• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

When will video game graphics become indistinguishable from reality?

I guess it depends on what you define to be "reality".

Tekken-2-Nina-Williams-CG-CGI-Ad-Campaign.jpg

The definition of horror.
 
It depends on the genre. In 5 years? No way. In 20 years? Maybe. Racing games will get there sooner since there are no wonkey animations breaking immersion.
 
http://www.fxguide.com/featured/maleficent/
The film industry has already done this, (more like is capable of doing it but will showcase it very rarely because of the cost). It's just super expensive to do and the hardware available is not as good as CGI. There are tons of "tells" in an animated character that immediately stick out. So we have to get past that barrier as well. On top of the the environments, physics, etc. It's just not possible right now without sacrificing an awful lot of factors that would take up memory, like complex AI. So maybe by the end of this generation we'll start seeing something that's close to indistinguishable from real life.
2kw3Dvm.gif

ZMPeQdu.gif

^ CGI

Yup, it's why GoT shows the dragons so little, that shit is expensive.
 
Let's say we get a 1000 x increase in processing power (nearly 2 petaflops), which may be possible with new materials in like 20 years. That would be like giving a PS4 engine which normally renders a frame in 1/30th of a second a little over 30 seconds to render a frame. CGI, which hasn't reached reality yet, though its closing in fast (well the problem is more artistic, I'll give you that) often gives hours to days to a frame. With very carefully constructed scenes you may match CGI without the IQ, but that's not within 5 years.
 
I think by 2025-2030 assuming there isn't some huge societal change and people still go in for it, home video game systems, be it PS6, Nintendo Wii Rn't Bankrupt Yet, Xbox Three, or PC will be very much capable of mimicking reality and some games will have to be pulled back from realism so people don't just spend all day in their own little world wearing 8K VR headsets and haptic suits.

It'll be great.
 
http://www.fxguide.com/featured/maleficent/
The film industry has already done this, (more like is capable of doing it but will showcase it very rarely because of the cost). It's just super expensive to do and the hardware available is not as good as CGI. There are tons of "tells" in an animated character that immediately stick out. So we have to get past that barrier as well. On top of the the environments, physics, etc. It's just not possible right now without sacrificing an awful lot of factors that would take up memory, like complex AI. So maybe by the end of this generation we'll start seeing something that's close to indistinguishable from real life.
2kw3Dvm.gif

ZMPeQdu.gif

^ CGI

This. Movie CGI is getting extremely close, and real time console video game graphics have traditionally stayed about 10 years behind CGI. I think in 10 years we'll be extremely close.

Never I hope, i have no interests in games that are hyper realistic. Its defeats the point of them imo.

The thing is, I think hitting the point of photorealism might eventually lead to MORE games that deliberately try not to be hyper realistic. At least when we reach a point where photorealism isn't extremely expensive.

It's kind of like what happened to art when photography was invented.
 
We know we will have gotten there only after it's already happened.

What I mean is that it won't happen in some hyped, obvious way like the lunar landing. It'll be something that most of us won't even realize has happened. Say like adding virtual overlays to real life (through technologies like Googleglass or Oculus). By the time those technologies have begun recreating and virtualizing the "reality" around us, there won't be much point in drawing a distinction between "game graphics" and "reality." It will be a moot point by the time we realize that it's already happened.

However, if you're just talking about verisimilitude, that's a different discussion. We will always know the difference between verisimilar graphics and reality. It's not like people confuse live action TV or movies for reality.
 
I play games to escape reality.

But yeah, we're never going to be able to say "this looks and feels real". If VR becomes mainstream, that's the closest we'll get visually I think, but even then, it's not even going to be close.

For games to become "indistinguishable from reality", we'd need VR + PowerGloves + Treadmill + Smell-o-vision to become mainstream and affordable, but then you'll still feel that you're wearing them gloves, the VR helmet, you'll feel the treadmill instead of whatever the game is telling you you're walking on, and Smell-o-vision is Sci-Fi.

Never, not gonna happen, ever.

I think five years is a bit ridiculous as a prediction.
But to say, flat out and definitively, NEVER? Nah, I don't believe that at all.

VR should eventually be an interface directly with the brain - no gloves and goggles and shit. At that point, embedded realities running at various speeds for various purposes should be possible. Smell-o-vision and all. Gaming will surely be one use of that.

I guess if you postulate an unavoidable extinction event to occur before our technologies have achieved the sophistication necessary then of course it will NEVER happen.
Seems a bit pessimistic to be so absolute about it though.
 
Getting past the uncanny valley is gonna take a long time. Not sure about the resolution/framerate either, refresh rates can still improve.
 
In a screenshot of a largely static scene? Sure 5 years from now it should be doable pretty easily with laser scanning etc. In motion with human characters during gameplay or even real-time cut scenes? Possibly never, but at least 10 years from now assuming we can get a replacement for silicon processors that accelerates the pace of computing power improvements.

The best movie CGI isn't quite there yet for non static or artificial objects. Sure you can point to a couple examples where they have digital actors scanned from real people being performance captured in real time but those use very controlled environments. Try that when you have actors walking around and interacting with each other and the environment for more than a few seconds and it will quickly fall apart. Unless you use other tricks to disguise it.
 
