Support NeoGAF

[bender]

yeah, but let's not act like this wasn't almost the norm back in Gen7.
drops to 20fps, full screen tearing, sub HD... these were expected, especially on PS3 but also on 360.

I didn't say that Remedy weren't the only ones doing this and like SlimySnake pointed out, they've continued to do this, often times much worse, in the generations of hardware that followed.

Your other point is fair.

 

[SlimySnake]

I just have zero interest in studios that want to use this additional horse power to just throw more effects at a 30fps mode. I am so completely done with 30fps as a concept, regardless of how shiny it might look. Exact same as Blizzard and their D4 enhancements, it's literally just a bump in fidelity for the existing 30fps mode which is garbage to play on.

Once again this just stands to elevate those studios who have been consumer focused with their PS5 Pro enhancements, Square Enix, Insomniac, Naughty Dog, Ivory Tower.

At first I thought that The Crew: Motorfest was not ideal, by only offering 1 mode on Pro but now I'm leaning towards that being the most consumer friendly option. So kudos to them.

Remedy can fuck off with this poor effort. It does not represent the capabilities of this console at all.

It's not Remedy's fault that you are getting RT only in 30 fps mode. It's sony's. The console is simply not powerful enough to run RT in this game at 60 fps. Lack of vram bandwidth, vram size, CPU upgrade all contribute to this. Maybe its powerful enough to run RT In less graphically demanding games like Spiderman 2 but there is a reason ND, SSM, GG and BP did not add RT in the Pro version for their cross gen games. The console is just not able to do both next gen visuals AND RT at 60 fps.

Remedy are just dumb idiots who decided to add more effects in the 60 fps mode instead of simply upping the resolution so that PSSR can finally shine. It's pro-consumer to ship the game with two modes. It's pro-consumer to give them better graphics settings on a brand new $700 mid gen refresh. It's just a dumb move to do it on a console clearly not capable enough.

Cannot compare ND who made a PS4 looking game with PS4 assets and PS4 quality lightings and released it as a next gen game. AW2 is pushing way better quality lighting, assets and visual effects. And that comes at a cost. if you want to continue playing PS4 looking games then dont even bother getting a pro because HFW, GOW Ragnorak, and TLOU1 already look and run at 60 fps at pretty decent resolutions on the base PS5.

 

[kevboard]

I didn't say that Remedy weren't the only ones doing this and like SlimySnake pointed out, they've continued to do this, often times much worse, in the generations of hardware that followed.

Your other point is fair.

yes but the 360 and PS3 gen specifically can not be used as an example here, especially when Alan Wake was one of the better games in that regard. (clean image quality, real time lighting tailored specifically for the GPU, ok-ish performance)

I edited my previous post, but I'll just include his here,



if a console exclusive that was designed from the ground up for this hardware runs like this... I think most games get a pass.
almost all games on that gen were way too much for the harsware, as hitting HD resolutions was just too much for those GPUs if we're honest.

 

[James Sawyer Ford]

I can tell you don't hang around AW2 or Control discussions on GAF if that's all you have to say.

There's a reason why AW2 won 63 GoTY awards last year and was third only behind the juggernauts BG3 and Tears of the Kingdom.

No reason to hang around when I personally don’t like the games that much

I think AW2 last year was probably a huge overrate. A lot of complaints came up about lacking gameplay and woke elements that probably fueled the reviews

 

[twilo99]

 

[SlimySnake]

As expected usual suspects are at near light speed to jump to definitive conclusions about overall capabilities of PS5 PRO based on this single, absolutely worse enhancement yet. Try to be a bit more subtle about your disingenuousness, we know you aren't interested in the objective truth/results in the slightest.

?? This is literally in line with what we've seen from other games. The base resolutions remains the same, pss4 is replaced by fsr, and you get some added RT effects. if anything they have upped settings in the performance mode which is not what other devs are doing. I am playing spiderman 2 and ratchet right now and insomniac literally greys out the new RT settings in the performance mode. its the same performance mode as the base ps5, just using pssr instead of their TAAU solution. FF7 rebirth does the same thing. Uses performance mode settings and base resolution instead of the higher quality fidelity mode settings.

