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

[zed]

VRS is a tradeoff, faster performance vs usually looking worse, yes in most cases the tradeoff is worth it as better performance is generally better for the end user.
heres MS's whitepaper about it
https://microsoft.github.io/DirectX-Specs/d3d/VariableRateShading.html

I will leave this here from the Program Manager on the Graphics Team at Microsoft. Speaking about tiers 2 VRS on Gears

Tier 2 VRS allows for a free boost in performance with minimal visual impact.

In saying that Im not sure if what Globaliser showed is VRS artifacts in the screenshots, though Doom eternal seemed to display VRS artifacts

 

[Lurkmass]

Do the Gears games ship FP16 optimizations for their shaders ? I'm pretty sure it can be applied for at least for SSAO and several post-processing shaders so they can potentially see some savings over there ...

 

[Karamazov]

yes VRS worsens IQ in some protions of the image, but noticed in still images and zoomed in, in fast spaced games like doom it's unnoticeable, what' noticeable is the perf boost though.
you can have the same perf boost with DRS, but the IQ of the whole image is impacted.

 

[snc]

PS5 and Xbox Series X both render at a native resolution of 2560x1440. PS5 and Xbox Series X render the UI at 3840x2160. Xbox Series S renders at a native resolution of 1600x900 and uses temporal upsampling to reconstruct a 1920x1080 resolution. PS5 and Xbox Series X have some graphical improvements over the Xbox Series S such as higher quality shadows and an improved draw distance.

Stats: https://bit.ly/2UDt0DA

 

[Globalisateur]

yes VRS worsens IQ in some protions of the image, but noticed in still images and zoomed in, in fast spaced games like doom it's unnoticeable, what' noticeable is the perf boost though.
you can have the same perf boost with DRS, but the IQ of the whole image is impacted.

What perf boost in the 60hz modes? Both games mostly locked and PS5 mostly with better IQ (cause res is the same in many cases apparently). You are telling me the Xbox can't sustain 60fps in RT and 4K mode without VRS (contrary to PS5)?

How the DRS would be noticed in fast paced game with so much post AA? Same thing than VRS. The only moment we would notice the VRS is, like DRS, in quiet / still moments, well it's in those moment both games run at max resolution.

You all need to stay consistent with your argumentation. As shown by PS5 in the 60hz mode, this game doesn't need VRS (in those modes) as PS5 overall gives better results than VRS powered Xbox. The 120hz mode could be another discussion as Xbox gives overall better results than PS5 (but does Xbox really need VRS to output a 20% higher res than PS5 in the 120hz mode? very odd).

 

[zed]

yes VRS worsens IQ in some protions of the image, but noticed in still images and zoomed in, in fast spaced games like doom it's unnoticeable, what' noticeable is the perf boost though.

Yes its why games do dynamic scaling, usually the user is not going to notice the lesser resolution but they will notice the drop in FPS.
Similar to VRS, One thing that will massively increase quality/performance is https://en.wikipedia.org/wiki/Foveated_rendering
though to see this on consoles you'll need some eye tracking cam on the TV (in VR its possible I expect) once they do this, its a generation leap, it will turn a PS5 -> PS6, xb series X -> xb series Y

 

[Clukos]

(but does Xbox really need VRS to output a 20% higher res than PS5 in the 120hz mode? very odd).

You are missing the point of VRS entirely. You use VRS to lower the shading rate in parts of your image that will be unnoticeable during general play so you can retain detail on parts of the image where such drop in detail would be noticeable.

You can think of it as DRS but for different parts of the frame. Whether DOOM makes good use of it is up for debate but the feature itself can be very useful when used right.

 

[Karamazov]

What perf boost in the 60hz modes? Both games mostly locked and PS5 mostly with better IQ (cause res is the same in many cases apparently). You are telling me the Xbox can't sustain 60fps in RT and 4K mode without VRS (contrary to PS5)?

How the DRS would be noticed in fast paced game with so much post AA? Same thing than VRS. The only moment we would notice the VRS is, like DRS, in quiet / still moments, well it's in those moment both games run at max resolution.

