Support NeoGAF

[Devilgunman]

Most likely because they couldn't. blending two frames together is ROP dependent and the XB1 may struggle with it.

I don't know if anyone has noticed but there is something wrong with specular reflections on the XB1 version. Like it is missing tone mapping or something.

Omg X1 AA..

 

[Durante]

I still think the temporal AA implementation on PS4 is simply buggy. Temporal AA which uses screen-space motion vectors correctly should not look like that.

 

[MilkyJoe]

Nvm

 

[Dictator93]

Yeah I know this game has issues but what I mean is whether or not performance of the game on consoles is affected because they are using AMD GPU.

From what I have heard about why it performs so poorly on AMD GPUs has more to do with the Driver set up on Windows PCs (poor driver multithreading for deferred contexts or whatever). Not because GCN is bad at tesselation or something.

Consoles have a different APIs which go around the limitations affecting AMD PCs right now.

I still think the temporal AA implementation on PS4 is simply buggy. Temporal AA which uses screen-space motion vectors correctly should not look like that.

Perhaps it is a homage to MGS2.

 

[Fafalada]

Perhaps it is a homage to MGS2.

Ironically that was considered a type of AA as well, even though it was just Blend Circuit accumulating frames.

 

[HTupolev]

I still think the temporal AA implementation on PS4 is simply buggy. Temporal AA which uses screen-space motion vectors correctly should not look like that.

Even Halo Reach's implementation, which doesn't reproject, has mechanisms to turn off blending when there's motion (said mechanisms occasionally missed things, but they were there).

 

[hesido]

I still think the temporal AA implementation on PS4 is simply buggy. Temporal AA which uses screen-space motion vectors correctly should not look like that.

That's what I thought but given the transparency of the devs, they never acknowledged it as a bug but a feature that enhances the image qualiy. They also say they have reduced the motion velocity because it would be too blurry along with the ghosting that comes with their temporal AA, reduced it to the point that eliminates any visible motion blur where you expect to have some, which reduces the sense of speed. So they acknowledge that there's ghosting and they have changed other settings to mitigate it, instead of trying to fix that bug. (So I think it may be a case of bad design choice with the worst temporal AA implementation of the last 10 or so years.)

They say they are also doing motion blur with as many samples at least as XboxOne, so they have all information at hand (motion vectors). All they had to do is not blend for any pixel above a certain threshold of motion. This would eliminate ghosting and since pixels that are moving above a certain threshold would already be subjected to motion blur, which wouldn't need that extra anti-alias pass anyway!

There's this small possibility that, if the devs decide against simply an option to completely turn of temporal AA, PS4 can have a proper temporal AA with minimal artefacting and with proper motion blur. This would help reduce jaggies where you need to (e.g. around the car edges from behind view, or the cockpit) while maintaining a blur free surrounding with proper motion blur.

Edit: Has the infamous sharpen filter come back for XboxOne?

 

[Insane Metal]

I'm actually surprised with the console versions (well, the PS4 one at least) as the PC version runs like shit on AMD hardware.

 

[Fredrik]

I still think DriveClub is the better looking game with its dynamic lighting and all

At 30fps. The real challenge this generation will be to get a 1080p racer on these consoles to look nice with dynamic weather and lighting etc at locked 60fps. Turn 10 will probably try to do it with FM6, possibly with some cheating, but I'm sure PD will go 10 steps extra as usual with GT7, hopefully without sacrificing the framerate this time.

 

[SMSRenderTeam]

Just a quick nod to Digital Foundry from the SMS Render Team for this article and their linking of the various posts that we did over the last few weeks. It is a very fair and accurate Face Off between the platforms and a great piece of technical Journalism as a result - we are working hard to address issues raised on the initial release. Stay tuned ... I can tell you that Xbox One will gain at least 4-7% performance in a patch shortly and address some of the controller issues. Lots more discussion over at forum.projectcarsgame.com.

 

[Metfanant]

Just a quick nod to Digital Foundry from the SMS Render Team for this article and their linking of the various posts that we did over the last few weeks. It is a very fair and accurate Face Off between the platforms and a great piece of technical Journalism as a result - we are working hard to address issues raised on the initial release. Stay tuned ... I can tell you that Xbox One will gain at least 4-7% performance in a patch shortly and address some of the controller issues. Lots more discussion over at forum.projectcarsgame.com.

How about the apparent discrepancy in the levels of AF found in the console versions?

I know you've addressed the issue in the past and gave us some hard numbers, there does seem to be some difference between the two

 

[Piggus]

Wiiu gonna be the worst one I bet

World's safest bet. I don't know how they're doing a Wii U version after they said the PS3/360 weren't powerful enough.

 

[nib95]

At 30fps. The real challenge this generation will be to get a 1080p racer on these consoles to look nice with dynamic weather and lighting etc at locked 60fps. Turn 10 will probably try to do it with FM6, possibly with some cheating, but I'm sure PD will go 10 steps extra as usual with GT7, hopefully without sacrificing the framerate this time.