At least Remedy used fidelity mode settings in performance mode. Now of course, they shouldnt have done that and followed what sony first party and other third party devs have done and simply used performance mode settings from the base ps5 but you cant say this is one game. this is one of many games.

all the other games with those 30% performance improvements? come on. they all say the same thing. the console just isnt performing like a 16.7 tflops gpu. and the RT which is supposed to be 2-4x faster is simply not performing like they said it would. at least not at 60 fps. sony's own leading RT studio disabled those rt effects in the performance mode.

 

[bender]

yes but the 360 and PS3 gen specifically can not be used as an example here, especially when Alan Wake was one of the better games in that regard. (clean image quality, real time lighting tailored specifically for the GPU, ok-ish performance)

I edited my previous post, but I'll just include his here,



if a console exclusive that was designed from the ground up for this hardware runs like this... I think most games get a pass.
almost all games on that gen were way too much for the harsware, as hitting HD resolutions was just too much for those GPUs if we're honest.

I guess we just have different expectations for what clean image quality refers to and I don't think anyone should get a pass. That's not to say these games are unplayable, but it is fair to call out shortcomings...hell, I've loved Souls games on consoles since DeSo. I'd also say that you can call at any generation for this problem as there are plenty of examples to point to, but that's probably a discussion for another thread.

 

[James Sawyer Ford]

all the other games with those 30% performance improvements?

How is running 60 fps at better fidelity than 30 fps modes a 30% improvement?

 

[kevboard]

I guess we just have different expectations for what clean image quality refers to and I don't think anyone should get a pass.

having 4x MSAA made the game look really clean and stable. it was a bit blurry sure, but the choices seemed very deliberate.
I think they could probably have hit 720p native with fxaa if they wanted, but opted for 544p and 4x MSAA

 

[Hugare]

Base resolution even lower than OG PS5 is wild

This was the game that I was most curious about knowing how it runs on the Pro, and it turned out to be a letdown.

having 4x MSAA made the game look really clean and stable. it was a bit blurry sure, but the choices seemed very deliberate.
I think they could probably have hit 720p native with fxaa if they wanted, but opted for 544p and 4x MSAA

They've made the game in a way where MSAA was mandatory. You cant disable it, its locked at 2x at least even on PC. If you try forcing it off, vegetation becomes all blocky and glitchy.

 

[daninthemix]

Remedy has to do better with there patch. Let us be patient with them. This is not the last patch and update. I m sure by next year will finally see significant increase of FPS and better image quality s with RT for AW 2.

You guys are always having to wait for stuff to get good aren’t you? The endless wait for the Pro, then all the apologetics around PSSR and how it’s sure to improve. What you have now never seems to satisfy.

 

[scydrex]

Well will not play this on Pro until they fix it. Remedy messed up this upgrade.

 

[James Sawyer Ford]

You guys are always having to wait for stuff to get good aren’t you? The endless wait for the Pro, then all the apologetics around PSSR and how it’s sure to improve. What you have now never seems to satisfy.

Hi Dan

You seem to be an expert on waiting. Any suggestions you can give?

 

[powder]

How is running 60 fps at better fidelity than 30 fps modes a 30% improvement?

No ray tracing in 60fps mode

 

[James Sawyer Ford]

No ray tracing in 60fps mode

No ray tracing on base PS5 at all.

 

[Desprado123]

AMD RDNA any series is weaker at higher resolution and more higher resolution more weak Example.

Just look at 2 games, which uses pure bandwidth of GPU and horse power due to older engines.

Stalker at 8K

RTX 3080 Ti beating RTX 4070 Ti Super and equal as RX 7900 XTX


Same happening in RDR 1 that RTX 3080 Ti is beating RX 7900 XTX

Turning and Ampere are superior on higher resolution compare to equivalent RTX 4xxx and RDNA 3.

On higher resolution 90% games of are depended on higher bandwidth.

 

[Neo_game]

But that benchamark is running Max Raster settings on top of Max RT, it actually shows how the Pro seems to be underperforming compared to the 7700XT when the Pro is running some settings on lower than lowest on PC.

