Support NeoGAF

[Arioco]

Yes because clearly the PS4 Pro is more powerful than the X1X. Developer platform prioritization and optimization played no role at all.

Xbox One X is running at twice as much resolution as PS4 Pro. Yes, they clearly favored PS4 Pro.

And yes, this must be the first time we see 1080p on PS4 and 900p on One S. That had never happened before. Ever. It's not like we've seen the same a thousand times at all.

The difference is exactly the same we saw in Sekiro: 900p on One S, 1080p on PS4, 1800 checkerboard on PS4 Pro and native 1800p on One X.

 

[S0ULZB0URNE]

I agree that it isn't finished but since it's a last gen title as you say it should be running far better on the most powerful hardware last generation. It is not. I doubt From will do much to use the strengths of any Xbox platform. I suppose we'll see.

Going by the CLOSED NETWORK TEST which is not a finished build... it does run better than the last gen.
Most powerful hardware? What are you saying is the most powerful hardware?

MICROSOFT has the marketing rights to this game so they will surely make sure it's a good representation.

 

[tommib]

Nope, all Xbox consoles are underperforming. The game is clearly optimized for Playstation.

The only thing overperforming here is your monstrous bias and inability to let it go.

 

[Arioco]

Exactly. There is no reason for the X1X performance issues here. This is not like the PS5 and XSX that are similar in specs. The X1X has more memory and an undeniably stronger CPU AND GPU yet amazingly the PS4 Pro just to happens to have a significant performance advantage. It goes to show that not matter how good the hardware is nothing trumps software optimization. It makes sense for From to focus on the platform with the higher installed base I suppose.

There's a reason and it's been pointed out several times: One X is running the game at twice as much resolution, while the hardware is not twice as powerful nor does it have twice the bandwidth. In that scenario you would expect One X to perform worse. It's nor a misery. We saw the same in Sekiro.

 

[SportsFan581]

X1 getting it done. LOL

The latest big releases have started to look rough on the old base console, much bigger difference with PS4 than when the generation started. Oh well, I had a lot of fun on mine, but it's time to put that system on ice (except for the streaming). Hopefully we'll start to get some current-gen only 3rd party stuff announcing. I can't believe we are this far in and don't really have anything yet.

 

[Rivet]

Base Xbox One version looks really terrible between the total lack of grass and the half rate 15 fps animated monsters (never heard about any other dev doing that pathetic trick). I know that console is weak, but there's no excuse here.

From Software are proving once again they're no tech wizards at all. I don't understand with all the money they made why they can't hire more advanced staff. Their games deserve a technical upgrade.

 

[Lysandros]

Xbox One X is running at twice as much resolution as PS4 Pro. Yes, they clearly favored PS4 Pro.

And yes, this must be the first time we see 1080p on PS4 and 900p on One S. That had never happened before. Ever. It's not like we've seen the same a thousand times at all.

The difference is exactly the same we saw in Sekiro: 900p on One S, 1080p on PS4, 1800 checkerboard on PS4 Pro and native 1800p on One X.

Not quite since it's 1800P when looking at the sky mostly, with lower bound of DRS at 1512P. PS4 PRO is not that far off from the lower bound, especially if one considers the added cost of reconstruction (which is not cheap) even with the support of hardware ID buffer.

 

[Arioco]

Base Xbox One version looks really terrible between the total lack of grass and the half rate 15 fps animated monsters (never heard about any other dev doing that pathetic trick). I know that console is weak, but there's no excuse here.

In fact it's a very common "optimization" trick. You can find it in Halo 5, the reflections of Miles Morales (RT Performance Mode) and many games. The difference is that it usually affects the characters far away from the camera, when you have those characters moving at 15 fps right in front of your face it looks horrible. If done well it's not supposed to be noticed. I thinks that was From's intention, but some kind of bug is causing the half rate animation to work when it shouldn't.

 

[Rivet]

In fact it's a very common "optimization" trick. You can find it in Halo 5, the reflections of Miles Morales (RT Performance Mode) and many games. The difference is that it usually affects the characters far away from the camera, when you have those characters moving at 15 fps right in front of your face it looks horrible. If done well it's not supposed to be noticed. I thinks that was From's intention, but some kind of bug is causing the half rate animation to work when it shouldn't.

Oh ok then, thanks, that's interesting... I must say I never noticed any enemy moving at half the framerate in the games I played, so yes I guess that could be a bug.

 

[Arioco]

Not quite since it's 1800P when looking at the sky mostly, with lower bound of DRS at 1512P. PS4 PRO is not that far off from the lower bound, especially if one considers the added cost of reconstruction (which is not cheap) even with the support of hardware ID buffer.

1512p is the absolute minimum, and still is much higher native rez than 1800p CB (a little over 4 million pixels vs 3.2 million in the worst case scenario for One X).

