One interesting point of reference is the initial switch to HD in the 2000s; I remember there was a bit of panic in the beginning because news studios had to adjust studio lighting and makeup; flaws that were not perceptible on a CRT were all of a sudden extremely noticeable blown up on a 48" flatscreen.
Higher resolution, better color replication, and frame rate make very obvious the fact that there seems to be a magical glowing orb following around the characters right behind the camera. Immersion breaking because you can get away with it with less quality, it's more difficult to notice.
Something that I've also found more and more irritating is the foley artists doing ridiculous things for sound effects, especially in nature documentaries, but all over the place really.
ex: Farscape - all the CG was rendered in low res, so even when you view a 1080p copy it looks silly. Imagine all the content that would have to be re-done and upscaled to make it watchable in 4k.
Nowadays people open up koi pond videos on phones and let their cats play with it. But like you mentioned, it would be interesting to see a study on it.
My point is: CRT did show a lot more colours, had much lower latency, and did not have a grid where you could see individual pixels.
No colour banding, no lag and no screen door effect. That really enhanced the content you saw when you had a CRT.
During that era, 640x480 video was the common high end.
Occasionally things will be slightly out of focus, but it wasn’t apparent on a SD CRT so they shot the scene. On an HD screen you can see it and it’s kind of distracting.
Apart from graphics most stuff in games is pretty rough. Animations are generally bad and ways before reaching the uncanny valley of "getting close"; they're still in the "abstract representation of concept" detail level. AI is dumb (largely for [perceived] player-acceptance reasons). Sound is generally poor; some games still don't use 3D sound. Physics are "abstract representation" level again, some games still have fly-through walls and vibrating items. etc. etc.
This hasn't changed. New "realistic" games still range from horrible to boring-looking. I don't know if it's collective delusion or if I'm missing something.
https://www.youtube.com/watch?v=isvWpUXgKgM
Take this as a compliment, not a criticism when I say you've rigged your proposition by holding up Nintendo first party games, and Shigeru Miyamoto for comparison. If you were to look back at he average Nintendo game, not made by Nintendo, they are mostly unremarkable.
For example, I can clearly see that ray tracing produces better results. But it's a bit harder to tell how better it is, to find the words that describes how better it is. Of course, one can say that, for example, photon tracing is more physically accurate. But still, what words can we use to describe how real (or not real) a still image is ("more realistic" doesn't count :-))
But also, the design often adapts to the capabilities. Games like GTA3 used to have thick fog just to hide the fact that they couldn't render stuff far away in time. You can say that's an artistic choice to give a smoggy big city atmosphere, but clearly it was a practical choice as well. Even today, game designers like to make things dark, blurry and rainy, so that the un-realism becomes less obvious.
The rest, not so much, sure.
It’s basically using the same technique as hand drawn animation where as long as you realize what’s being represented you can automatically fill in missing details. However, this fails as soon as you focus on any one detail.
Honestly, it’s not bad for what it is. I mean the physics engine was probably the worst part, but as a visual demo that’s fine.
Which is something really obvious with the PS5 UE 5 tech demo [0]. The environment looks great and very realistic, while the character is very stylized and looks more like a comic than realistic.
We are getting better. Like snow deformation in RDR2, for instance(works even in a PS4).
But random bits of debris that can get kicked around - and subsequently inspected - no. That's a problem.
> No video game will let you inspect the leaves of plants to look for little bugs, the way reality will
That's an easier problem than tracking all the debris. Before you inspect, you have no idea what will be there - the computer also doesn't have to and it can be optimized away until there's an observer.
Think heisenberg uncertainty principle but for virtual worlds.
You could implement little bugs without all that much trouble, true. Maybe a better way for me to think about it is more like, the real world has an incredible diversity of appearances at a small scale, that's hard to reproduce in a video game. The bug isn't the part that causes trouble, it's the fact that every leaf looks different, some leaves are oddly discolored, some leaves have bug bites on them, some got a little torn up, et cetera.
Sounds like a fun project for the next generations. Something to play when I am older.
Now suppose I want to take some fruit from the trees. And I want to slice and dice the different fruits. Peeling the skins when appropriate.
Or I want to plant a new tree. Or I want to remove it, with say, a lightsaber.
That's a couple generators more. Now add birds, dogs, insects, moss, moist surfaces, and so on. Hence, my point on needing several generators.
There are lots of things to be improved in games, but what you're describing are improvements to a real life simulation.
Why not? With procedural generation and LOD rendering it's not impossible. Not that it's easy, but not impossible?
Also, some games are intentionally cartoonish as an artistic choice so photorealism isn't always the goal.
We've probably reached the point where human cost exceeds computational cost, which is to say that developing and QA testing such a feature would probably cost more money than it's worth. How many users of software would gain from such minutiae?
- Hair. Up until very recently hair was downright awful. Nowadays it’s acceptable-ish, but there’s still lots of room for improvement, in particular in natural motion of long hair.
