Current Generation Games Analysis Technical Discussion [2020-2021] [XBSX|S, PS5, PC]

Status
Not open for further replies.
;)
The reflection we're seeing in control, but the mirror is on the wrong side.
spiderman.jpg

There is so much wrong with that reflection, too. Don't let me start ^^, but it looks cool.

We really shouldn't forget that those are just games and they used to have reflections because they look cool and not because they are implemented correct. There are so many perfect mirrors (even puddles) in those games. Funny thing is (IMHO), SSR reflections looks most times more realistic than the RT reflections in games.
 
There is so much wrong with that reflection, too. Don't let me start ^^, but it looks cool.

We really shouldn't forget that those are just games and they used to have reflections because they look cool and not because they are implemented correct. There are so many perfect mirrors (even puddles) in those games. Funny thing is (IMHO), SSR reflections looks most times more realistic than the RT reflections in games.
I'm more concerned with the plane of reflection than correctness.

But, I'll need to see Dictator's photos uncropped, but I honestly don't see how that could reflect onto the right side with the mirror being in front. We reflect into mirrors. It would help to see the uncropped photos. Looking over control videos, everywhere else looks correct honestly. It doesn't look flipped or anything.

But without a good feeling for where the camera is, I can't actually say/know if those angles are correct.
 
There is so much wrong with that reflection, too. Don't let me start ^^, but it looks cool.

We really shouldn't forget that those are just games and they used to have reflections because they look cool and not because they are implemented correct. There are so many perfect mirrors (even puddles) in those games. Funny thing is (IMHO), SSR reflections looks most times more realistic than the RT reflections in games.

A good combination of RT and SSR strikes a good balance between performance and IQ. Heck, CP2077 uses SSR and RT heavily.

Edit: Brainfart. While CP2077 uses SSR as well, however, SSS only has a psycho mode.
 
Last edited:
I can't replicated PS5s reflections either, they look like incorrect angles as PC, but shifted to the left.
This is what I think is happening.
Remedy gotta resolve. I think the PS5 one is same, but shifted left and way deep


I think the reflection could be correct, but when they rendered out to the texture, they flipped it on the vertical axis or something. Whatever it is, Remedy should address.

Why should Remedy address it? Is the game unplayable? Is it less fun? Will they make more money by fixing it? How tarnished will Remedy's reputation be by not resolving it?

I would dare to say, nope nada, let it be and move on. Especially when we compare it to the CP2077 buggy release, this is just minor nitpicking which photo mode was needed to find?
 
Why should Remedy address it? Is the game unplayable? Is it less fun? Will they make more money by fixing it? How tarnished will Remedy's reputation be by not resolving it?

I would dare to say, nope nada, let it be and move on. Especially when we compare it to the CP2077 buggy release, this is just minor nitpicking which photo mode was needed to find?
You're right, nothing changes honestly. They don't have to, that was a poor choice of words. But it's worthwhile to resolve for their next title I guess. It should be brought up to them though.
 
Last post for me, and I'll edit this one with new graphs. Just finished out of the shower from working out and it came to me. I should have paid more attention when @Dictator said that this is not a path tracer.

Ray tracing single bounce means your camera projects onto any surface and it traces itself back to the light. Anything in it's path to the light will show up as a reflection on the surface. So this is why we're seeing the wrong plane of reflection, because the light source is back and behind the camera and to the left. So the rays on the right side of Jessie are tracing back to the light source on the left and back of the camera giving you the wrong plane of reflection for real life.

this problem cannot be avoided unless they path trace light to the camera and then back.

TLDR; I think the Ray Tracing is correct on PC & XSX. Shifted incorrectly on PS5 (ie, the lightning bolt is not the same place as PC and XSX, but is working fine). But Ray tracing does not mimic real life physics. So there are incorrect angles etc happening that people will notice.

ivXzszS.png
 
Last edited:
Ray tracing single bounce means your camera projects onto any surface and it traces itself back to the light. Anything in it's path to the light will show up as a reflection on the surface. So this is why we're seeing the wrong plane of reflection, because the light source is back and behind the camera and to the left. So the rays on the right side of Jessie are tracing back to the light source on the left and back of the camera giving you the wrong plane of reflection for real life.
[/QUOTE]

I don't think this is right -- I believe you're mixing some ideas together. maybe somebody who knows more rtx details could correct, but my impression of rtx raytracing is that it's fairly normal raytracing.

General (monte carlo) raytracing refresher: Rays are projected from each pixel on the "camera" and bounce randomly in a hemisphere away from the surface normal they hit. The degree of randomness (size of the hemispherical lobe) varies depending on how reflective the surface is, up to completely random for diffuse surfaces and not random at all (straight reflective bounce along the normal) for mirror surfaces. With those random distributions, many raytracers (i assume all the ones we see running realtime) will bias that randomness towards the light. Due to this biasing, it will take far fewer samples to converge on a realistic lighting situation, at the cost of maybe less precise GI. There's no random component to mirror reflections, so they aren't biased at all.

One common step in raytracers is casting shadow rays which behave like the rays in your picture -- they bounce directly towards the light. By doing this, they can check whether they "hit" a light to determine whether the surface they bounce from is in shadow or not, and then this can be added to the shading.

