Support NeoGAF

[ALLseeingEYE]

Meh i wish they optioned to play in 60 fps since this isnt a multiplayer game there wouldnt be problem

I'm not a fan of the trend to simply 'Downsample' games from 4K when in 1080p mode. It seems like it is a huge waste of the PS4 Pro's Capabilities.

If the PS4 Pro really is close to a Radeon RX 470 or RX 480, it should be a 1080p beast on PS4 games, giving excellent image quality, and 60fps at 1080p. Using downsampling and 30fps doesn't make a lick of sense, unless PS4 Pro support was an afterthought, and the game was never designed or considered for 60fps. It's possible the game is CPU bound, and could not hold a steady 60 due to all the asset streaming from their open-world engine.

Either way, the game looks gorgeous, and even if 4k checkerboarding isn't exactly impressive sounding, I think the end product will look amazing in 4K.

 

[zewone]

Once Scorpio hits and there are (presumably) multiplatform games that are native 4K on Scorpio and checkerboard 4K on the Pro we can actually see how big the difference is.

You didn't get the memo that Scorpio is using a similar 4K method and not native either.

 

[Spukc]

creorder panceled

Spoiler

day1baby

 

[ALLseeingEYE]

You didn't get the memo that Scorpio is using a similar 4K method and not native either.

Umm, no. Phil Spencer has stated numerous times that first party MS titles will be native 4K on Scorpio. Scoprio is absolutely set to be an actual 4K system, and seems like it will be significantly more powerful than PS4 Pro from the info we currently have.

MS did mention that developers have the option to use 4k checkerboarding though, and to be honest, I would love to see some devs go crazy with that. Games could look mind blowing if they really push scoprio, and use smart rendering techniques like that.

 

[luffeN]

The former shades 4.1 million pixels each frame, the latter shades 8.3 million pixels each frame.

Ah, like the eshifting "4k" projectors.

 

[lightchris]

No, it looks fine in both static and moving scenes. RotTR implementation introduced barelly visible artefacts on very thin objects [like sparks after grendades].

Do we have good comparison images between PS4 Pro and PC at 4K?

I only know of this one, and I'm not sure whether it is still accurate. The difference was pretty huge in this one.

 

[ethomaz]

Cerny is correct. But a checkerboard resolution can't really be expressed in a base resolution as it renders a picture with voids in it. So 1920x2160 is only fit to explain the number of actual shaded pixels (~4.1m), not as the image size which is still 3840x2160 but with holes in it.

You're misunderstanding what cerny said. There's no base resolution, you start with the native resolution you're trying to render but half filled with traditionally shaded pixels while the other half is filled in from information from previous and current frames.


Take this checkered pattern is 4K but only the white ones are traditionally shaded. The blacks are filled in.

I understood the tech used but that didn't change the game has a native 1920x2160 frame buffer.

You can say the use of old frames to guess half of the pixels is more accurate than for example a simple upscale that will in most cases decide to repeat the color of the pixel close to it.

Being more accurate to construct the final image didn't make it native... it is still a checkerboard from a native 1920x2160 frame buffer that give you a better overall image than a simple upscale from the same 1920x2160 framebuffer.

Half of the image is being dynamically created based in the info they have while the others half is being rendered natively.

 

[Tripolygon]

I understood the tech used but that didn't change the game has a native 1920x2160 frame buffer.

You can say the use of old frames to guess half of the pixels is more accurate than for example a simple upscale that will in most cases decide to repeat the color of the pixel close to it.

Being more accurate to construct the final image didn't make it native... it is still a checkerboard from a native 1920x2160 frame buffer that give you a better overall image than a simple upscale from the same 1920x2160 framebuffer.

Half of the image is being dynamically created based in the info they have while the others half is being rendered natively.

The frame buffer is 4K. Seriously its not hard to understand.

 

[grandwizard]

actually, still images is where you'd be most likely to perceive said artifacts.