I honestly don't think the extra rez is worth the performance hit (the reconstruction is very good and the difference in IQ is subtle, as was in Sekiro), but it's clear that if One X ran the game at the same rez as PS4 Pro it would perform better, as it does in many other games.

 

[Darsxx82]

Xbox One X is running at twice as much resolution as PS4 Pro. Yes, they clearly favored PS4 Pro.

And yes, this must be the first time we see 1080p on PS4 and 900p on One S. That had never happened before. Ever. It's not like we've seen the same a thousand times at all.

The difference is exactly the same we saw in Sekiro: 900p on One S, 1080p on PS4, 1800 checkerboard on PS4 Pro and native 1800p on One X.

No and no.

XBO X on ELDER is running at a noticeable lower average resolution and framerate than on Sekiro compared to PRO.

The 1800p on XBO X in this case is only when you look at the sky. They use dynamic resolution that goes down to 1530p, in Sekiro was 1800p lock. There are never double the pixels as if it happened in Sekiro.

But it is also that in the average framerate there is more difference despite the cut in resolution. Not to mention that on Sekiro even XBO X had better shadows.

Then NO, it is not the difference between XBO vs PS4 either.
In Sekiro it was reduced to only 900p vs 1080p. In Elder you have that plus huge cuts in vegetation, shadows, AO, AF, half fps on enemies....and lower average framerate.

That is, it is not the same as what was seen in Sekiro like you said, not even close but there is a negative evolution in the XBox platforms optimization.
Not to mention more details and indications that clearly indicate which platforms (at least right now) are one step ahead in attention and optimization such as loading times. Which is another example that the xbox versions have reverted in attention and optimization compared to what was seen in Sekiro.

PS.This talking about the beta of course, we will see in the final game.....

 

[Pedro Motta]

Nope, all Xbox consoles are underperforming. The game is clearly optimized for Playstation.

Nope, Checkerboarding is a very inteligent technique that puts a weaker console with almost the same IQ and way better performance than the one that is pushing native.

 

[Tchu-Espresso]

The game just massively favors Playstation, no matter the console. Makes the other thread look a bit clownish in hindsight

 

[SF Kosmo]

RIP Last Gen. It's interesting to see how much better Series S looks next to X1X. But this is a game where I think 60fps is going to add a lot even beyond that, so PC for me.

 

[SportsFan581]

Base Xbox One version looks really terrible between the total lack of grass and the half rate 15 fps animated monsters (never heard about any other dev doing that pathetic trick). I know that console is weak, but there's no excuse here.

From Software are proving once again they're no tech wizards at all. I don't understand with all the money they made why they can't hire more advanced staff. Their games deserve a technical upgrade.

They might work on it more before it releases. X1 has tons of amazing looking games on it, like RDR2 and RE 2 and 3, etc. It should be able to handle similar settings as the PS4 with a little less resolution. That's the way it worked out most of the time. Bit of a crazy difference on this one.

 

[MonarchJT]

The game just massively favors Playstation, no matter the console. Makes the other thread look a bit clownish in hindsight

i said from the start ...the only reason why series x perform worst is the famous technical incapability of the dev team
This results make some comments in the other thread obviously ridiculous

 

[Ezekiel_]

Even the loading times on the XOX is twice that of the PS4 Pro, despite having the more powerful CPU.

Dat Kraken compression though

 

[DForce]

The game just massively favors Playstation, no matter the console. Makes the other thread look a bit clownish in hindsight

i said from the start ...the only reason why series x perform worst is the famous technical incapability of the dev team
This results make some comments in the other thread obviously ridiculous

@13:40

The best theory as to why? Well, despite the 1X being the more powerful 6TF machine against PS4 Pro's 4.2TF - the fact is, Pro's reconstruction image is less taxing overall to render.

PS4 Pro is rendering at a lower resolution. That's it.

 

[MonarchJT]

@13:40



PS4 Pro is rendering at a lower resolution. That's it.

the series x is a 12tf machine the PS5 is a 10tf variable one...at least i would hope for a parity in performance if they really don't want to optimize . Let's hope they fix this shit show before the launch

 

[DForce]

the series x is a 12tf machine the PS5 is a 10tf variable one...at least i would hope for a parity in performance if they really don't want to optimize . Let's hope they fix this shit show before the launch

You wanted to use PS4 Pro and Xbox One X comparison to prove your point. You obviously didn't watch the video.

 

[hemo memo]

Maybe they'll learn how to do proper frame pacing by the time Dark Souls 16 comes out. I just can't be bothered with this kind of thing anymore, I just avoid games with constant frame pacing issues.