I think five years is a bit ridiculous as a prediction.
But to say, flat out and definitively, NEVER? Nah, I don't believe that at all.

VR should eventually be an interface directly with the brain - no gloves and goggles and shit. At that point, embedded realities running at various speeds for various purposes should be possible. Smell-o-vision and all. Gaming will surely be one use of that.

I guess if you postulate an unavoidable extinction event to occur before our technologies have achieved the sophistication necessary then of course it will NEVER happen.
Seems a bit pessimistic to be so absolute about it though.
I agree, I'm pessimistic about it.
I think there's always going to be some limiting factor, unless you eventually don't even need to move in real life to move in the game, but still feel like you're moving "for real". Maybe I'm thinking too short term, and in too many modules. If one device could beam stuff to your brain while you're laying/sitting somewhere, that obviously be simpler than having all the separate peripherals I listed.

I have no understanding of how that would be possible though, but would love to laugh about it 20 years from now :)
 
When we welcome...

THE WORLD OF TOMORROW!

So essentially, jump into your local cryogenics lab and wait. 1000 years is a good benchmark.
 
They already are.

a88uef.gif

This is uncanny. And it us used as an example so often because it is a blurry gif of a blurry home video type scene. The mind fills in the blanks and makes the viewer think "Wow that's so real!"

But once you see it, you can't unsee and you'll notice it in all CG henceforth.

It's the blurry home video technique that hides CG trickery in thousands of youtube videos i.e. eagle(falcon?) carrying the human baby away in a park which went viral and tricked many. Home/phone video effects should qualify as AA techniques.

That said, the movements are well done. It is a good job with its faults blurred and camera-shaked out. But we're gonna be uncanny for a couple of decades or more in regards to console gaming. If gaming even exists as we know it 20 years from now. So many directions it could take, backwards or forwards.

Hopefully like looking through a giant window and seeing another world. Finally. Couches or recliners still heavily involved. And controllers. That's all I desire. No head tracking required for me.
 
Depending on the talent of the artists and animators working on the project, we did it 10 years ago. Don't make the mistake of thinking this is a technology issue. It's not. it's a talent issue. Photo realism is freaking hard to nail with any consistency the instant things start moving.
 
For environments, static objects -- it's pretty much indistinguishable now.

For humans, not in our lifetime. Our brains are supercharged for reading even the most minute details on human facial movements. Just think, a computer model has to be complex enough to simulate all those details, in real time. It's like capturing a soul.
 
One thing I don't understand is how real life video looks prefect real on a 480 set or even at 240p on youtube. Do we really need 16k for real lifeness I be achieved. Why can't we emulate what we see on a standard definition tv? Anyone know why or why not?
 
Going from 1080p to 16K is going to delay the advent of photorealistic graphics more than anything. 64 times the pixels requires 64 times the processing power at today's level of realism, let alone the level required for graphics indistinguishable from real life.

To reach a factor of 64 will take more than a decade at today's pace. After that, we will still have to transition to path-tracing to reach consistent and dynamic photorealism that really fools the eye without any artistic trickery. That will require a two or three orders of magnitude leap in processing power.

By that time, VR will probably be widespread and also support 16K. That means upping the frame rate to at least 120fps, probably more. Add another factor of two. The FoV will be increased significantly from today's games, which will further add to the requirements. And, of course, everything will need to be rendered in stereo.

My best guess: 20-30 years.
 
Yeah 20-30 would be my guess. So many things to figure out. Especially in motion. Not just animations, but pop-in, aliasing etc.
 
If you have a pt, like a cat or a dog, look at them. When graphics can account for each and every single hair on that cat, with each hair being fully represented on the color of that fur, from the outside and in, when we have a perfectly made skeleton for these things in the games, and their eyes can fully be captured, etc. then we start to talk
 
well we need to light a scene accurately, so we're probably going to need some sort of path tracing kind of thing to trace a sample of all of the potential paths light can take between a light source and the camera.

we're going to need better modeling of how that light interacts with surfaces. multiple specular lobes, sampled brdfs, etc.

we're going to need to lock down animation which is likely not going to be possible through hand animation, so it seems like super high quality physics will be required.

we're going to need to create scenes with the density of detail and variation we expect from real life.
 
That depends on your definition of realism and if absolute realism a wall we have to knock down... could take forever to bash on that wall. Always so close but never truly there.

We all have been fooled by movies already, we've seen things without knowing they were CGI. But I'd be surprised if that CGI was the main actor in a movie, over the length of a movie. Because as far as I know that is still incredibly hard to make indistinguishable from real humans.
 
I hope somebody does it once so we can get it out of our system, then go back to making games with good art direction. Which, btw, hyper realism ain't.
 
Well if and when graphics get to a photorealistic level I know I won't be playing any games with heavy violence anymore.
 
What about movies? When movies achieve photorealism, then games will follow a few decades later perhaps. Then again the whole virtual camera and dynamic movement of characters makes it almost impossible to achieve ever.
 
Top Bottom