The more motion, the more artifcats, the less you'll be able to see them because - well - the image is moving.

I don't know about this. In Rainbow Six Siege and Watch Dogs 2 if you turn checkerboarding on it's most noticable in motion and really fuzzes up things.

 

[Liabe Brave]

What I posted came from Cerny's interview with DF.

Cerny said 1920x2160 buffer for 2160p checkerboard.

Sorry to take Cerny words over yours.

Check the article again. That framebuffer size is not a quote from Cerny, it's a statement from Digital Foundry. If you wish to believe them over me, that's still your prerogative.

But they're wrong.

So if it's only circa 4 million pixels v circa 8 million how can that be classed as basically the same?

I don't get it, I've read the thread and seen some comments but it still makes no sense to me.

You could see if my thread about how CBR works helps explain it any.

QUOTE=lyrick;229381414]"for certain" - There's those fucking weasel words again.[/QUOTE]
It seems you don't understand what "weasel words" means. They add vagueness, to escape being trapped in a lie. By specifying the exact conditions I referred to, I was adding precision, not obscuring it.

QUOTE=lyrick;229381414]What's the percentage of error from a native 4K image. How many pixels are drawn incorrectly when comparing checkerboarding vs native?[/QUOTE]
It depends on local conditions from frame to frame. Perhaps a dev could give us some statistics?

Because logically the word 'basically' can be interchanged with the term 'basically not' and not change in actual meaning one bit.

This is not correct. "Basically" is a probabilistic term. It does not equal its own negation. And certainly natural use doesn't supoort you either.

Pretty sure theres a PS4 Pro ad out there on youtube, that says dynamic 4K gaming. If that has changed well then who cares i guess.

"Dynamic" in that case could be an adjective about "gaming", not about "4K".

But it sounds like checkerboard rendering = 2160i instead of 2160p.

Yeah I think of it like 4KI or interlaced....

It's not like interlacing. CBR sends exactly as many pixels, and in the same manner, as any other progressive method.

After perceptual 60fps comes perceptual 4k.

Both of these are real concepts. Or are you claiming to be able to tell, by eye, the exact framerate and resolution of all media?

 

[dr_rus]

QB didn't use checkerboard rendering
Most people disliked the TAA implementation due to ghosting artifacts, not how their super resolution method worked.

It does actually, if you consider all such "MSAA tricks" as one and the same approach to temporal resolution reconstruction. The difference is that while "checkerboarding" is using MSAA 2x (which results in a checkerboard pattern), QB used MSAA 4x with a much more advanced accumulation technique which would probably be hard to describe with just one word (like "checkerboard").

TAA implementation and ghosting are both tied to this technique as usage of four different shading grids means that you have to (or should I say "choose to") accumulate frame data over four previous frames which makes all kinds of temporal artifacts twice as visible as with a "checkerboard" solution.

 

[Samurai G0SU]

The frame buffer is 4K. Seriously its not hard to understand.

ninja edit. damn! i was going to ask ..

 

[Lactose_Intolerant]

I don't know about this. In Rainbow Six Siege and Watch Dogs 2 if you turn checkerboarding on it's most noticable in motion and really fuzzes up things.

Yeah, if it is using previous frame information to help fill in the gaps, it will have more artifacts in motion.

 

[Brix]

My Pro Is Ready.

 

[WadeitOut]

Didn't we already know this? I thought they said this at the Pro event.

 

[icespide]

let me guess, this thread has turned into some people handwaving checkerboarding away as "upscaling" while others explain why its actually quite good

 

[timberger]

Umm, no. Phil Spencer has stated numerous times that first party MS titles will be native 4K on Scorpio. Scoprio is absolutely set to be an actual 4K system, and seems like it will be significantly more powerful than PS4 Pro from the info we currently have.

Okay, but this is completely irrelevant to the post you're responding to which is addressing the statement of multiplatform games being "native 4k" on Scorpio and checkerboarded on PS4 Pro. Something which we have seen recent evidence of being a fairly unlikely scenario as MS themselves have been advising third parties to use checkerboarding instead of native.

