylistic leanings.
doom 3 is a good example. it would n't fool anyone that it was totally real. it does n't even sell that really.

If graphics became so real that it fooled us then that would be fun for say 5 secs until we got bored with it.

Stylistic work will always win out in the end so stop worrying.

Didn't Half Life 2 tried to go for that "realistic" look?

I wouldn't say so, no. I'd say HL2 went for the bleak communist urban look. And did quite a good job of it in my opinion.

I do see advantages of photorealism in games, even if I also agree with many of the arguments against it.

Quite frankly, I'd like to place myself in a photorealistic immersive environment, and cause utter havok. I'd like to make cars explode, land a 747 on a freeway, etc. And I think seeing that presented with photoreal graphics would heighten the visual impact and immersion.

To the kind of people that say "oh, I have a real life, I don't need to see the same thing on the computer screen" I must ask, "do you regularly perform the kind of actions you see in videogames in real life?".

Yes, photorealism can also be boring, we do see it every day. Films somehow survive having "photoreal graphics" because they have other priorities. Plot, characterisation, etc. Once photoreal graphics in videogames have become pedestrian, perhaps we'll see similar emphasis from them. To continue the film comparison, filmmakers don't just use videocameras to film the real world. They use a raft of techniques and art styles to give us a different visual style, often not representing reality at all. Again, games do this too, and still will.

To the kind of people that say "oh, I have a real life, I don't need to see the same thing on the computer screen" I must ask, "do you regularly perform the kind of actions you see in videogames in real life?".

I get that same feeling whenever someone uses that argument. Nobody is trying to simulate your life, they are trying to simulate the appearance of objects in your life (glass, refraction, radiosity, etc). The experiences you have in the game are going to be more exciting than going to work and paying bills.

Non-realism adds opportunities. In a Mario-esque platformer you could collect items with which you can jump higher (meters high, in realistic terms)..

Anyone fancies a photorealistic person jumping ~5 meters up to the next platform? :P

Why are you equating realistic graphics with realistic behaviors/physics?

To me, they seem somewhat related. As I wrote, I can't imagine 'real person' jumping 5m into the air, but I can imagine a Mario does so.

So, to me moving towards realism more or less cancels out all these classic movements and other weird unnatural things we had in yesteryear's games. But that's all imho ofcoz..

Look at the Matrix. They do completely unnatural things but are set in realistic environments. I think the two things are unrelated.

But, as I said earlier, simulating realistic environments isn't really the goal as much as creating an authentic feeling game world is. Being able to accurately simulate the various effects in the real world and the light effects that hit your eyeball allow game designers to create more believable worlds.

What happens in those worlds doesn't have to be, in any way, realistic.

The photo-real painting movement ended after a brief period in the late 1960's for exactly the same reason - it was boring to look at.


Actually the movement is still very much alive and highly collected... People admire the technique. I used to teach a class in photorealistic painting methods at the Art Institute of Houston.

And as far as Realism in games... Eve Online... Space game... it looks like it should... space.

While we're discussing photo-realism, lets actually discuss the different methods game programmers use to "duplicate reality"

Normal mapping - although it has its uses, I've grown very tired of it. Oblivion, Gothic 3, NWN2, you name it - they use it on every available surface, and it makes it all look gritty, but also low-res (to save memory space)

Parallax mapping - wow, there's a feature I could do well without. Sure, the walls and stones look great from that perfect angle, but from everywhere else, the colors bleed and it just looks like a shiny, wobbly glass surface coat.

Bloom filter / HDR - they pretty much look the same to me, just HDR is a tad brighter, and bloom is often more yellow tinted. They absolutely make it hard to look at the screen - HDR on a mostly white surface will blow the colors and you'll end up with a white blob filling big chuncks of the screen.

Water shaders - love them. Especially the Shader model 3 version, with depth and fog.

Please, list some of your "realistic techs" and your opinions on them.

Relief mapping :)

Ah, I wish it was used more often.
I think UT 2007 uses it, and there is a relief mapping mod for Doom 3, but other than that there isn't much.

It is superior to parallax mapping because it can work at absolutely any angle and even on curved surfaces, to draw all sorts of detail on top of a mesh that would otherwise be done with tons of polygons.

Relief maps are so effective, in fact, that you could theoretically render a believable game of Checkers using a single 2 polygon plane and a detailed relief map.