- Water. I spend a big chunk of my time on the water, so I’m probably more attuned to how it moves than most. Games just don’t have it down. In particular, I think a lot could be gained by embracing it’s fractal nature: in my experience, at every human scale (mm to dam and everything in between) very similar wave patterns exist, but games tend to have just a small fixed number of “wave-layers” at various scales stacked together.
- Clouds. I can easily spend hours just observing clouds, looking at things like their shape, overall motion, internal motion, composition, edge behaviors, etc, and how they change over time. Game clouds are lacking in all these regards, particularly the time-sensitive nature of a cloud.
- Foliage. In games I’ve seen, individual plants/etc. in isolation generally look really quite decent. But the second an physical object interacts with them, they very clearly don’t respond in the right ways. There’s a lot about how branches bend and leaves rustle and more that is lost. Additionally, in groups of plants it’s often clear that some small number of models are being reused, possibly with some generated randomness added. But the variety doesn’t come close to matching what one would really see.
- Human faces and expressions. These are generally really bad, especially in normal gameplay (cut-scenes are sometimes better)
Again, this is probably all just really weird stuff I notice because I spend the vast majority of my time outdoors and only see “HiFi” games being played very infrequently. I don’t think games are worse for not implementing these, but I am very interested in what they’ll look like 10-20 years down the road.
What I think will be a huge boost is proper raytracing. You can see animated movies and blender renders look stunning but they take minutes per frame.
Foliage seems like something that will be difficult to get realistic for any close inspect, at least from a layperson's point of view.
It was so noticeable because it was such a huge increase in quality compared to what passed for water in games before - usually some kind of blue-grey translucent texture. For the first time, pixel shaders produced water that was clearly an attempt to imitate water IRL, not the cartoonish representation of it.
My vote would be cloth simulation and clothing clipping. I've yet to see a game that comes even close to doing this realistically. Imagine what happens to your sleeves when you lift your arms above your head for example, or how the plates of a suit of armor naturally slide over each other. In every game I can think of, clothing is rigidly attached to the underlying skeleton and it just stretches/clips as the character moves.
I guess fidelity of everything else has gotten much better, because I recently started noticing this and now I find it very distracting in any game that has in-engine cut scenes involving character closeups.
http://www.retroadv.it/images/03082019/Home_Computer_Magazin...
ML creative tools stand to automate a lot of this imo.
[1] random example: https://snapperstech.com/
> Makes me wonder - what are we "not" seeing right now that will make us think this way 10-20 years from now?
Yes - well said. That's what I was trying to convey in my comment.
Then there’s the more obvious stuff that isn’t done well even today: skeletal animation is still lacking and feels unnatural, physics systems are still very approximate and constrained - often times most things are indestructible in games, fluid dynamics are still very slow/limited. Human models still don’t look real though, and the voice acting never quite matches the mouth movement or body language.
I do really feel like we’ve crossed the uncanny valley when it comes to natural scenery rendering. But a lot of what makes things feel real are still missing from games.
It's very cool tech. But it doesn't look real.
Take a look at The Mandalorian series, almost all of the outdoor scenes were shot using the video wall technology.
I've worked in computer graphics and I didn't realise the sets were fake until after I finished the whole series.
How about ray tracing?. If we get real-time 60fps+ ray-traced computer graphics in games, that would blow what we have now out of the water
Graphics — even movie quality graphics — don’t look anything like what a camera produces. The sole exception is when you’re mixing in real video with CG elements. But try to synthesize an entire frame from scratch, then put it next to a video camera’s output, and people can tell the difference.
Also, screenshots are misleading. You have to test video, not screencaps. Video is processed differently by the brain.
10 out of 10 times, all the graphics engineers come out of the woodwork going “but actually we do know how! It’s a matter of using so and so calculations and measuring BDRF and” none of those equations work.
The reverse-problem is a pet-peeve of mine: It seems many people have been accidentally brainwashed by Hollywood into thinking that film-camera effects are signs of "realism."
So then the first-person game introduces something like lens-flares, and everyone goes: "OMG it's so realistic now", even though the truth is the exact opposite. If you were "really there" as Hero Protagonist, you wouldn't have camera-lenses for eyeballs except in a cyber-punk genre.
Plop a nature video next to your forest rendering and it’ll become apparent just how unsolved trees are. And everything else, for that matter.
The precise claim is this: viewers should be able to identify a rendered video no better than random chance. If you conduct this experiment, you’ll see that real videos from actual video cameras wipe the floor.
To be fair: have you run such an experiment yourself, or are you just assuming that this conclusion will always result?
(But note this is only replicating video, not reality. Truly realistic motion blur requires ultra-high displayed frame rates beyond the capabilities of current hardware.)
If you're driving at 200 mph, and there's a car next to you also going 200 mph, it shouldn't be blurry.
Also, the length of the blur should not exceed the distance an object travels on your screen in 1 frame. In other words, if an object moves 30 pixels from one frame to the next, then the blurred image shouldn't be more than 30 pixels wide.
