Is this a rare drop or expected performance?
-
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.
Digital Foundry: Resident Evil 7 PS4/Pro/XB1/PC Comparison
- Thread starter adamsapple
- Start date
Is this a rare drop or expected performance?
LukasTaves
Member
They address it. The game on pc and ps4 has a post process aa with a temporal component to handle the shimmering, that temporal component is missing from xbone.The difference is very clear, and the Xbox One is definitely running at a different resolution, not the same as PS4. Just look at the shimmering on the foliage in this shot (click on the image for a bigger version):
But the absolute proof is when they zoom in. Click here for a comparison.
I think it's literally impossible that Xbox One is running at the same resolution as PS4 there. So there are only three scenarios:
1. Candyland is comparing using the demo, which has different settings than the retail version Digital Foundry is looking at.
2. Digital Foundry is right about 1080p on Xbox One, but didn't notice the resolution is dynamic and sometimes less.
3. Digital Foundry are just wrong, and the game isn't 1080p on Xbox One.
When chandoog or others post screenshots from Pro 4K and Xbox One, we should be able to push toward the real answer.
simatron3000
Member
Dude you will be running around 100fps at 1440p on that.
adamsapple
Or is it just one of Phil's balls in my throat?
But multiple people still getting in the count of 16xx is confusing lol. Ah well, whatever the case, they managed to make a 1080/60 game across the board, which is a rarity, not the norm.
Good on them.
adamsapple
Or is it just one of Phil's balls in my throat?
No.
thelastword
Banned
Where are the receipts on Capcom confirming 1260P? Seems like people can just say such and such dev said whatever without anything to prove it. Especially when hard counts suggest otherwise....
I also don't believe the XB1 is at 1080p, the IQ is far worse...It's just like when Guerilla said Shadowfall MP was 1080p. At this point, something is just really off with the XB1 image, worse effects, more morphing on effects and edges, worse ground textures, missing effects, which are all attributable to worse graphical settings, yet, even the DF guy didn't spell out too much of these differences as we know it.
If this is indeed 1080p on XB1, the lower resolution effects and AA solution on the XB1 is really hacking the final image. Usually, effects based games are the ones XB1 suffers most over the PS4, it makes no sense to have parity of resolution in that case......Even games like Project Cars showed framerate leads of 15-20fps over the XB1 version at 1080p vs 900 rez differences.....The XB1 GPU is definitely struggling with the PP and also has some noted performance issues in forested areas "with rendering foliage" et al.....It just makes no sense to have this game run 1080p on XB1 with all these bandwidth related issues.
I also don't believe the XB1 is at 1080p, the IQ is far worse...It's just like when Guerilla said Shadowfall MP was 1080p. At this point, something is just really off with the XB1 image, worse effects, more morphing on effects and edges, worse ground textures, missing effects, which are all attributable to worse graphical settings, yet, even the DF guy didn't spell out too much of these differences as we know it.
If this is indeed 1080p on XB1, the lower resolution effects and AA solution on the XB1 is really hacking the final image. Usually, effects based games are the ones XB1 suffers most over the PS4, it makes no sense to have parity of resolution in that case......Even games like Project Cars showed framerate leads of 15-20fps over the XB1 version at 1080p vs 900 rez differences.....The XB1 GPU is definitely struggling with the PP and also has some noted performance issues in forested areas "with rendering foliage" et al.....It just makes no sense to have this game run 1080p on XB1 with all these bandwidth related issues.
thelastword
Banned
Sad Affleck
Member
Dictator93
Member
Are you implying that Dark is lying?
What does he have to accomplish in such an action? Why does his character even need such defamation?
thelastword
Banned
Does that look like 1080p to you? As I've said, if there's a remote chance it's 1080p, graphical settings have been butchered......
I'm not implying anything, I simply asked for a dev quote on the resolution. Up to this day, I don't believe Guerilla Games when they said Shadowfall MP was 1080p......
Pixel counts show one thing, dev says another, at least we had the receipts from GG...Where are those Capcom receipts?
Sad Affleck
Member
I don't care man, I really don't. I play many of my games at 720p because I like high framerates. But you have to realize the irony of first saying that anyone can claim anything without proof and then immediately making a claim without proof.
Liabe Brave
Member
I simply don't believe that this can be the whole answer. While Capcom's info is very precise, keep in mind how often pixel counting finds games running at a different resolution than claimed by publishers. (Though of course it's usually in the opposite direction!)