It does self-shadowing, too :)


Normal mapping, I tend to agree with Ashmantle. For some reason, people just can't figure out how to use it...
Parallax mapping, of course, is nothing compared to relief mapping. It looks too exaggerated or something.

HDR has its places. HL2 Episode 1 makes excellent use of it for the reactor core scene; the blinding HDR really shows the amount of energy running through that thing, and it wouldn't be possible any other way.
With a bloom filter it would just get in the way and be ugly, because it wouldn't fade away. The HDR shows the thing's power, then it gets out of your way until it sees it fit to remind you again.
Without any of that, it would just look like a blue ball and a weirdly shaped room with a few windows.

I liked how most scenes aside from the reactor core were more conservative about HDR (I only really noticed when I thought about it, usually), with the reactor core really standing out as one of those strangely intense "holy crap!" moments.

Well, I'm working on a text based MUD, so I guess that says something about my opinion. (or maybe just my coding skills? ;)

I do think that as graphics become more photo-realistic graphics the burden of immersing the player in the environment is greatly lessened since there is a much stronger correlation to environments already known. Of course the flipside is that as graphics become more photo-realistic, the player will intuitively expect other things to more closely mimic reality - or at least more closely mimic the complexity of reality. It's just a matter of consistency. Consistency of depth.

If you give me the original Final Fantasy with 8 bit graphics - it's consistent with the depth of the rest of the technology that the NPC's just move around town at random and repeat the same canned line until the next major event. If you give me a photo-realistic fantasy environment, those same behaviors would be completely jarring because it's just not consistent with the expectations generated.

I think we're at a point graphically where a lot more time and resources, and consequently CPU share, should be put towards supporting the other aspects of immersion. Sound has come a fair ways, but AI is still sorely lacking, IMO.

Of course, no technology inherently guarantees a better game. Superior technology obviously has it's upsides, but it has it's downsides as well."Inferior" technology often allows more rapid development (take a text based game for example) - such that the increased development allows features that may not be reasonably (cost-wise, development time, etc) implemented with the superior technology, thus negating or surpassing any of the benefits inherent to the superior technology itself.

The fact that some games now lack any creativity is not due to photo realism, its due to the lack of creativity from the game designers themselves


I don't know if I'd blame the game designers. I'd be more inclined to point the finger at the publishers/investors, and, the market itself.


Please, list some of your "realistic techs" and your opinions on them.



CSM - cascaded shadow maps... nice shadows are one the most important features for making things look "real". And I'm not talking about stencils, because that is so far from any reality... soft shadows cast and received by EVERYTHING in the scene, that's more like it.

Thinking back to how thing used to be done in computer games, in the retro period and the early days, there was only a very loose association between what you did in the game and something you do in real life. There was no advanced physics engines and texture mapping or correct lighting. You could draw a few pixels on the screen roughly in the shape of a `something` and you were able to figure that it actually was that thing. The old 8-bit games and even prior to that were not exactly running a simulation of planet earth. It's been very popular in games to move beyond the limitations of accurate physics.

Think of a game like worms where you have pieces of land that float in the sky. Clearly it's not physically accurate or even realistic looking but it's fun. Or a platform game where the main character jumps extra high or far with non-realistic gravity. It would be all very well to implement totally realistic physics and realistic graphics in these games but then some elements of the gameplay would be near enough impossible. Physics, if done correctly, clamps down on `unreality` and requires that everything work like it does in `real life`. That would go a long way toward obliterating many many activities and behaviors that defy reality that you would otherwise have fun doing in a game, and indeed maybe games are fun because of that very ability to defy gravity or pass through walls or do things that physics can't do. Realistic representations of real life behavior is not necessarily a good thing for games. Maybe for some kinds of games, that are basically simulations, and try to make things happen how the would happen in real life. But that's not all games by any means and it'd be a very sorry state if it were.

Things go in 3 steps - thesis, antithesis and synthesis. The thesis was what we started out with where computers very strictly defined the form of a game due to hardware limits and the player had to use lots of `interpretation` and imagination. The antithesis is to totally abstractify everything and center everything around the player and be totally `realistic` with physics so nothing is left to the imagination other than that models may be of things that don't really exist. I think the synthesis, where these two extremes come together, will be that you have some form of accurate physics and graphics, but with artistic styling, and the ability to perform `magic`, which means defying the physics and and using it creatively knowing that it's all a controlled illusion. Some genre's already do that.

Uncanny valley - do a search, interesting stuff :)