You all need to stay consistent with your argumentation. As shown by PS5 in the 60hz mode, this game doesn't need VRS (in those modes) as PS5 overall gives better results than VRS powered Xbox. The 120hz mode could be another discussion as Xbox gives overall better results than PS5 (but does Xbox really need VRS to output a 20% higher res than PS5 in the 120hz mode? very odd).

but ps5 has to drop lower without VRS.

You would tend to notice DRS more in action as it lowers the whole image resolution instead of some part of it.

 

[snc]

but ps5 has to drop lower without VRS.

You would tend to notice DRS more in action as it lowers the whole image resolution instead of some part of it.

But this part is definitly more impacted than just drs with lower range

 

[mr magoo]

But this part is definitly more impacted than just drs with lower range

Yeah this part only vs whole screen. And part which is affected is selected on the region where it makes most sense and theoretically should affect picture least. This allows drs to hit higher bounds and triggers less often. I think the id dev summarized it perfectly it increase overall picture by allowing to render at higher resolution more often.

 

[Proelite]

So with X holding around 30% res advantage in the 120hz mode, VRS accounting for 10-15% of that, X would have a 12-17% advantage in 120hz mode without VRS.

For 60hz mode, according to VG Tech stress points have X holding a 20% advantage, which would be 5-9% advantage without VRS.

The difference in the gap is the result of variable clocks on PS5, no doubt about that.

 

[snc]

So with X holding around 30% res advantage in the 120hz mode, VRS accounting for 10-15% of that, X would have a 12-17% advantage in 120hz mode without VRS.

For 60hz mode, according to VG Tech stress points have X holding a 20% advantage, which would be 5-9% advantage without VRS.

The difference in the gap is the result of variable clocks on PS5, no doubt about that.

nah rather bandwidth difference is more impactful in 120fps mode

 

[PSman1700]

So with X holding around 30% res advantage in the 120hz mode, VRS accounting for 10-15% of that, X would have a 12-17% advantage in 120hz mode without VRS.

For 60hz mode, according to VG Tech stress points have X holding a 20% advantage, which would be 5-9% advantage without VRS.

The difference in the gap is the result of variable clocks on PS5, no doubt about that.

XSX is having that 18% or whatever it is more capable GPU, lines up quite well there. Im sure the variable clocks do play a role whenever they kick in though, i guess that would be more apparant in later generation games when devs start pushing those boxes more.

 

[Globalisateur]

Someone finally did compare VRS ON / OFF on Gears 5. VRS ON 116fps, VRS OFF 113fps in one scene (average 9%). The sharpness difference is very noticeable (just compare his raw images, this is very easy to see), the guy compares VRS image (in performance mode to have a notable fps increase) like a PS1 game, similar to Doom in his experience.

So with X holding around 30% res advantage in the 120hz mode, VRS accounting for 10-15% of that, X would have a 12-17% advantage in 120hz mode without VRS.

For 60hz mode, according to VG Tech stress points have X holding a 20% advantage, which would be 5-9% advantage without VRS.

The difference in the gap is the result of variable clocks on PS5, no doubt about that.

Sure but it won't look 30% higher resolution anymore with VRS blur applied on the image. In the identical images I saw (including one from DF) the game does look sharper on PS5 in the 60fps modes! We are talking about perceptual resolution here, something @iroboto should be able to test with his tools.

Based on how the dynamic clocks works In most gameplay moments the clocks would stay at max. Just look at the power consumption during gaming. As long as the PS5 is running <200W, the clocks would be running at their max because this is how the system has being designed. AFAIK the only moments we know PS5 power consumption is sustained at 200W is during cutscenes (or similar scenes like photo-mode). Just show me a game where the power consumption is hovering at 200W during gameplay then we'll talk about the impact of dynamic clocks in specific scenes.

 

[PSman1700]

Based on how the dynamic clocks works In most gameplay moments the clocks would stay at max. Just look at the power consumption during gaming. As long as the PS5 is running <200W, the clocks would be running at their max because this is how the system has being designed. AFAIK the only moments we know PS5 power consumption is sustained at 200W is during cutscenes (or similar scenes like photo-mode). Just show me a game where the power consumption is hovering at 200W during gameplay then we'll talk about the impact of dynamic clocks in specific scenes.