We have seen with PS5 and SX as well some games actually are par on RTX 2080 or 2070S but some games barely can match 3060. If you look at the Pro specs 60CU, 16tlflops, 576gbps very similar to RX6800 and 7700XT. It depends on the dev and optimization. There are some videos without any patches Pro runs around 30% more faster than PS5. With optimization it should be close to 45% raster and apprently RT should be much bigger upgrade as well. We need better sample size with different games and AMD cards as refernce to compare.

the 4070 was probably used because nearly everyone ahead of launch compared the Pro to the 4070... now in hindsight, yes, a 4060ti might have been more appropriate.

but if you compare FSR to PSSR you then have the issue that PSSR might actually use more GPU resources than FSR does.

best comparison I guess would have been an AMD GPU using XeSS, because XeSS should have a similar GPU impact to PSSR on the same architecture.
but as far as I know Alan Wake 2 sadly doesn't support XeSS (which is kinda weird tbh)

there will always be some things that don't compare perfectly when comparing to PC.

Not sure they had those AMD cards to test. Someone mentioned 4070 is 40% faster in this game which means Pro which is performing like RTX 2080 ?

 

[Desprado123]

We have seen with PS5 and SX as well some games actually are par on RTX 2080 or 2070S but some games barely can match 3060. If you look at the Pro specs 60CU, 16tlflops, 576gbps very similar to RX6800 and 7700XT. It depends on the dev and optimization. There are some videos without any patches Pro runs around 30% more faster than PS5. With optimization it should be close to 45% raster and apprently RT should be much bigger upgrade as well. We need better sample size with different games and AMD cards as refernce to compare.



Not sure they had those AMD cards to test. Someone mentioned 4070 is 40% faster in this game which means Pro which is performing like RTX 2080 ?

Performing as RTX 3060 Ti with RT on and if Raster than best case RTX 3070/Ti.

 

[kevboard]

They've made the game in a way where MSAA was mandatory. You cant disable it, its locked at 2x at least even on PC. If you try forcing it off, vegetation becomes all blocky and glitchy.

yeah, and modern games on UE5 or Northlight engine need to use TAA or they will look pixelated and filled with dithering artifacts, so you usually can not turn TAA off.

but that's also a deliberate choice by the devs...

I do find it weird tho that turning off, what is essentially simply a method to run geometry at a higher resolution, would result in graphical issues especially graphical issues related to vegetation...
that sounds weird... doesn't really make sense...

what exactly did they do that the game has these issues?

 

[Bojji]

As expected usual suspects are at near light speed to jump to definitive conclusions about overall capabilities of PS5 PRO based on this single, absolutely worse enhancement yet. Try to be a bit more subtle about your disingenuousness, we know you aren't interested in the objective truth/results in the slightest.

Show me some "objective" results then. We have numbers from old games running around ~30% better than PS5, we also have PS5 Pro native app AW2 that has performance measured.

GPU that is the closest to Pro is 3070, if it wasn't vram limited in many new games it would almost be a perfect match:

How is running 60 fps at better fidelity than 30 fps modes a 30% improvement?

How you measure that?

Last of Us Pro mode is 1440p 60 just like base mode but with PSSR
God of War Pro mode use quality settings in 60fps (unlocked) but is 1440p instead of 4k from base ps5 quality mode.
Demon's Souls is 1440p 60 just like PS5 but with PSSR? etc.

We have seen with PS5 and SX as well some games actually are par on RTX 2080 or 2070S but some games barely can match 3060. If you look at the Pro specs 60CU, 16tlflops, 576gbps very similar to RX6800 and 7700XT. It depends on the dev and optimization. There are some videos without any patches Pro runs around 30% more faster than PS5. With optimization it should be close to 45% raster and apprently RT should be much bigger upgrade as well. We need better sample size with different games and AMD cards as refernce to compare.



Not sure they had those AMD cards to test. Someone mentioned 4070 is 40% faster in this game which means Pro which is performing like RTX 2080 ?

Optimizations for Pro would also make base PS5 version perform better, it's still the same architecture. So far delta seems around that 30% and if games will start to show more than that we will know that something was wrong at launch.

 

[James Sawyer Ford]

Last of Us Pro mode is 1440p 60 just like base mode but with PSSR
God of War Pro mode use quality settings in 60fps (unlocked) but is 1440p instead of 4k from base ps5 quality mode.
Demon's Souls is 1440p 60 just like PS5 but with PSSR? etc.