So... what is this "actual 4k system" to which you are referring?

let me guess, this thread has turned into some people handwaving checkerboarding away as "upscaling" while others explain why its actually quite good

Solid guesswork.

Also: Scorpio!

 

[Liabe Brave]

I don't know about this. In Rainbow Six Siege and Watch Dogs 2 if you turn checkerboarding on it's most noticable in motion and really fuzzes up things.

I haven't checked out Siege, but Watch_Dogs 2 has a particularly artifact-prone implementation of CBR. It's possibly related to Ubisoft's proprietary HRAA method. The same artifact problems seem to be present in the For Honor beta too, based on my initial look at it.

 

[Ushay]

The game will be rendered in 2160p checkerboard (meaning, basically 4K)

The game will be rendered in 2160p checkerboard (meaning, not actually 4K)

I can see why they went with the first one

I know right, so confusing. As long as it looks ace, that's what counts.

 

[BigEmil]

Do we have good comparison images between PS4 Pro and PC at 4K?

I only know of this one, and I'm not sure whether it is still accurate. The difference was pretty huge in this one.

Was it with the same texture quality and settings to the console version then do the 4K comparison was the Pro version also 2160c

 

[-shadow-]

Not really surprised it won't be actual 4K in any way shape or form. But I wonder how the game will compare on a 1080p screen. Guess we'll see soon.

 

[Planet]

Not really surprised many people still have no clue how checkerboard rendering works, or what the benefits are.

 

[routerbad]

This game looks incredible. The resolution supersampling on OG PS4 and higher res on the Pro will clean up the experience a bit, make things slightly more detailed and crisp, but either way, beautiful game.

 

[Timu]

Not really surprised many people still have no clue how checkerboard rendering works, or what the benefits are.

Of course, heck do we know the true resolutions for checkerboard rendering PS4 Pro games besides just saying "checkerboarded 1800p, checkerboarded 2160p" and such?=O

 

[TetraGenesis]

More """""""4K"""""" resolution. I'm really not impressed with the PS4Pro.

Do you have one and a 4k TV? Because if you don't, I can assure you it's impressive in action. Full 4k or no.

 

[Chaostar]

There's a surprisingly detailed but straight talking explanation of some of the techniques PS4 Pro uses for rendering 4k here...

http://www.stuff.tv/features/you-wont-be-able-tell-difference-between-ps4-pro-games-and-proper-4k

On difference between native 4k and checkerboard:

"...I’ve seen the side-by-side comparison, and were it not for the fact that I was standing a foot away from the screen and had Mark Cerny (you know, the bloke who designed the PS4 and Pro) pointing out the differences, I would never have spotted them myself."

 

[Goose_Se7en]

let me guess, this thread has turned into some people handwaving checkerboarding away as "upscaling" while others explain why its actually quite good

Bingo. Don't care if it's upscaled 4K, this game is the reason I purchased a good 4K TV and I cannot wait to play it.

 

[ALLseeingEYE]

Okay, but this is completely irrelevant to the post you're responding to which is addressing the statement of multiplatform games being "native 4k" on Scorpio and checkerboarded on PS4 Pro. Something which we have seen recent evidence of being a fairly unlikely scenario as MS themselves have been advising third parties to use checkerboarding instead of native.

So... what is this "actual 4k system" to which you are referring?



Solid guesswork.

Also: Scorpio!

Do you have any actual experience with 4K rendering and the computer hardware it takes to hit 4k at 30fps? There are some very specific video cards that can absolutely hit that or higher on basically every AAA 3rd party title out these days.

If MS can hit those video card targets, and if they are targeting 4k at 30fps on Xbox One quality games, it will absolutely be able to hit 30fps on 3rd party titles. That's not even considering how developers can scale video settings to make it easier to hit that target. Whether or not 3rd parties choose to shoot for 4k is another matter entirely. Another awesome option could be developers targeting 1440p/ 1800p at 60fps, with cranked visual quality. I would personally love to see both Scorpio and PS4 Pro games offer settings options, so we can tweak it how we like.