But we don't need to be conspiratorially minded. There's plenty of actual evidence that something else is going on. For example, here's a crop 24 pixels wide I took:
If the game is 1260p, then a line that runs across 24 pixels as this one does should have exactly 14 steps. I simply do not see any intelligible way to find 14 steps in this crop. How do you do it?
Also, there's this:
Since this is such an interesting case I just recounted using a different method: local brightness peaks generated by anti-aliasing rather than actual corners:
This rather clearly gives me 13 peaks across 17 pixels, or 1650p. Curious. And this is in the same image as my previous result, so there's obviously not any dynamic shenanigans involved!
Finally, if you didn't pixel count the Pro version yourself, did you also rely on Capcom's word about standard PS4 and Xbox One both being 1080p? Because that seems it may not be the case either. Here's a couple of zooms on background detail of the two versions (click here to get a mouseover comparison):
While there's extra blur here from video compression, the differences are obvious. I don't see any way to interpret those shots where Xbox One is the same resolution. And note that these come from footage where the player is standing still, so accumulation effects should be at their maximum efficacy.
I can't imagine why Capcom would be deceptive, especially in a way that minimizes their results. But there's simply solid visual evidence that "1260p" can't explain by itself. Perhaps Durante is correct here:
paperspace
Member
Last edited:
thelastword
Banned
I gave both arguments on the IQ issues on the XB1, either it's not 1080p or some settings are really low, but some things are off that was no explained thoroughly in the analysis...
What I asked for was different though, pixel counts show one thing and it was said that the devs said it's 1260p. I just wanted to see a quote from the dev? Is that outlawed?
Go deep or go home.
Liabe Brave
Member
If you count 14 steps across 32 pixels, that gives the ratio .4375. Since it's a near-vertical line, you're measuring horizontal resolution. So your answer is .4375*3840, which is 1680. All this is well and good.
But that 1680 number is the true rendering width of the frame. The usual "p" numbers people use are the height of the frame: 1920x1080 is 1080p, 3840x2160 is 2160p, and so on. If you got a width result of 1680 pixels on a 16:9 image, then the height would be 945p. (1680/16*9)
It's not rare. Every combat sequence they analyzed was below 55fps, and several dipped below 50. There were also 3-4fps drops just walking around in more detailed areas.
Not checkerboard--there's no evidence of that here--but I do think there's something else involved (as does Durante).
Check out the still image comparison in the same post (which was just a link, so I've embedded the shots in another more recent post). That doesn't look like an AA difference, but a true resolution gap.
Zoom is to highlight interesting features, and allow others to analyze them. The details seen in such shots may or may not be visible at normal viewing distances. Keep in mind that if you sit 12 feet from your TV, a 400% zoom is like standing 3 feet away from it. It's not like putting your nose up on the glass.
adamsapple
Or is it just one of Phil's balls in my throat?
Enhance !
This comparison actually makes it seem like the Pro version is running at a similar resolution to the PC. There's some very faint extra blur but that could be down to the AA solution or post process, the edges and geometry looks nigh identical.
Liabe Brave
Member
I agree about the relative clarity of these shots, PC even at 1440p looks sharper. That does lend credence to the 1260p number.
But I can't get past the other points I've brought up above. My own counts could be wrong of course, but it's not just me. NXGamer and Candyland both got 16xx results independently, and they've been pretty consistent with counts previously. And in this thread Durante got a 16xx result using a different methodology, even though he had already counted the game as 12xx and 13xx before.
I'm willing to believe that the native rendering resolution is 1260p, but if so it seems there are temporal accumulation effects going on to improve the image over that level, at least selectively.
We don't imply that DF is lying, we are supposing that the info is not accurate or not well communicated.
Perfect, very interesting.
Would you accept confirmation from wbacon as "a quote from the dev" or argue that his position as a community manager is grounds for invalidation because although he is in a position to be privy to developer-level information, he himself isn't a dev? I've yet to PM him, just to be clear, but I figured it prudent to first clarify the boundaries.
Edit: Feel free to ignore this as I've sent him a PM. Let's clear this up once and for all.
Edit edit: He said he can ask but he's not sure if the info will be shareable.
PLASTICA-MAN
Member
If the PS4 PRO version is really 1260p, then what is the purpose of the PS4 PRO version then, especially if both games run at 60 FPS and share most of the same settings: textures, effects? I think the PS4 PRO is more capable than just a slight bump in resolution especially when this game isn't open world or has very demanding effects and physics.
