0
$\begingroup$

I have to admit that, while watching footage of the enhanced 2020 version of the 2018 game Forza Horizon 4, when they are driving fast, I have to sometimes remind myself that I am watching a video game and not a video recording of reality. It's the closest I've ever seen to photo realism, but as soon as the car stops, or they go too slowly, or they do something slightly not expected, the illusion quickly falls apart.

It's just a bunch of smoke and mirrors and dead polygon objects after all. Not an actual photo-realistic 3D version of England where I can "drive anywhere and do anything". But for small moments, it can fool me. Example gameplay footage: https://youtu.be/PqwQDRersIM?t=42

What exactly is the "missing piece" at this point? It's impossible for me to not be impressed by the game, but I still don't want to play it. I can tell that I would very quickly get bored of the illusion, and just grow ever more frustrated by the limitations of the reality as presented in the game.

For me, it's largely knowing that it's all dead 3D objects. There are no detailed rooms inside all of those buildings which I could enter and go through. There are no people either, or, if there are, they don't look or behave anything like real humans would. There are many visual details which just look "off", and the more photo-realistic it looks overall, the more demanding my brain becomes and the more eager I am to point out any flaw.

If you then take a look at this random actual video, even that game looks like garbage in comparison: https://youtu.be/FMOb5-e3J7Y?t=49 (yes, I know that the settings/locations are quite different, but that doesn't really matter).

Something about reality is so much "richer" and "real" than what they can do with a 3D world, but I don't quite understand what exactly. I feel as if I should be more impressed than I actually am by that car game. It looks objectively amazing, yet at the same time, I have very little interest in actually playing it. And keep in mind that I'm a person who has been hunting for photo-realistic 3D simulations my entire life and who is endlessly fascinated by the concept.

One thing I can say for sure: it has almost nothing to do with the resolution of the game/camera. Even if you take a very old and cheap camera and go filming reality, it will look a million times better than the 8K game.

Can you explain to me why we don't have a simulation that looks and behaves just like the actual video I just linked to?

Improve this question
$\endgroup$
3
  • 2
    $\begingroup$ Because we don't have the compute power to simulate an entire city worth of molecules and their physical interactions real time. Until then we will have to make do with approximations, that while good, are still approximations. $\endgroup$
    – AnnoyinC
    Nov 29, 2020 at 14:03
  • 2
    $\begingroup$ Just to nitpick a bit, a whole lot of what you see around you is 3D simulation that works. There is a lot more CGI thats entirely invisible for people on TV and movies in scenes that you dont expect. Yes sometimes they fail in the harder shots but still you dont notice. Did you know that most stuff you see in a Ikea catalog is a render? Well the answer is we arent missing anything as such just ant do it fast enough for realtime renders. $\endgroup$
    – joojaa
    Nov 29, 2020 at 15:35
  • $\begingroup$ Fine-scale geometry, lighting, and texture details are a big part of it. For instance, in the thumbnail frame of the Forza video, what jumps out to me as unrealistic is the comparatively low-res textures and geometry of the building and sidewalk at left. You can see that they are flat images stretched across boxes. The car is a lot more realistic but you can still see a lack of material detail around the bumper, tailpipes, and tires. $\endgroup$
    – Nathan Reed
    Nov 29, 2020 at 23:41

2 Answers 2

Reset to default
2
$\begingroup$

As @AnnoyinC already mentioned in his comment, we are simply lacking computational power and storage requirements (RAM).

For the visual effects, you need to calculate every ray of light that somehow ends up in the observer's eye after an arbitrary number of reflections on different surfaces. The reflections depend on the surface's microstructure. This alone is so extremely demanding that is far from possible to calculate it in real-time with current consumer hardware. The first steps into this direction are already taken with ray tracing support of modern graphic cards but it is still very limited. And that's just the "easy part".

It gets way more complicated if you want to have objects and fluids that behave physically correct. While each ray of light is more or less independent of other rays, physical objects influence each other. Also, simulating the correct reaction of an arbitrary object to an external force can already take days and weeks for a single timestep/frame since you need to solve extremely large equation systems that are often non-linear. If you want to know why I would suggest searching for a "Finite-Element-Method" or "Computational Fluid Dynamics" tutorial. These techniques are commonly used by engineers and separate a structure into a mesh of smaller interacting substructures. The results get closer to the behavior of the real structure/fluid the denser the mesh. However, doubling the mesh density in each of the 3 spatial directions comes at the price of needing 8 times more computational power and RAM. Now think of a tree waving in the wind with thousands of leaves where each leaf alone is a complex physical object... I think you get the picture.

To get around all these problems, a lot of approximations are used. While most of them are quite good illusions they will give wrong results at some point. However, that doesn't necessarily mean that we realize it.

Improve this answer
$\endgroup$
0
$\begingroup$

One of the thing that breaks the illusion in that video for me is the cinematography. During the fast moments, the camera is more physically plausible, in that it could have been filmed from a camera mounted to a vehicle following the hero car. When stopped, the player moves the third-person camera around in a way that doesn't mimic physical camera rigs.

It's also worth noting that motion blur is blur, and the loss of resolution caused by high speed motion covers up a lot of sins.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.