Photorealistic Games – By 2024?

TheBluePill

2[H]4U
Joined
Aug 17, 2005
Messages
3,773
Photorealistic Games – By 2024?

Eventually technology will reach the “holy grail” level of photorealistic real-time 3D rendering at an affordable price. Once we arrive there, there really isn’t anywhere new to innovate, convergence will happen between all platforms, PCs, Consoles, etc will all have the same detail level and capabilities and graphics will become a commodity item. In a nut shell, we are racing to the “end” of 3D rendering as a prized technological advancement made regularly. At a given point, there will be highly diminishing returns for rendering advancement, but when?

At our current rate of progress, we really are following Moore’s law, we see nearly double the rendering power on the desktop every 18 months. At a point every few years, consumer devices like console game systems sample the level of advancement and advance themselves. Likewise, once photorealistic, real time rendering reaches the Desktop, soon consoles will follow and then Portables, phones and so on and so forth.

But when?

That’s is the question that is up for debate.. 15 years (10 Cycles of advancement)?
15 years ago, (1994), we saw the first real consumer 3D acceleration hardware for the desktop, The GLINT 300SX put out 2.5 Gflops.. Quite a powerhouse for its day.. But today, we have 1 Tflop cards in the desktop… That 400x the power in only 15 years.. If the same holds true, surly 400 Tflops of rendering could achieve Photorealism in Real time?

Perhaps more?
 
The question I think we need to answer is whether we really need photorealistic games or not. Because in games, graphics is not everything. I am sure we'll have the power by then, but maybe developers will prefer non-photorealistic approaches not only due to simplicity, but also for other reasons. I am not quite certain that everybody will find it so entertaining to see photorealistic intestines popping out left and right when playing Call of Duty 17, so developers might not stick to it... ;-)
 
Can't wait for AIs to be bitching about weapons of the game server chat! Imagine them actually talking over coms too lol.
 
The question I think we need to answer is whether we really need photorealistic games or not. Because in games, graphics is not everything. I am sure we'll have the power by then, but maybe developers will prefer non-photorealistic approaches not only due to simplicity, but also for other reasons. I am not quite certain that everybody will find it so entertaining to see photorealistic intestines popping out left and right when playing Call of Duty 17, so developers might not stick to it... ;-)

Speak for yourself. :p I've experienced 5 generations in video game technology. I was blown away during the great video game boom. I kept looking forward to the next gen every time.

Nintendo - The days of Duck Hunt and Mario Bros. It was my first video game console played on my grandparents old TV set. I remember having to fiddle with the antenna box to get the signal just right. To me it was amazing. I couldn't believe I was shooting ducks through my TV set! Mind you I was only 4.

Super Nintendo - My god! The graphics! Mortal Kombat! Super Mario Bros.! POWER RANGERS! It was like comparing Counter-Strike 1.6 to Source. I had never seen so much color. I thought it couldn't get any better.

___Nintendo begins to slow down____

Nintendo 64 - Sweet jesus! The Crysis of the console world! Super Mario 64 was one of the best games I had ever played. I could finally move in three dimensions! HOLY SH--! Did I just shoot someone? I FREAKING SHOT SOMEONE IN THE ASS AND HE REACTED! WAS THAT A HEADSHOT!? YOU DAMN RIGHT IT IS!! I honestly thought nothing could get better than this. I was wrong.

Playstation - What!!? Explosions! Sprites! MOFO RAGDOLLS!? *HEAD EXPLODES*

Gamecube - Nice try Nintendo

Playstation 2/Xbox - Whoa this is pretty nice. Higher quality models, better light and shading. Damn splinter cell? I can stalk someone in the dark? Freakin sweet.

Wii - Meh

Xbox 360/PS3 - This is cool too.
 
2060, rendering a taxi cab through NY City would require a huge amount of power; Graphically, Processing, I/O Sub system, Hard Drive's throughout put with Displays capable of insane resolutions like 38400x24000 with 1000hz+

Your talking about every single person on the sidewalk having its own advanced AI, a single wind gust would effect the slightest details....

2024 is way too soon imho
 
We'll probably have photorealistic scenes in games pretty soon. The trouble will be that faces/expressions/emotions will still suck and AI will remain unrealistic. We may be only a few hardware generations away from the creepy close-but-not-quite-real creepy weirdness seen in things like that Beowulf film.
 