The PSSR is putting in the work to enable this. 1440p without PSSR is not the same as 1440p with PSSR

I thought you had a gaming pc with DLSS?

 

[Bojji]

The PSSR is putting in the work to enable this. 1440p without PSSR is not the same as 1440p with PSSR

I thought you had a gaming pc with DLSS?

This is subjective. Can PSSR/DLSS look close to native 4k? Yeah, why not. In some games it does.

Is 1080p/1440p game with PSSR/DLSS as heavy to render as native 4k? NO.

When measuring console performance uplift subjective things don't really help in any way. So far average uplift for Pro looks to be around that 30%.

Only game I know that is 100% better on Pro is empire of ants, and that's purely developers decisions to fuck up millions of players. Both PS5 and XSX could run this game in 60fps.

 

[James Sawyer Ford]

This is subjective. Can PSSR/DLSS look close to native 4k? Yeah, why not. In some games it does.

Is 1080p/1440p game with PSSR/DLSS as heavy to render as native 4k? NO.

When measuring console performance uplift subjective things don't really help in any way. So far average uplift for Pro looks to be around that 30%.

Only game I know that is 100% better on Pro is empire of ants, and that's purely developers decisions to fuck up millions of players. Both PS5 and XSX could run this game in 60fps.

Huh?

Sorry, no, and PSSR can look better than native 4k - it already does in many titles

This hand waiving away of the performance leap just because it’s not in raster is disingenuous AF

 

[Bojji]

Huh?

Sorry, no, and PSSR can look better than native 4k - it already does in many titles

This hand waiving away of the performance leap just because it’s not in raster is disingenuous AF

We already have plenty of uplscalers: DLSS, XeSS, FSR, TAAU, TSR... PSSR is just another one.

Raw power is the most important metric and always will be, you can do whatever you want with that power but looks like jump between PS5 and PS5 Pro is small - devs have to make hard decisions: better performance, better iq or better overall quality settings?

I was told on this forum when we discovered that AW2 is 800p that "native resolution don't matter" and stuff like that, turns out 800p is too much for PSSR...

So yeah, native resolution and raw hardware power will always be important.

 

[James Sawyer Ford]

We already have plenty of uplscalers: DLSS, XeSS, FSR, TAAU, TSR... PSSR is just another one.

Your disingenuous arguments are getting really tired

 

[adamsapple]

Your disingenuous arguments are getting really tired

???

What's disingenuous about the quoted statement? PSSR *is* an up-scaler. And all upscalers work better the more native resolution they have to work with. For comparison, Rebirth is DRS 1296p before PSSR and it has a lot less complex geometry and foliage etc than AW2, so the results are better there for up-scaling owing to those things combined.

 

[James Sawyer Ford]

???

What's disingenuous about the quoted statement? PSSR *is* an up-scaler. And all upscalers work better the more native resolution they have to work with. For comparison, Rebirth is DRS 1296p before PSSR and it has a lot less complex geometry and foliage etc than AW2, so the results are better there for up-scaling owing to those things combined.

????

He’s making the claim that it’s just “another” upscaler like FSR or TAA

as if there’s not dedicated ML hardware with 300 tops that actually allows it to upscale far better than those solutions and bringing it roughly in line with top tier solutions like DLSS not available on consoles or certain types of gpus

I do not care that the crisp resolution is achieved successfully by methods other than pure raster, all I care about is the end result

Games like stellar blade run better with PSSR at higher frame rates than the native 4K mode *AND* have a better image quality to boot

 

[kevboard]

in line with top tier solutions like DLSS not available on consoles or certain types of gpus

technically, XeSS is an ML based solution that is available on anything past RDNA2 on AMD, and GTX10 on Nvidia. and therefore, technically, also on Series X and S... should Intel finally fulfill their promise to make it open source and thereby available for all developers to use however they want.

it already does wonders in some games on the Steam Deck for me. not great at such low resolutions, but better than FSR for sure

 

[Bojji]

????

He’s making the claim that it’s just “another” upscaler like FSR or TAA

as if there’s not dedicated ML hardware with 300 tops that actually allows it to upscale far better than those solutions and bringing it roughly in line with top tier solutions like DLSS not available on consoles or certain types of gpus

I do not care that the crisp resolution is achieved successfully by methods other than pure raster, all I care about is the end result

Games like stellar blade run better with PSSR at higher frame rates than the native 4K mode *AND* have a better image quality to boot

All of them do the same thing, quality differs.