I doubt Driveclub's frame rate is that far removed from ProjectCars to be honest. If the frame rate were unlocked, I'd imagine it'd reach 60fps quite often. Even with all cars on screen at a time, plus snow, rain, a storm, lightning, massive backdrops and everything else, it doesn't ever drop a single frame below 30fps, ever. Lest we forget ProjectCars can drop to the 30's in the most extreme of situations.

ProjectCars is a fine looking game, but I think it being multi-platform probably held it back.

 

[dreampeppers99]

you're mean! I only own a wii u, I was dreaming at least something like.

 

[Sassen]

Consoles have broken wipers, how can one drive a supercar in the rain without wipers?!?

Rain-X! <_<

 

[Elfforkusu]

I'm pretty interested to see how far they'll be able to push the Wii U version.

As for the results, nothing surprising -- except perhaps that AMD struggles so much on the PC side. But is that really a surprise?

 

[RoboPlato]

Seems odd that they estimate AF to be at 4x on both consoles but the PS4 version looks like it has a lower amount. How is that possible?

That temporal AA on PS4 looks way off as well. Haven't seen ghosting like that since Halo Reach.

 

[Brad Grenz]

Can't imagine how this game ended up poorly optimized for AMD on PC...

 

[Qassim]

Can't imagine how this game ended up poorly optimized for AMD on PC...

From the Head of Slighty Mad:

&#8220;We&#8217;ve provided AMD with 20 keys for game testing as they work on the driver side. But you only have to look at the lesser hardware in the consoles to see how optimised we are on AMD based chips.

We&#8217;re reaching out to AMD with all of our efforts. We&#8217;ve provided them 20 keys as I say. They were invited to work with us for years, looking through company mails the last I can see they (AMD) talked to us was October of last year.

Categorically, Nvidia have not paid us a penny. They have though been very forthcoming with support and co-marketing work at their instigation. We&#8217;ve had emails back and forth with them yesterday also. I reiterate that this is mainly a driver issue but we&#8217;ll obviously do anything we can from our side.&#8221;

 

[Shpeshal Nick]

Clear win for PS4 (on consoles)

Didn't quite read that way to me.

But I guess that depends on which takes precedence for any given individual resolution bump with some frame drops instead of a few more frame drops with lower res vs worse AF and ghosting or better AF no ghosting.

Which do you care about more?

 

[nib95]

Didn't quite read that way to me.

But I guess that depends on which takes precedence for any given individual resolution bump with some frame drops instead of a few more frame drops with lower res vs worse AF and ghosting or better AF no ghosting.

Which do you care about more?

Still seems like a clear win overall.

PS4

+ Higher resolution
+ Much cleaner image (AA)
+ Better frame rate
+ Better shadows
+ Less tearing

Xbox One

+ Slightly more AF
+ No ghosting

 

[impact]

I'm pretty sad to hear about such big Nvidia advantage on PC. That's unacceptable.

Isn't this commonplace by now with most new releases?

It's like being sad to hear about such a big PS4 advantage on console

 

[SmartBase]

Great, going to have to wait for AMD's lightning quick driver team to fix performance before I even go near the game.

 

[JAYSIMPLE]

I wouldn't even bother puttingbthings like Clear win etc. A win is a win tho. The ps4 stumbles over the line with an xbox riding a jaggi following a distance behind. The ps4 raises his hands in victory to be told the pc finished a lap ahead lol

This is a joke please don't take this post too seriously

 

[Brad Grenz]

From the Head of Slighty Mad:

Right, so they had a financial arrangement with nVidia, just not one where cash was given directly to Slightly Mad, and AMD has had keys for maybe a week. What do you think was the content of that email from October? "We've accepted a co-marketing deal with nVidia and will no longer be able to work with you on optimizing our game until after release."

 

[Beautiful Ninja]

I wouldn't even bother puttingbthings like Clear win etc. A win is a win tho. The ps4 stumbles over the line with an xbox riding a jaggi following a distance behind. The ps4 raises his hands in victory to be told the pc finished a lap ahead lol

This is a joke please don't take this post too seriously

Well maybe if the PC is running an nVidia GPU. An AMD user sadly spun out on the first turn and is waiting to be towed back to the pit where they can hopefully fix the car.

 

[visiblemode]

Black Painted Fence ps4 exclusive.

Hah! Are you telling me we are about to embark on fencegate?...because, well fences usually have gates to begin with and it'll just get confusing.

 

[Qassim]

Right, so they had a financial arrangement with nVidia, just not one where cash was given directly to Slightly Mad, and AMD has had keys for maybe a week. What do you think was the content of that email from October? "We've accepted a co-marketing deal with nVidia and will no longer be able to work with you on optimizing our game until after release."

They were invited to work with us for years

It's really hard to decide who to believe, you, with your random and baseless conjecture, or those with actual direct experience of the matter.

 

[Brad Grenz]

It's really hard to decide who to believe, you, with your random and baseless conjecture, or those with actual direct experience of the matter.