You might be dead by then, so who cares about something that far in the future? I'd say games are already approaching photorealism if you're just talking about it in cutscenes and stills. The problem is rendering it in real time in actual video game play without stuttering framerates.
 
Photo-realism is only partially dictated by the hardware. A big part of it is also environmental design and art assets. Those things are expensive so in order to get there budgets will have to increase significantly. The other part of it is that by 2024 we would've only gone through two more console cycles at most so advances on the PC won't be reflected in games (as it is today).
 
Photorealism is a very broad term. There's graphical realism, then there's physical/logical realism, and they can get lumped together under "photorealism". For example:

I am not quite certain that everybody will find it so entertaining to see photorealistic intestines popping out left and right when playing Call of Duty 17, so developers might not stick to it... ;-)

There are two aspects to this - the scripting/physical modelling which says "when shot/stabbed/blown up in this way, there should be intestines all over the place" and the graphical work which says "the intestines should look like this, and thus be photorealistic".

If you are using the term "photorealism" as in "a perfect model of the world, in which everything that should happen, happens and it looks right when it does" then obviously that's going to take a great deal longer to achieve.

On the other hand, if you set aside physics and scripting and just mean photorealism as in "whatever the developer is choosing to depict, it looks how it should look" then that's still not entirely straightforward. I'm sure a photorealistic cardboard box in a featureless white room with a single light source could almost be rendered in real time right now. Whereas a view of 50,000 photorealistic soldiers fighting on a huge photorealistic battlefield (or, as someone else mentioned, driving through New York at night) is much larger scale and more complex, there's a lot more stuff to render, and so a photorealistic real-time depiction of it is not going to happen very soon.

It's also going to be a matter of diminishing returns. Once something looks "almost" photorealistic (however you choose to define "almost"), few will be able to tell (or care) that it isn't completely photorealistic... but the jump from "almost" to "completely" photorealistic might require orders of magnitude more processing power (and development time?). In which case, perhaps there will come a point where we just stop trying to improve the graphics of our games because it just isn't worth the cost and effort.
 
It's also going to be a matter of diminishing returns. Once something looks "almost" photorealistic (however you choose to define "almost"), few will be able to tell (or care) that it isn't completely photorealistic... but the jump from "almost" to "completely" photorealistic might require orders of magnitude more processing power (and development time?). In which case, perhaps there will come a point where we just stop trying to improve the graphics of our games because it just isn't worth the cost and effort.

I think you really hit the nail on the head there, At what point are we going to see that level that people just cant seem to justify needing/wanting more?
 
Within the next 5-7 years, 3D TV, Projection and the like will be the norm.
We already see it becoming oh-so popular with Movies as of late, the home market is about
to explode with it. Couple real time "near" photorealistic 3D rendering for games and it will
be ultra intense and make game play all the better.
 
Photorealistic Games – By 2024?

<snip>

Perhaps more?

Sooner I believe. Just look at games like Crysis or better yet, the engine that powers it: CryEngine 2.

It achieves a level of realism never before seen on today's hardware. Things will improve year after year, although with the current "console prioritized over PC" methodology, it will go slower than it would if the PC was still the main platform for developers. The next iteration of engines (in 3-4 years), will certainly bring even more realism than CryEngine 2. And 3-4 years after that, I believe we will get very close to full on photo realism. So, 2015-2017 is my bet for truly photo realistic games.

As for the rest of "realism", we're also coming closer to the real thing. The fact that hardware evolved to allow physics calculations to be performed much faster, thus allowing for much more complex and realistic effects, is one of the steps that will lead to (at last!!) realistic and fully destructible environments, but I believe this one will take more time than the photo realistic graphics, since this "problem" isn't tied to the GPU only, but with storage space and memory aswell.
 
Imagine: killing enemy soldiers in games (hell, TKing) which AI is so advanced, they are self-aware!
Oh, the moral questions!

:D

Anyways, have you seen what can be pulled off with Cryengine 2? Crysis barely scratchs the surface of that engine graphicly.
 
It's going to take more than sophisticated hardware to power photo-realistic games. It's going to take breakthrough upon breakthrough in physics and rendering methods for starters. And then breakthroughs in AI to make your photorealistic game more than a pretty picture.
 
It's going to take more than sophisticated hardware to power photo-realistic games. It's going to take breakthrough upon breakthrough in physics and rendering methods for starters. And then breakthroughs in AI to make your photorealistic game more than a pretty picture.


