• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

NXGamer: Resident Evil 4: REmake - PS5, Xbox Series X, PC, Steam Deck Technical Review

People saying there's no need for a PS5 Pro.

Bt71dQY.gif
 

DenchDeckard

Moderated wildly
Gonna double dip on PC. Was planning to anyway but PC is just dominating in this game.

Need some steam deck analysis for the lulz

Kb and mouse squad rise up!
 
Last edited:

rofif

Banned
Beware as deadzone on PC is twice worse than on PS5 if you use a Microsoft controller.
now... is it the same with re2r, 3r and village on pc when using a controller?
Are PS versions of these the same? I played all of these on pc with mouse and keyboard but I wanted to do replays with controller
 

01011001

Banned
- Input latency is very high on both (204 vs 219 ms on PS5 and SX respectively)
- NXG thinks this is an intentional decision.

I said this in the other thread already, but that simply can not be the case... I'll just quote myself

there is absolutely no way that is the case. and also no way it is intentional, intentional input lag doesn't exist, animation lag yes, not input lag. a game has input lag due to technical reasons like bad Vsync with too many pre-buffered frames, or too much GPU load. especially given it 100% is not present in the PC version.
I know what 200+ms latency feels like, this is simply false.

does he say how he tested this? he has no timestamps in his video and I really don't wanna watch it all just to get to that point.

I wanna bet what he did is look for an animation start instead of actually testing the latency at which an input is registered by using something that isn't bound to an animation wind-up like opening a menu or similar inputs.
 
Last edited:

01011001

Banned
Surprised they dont drop the res for RT to something like 1440p.

yeah it's weird in general that RT is not a separate mode but instead an on/off toggle, same with the hair.
if they do that shit they should also add another SUPER Performance mode that will lock the res to 1440p
 





edit: he will apparently do a full review on launch as well even though he says the final review game (on consoles at least) is give or take the same, minus the missing RT reflections being fixed on PS5.


- No sign of DRS in his testing but uses Capcom's RE engine checker-boarding (interlacing) reconstruction
- Both PS5 and SX have same resolution metrics (2160p in resolution and 1944p in performance)
- While the pixel count is the same, the image quality between the two is not similar.
- PS5 has noisier image, post effects, worse aliasing, temporal stability etc
- NXGamer thinks the issue is in difference of AA working with the reconstruction/interlacing
- PS5 looks similar to FSR1 and SX looks similar to FSR2 in PC equivalent terms.

- In resolution mode both drop lots of frames with RT and Hair on
- SX is often 6~10 FPS lower in this mode. It is often below the VRR range here. PS5 can stay in the VRR range more often here.
- In Resolution mode with Hair and RT off, PS5 is closer to 60 more often but can dip in heavy bandwidth scenarios, and SX can dip a further 10-15% over that still.

- In Frame Rate mode with RT/Hair on, the performance is nearly identical with SX having approx 5% worse performance.
- Turning RT and Hair off makes this a locked 60 FPS mode on both barring some streaming stutters.

- Cinematics are GPU bound and game play are CPU bound which contributes to the above in Xbox having a small advantage in cut-scenes and PS5 in game play.

- Input latency is very high on both (204 vs 219 ms on PS5 and SX respectively)
- NXG thinks this is an intentional decision.
- NXG thinks it helps to increase acceleration on Xbox pad high to remedy some of the deadzone issues on Xbox.

- 800p / 30 FPS works great but can be very soft with acceptable settings
- very CPU bound and also GPU RAM dependent, can crash if overloaded.
- No pre shader compilation so 1st run can cause shader compilation stutters
- His PC equivalent settings on a 2070
- Consoles generally have High equivalent. Volumetrics = medium, Hair = normal, RT = normal.
- 4K with interlaced on doesn't hold close to 60 FPS. Dropping to 70% resolution scale gives equivalent performance to PS5/SX.
- Memory issues are a problem on 8GB. Can cause crashes when overloaded. Village fight would always crash for him at 4K. Only fixed by dropping to 1620p at 80% resolution scale.
- The game is very demanding on GPU (and CPU). Don't expect 8GB cards to run this at 4K



They really need to fix the mud filter on PS5, I may hold of on picking this up until they do so.
 
why performance mode is almost solid 60fps anyway

Yes but maybe to avoid the IQ issues caused by interlacing and/or FSR?