Ftfy

 

[bender]

PS3

 

[MonarchJT]

You wanted to use PS4 Pro and Xbox One X comparison to prove your point. You obviously didn't watch the video.

my point as always been that from soft are the dev to take less into consideration when it comes to technique and certainly among the last in terms of platform optimization, to this add that up more used to develop for playstation. Here the results

 

[Hendrick's]

Poor Xbox keep getting mistreated by devs who only click a few buttons for all 4 versions.

They're using Unity now?

 

[DForce]

my point as always been that from soft are the dev to take less into consideration when it comes to technique and certainly among the last in terms of platform optimization, to this add that up more used to develop for playstation. Here the results

This results make some comments in the other thread obviously ridiculous

Those are your words. This video doesn't prove your point. You're not trying to change what you meant instead of admitting that you didn't watch the entire video.

 

[Md Ray]

the series x is a 12tf machine the PS5 is a 10tf variable one...at least i would hope for a parity in performance if they really don't want to optimize . Let's hope they fix this shit show before the launch

"Comparing the theoretical FLOP rates of GPUs doesn't give you the full picture".



-Unity principal engineer, creator of Claybook.

 

[3liteDragon]

I see we're still stuck on the TFLOPs here after everything that's happened so far…

 

[tommib]

I see we're still stuck on the TFLOPs here after everything that's happened so far…

Xbox One X has more TFLOPs than Series S but ignorant people are still hanging on these numbers to try to bait console warring between the next-gen premium consoles. Series S is a next-gen console. XBox One X is not.

We all know by now that TFLOPs don't tell the full story but what can we do.

 

[tommib]

.

 

[tommib]

the series x is a 12tf machine the PS5 is a 10tf variable one...at least i would hope for a parity in performance if they really don't want to optimize . Let's hope they fix this shit show before the launch

 

[Riky]

Tflop comparisons between different generations is silly, in Polaris terms the Series S would be at least a 5tflop machine.
In the same generation it's a more viable comparison as it includes clock speeds.

 

[Bogroll]

Cap them at 30fps I'll take native over checkerboarding all day everyday.

 

[Zathalus]

@13:40



PS4 Pro is rendering at a lower resolution. That's it.

The difference is about a 40% pixel count. The XOX GPU is 40% faster with over 50% more memory bandwidth. If anything they should be performing the same with a slight advantage to the XOX. Not a 30% performance advantage for the PS4 Pro at times.

 

[MonarchJT]

@13:40



PS4 Pro is rendering at a lower resolution. That's it.

the PS4 pro perform better with all that is significantly weakened compared to the 1x ...everything in that game scale like a shit

 

[MonarchJT]

The difference is about a 40% pixel count. The XOX GPU is 40% faster with over 50% more memory bandwidth. If anything they should be performing the same with a slight advantage to the XOX. Not a 30% performance advantage for the PS4 Pro at times.

they are deaf who do not want to hear

 

[MonarchJT]

"Comparing the theoretical FLOP rates of GPUs doesn't give you the full picture".



-Unity principal engineer, creator of Claybook.

yep we all know and in fact in lots of multiplats cross gen games where the distribution on so many CUs is not optimized ...the PS5 can close the power gap and basically perform the same as the more powerful series x ..or at most with a 0.something of advantage in fps. Here the thing is different Have you seen how much better the PS5 performs in this game? I repeat this is simply a technical deficit from the developers nothing more nothing less.

 

[DenchDeckard]

I can't believe we are arguing over a beta and the Xbox one x over the PS4 pro. The one x if developed for properly absolutely wipes the floor with the PS4 pro specs wise.

has The war become so boring we are now doing cross gen wars, arguing over last gen versions of games.

theres Some amazing games out right now. Why not go play them lol

 

[Mr Moose]

The difference is about a 40% pixel count. The XOX GPU is 40% faster with over 50% more memory bandwidth. If anything they should be performing the same with a slight advantage to the XOX. Not a 30% performance advantage for the PS4 Pro at times.

1600 x 1800 v 3200 x 1800 is not 40%, it's 100%.
Oh, you're assuming the One X is at 2688 x 1512 all the time? Why? Pro has DRS, too. Dragon part is where it drops to that, which is the taxing area.

 

[RoadHazard]

1512p is the absolute minimum, and still is much higher native rez than 1800p CB (a little over 4 million pixels vs 3.2 million in the worst case scenario for One X).

I honestly don't think the extra rez is worth the performance hit (the reconstruction is very good and the difference in IQ is subtle, as was in Sekiro), but it's clear that if One X ran the game at the same rez as PS4 Pro it would perform better, as it does in many other games.

4 is 25% more than 3.2. 6 is 43% more than 4.2. The X1X is underperforming (as is the base X1, it shouldn't need to have all those visual downgrades in addition to the lower resolution).

 

[Little Chicken]

I feel for anybody who has to endure the GOTY on last-gen systems.