Even if your argument is essentially an assumption that the rumor of MS not using the absolute latest CPU tech from AMD will somehow hinder them from hitting 4K 30 on a huge amount of games, you would still be wrong, since it's the GPU, not the CPU that would be bottlenecking them at that resolution, especially at 30fps.

I think all you'd have to do, is look at PC and the state of 3rd party games that can render at 4k to see your answer. I think this is one area where 3rd parties and Microsoft are likely several steps ahead of Sony.

 

[iHaunter]

I'd prefer 60fps over 4K tbh.

 

[hardcastlemccormick]

Of course, heck do we know the true resolutions for checkerboard rendering PS4 Pro games besides just saying "checkerboarded 1800p, checkerboarded 2160p" and such?=O

It's difficult to come up with a "true resolution" for these sorts of temporally assembled post-process methods. Perhaps in the future, we will look back on checkerboard 4K as poor image quality (kind of like how Field Rendered PS2 games don't look too hot these days) but for now, it doesn't really matter.

 

[clarknova156]

Was anyone expecting native 4K... Lol

Based on the thread reactions so far it seems so. Personally I don't care I want smooth performance and I just really want the game to be good. Graphics are great looking and all but if the story and gameplay don't live up that'll be a bummer.

 

[Planet]

The true resolution for checkerboard 2160p is 2160p. Just because half of these pixels are approximated doesn't mean they aren't there. They don't get omitted, they are just calculated in a relatively simple but clever way.

In demonstration sessions, Sony had displays of the very same game running at 4K native and checkerboard side by side (at a matching lower framerate to achieve the former). The journalists had to look real close to tell which is which. I wish Sony would give this demo to the public.

The slower the movement, the less discernible the method is from the real thing. Faster irrational movement can generate less favorable artefacts. But you can safely assume that in sum it is better than rendering at half resolution and then upscale traditionally. And of course leaps and bounds better than upscaling 1080p material. Quite impressive for a hardware that is on paper not fast enough to deliver a 4K image at this quality if you ask me.

But all that has been explained ad nausea many times, and still is ignored by the crowd that loves to hate.

 

[adamsapple]

My 4K TV is ready

Bring it on !

 

[belvedere]

It's been said time and time again by DF, Linus Tech Tips and others that the difference between checkerboard rendering and native 4k is only truly noticeable when standing directly in front of the display.

 

[Aztechnology]

So we don't know what the base resolution for the Pro is yet?

The only time it's really bothered me personally was in FFXV. A lot of the particle effects, caustics and shaders had quite noticeable artifacts. I'm all for checkerboarding, I prefer performance over graphics so It's a nice compromise.

 

[ALLseeingEYE]

It's been said time and time again by DF, Linus Tech Tips and others that the difference between checkerboard rendering and native 4k is only truly noticeable when standing directly in front of the display.

I'm curious to know how true that actually is. You could almost argue the same thing for 4k in general. If you stand 8 feet back from a certain sized 4K display, it won't look much different than 1080p.

If this technique is actually that great though, that is awesome. I was a fan of the rendering techniques in Quantum Break, and they looked even better on PC at 1440p or 4k, even if you could definitely still see artifacting and ghosting.If this technique is that good, perhaps we will see it used often, and see developers truly pushing Scorpio and PS4 Pro. Like, why not shoot for 4k checkerboard at 60fps on Scorpio?

 

[bonej]

I don't know about this. In Rainbow Six Siege and Watch Dogs 2 if you turn checkerboarding on it's most noticable in motion and really fuzzes up things.

cb quality depends on the implementation. Some looks better some looks worse

 

[Curufinwe]

I'd prefer 60fps over 4K tbh.

You should have invented a better CPU than AMD then.