It seems like the PS4 Pro version runs at straight 2240x1260 unlocked

On PS5 that unlocked becomes locked 60 fps
 
Last edited:
Deadzone on PS5 is noticeable but not gamebreaking for me. But I'm getting it on the PC so it ultimately doesn't matter to me. If I did get it on the PS5, I would've gotten over it quick.

There's nothing wrong with deadzone on ps5, many games have similar deadzone %. Series X Def has an issue
 

01011001

Banned
There's nothing wrong with deadzone on ps5, many games have similar deadzone %. Series X Def has an issue

well deadzones should ALWAYS be changeable in the options of a game that has heavy shooter elements. so there definitely is something wrong there right away.
but even at that, 12% isn't good. it's at best serviceable but not good, especially paired with such an extremely WIIIIIIDE reaction curve that they are using.
with a reaction curve as wide as theirs I'd use a 0% to 3% deadzone if I could chose.
 
They really need to fix the mud filter on PS5, I may hold of on picking this up until they do so.

That's the only logical thing to do if you care about image quality. I canceled my order. Series X version is out of the question too because of the aiming. Re4 on ps5 looks fine indoors but as soon as you go outside and it's daytime it looks bad.
 

adamsapple

Or is it just one of Phil's balls in my throat?
FYI here is a direct-feed comparison between Resolution mode and Frame Rate mode in the same place.

They are extremely comparable, the IQ looks very close. But besides the resolution, there's also some other small visual cut-backs on the Frame Rate mode. There's either less or missing god rays in the fence close to Leon, and the trees in the distance have less branches/vegetation.




Resolution Mode:



Frame Rate Mode:




It also looks to me that the Xbox Frame Rate mode looks better than the PS5 Resolution Mode because of all the IQ differences.
 
Last edited:

Imtjnotu

Member
FYI here is a direct-feed comparison between Resolution mode and Frame Rate mode in the same place.

They are extremely comparable, the IQ looks very close. But besides the resolution, there's also some other small visual cut-backs on the Frame Rate mode. There don't seem to be any god rays and the trees in the distance have less branches/vegetation.




Resolution Mode:




Frame Rate Mode:





It also looks to me that the Xbox Frame Rate mode looks better than the PS5 Resolution Mode because of all the IQ differences.
when i booted up the demo, the ps5 version was the worst looking PS5 game ive played. everything was shimmering, EVERYTHING
 

SlimySnake

Flashless at the Golden Globes
There is no need for ps5 pro. They wouldn't be able to replace that gpu or cpu with much better stuff anyway
There is a need for better development
Why does everyone keep saying this? The 24.5 tflops 6950xt was their 300 watt card on 7. Their 5nm 52 tflops chip is the same 300 watt tdp. They can easily get 2x more graphics power for the ps5 pro if they target 6950xt specs and don’t skimp on the vram upgrade.

It will cost $599 minimum but someone said they will ship it without the bluray drive and charge $99 for it so they will definitely make profit on day one.

The key is not to make the same mistake they made with the ps4 pro.
 

01011001

Banned
Frame Rate Mode:

residentevil4chainsawaoc4g.png




It also looks to me that the Xbox Frame Rate mode looks better than the PS5 Resolution Mode because of all the IQ differences.

framerate mode here actually is 1944p from my count looking at the one of the edges of that wood on the left.
27 steps on a 30 pixel ruler.
I might need to take a snap on PS5 from that exact angle later (or someone else could upload one) just to see the difference.
 
Last edited:

adamsapple

Or is it just one of Phil's balls in my throat?
framerate mode here actually is 1944p from my count looking at the one of the edges of that wood on the left.
27 steps on a 30 pixel ruler.
I might need to take a snap on PS5 from that exact angle later (or someone else could upload one) just to see the difference.