He's talking about the variable clocks, you cant see on the power usage how much the CPU is trading in for the GPU or vice versa. Unless the PS5 is clocking down both at the same time.

 

[higherARC]

What perf boost in the 60hz modes? Both games mostly locked and PS5 mostly with better IQ (cause res is the same in many cases apparently). You are telling me the Xbox can't sustain 60fps in RT and 4K mode without VRS (contrary to PS5)?

How the DRS would be noticed in fast paced game with so much post AA? Same thing than VRS. The only moment we would notice the VRS is, like DRS, in quiet / still moments, well it's in those moment both games run at max resolution.

You all need to stay consistent with your argumentation. As shown by PS5 in the 60hz mode, this game doesn't need VRS (in those modes) as PS5 overall gives better results than VRS powered Xbox. The 120hz mode could be another discussion as Xbox gives overall better results than PS5 (but does Xbox really need VRS to output a 20% higher res than PS5 in the 120hz mode? very odd).

res in all modes aren’t always locked to top only XSX manages to hold the cap res while ps5 drops per DF and vgtech. VGtech took random pics and in balanced mode shot, which maxes out at 4K, he took showed a 23% advantage for XSX. XSX holds sharper image in all modes

Balanced
PS5 - 6.71 million pixels
XSX - 8.29 million pixels
23% difference

vrs set to highest will cause some blur but you’re getting a lot more FPS it’s like DLSS. Best is balanced which gets you about 15% and barely image quality loss. Devs are just using it too they’ll learn to get more out of it over time like DLSS. They can even do mixture of vrs with other effects like DoF to get more performance whenever DoF is used.

here’s what gains you get in Gears 5 using VRS

 

[see colon]

Based on how the dynamic clocks works In most gameplay moments the clocks would stay at max. Just look at the power consumption during gaming. As long as the PS5 is running <200W, the clocks would be running at their max because this is how the system has being designed. AFAIK the only moments we know PS5 power consumption is sustained at 200W is during cutscenes (or similar scenes like photo-mode). Just show me a game where the power consumption is hovering at 200W during gameplay then we'll talk about the impact of dynamic clocks in specific scenes.

Are you sure about this? Power consumption can change based on workload independent of clock speed. PS5 was designed to maintain maximum clock speed within it's power budget. So it should be often hitting it's power budget of 200W unless it's maximum clocks can be achieved under that limit based on the workload.

 

[zed]

Someone finally did compare VRS ON / OFF on Gears 5. VRS ON 116fps, VRS OFF 113fps in one scene (average 9%). The sharpness difference is very noticeable

hmmm I looked at the images and the differences are very large, VRS looks a lot blurrier, this was the poster child title wasnt it?
I would say the screenshots are fake cause the downgrade is huge Performance VRS vs Off

Balanced or Quality VRS vs Off is much better there is some downgrade but much less

Im not convinced these are real (maybe the poster is making VRS look bad on purpose) cause the differences are so big, we need other ppl also posting raw screenshots for comparisons, Surely other ppl on these forums have the game (I dont, plus Im not even sure if my GPU supports it)

 

[see colon]

Don't pay too much attention to the performance stats in the corner. After changing the settings the game took a few seconds to normalize. "Performance" mode definitely takes a hit in image quality, but "quality" I couldn't tell the difference. "Balanced"... Only if I was really looking.

Screenshots were taken at 1080p with all settings at maximum.

 

[zed]

Thanks see colon nice one, thats a very good area to take a screenshot from.

but is there any way to have the camera in the same spot.
not sure how this game works,
can you pause, take screenshot, change VRS take another screenshot etc.
perhaps after a save game the camera is in the same position?
or record a gameplay video, change the settings, replay the video, then take screenshot.

Not sure what the best method is as I dont have the game
As all we can conclude from the screenshots is performance VRS looks terrible and should not be turned on, you're prolly better off playing at a lower resolution