This is also third ML upscaler so "another" one of them (after DLSS and XeSS). Oh... I forgot about Apple Ai upscaler and Auto SR from MS so it's the fifth one.

Some developers still use FSR/TAAU/TSR in their pro patchet games for whatever reasons.

 

[Senua]

James Sawyer Ford vs the world

 

[James Sawyer Ford]

James Sawyer Ford vs the world

Versus

 

[James Sawyer Ford]

All of them do the same thing, quality differs.

This is also third ML upscaler so "another" one of them (after DLSS and XeSS). Oh... I forgot about Apple Ai upscaler and Auto SR from MS so it's the fifth one.

Some developers still use FSR/TAAU/TSR in their pro patchet games for whatever reasons.

Yeah, quality differs ….a lot ….that’s the point of them adding the hardware they did … the results speak for themselves in games that use PSSR well. It isn’t a 30% increase

 

[Desprado123]

Anyone who thinks that PSSR performance will increase by just a magic update is extremely stupid or Nvidia is scamming the world by Tensor cores.

DLSS=Tensor core, more core lower DLSS lower hit on RTX GPU's and AI cannot be magically update by a patch. It is totally depended on Hardware and only a hardware update can solve PSSR 2ms or 3ms issues.

 

[James Sawyer Ford]

Anyone who thinks that PSSR performance will increase by just a magic update is extremely stupid or Nvidia is scamming the world by Tensor cores.

DLSS=Tensor core, more core lower DLSS lower hit on RTX GPU's and AI cannot be magically update by a patch. It is totally depended on Hardware and only a hardware update can solve PSSR 2ms or 3ms issues.

It has 300 tops ML hardware equivalent to what tensor cores are.

There is always a raster cost to AI up scaling independent of dedicated hardware that performs matrix ops

Yes, there will be software updates that improve PSSR…

 

[Desprado123]

It has 300 tops ML hardware equivalent to what tensor cores are.

There is always a raster cost to AI up scaling independent of dedicated hardware that performs matrix ops

Yes, there will be software updates that improve PSSR…

Please Shut up with BS and please go to google learn what is Tensor core and AI upscales. BS on neogaf won't solve issues of PS5 Pro or PSSR.

 

[Salty Pickle]

Anyone who thinks that PSSR performance will increase by just a magic update is extremely stupid or Nvidia is scamming the world by Tensor cores.

DLSS=Tensor core, more core lower DLSS lower hit on RTX GPU's and AI cannot be magically update by a patch. It is totally depended on Hardware and only a hardware update can solve PSSR 2ms or 3ms issues.

So you are telling me nvidia did’t improve anything on existing hardware going from dlss 1 to 2 and then onwards to 3?

I guess you’re not old enough to have played the first dlss games when they arrived with the 20-series rtx.

Jeez.

 

[Salty Pickle]

Don’t AW2 use hw mesh shaders on pc and xbox while using sw version on ps5 base? We know the pro have hw support for in but maybe the devs aren’t using that on the pro.

I’m starting to believe this is just lazy and unoptimised.

 

[nnytk]

Is there any news from Remedy regarding this performance?

I mean, I'm interested in getting Alan Wake 2 for the Pro, but is Remedy working on improvements here? All I want is a 60fps, stable mode. RT isn't important to me.

 

[Bojji]

It has 300 tops ML hardware equivalent to what tensor cores are.

There is always a raster cost to AI up scaling independent of dedicated hardware that performs matrix ops

Yes, there will be software updates that improve PSSR…

We don't even know if this console is using some new ML hardware or it's just WMMA instructions. If it's the latter then PSSR is taking away more resources from GPU than DLSS is for Nvidia cards.

So you are telling me nvidia did’t improve anything on existing hardware going from dlss 1 to 2 and then onwards to 3?

I guess you’re not old enough to have played the first dlss games when they arrived with the 20-series rtx.

Jeez.

Latest DLSS runs fine on RTX 2xxx. It's all software improvements.

Tensor cores are more and more powerful but this power is not exactly required by DLSS.