I got you fam.

RT on in all below, Hair off and all Lens Distortion and DoF shit is off.



PS5 Resolution Mode:

residentevil4chainsaw5zc64.png






PS5 Frame Rate Mode:

residentevil4chainsawq3ezl.png
 

rofif

Banned
Why does everyone keep saying this? The 24.5 tflops 6950xt was their 300 watt card on 7. Their 5nm 52 tflops chip is the same 300 watt tdp. They can easily get 2x more graphics power for the ps5 pro if they target 6950xt specs and don’t skimp on the vram upgrade.

It will cost $599 minimum but someone said they will ship it without the bluray drive and charge $99 for it so they will definitely make profit on day one.

The key is not to make the same mistake they made with the ps4 pro.
But it’s not soc. you can’t compare desktop gpu
 

01011001

Banned
I got you fam.

RT on in all below, Hair off and all Lens Distortion and DoF shit is off.



PS5 Resolution Mode:

residentevil4chainsaw5zc64.png






PS5 Frame Rate Mode:

residentevil4chainsawq3ezl.png

wow, this looks really rough in some spots.
it seems to me to almost run at half horizontal resolution, which is typical for Checekrboarding, but usually you do not see that when checkerboarding works correctly

I get 2048x1944 here in some spots... the vertical resolution seems to be 1944 for sure, but the horizontal res depending on where you look is definitely in the 2048 range.

2048 horizontal resolution would be 1152p in a normally rendered 16:9 image. I should get 3456 horizontally if the game targets 1944p.
but 2048 horizontally is still higher than the 1728 horizontal res you should expect in a 1944p checkerboarded resolution, should the reconstruction just fail.

we did see something like this in Red Dead Redemption 2, where the PS4 pro version didn't use checkerboarding but instead just rendered at half horizontal res and then just stretched the image to a 16:9 output.

but some edges here look better, where the horizontal res looks higher than on that low res edge I counted... most looks low res tho.

testpvih4.png


but even by eye, if you compare how edges along horizontal edges look like, and then compare to vertical edges, the vertical edges look way lower res
 
Last edited:

adamsapple

Or is it just one of Phil's balls in my throat?
wow, this looks really rough in some spots.
it seems to me to almost run at half horizontal resolution, which is typical for Checekrboarding, but usually you do not see that when checkerboarding works correctly

I get 2048x1944 here in some spots... the vertical resolution seems to be 1944 for sure, but the horizontal res depending on where you look is definitely in the 2048 range.

2048 horizontal resolution would be 1152p in a normally rendered 16:9 image. I should get 3456 horizontally if the game targets 1944p.
but 2048 horizontally is still higher than the 1728 horizontal res you should expect in a 1944p checkerboarded resolution, should the reconstruction just fail.

we did see something like this in Red Dead Redemption 2, where the PS4 pro version didn't use checkerboarding but instead just rendered at half horizontal res and then just stretched the image to a 16:9 output.

but some edges here look better, where the horizontal res looks higher than on that low res edge I counted... most looks low res tho.

Yeah I'm not surprised at all.

In direct comparison the PS5's Resolution Mode looks fuzzier and softer than SX's Frame Rate mode.


PS5 Resolution:

5bne67.png




SX Frame Rate:

681d93.png
 
Last edited:

01011001

Banned
Yeah I'm not surprised at all.

In direct comparison the PS5's Resolution Mode looks fuzzier and softer than SX's Frame Rate mode.


PS5 Resolution:

5bne67.png




SX Frame Rate:

681d93.png

same issue with the low res edges on vertical objects, hence the low horizontal edge count I got (yeah it gets kinda confusing talking about edges and resolution orientations because you count horizontal res with vertical edges and vertical res on horizontal edges...)

vertical edges look like half res... so I assume a broken checkerboarding or something.
 