What CG videos are you considering, what specifically have you looked at? Can you show some good faith examples of the best CG forests ever made, compared to some specific nature videos? Are you talking about attempts to match a nature video, and saying it's not possible regardless of what's in the shot?
Are you looking at the best examples of CG forests lately? There are some CG full frame video examples of forests I don't believe people would reliably identify as CG, if they didn't know before hand and you left out the explosions & spaceships.
The problem is that we can't do that in 1/60 second on consumer-priced hardware, and also both scanning real objects and manually modelling are expensive.
"The Dress" illustrates how easy it is to fool people with still images. Movement gives crucial context. Our visual system has evolved for millions of years specifically to exploit that context.
That were capable of getting pretty close isn't that surprising, because most photos have already had N layers of digital effects applied, moving them closer to renders, rather than the other way around.
The point is, currently it only works for simple scenes and needs tons of manual work otherwise. Reality is realistic effortlessly.
If you think the answer is obvious, then CGI definitely has more work to do. If not ... ??
If this is CGI, then I'm impressed and want to see more from where this came from.
Agreed with you that I'd be super impressed if anyone could render at that quality, but I haven't seen it yet.
But in the full version[0], it's clearly a photo. But it has been HDR'd which can sometimes create shading that gives a somewhat rendered look.
(Although, maybe you could argue that if the challenge is to make photorealistic renders rather than realistic renders, nailing the camera artifacts is part of the challenge.)
> we do not understand how to make graphics look real > none of those equations work
Why do you claim this, and what do you mean exactly? Surely you aren’t claiming we can’t make any graphics look real at all. Some people are getting pretty good at understanding how to make some graphics look real, even if we aren’t. The math for approximating materials and lighting is getting pretty good these days, and the equations have been changing over time. It’ll continue to improve, but it’s certainly not the main thing holding CG back today. IMO less realistic CG comes much more from low resolution or low fidelity sampling / geometry / textures / animation / input data in general. Most of that is due to lack of time & money.
I’d say we understand how to make graphics look real in many cases, but not in all cases. We understand how to make good looking graphics, but how to do it quickly and efficiently enough to be practical and done with limited time and effort is an open problem.
Making graphics that look real is almost equivalent to the Turing Test, I think plenty of people are willing to acknowledge that it's unsolved.
Amusingly relevant yet slightly off-topic: https://existentialcomics.com/comic/357
When people are saying something is photo-realistic, they aren’t comparing to reality, they’re comparing to photos or videos as viewed on the same display.
By some metrics, even extremely expensive modern hardware is very far from reality. Pretty much all games show Sun occasionally, to reproduce same luminosity at 1m distance you need kilowatts of light (assuming 180 degree viewing angle; the surface of 1m half sphere is about 2 m^2 and the flux of visible spectrum is about 550w/m^2), these levels are simply unsafe for home use. For example, such display is going to ignite or at least melt a keyboard if it’s too close.
Similar for dynamic range. Reality has extreme dynamic ranges, i.e. both very light and very dark areas on the same scene.
At least modern hardware generally delivers realistic resolutions, and color accuracy.
I don't know, look at some flowers... most screens can't show their colors. Stuff like the intense, super-saturated reds and purples are basically impossible to get right in sRGB, with very obvious artefacts, yet in real life there are no artefacts, there is texture, detail and color, where the picture only has a smear of four different reds. P-3 and Rec.2020 might reduce the issues there, as would 10 bit color.
I agree, but professional monitors are close to Adobe RGB, and have been for decades. Adobe RGB is close to DCI-P3, and not much worse than BT.2020.
P.S. Doesn’t help with red/purple though, Adobe RGB extends sRGB mostly in green direction.
My naive questions.
I thought Rec.2020 was already insanely good? I am reading this as it meant Rec.2020 is good, but far from perfect ?
Then down the road came 3dfx with Voodoo and that to me was the next great leap forward. Each iteration has been leading to ray tracing which is the next great leap.
Now just for screen tech to become as affordable as the cards that can drive them, the LG OLED we have is stunning but that is "just" 4K.
just for fun, the story of 3dfx voodoo https://fabiensanglard.net/3dfx_sst1/index.html
Imagination. Same as the people who grew up with the original Atari and Sinclair Spectrum and Commodore 64.
Sadly, now when I look at a DVD, it's unbearably blurry.
I suppose part of that is that we're now viewing on a much larger/higher resolution screen.
I had opportunity to test Varjo VR-2 Pro ($5,995.00) https://store.varjo.com/varjo-vr-2-pro and now all consumer VR products feel like total crap.
I remember playing Final Fantasy 9 entirely on an emulator and how much better it looked compared to a "Real" PlayStation.
I remember as a kid my younger sister watching cartoons and realising she couldn’t tell the difference between them and live action.
I think as graphics get more complex our ability to distinguish increases. But we’ll probably hit a limit in our ability to keep up sooner or later.