 

[bonej]

sony should have waited til vega like ms does. The pro is an underpowered 4k machine

 

[Liabe Brave]

Of course, heck do we know the true resolutions for checkerboard rendering PS4 Pro games besides just saying "checkerboarded 1800p, checkerboarded 2160p" and such?=O

So we don't know what the base resolution for the Pro is yet?

There is no "base resolution" for CBR. It is not upscaling. If a game is 1800c, it renders just as many pixels as 1800p. If it's 2160c, it renders just as many pixels as 2160p. Both 1800c and 1800p are upscaled on 4K displays; both 2160c and 2160p are not.

 

[Curufinwe]

Here's the thing.

The difference won't be big. Same way it was never big between Xbox One and PS4.

RE7 is just the latest of many games where the difference is actually quite noticeable, and anyone with both consoles would be better off getting the superior version.

 

[Durante]

Define "shades", please.

Calculating the color of a pixel based on the surface properties at that particular point, and the scene lighting.

Of course, heck do we know the true resolutions for checkerboard rendering PS4 Pro games besides just saying "checkerboarded 1800p, checkerboarded 2160p" and such?=O

We know exactly how many samples are shaded: half as many as for a full resolution image.

There is no "base resolution" for CBR. It is not upscaling. If a game is 1800c, it renders just as many pixels as 1800p. If it's 2160c, it renders just as many pixels as 2160p. Both 1800c and 1800p are upscaled on 4K displays; both 2160c and 2160p are not.

You really are hell-bent on confusing people by using "rendering" for something which is completely unlike anything anyone would traditionally call "rendering" in 3D graphics.

Why?

 

[ElfArmy177]

More """""""4K"""""" resolution. I'm really not impressed with the PS4Pro.

That's unfortunate cause games on the pro look amazing.

With that said my 1080 GTX struggles to get 4k with all bells and whistles at 60fps and it's a 8 or 9 TF graphics card.

Anyone who isn't impressed that a 300-400$ console is pushing "4k" with improved visuals is absolutely ignorant to technology and how it works. Sorry...

 

[charsace]

It's difficult to come up with a "true resolution" for these sorts of temporally assembled post-process methods. Perhaps in the future, we will look back on checkerboard 4K as poor image quality (kind of like how Field Rendered PS2 games don't look too hot these days) but for now, it doesn't really matter.

I know checkerboard doesn't look as good as true 4k. NBA 2k17 looks sharper and cleaner than WD2 because it. It doesn't fool me.

 

[Lord Error]

It does actually, if you consider all such "MSAA tricks" as one and the same approach to temporal resolution reconstruction. The difference is that while "checkerboarding" is using MSAA 2x (which results in a checkerboard pattern), QB used MSAA 4x with a much more advanced accumulation technique which would probably be hard to describe with just one word (like "checkerboard").

TAA implementation and ghosting are both tied to this technique as usage of four different shading grids means that you have to (or should I say "choose to") accumulate frame data over four previous frames which makes all kinds of temporal artifacts twice as visible as with a "checkerboard" solution.

QB method was really inferior in practice than Checkerboard used on PS4Pro. In QB it all falls apart in motion. Maybe the method itself was more complex, but the end result was worse regardless.

 

[dr_rus]

It's been said time and time again by DF, Linus Tech Tips and others that the difference between checkerboard rendering and native 4k is only truly noticeable when standing directly in front of the display.

The most obvious difference between a temporal resolution reconstruction approach and native rendering is the significant increase in shimmering on both polygon edges and surfaces which I'd say could be visible from any distance - if you know what you're looking for. The game does need to be in motion though as in a perfectly stable frame there would not be any difference at all.

QB method was really inferior in practice than Checkerboard used on PS4Pro. In QB it all falls apart in motion. Maybe the method itself was more complex, but the end result was worse regardless.