 

[Mr Moose]

4 is 25% more than 3.2. 6 is 43% more than 4.2. The X1X is underperforming.

2880000 pixels (Pro max) v 4064256 (One X min). It's ~41%. They have the incorrect number, not sure how they got 3.2M.

 

[RoadHazard]

2880000 pixels (Pro max) v 4064256 (One X min). It's ~41%. They have the incorrect number, not sure how they got 3.2M.

Alright, I was just trusting that guy's numbers, but yes, your math is correct. But ok, at that level it should pretty much perform the same as the Pro, but it performs worse. And CB isn't completely free either.

 

[Mr Moose]

Alright, I was just trusting that guy's numbers, but yes, your math is correct. But ok, at that level it should pretty much perform the same as the Pro, but it performs worse. And CB isn't completely free either.

Pro uses DRS too, apparently, which would explain why it's ahead in some areas. Could have the same lower bounds but with CB (1344 x 1512).


Though I'd wait for VGTech, because I don't really trust Tom lol.

Edit:

That's not a huge upgrade, Tom.
It's the PS4 version used twice

 

[Md Ray]

yep we all know and in fact in lots of multiplats cross gen games where the distribution on so many CUs is not optimized ...the PS5 can close the power gap and basically perform the same as the more powerful series x ..or at most with a 0.something of advantage in fps. Here the thing is different Have you seen how much better the PS5 performs in this game? I repeat this is simply a technical deficit from the developers nothing more nothing less.

Some of the units of the PS5 GPU can be up to 22% faster than XSX GPU, you know? It's totally possible for PS5 to have higher fps due these differences, and close the gap in other scenes/titles.

Are we forgetting that there are literally scenes (in the same game) where XSX can also pull ahead leading to higher framerate than PS5? I guess the game is magically optimized for XSX in those scenes? Depending on what's happening on screen and the engine workload, a scene can favor one hardware over the other. Some scenes will run faster on XSX, some will run faster on PS5. Some will have no difference at all. XSX GPU isn't faster than PS5 in every way. Nope.

And the results speak for themselves.

 

[DForce]

The difference is about a 40% pixel count. The XOX GPU is 40% faster with over 50% more memory bandwidth. If anything they should be performing the same with a slight advantage to the XOX. Not a 30% performance advantage for the PS4 Pro at times.

They never said what resolution the game is running at when comparing the scenes together. The lowest is 1512p native and PS4 Pro is 1800p. They never told us the lowest/highest resolution found in those scenes.

the PS4 pro perform better with all that is significantly weakened compared to the 1x ...everything in that game scale like a shit

It performs better at a lower resolution. That's something that you love to ignore.

 

[Shmunter]

Xbox out to lunch on a regular basis these days

 

[Lysandros]

They might work on it more before it releases. X1 has tons of amazing looking games on it, like RDR2 and RE 2 and 3, etc. It should be able to handle similar settings as the PS4 with a little less resolution. That's the way it worked out most of the time. Bit of a crazy difference on this one.

Most of the time PS4 pushed 42-46% higher resolutions with more stable framerates on top of that with occasional higher settings like anisotropic filtering or shadows. Now XboxOne struggles a bit more, it aged badly compared to PS4, this game is not the sole example.

 

[Filben]

Man, the lack of spotlight shadows from dynamic light sources such as particle lighting (e.g. from spells) and torches is jarring. I've noticed that in official trailers/gameplay even captured on PS5 and in the previous DF video and already mentioned it here I think. This is something other games could render ten years ago. In fact I saw shadows cast by carried torches back in 2006 in Gothic 3. Also, the lighting looks flat most of the times and the ambient occlusion could be better as well. I'd be okay with 2018ish visuals but this could have been Dark Souls 3... The game already looks old; technically, a very disappointing outlook. Artistically, the game looks great though.

 

[Zathalus]

Man, imagine this game if the technical bits were left to Bluepoint. Sure it would likely have been PS5 exclusive, but it would have been a much better looking and performing game.

Oh well, if the performance difference between the PS5 and XSX does not change then I'm obviously going to pick it up on PS5, only after VRR is out though.

 

[Lysandros]

For those who still think checkerboard solution is free/cheap, these slides are from Dark Souls remastered. PS4 PRO and X1X versions are 1800pCB and XboxOne X is 1800P native. Despite the very large difference in pixels compared to a 'native' 1800P (%100) they only managed to save 27% (2.6 ms) of rendering time.

Edit: This is why i dislike statements like "is pushing twice the number of pixels" when one machine use 1800pCB and the other 1800P in a game. This is quite far from the truth on processing cost and IQ basis. If a game is 1600x1800 in one machine and 3200x1800 on the other using simple upscale now we could talk about 'twice number of pixels pushed' since performance cost and IQ difference would be much higher.