However, under no circumstances should reflective surfaces bounce towards light. That would give you a wildly inaccurate render with no gains to performance. Mirror reflections should be straight reflections: I - 2 * dot(I, N) * N -- there are some illustrations of this concept here: https://www.scratchapixel.com/lesso...tion-to-shading/reflection-refraction-fresnel
 
I don't think this is right -- I believe you're mixing some ideas together. maybe somebody who knows more rtx details could correct, but my impression of rtx raytracing is that it's fairly normal raytracing.

General (monte carlo) raytracing refresher: Rays are projected from each pixel on the "camera" and bounce randomly in a hemisphere away from the surface normal they hit. The degree of randomness (size of the hemispherical lobe) varies depending on how reflective the surface is, up to completely random for diffuse surfaces and not random at all (straight reflective bounce along the normal) for mirror surfaces. With those random distributions, many raytracers (i assume all the ones we see running realtime) will bias that randomness towards the light. Due to this biasing, it will take far fewer samples to converge on a realistic lighting situation, at the cost of maybe less precise GI. RTX raytracing is broken up into several kinda different calculations and passes, and I'm not sure what they are or how they all work.

One common step in raytracers is casting shadow rays which behave like the rays in your picture -- they bounce directly towards the light. By doing this, they can check whether they "hit" a light to determine whether the surface they bounce from is in shadow or not, and then this can be added to the shading.

However, under no circumstances should reflective surfaces bounce towards light. That would give you a wildly inaccurate render with no gains to performance. Mirror reflections should be straight reflections: I - 2 * dot(I, N) * N -- there are some illustrations of this concept here: https://www.scratchapixel.com/lesso...tion-to-shading/reflection-refraction-fresnel

hmm you might be right, I might be mixing some terms; I thought that's why Path Tracers are invented. Because with Ray Tracing you have to run the path from camera to the light, and you ignore everything on the way to the reflective surface, and any intersections on the way to the light source you collide with and stop.

With Path Tracing once you reach the light, you need to go back in reverse. It would tell you which rays are technically allowed to reflect and which ones can't. The reason we don't see Jessie reflected on the left side is because their Ray Tracer is passing through Jessie, hitting the glass and heading towards the light. On the reverse path tracing, it would hit the mirror and hit jessie. You'd see the correct reflection there.

but my google fu is showing me that path tracers do not do that.

hmm.. so perhaps it's just a bug entirely in their RT algo.
 
Last edited:
I think that's why Path Tracers are invented. Because with Ray Tracing you have to run the path from camera to the light, and you ignore everything on the way to the reflective surface, and any intersections on the way to the light source you collide with and stop.

With Path Tracing once you reach the light, you need to go back in reverse. It would tell you which rays are technically allowed to reflect and which ones can't. The reason we don't see Jessie reflected on the left side is because their Ray Tracer is passing through Jessie, hitting the glass and heading towards the light. On the reverse path tracing, it would hit the mirror and hit jessie. You'd see the correct reflection there.

In the context I'm speaking, path tracing is basically an interchangeable term to raytracing, at least in terms of the paths the rays take. I believe the term was coined around the first monte carlo raytracers. I'm not sure why the terms are used so divergently in different contexts.

Sorry for any terminology confusion -- I don't think you're correct about the routes the rays take, though.

It's hard for me to explain, but it would look extremely different than what we're seeing here -- there would be severe distortions in every surface.

Around somewhere there ought to be implementation examples for DXR/RTX reflections that we could look at to see for sure how it works in general, but I'm at work and shouldn't really be taking this time for posting, let alone digging around code examples :p
 
In the context I'm speaking, path tracing is basically an interchangeable term to raytracing, at least in terms of the paths the rays take. I believe the term was coined around the first monte carlo raytracers. I'm not sure why the terms are used so divergently in different contexts. Sorry for any terminology confusion -- I don't think you're correct about the routes the rays take, though. It's hard for me to explain, but it would look extremely different than what we're seeing here -- there would be severe distortions in every surface.
Agreed. Need to call in the senior members here to weigh in on what's happening
 
I don't know, anything is possible. Are there any other third-party titles on XBSX that exhibit similar stuttering issues even with similar installation sizes? If the answer is yes, then it's something else that's hardware related.
I'm sure we can find some lol. I guess the question just comes down to why and whether the developers can... remedy the situation for XSX.
(•_•)
( •_•)>⌐■-■
(⌐■_■)
 
Last edited:
Regarding I/O discussions and control - I think it is good to look at the last gen on disc size across the last gen systems before assuming that this Ultimate edition is doing something radically different or using PS5 specific compression.
Listing on Store for base Xbox One (not ultimate edition): 34.77GB
Listing on PSStore for base PS4 (not ultimate edition): 24.4 GB
 
Regarding I/O discussions and control - I think it is good to look at the last gen on disc size across the last gen systems before assuming that this Ultimate edition is doing something radically different or using PS5 specific compression. Listing on Store for base Xbox One (not ultimate edition): 34.77GB. Listing on PSStore for base PS4 (not ultimate edition): 24.4 GB

Somebody, and I can't remember who, explained this disparity was because the Xbox versions included all versions of the game.
edit: I misread that, the post I referred was referring to the difference between the PS5/XBS sizes version sizes.
 
Last edited by a moderator:
Somebody, and I can't remember who, explained this disparity was because the Xbox versions included all versions of the game.
That wouldn't explain the difference. PS4 version has about the same size as XB1 and XSX. I think they use oodle textures on PS5 and maybe deduplication of data (without Kraken compression which is possible as oodle textures are possible on base PS4).
 
Status
Not open for further replies.
Back
Top