Last edited:

SlimySnake

Flashless at the Golden Globes
But it’s not soc. you can’t compare desktop gpu
Sure you can. Sony used a AMD 480 and MS used a card very similar to a 580 to make their midgen consoles. Hell, just look at the RDNA 2 GPUs and see how they match 1:1 with the PS5.

Bolting on a zen 2 or zen 3 CPU on these GPUs is something they've done before.
 

Mr Moose

Member
same issue with the low res edges on vertical objects, hence the low horizontal edge count I got (yeah it gets kinda confusing talking about edges and resolution orientations because you count horizontal res with vertical edges and vertical res on horizontal edges...)

vertical edges look like half res... so I assume a broken checkerboarding or something.


Hopefully they can fix that shit, it shouldn't be hard once they notice it.
Yeah I'm not surprised at all.

In direct comparison the PS5's Resolution Mode looks fuzzier and softer than SX's Frame Rate mode.


PS5 Resolution:

5bne67.png




SX Frame Rate:

681d93.png
Some parts look worse on the 2nd pic but others look better.
 
Last edited:

01011001

Banned


Hopefully they can fix that shit, it shouldn't be hard once they notice it.


I hate to bring it up again... scratch that, I love bringing it up...
but it should be just as easy to not only feel that the deadzone on Xbox is wrong, but also as easy to fix as changing literally 1 fucking number in the code.

so if that was never fixed from production, to releasing a demo, to the full game with day 1 patch... then OH BOY...
 

Mr Moose

Member
I hate to bring it up again... scratch that, I love bringing it up...
but it should be just as easy to not only feel that the deadzone on Xbox is wrong, but also as easy to fix as changing literally 1 fucking number in the code.

so if that was never fixed from production, to releasing a demo, to the full game with day 1 patch... then OH BOY...
They fixed the shite RT reflections on the PS5, hopefully they'll fix the controller shit on Xbox and the IQ issues on PS5, if they don't I might skip it or get it later on PC.
 

01011001

Banned
They fixed the shite RT reflections on the PS5, hopefully they'll fix the controller shit on Xbox and the IQ issues on PS5, if they don't I might skip it or get it later on PC.

let's hope they will. they did eventually fix the One X version of 3... but that was the only instance I can think of where they drastically changed anything after launch that wasn't DLC related.
 

adamsapple

Or is it just one of Phil's balls in my throat?
let's hope they will. they did eventually fix the One X version of 3... but that was the only instance I can think of where they drastically changed anything after launch that wasn't DLC related.


I still find it weird that there's at least one person on Era whose swearing up and down that the controls feel much more responsive on the retail compared to the demo.

Guess we'll find out in 4 days.
 

01011001

Banned
I still find it weird that there's at least one person on Era whose swearing up and down that the controls feel much more responsive on the retail compared to the demo.

Guess we'll find out in 4 days.

I hope he is right, but DF barely ever talk about bad controls like that (it did help that there are multiple videos about it tho) and if they still feel the issue being there I kinda tend to believe them rather than a guy on RE that says "it feels smoother" if I remember the terminology correctly... like I said in that other thread, I tend to not trust people that use words like that instead of narrowing down exactly what improved... is it the deadzone? the response curve? the aim acceleration?... or maybe the outer edge at which point acceleration is applied?
 
Last edited:

SlimySnake

Flashless at the Golden Globes
lol PC version is trash too. Turned on RT and hair strand to play at 1440p and it crashed as soon as I went into the village and started recording. I had 2.5 GB left in VRAM and 32 GB of ram available but whatever.

Then I decided not to record, and saw that the performance went from 120 fps all the way down to 74 fps during gameplay. Fuck that.

The last cutscene was insane. The game went all the way down to 55 fps with the burning dude as the titles appeared. With RT and hair off, its a locked 60 fps at native 4k and locked 120 fps at 1440p in that cutscene with 10-15% of the GPU to spare. in the RT mode, the GPU was maxed out at 100% in that cutscene.

The game is clearly not optimized around RT and hair strand. Turn that off.
 
Top Bottom