Higher resolution with identical 60fps performance is pretty much exactly what I would hope for from the PS4 Pro version of a game.
Caayn
Member
The increase in resolution is indeed very minor, you'd expect something more in line with 1440p if Capcom would solely go for a resolution increase. But it's still an increase over the 1080p resolution from the base version if 1260p is true.
This is one of the weak points of console refreshes, you can't control what is improvement and how much.
PLASTICA-MAN
Member
Still we got more demanding games with even higher resolution and on the PRO before, i.e Uncharted 4 : 1440p and more with even much better effects.
IMACOMPUTA
Member
Is that wide shot opening of your car driving down the road actually real-time? It looked like real video to me. Holy shit.
Yeah, I thought the upgrade to Uncharted 4 was modest but this is bordering on pathetic. However, the difference between PS4 and PS4 Pro with PSVR on for RE7 is enormous (at least going by the demo). It looked like a completely different game when I compared the two.
This is a very good fucking question.
I just paid $575cdn for my Pro, and this is the best Capcom can do with that much extra horse power?
Devs should be called out for this bullshit.
Hockeyguy412
Banned
This game was in final stages of development before they even received ps4 pro hardware to work on so chill out
PLASTICA-MAN
Member
I have to see the PS VR version running on PS4 PRO. Are there any comparison videos already?
Liabe Brave
Member
Yes, here's one. The removal of the foveated rendering (where resolution is lower at the edges than in the middle) is extremely clear in this video--much more so than in other game examples I've seen.
PLASTICA-MAN
Member
Wow, the shimmering is gone on PS4 PRO which eases the eyes.
LukasTaves
Member
Any temporal AA works *best* when the next image is closer to the previous one. If the player is not moving then it's basically the best case scenario for the AA xbone lacks.
The game is also full of some other effects that can be lower res on xbone (like that with volumetric lights from the police car), so who knows if it's not an effect of that either.
It's not just a simple case of lower-resolution though - they seem to be using HTile-sized checkerboard and temporal reconstruction on the edges, which actually looks noticeably worse than simple lower-resolution render would (see other games that DO render lower-res on the edges for comparison).PLASTICA-MAN said:
Granted it's only a big deal in the bright outdoor scenes, indoors those same edges are too dark to see much in RE.
paperspace
Member
Last edited:
letmetrythisname
Member
I've played through the game twice on XB1 and the game has never dipped for me (noticeably) during combat. The game dips the hardest for me when there are a lot of lighting sources on screen, like the section where you're holding a lit candle.
Liabe Brave
Member
Of course. My point is that the differences in those two shots are not at all confined to the edges of objects (where temporal AA would be having an effect). The detail of the crinkled foil in the back, as well as reflections on the faucet and cutting board, are noticeably sharper on PS4. These aren't very plausible as effects of temporal AA, or of differences in volumetric lighting.
The most obvious explanation for variable surface detail on separate platforms is lower-resolution textures, but that doesn't seem to apply either. In other areas it's very clear that textures truly are identical on Xbox One and PS4. And while RE VII does have texture streaming issues, that this comparison comes from a shot held for a long time should iron out any contribution of that.
What are we left with? To my mind, there are only two possibilities. One is that the Xbox One is simply rendering at a lower resolution than PS4. Alternately, there is some issue with temporal reconstruction/reprojection. That is, either PS4 is native 1080p and Xbox One is temporally reconstructed 1080p; or both machines use reconstruction but the PS4 is using this method less often, or to cover a smaller "under resolution" gap.
Black Knight
Member
I think it's fantastic that the game looks as well as it does and runs great across all platforms. The team has done a phenomenal job and possibly might have set a standard with this one.
I had the PC version, but holy shit is there some intense frame hitching and stuttering. I have a 980 ti and at least 8 gb vram, and the graphics just aren't that demanding so it seems like an optimization issue. So I got it on the PS4 Pro and it looks 90% as good (lacks motion blur and no texture pop-in, but overall it matches PC quality, even looking better at times, especially in regards to the overly shimmery walls in the basement on the PC version - it's toned down on Pro).
Dictator93
Member
The 980Ti has 6 GB of VRAM, surely you as its owner must know that?
Also, any shimmery walls as you may mention it would be fallout from opting to not use TAA.