They really go hand in hand.. I see physics and rendering methods are advancing at a huge pace as well..
 
I don't think that 2024 is an unfair estimate, but I believe it it will take a while longer.
 
would prefer better AI over better graphics

I swear AI code has taken a step backward since Halo 1
 
I think you really hit the nail on the head there, At what point are we going to see that level that people just cant seem to justify needing/wanting more?

never. people will always want more and find a reason to push further.
 
i agree that AI really needs to step up a few notches to keep up with the graphics of things. its the only reason that i can play crysis and not die early on in the game (i suck at most real games without cheating). you can outsmart the AI sooo easily.
 
It's going to take more than sophisticated hardware to power photo-realistic games. It's going to take breakthrough upon breakthrough in physics and rendering methods for starters. And then breakthroughs in AI to make your photorealistic game more than a pretty picture.

+1

I think photorealism would be around late 2020's. Reaching photorealism won't be much of a problem. But I think the amount of power it will require from hardware would be a problem.
 
One way to see where we are headed is to look back ~10 years. Half life (technically 1998) vs Crysis. Go head and google image search the two of them to wrap your head around that.

Ten more years should get us most of the way there, but I agree, AI and physics need to advance if we want these environments to feel real.
 
I get excited just reading this thread and thinking about the possibilities.

When I'm 70 years old and photo-realistic games finally exist, I'm going to be such a gamer geek.
 
I get excited just reading this thread and thinking about the possibilities.

When I'm 70 years old and photo-realistic games finally exist, I'm going to be such a gamer geek.

lol


I could imagine playing Call of Duty Modern Warfare 10 in the future and playing it will be so lifelike, it would be as if you are actually in war and part of the ingame. :D
 
we will move more towards immersion, such as 3D or virtual reality. Then incorporating better graphics will play in to the mix.
 
The REAL question is do we even want photo realism? In case no one noticed, the real world is pretty, well, BORING. Hence, games. I don't want to play RL, I want to play make believe. For example, look at most space games. They are set in a colorful, brightly lit space with stars and nebulae in every direction. If they were photo realistic, it would be an all black canvas that is very dark with a bunch of white dots. Or Braid - that game looks fantastic, and has no intention of photo realism whatsoever, far from it.

Real != Better.
 
i agree that AI really needs to step up a few notches to keep up with the graphics of things. its the only reason that i can play crysis and not die early on in the game (i suck at most real games without cheating). you can outsmart the AI sooo easily.
I think that was the point in Crysis to be able to easily evade enemies. I thought Stalker had pretty believable ai but nothing that I've played had exceptionally realistic ai.
Posted via [H] Mobile Device
 
I dont know i love that girl in the "movies from space" and i cant tell if she is CG or green scrrened, i shure would like a desktop pic to examine though
 
I think that was the point in Crysis to be able to easily evade enemies. I thought Stalker had pretty believable ai but nothing that I've played had exceptionally realistic ai.
Posted via [H] Mobile Device

Thats when Multiplayer games come in if you unbelievable Ai :p
 
The REAL question is do we even want photo realism? In case no one noticed, the real world is pretty, well, BORING. Hence, games. I don't want to play RL, I want to play make believe. For example, look at most space games. They are set in a colorful, brightly lit space with stars and nebulae in every direction. If they were photo realistic, it would be an all black canvas that is very dark with a bunch of white dots. Or Braid - that game looks fantastic, and has no intention of photo realism whatsoever, far from it.

Real != Better.

And you WILL play "make believe", only with far more realistic models/animations and textures. As graphics and physics advance, you can already see realistic cloth movement around a body and almost realistic hair movement.

Having photo realistic graphics and realistic physics effects is not about making a game "real" i.e. to mimic or replace real life. It's about creating a totally different feel of immersion, into something we usually know from the start is just a game. We will still have medic kits or be able to take lots of hits, even without armor, but the realistic look of things and the realism of the physics simulations, will make us feel like we are in the game itself, as if we were part of a movie, where all the action revolves around us.

Although I would limit this to First Person games, since it doesn't quite work with other types of games.
 
I think a part of the graphical wall alluded to is the assumption that we will be using the same displays next year that we are today. With multitouch display, larger displays capable of larger resolutions, 120MHZ Stereoscopic 3d capable monitors, and mallible OLED displays all just around the corner theres no telling what kind of GPU's we'll need to power these next year.
 
Back
Top