 

[Mayar]

Basically, what many have written about happened. That in fact, a lot will depend on the hands and desire of the developers themselves to use what is in PRO. In fact, this is very interesting, because it will really show the situation on the market and how lazy the developers really are in terms of optimizations and use of technologies.

Is there any news from Remedy regarding this performance?

I mean, I'm interested in getting Alan Wake 2 for the Pro, but is Remedy working on improvements here? All I want is a 60fps, stable mode. RT isn't important to me.

What are we even talking about, the Pro currently can't run even Elden Ring in stable 60 in Performance mode, and the game is already 2 years old, and in theory with the promised performance it should pull it without any patches. But apparently we are running into a bottleneck - an old processor.

 

[rofif]

Alan Wake 1 also ran at 540p on Xbox 360.

These guys are obsessed with low res graphics!

But aw1 had clean smooth stable image that was very impressive. I don’t think aw2 warrants it’s image instability. There are better looking games that run faster

 

[Salty Pickle]

We don't even know if this console is using some new ML hardware or it's just WMMA instructions. If it's the latter then PSSR is taking away more resources from GPU than DLSS is for Nvidia cards.



Latest DLSS runs fine on RTX 2xxx. It's all software improvements.

Tensor cores are more and more powerful but this power is not exactly required by DLSS.

Yes and that is my point. The PS5 pro should have plenty enough of TOPS for ML-based upscaling to work really well so it will only get better and more efficient moving forwards also in terms of milliseconds.

With that said I’m also not surprised to see devs implementing it on old games and engines with varying results just like how Dlss worked in those early days before devs became more used to it.

The whole point of pssr on PS5 pro is to get the devs’ feet wet before PS6, or at least that is what I believe.

 

[mansoor1980]

Is there any news from Remedy regarding this performance?

I mean, I'm interested in getting Alan Wake 2 for the Pro, but is Remedy working on improvements here? All I want is a 60fps, stable mode. RT isn't important to me.

dont buy it , untill they get this shit sorted out

 

[Desprado123]

So you are telling me nvidia did’t improve anything on existing hardware going from dlss 1 to 2 and then onwards to 3?

I guess you’re not old enough to have played the first dlss games when they arrived with the 20-series rtx.

Jeez.

They did and improved performance but i am talking impact related to 2ms to 3ms. It is the same and it depends on GPU.

 

[Salty Pickle]

They did and improved performance but i am talking impact related to 2ms to 3ms. It is the same and it depends on GPU.

Nvidia went from 2-4 ms/frame with Dlss 1 to 1-2 ms/frame with Dlss2.

There is nothing that says Sony couldn’t improve theirs over iterations on the same hw just like Nvidia have.

 

[Zathalus]

Seems the major issue with PSSR at the moment is temporal stability, which is made worse at lower rendering resolutions. Hopefully Sony is working on that aspect.

 

[SKYF@ll]

A few people have concluded that the PS5 Pro only has a 30% boost, but they are wrong.
The 250 watt PS5 Pro (suitable for TV) has many limitations when it comes to benchmarking the GPU: 60fps or 120fps cap, CPU bound, memory bandwidth and V-Sync.
Remedy's Pro patch is a failure and developers should have improved the resolution and framerate.
Let's wait for better updates for other games in the future.

 

[Lysandros]

Don’t AW2 use hw mesh shaders on pc and xbox while using sw version on ps5 base? We know the pro have hw support for in but maybe the devs aren’t using that on the pro.

I’m starting to believe this is just lazy and unoptimised.

No, AW2 uses mesh/primitive shaders on both PS5/XSX using the same AMD enabler hardware which maps 1:1 to primitive shaders (AMD terminology for the feature). XSX doesn't have a specific "mesh shader hardware", nuances are related to directx12U interpretation of the implementation.

 

[twilo99]

Optimizations for Pro would also make base PS5 version perform better, it's still the same architecture. So far delta seems around that 30% and if games will start to show more than that we will know that something was wrong at launch.

Wouldn’t those extra optimizations help AMD GPUs on PC as well? Maybe not to the same extent, but something should trickle down.

I really wish devs would put the same amount of optimizing work into their PC projects.

 

[Fortress of Regrets]

Sorry, no, and PSSR can look better than native 4k - it already does in many titles

Which titles? Proof?