That depends on what you are comparing. It was worse in spatial resolution definitely but it was better in general image stability. I'd say that QB's aim at a "cinematic" presentation fit this method nicely as QB is one of those rare games where it's really hard sometimes to decipher if what you're looking at is a video or it's being rendered right now. For a vast majority of other games though combining only two frames in a straight "checkerboard" would be a better approach.

 

[LukasTaves]

The true resolution for checkerboard 2160p is 2160p. Just because half of these pixels are approximated doesn't mean they aren't there. They don't get omitted, they are just calculated in a relatively simple but clever way.

In demonstration sessions, Sony had displays of the very same game running at 4K native and checkerboard side by side (at a matching lower framerate to achieve the former). The journalists had to look real close to tell which is which. I wish Sony would give this demo to the public.

The slower the movement, the less discernible the method is from the real thing. Faster irrational movement can generate less favorable artefacts. But you can safely assume that in sum it is better than rendering at half resolution and then upscale traditionally. And of course leaps and bounds better than upscaling 1080p material. Quite impressive for a hardware that is on paper not fast enough to deliver a 4K image at this quality if you ask me.

But all that has been explained ad nausea many times, and still is ignored by the crowd that loves to hate.

If they were artificially lowering the frame rate then they were skewing the results. Because that makes the checkerboard looks better in the demo than it will be on real case tests.

I think it's tomb raider someone posted pics. The game gets noticeably sharper when you are not moving.

 

[lyrick]

This is not correct. "Basically" is a probabilistic term. It does not equal its own negation. And certainly natural use doesn't supoort you either.

basically "Thing" => at its truest sense implies that the subject does not meet all qualifications of "Thing"
basically not => at its truest sense implies that the subject does meet some qualifications of "Thing"

You're right, the negations are not perfectly equivalent. Though the phrases do not contain any metric of quantification, both simply imply that the the subject contains at least some but not all of their comparative criteria.

Saying this games rendering method is basically 4K, is no more true than saying this games rendering method is basically not 4K, As the term "basically" has no actual quantification.

 

[Lister]

I'd take sparse rendering/checkboarding at 4K over native 4K any day fo the week, considering the impact on IQ and of course the impact on performance at shading all 8+ million pixels.

Now, sure, If I'd own 2 GTX 1080's and a 4K monitor I'd pick full rendering, but I only own 1

My Monitor is 3440x1440p, utlra wide and not quite as pixel dense as a full 4K TV (thank goodness because it's hard to run at ultra and 60 FPS at 4K in some games even on a 1080).

Watchdogs 2 with sparse rendering turned on looks nearly identical to it running without it, on my PC. The main issues are: barely noticeable artifacts, especially around edges/shadows when the camera is moving, and a small loss of sharpness. But it cuts down shader load on ym GPU so that I cna run the game at near ultra settings at 60 FPS.

When you're talking about sitting 6+ feet away from a TV, I really don't see the point of shading all 4K's pixels every frame. Odds are, aside from the artifacts, that you wouldn't notice the difference in softness unless you had a relaly big TV or were sitting very close (or had very, veyr good eyesight - lucky bastard).

Pretty much any PC game that offers it, and where my GPU can't max at 60 FPS, I'll be turning it on.

I think the confusion over image quality on the PRO and sparse rendering/checkerboarding stems from MANY Pro games using the technique NOT actually targetting a 4K frame. Many are running at 1800p, or lower, or using dynamic resolutions on top of sparse rendering.

This is going to significantly soften the frame, as you are losing image detail, and in some instance might make artifacts more pronounced. You have to remember that the main performance boost of checkerboarding is a lower requirement on the GPU compute/shaders (IIRC), but the GPU still needs to do a lot fo other things. That 4.8 TF GPU has it's limitations.

 

[RoboPlato]

I know checkerboard doesn't look as good as true 4k. NBA 2k17 looks sharper and cleaner than WD2 because it. It doesn't fool me.

Watch Dogs 2 is 1800c so it's a step down res wise anyway. Rise of the Tomb Raider is the only game I know for sure is 2160c without a dynamic res implementation.