Well, my statement isn't a self serving attempt to disguise the fact that I sold out a third of my paying customers on PC for financial gain, and is informed by many years of malfeasance in nVidia's developer relations strategy.

 

[Qassim]

Well, my statement isn't a self serving attempt to disguise the fact that I sold out a third of my paying customers on PC for financial gain, and is informed by many years of malfeasance in nVidia's developer relations strategy.

Oh so you have no evidence, that's a shame, I was looking forward to some actual insight.

 

[hesido]

Well, my statement isn't a self serving attempt to disguise the fact that I sold out a third of my paying customers on PC for financial gain, and is informed by many years of malfeasance in nVidia's developer relations strategy.

It could also be that AMD might have stretched too thin, they had been hard at work for Mantle, maybe they are still providing some sort of support for the consoles.. But it may also be seen like a lost cause to them after learning they have this deal with NVidia, maybe that's why they weren't so much interested in helping out.

 

[Brad Grenz]

Oh so you have no evidence, that's a shame, I was looking forward to some actual insight.

All the evidence I need is in the carefully qualified statement he made about "not receiving a cent" which leaves the door open for millions in free marketing and engineering support, and him going on and on about 20 keys that we can reasonably assume were generated at the time of the game's actual release (last week). Perhaps you take pleasure in being snowed under by careful PR statements, but nVidia's entire TWIMTBP program is built on keeping final code out of AMD's hands until the last minute so everyone keeps wondering why AMD's drivers are "so bad".

 

[Qassim]

All the evidence I need is in the carefully qualified statement he made about "not receiving a cent" which leaves the door open for millions in free marketing and engineering support, and him going on and on about 20 keys that we can reasonably assume were generated at the time of the game's actual release (last week). Perhaps you take pleasure in being snowed under by careful PR statements, but nVidia's entire TWIMTBP program is built on keeping final code out of AMD's hands until the last minute so everyone keeps wondering why AMD's drivers are "so bad".

AMD claim this, NVIDIA claim the opposite. NVIDIA say they do not stipulate how competitors have access to the games builds pre-release on NVIDIA GameWorks (or otherwise) games and that it is entirely up to the developers. I'm taking the stance that we have no evidence either way, but AMD are the ones making those claims that aren't backed up by evidence. Why are you choosing to believe AMD's careful PR statements but not the developers of the game themselves?

Given data we have to actually support the fact that AMD's DX11 drivers introduce more CPU overhead resulting in reduced performance especially on games that are more CPU dependant, it's hardly all that difficult to believe that AMD just simply struggle to keep up without the competitor sabotaging them in matters relating to their drivers.

 

[Conduit]

Just a quick nod to Digital Foundry from the SMS Render Team for this article and their linking of the various posts that we did over the last few weeks. It is a very fair and accurate Face Off between the platforms and a great piece of technical Journalism as a result - we are working hard to address issues raised on the initial release. Stay tuned ... I can tell you that Xbox One will gain at least 4-7% performance in a patch shortly and address some of the controller issues. Lots more discussion over at forum.projectcarsgame.com.

Xbone version to be superior confirmed!

 

[Shpeshal Nick]

Still seems like a clear win overall.

PS4

+ Higher resolution
+ Much cleaner image (AA)
+ Better frame rate
+ Better shadows
+ Less tearing

Xbox One

+ Slightly more AF
+ No ghosting

Higher res for most people isn't all that noticeable, the frame drops are on both, just less on PS4, they actually mention some shadows being better on Xbox don't they? Tearing? Fair enough.

Having said that, you played a game with constant ghosting? It's easily for me, the worst thing on the list of issues. Halo Reach was really fucking annoying with it.

 

[Javin98]

I doubt Driveclub's frame rate is that far removed from ProjectCars to be honest. If the frame rate were unlocked, I'd imagine it'd reach 60fps quite often. Even with all cars on screen at a time, plus snow, rain, a storm, lightning, massive backdrops and everything else, it doesn't ever drop a single frame below 30fps, ever. Lest we forget ProjectCars can drop to the 30's in the most extreme of situations.

ProjectCars is a fine looking game, but I think it being multi-platform probably held it back.

Agreed completely. It's amazing how Driveclub holds a completely solid 30FPS no matter what happens on screen. I'm willing to bet in heavy storms, the frame rate is hovering around 30-40, which is very similar to what PCars runs on PS4 in a storm. No offense to SMS, but DC is a far better looking game running at close to PCars' frame rate IMO.

 

[HTupolev]

It's easily for me, the worst thing on the list of issues. Halo Reach was really fucking annoying with it.

Reach doesn't even have constant ghosting. It disables the blending at pixels where motion is detected; this sometimes breaks on dynamic objects, especially in cutscenes, but usually works fine.

 

[Three]

Higher res for most people isn't all that noticeable, the frame drops are on both, just less on PS4, they actually mention some shadows being better on Xbox don't they? Tearing? Fair enough.

Having said that, you played a game with constant ghosting? It's easily for me, the worst thing on the list of issues. Halo Reach was really fucking annoying with it.

None of the shadows are better on the XB1, not sure where you read that.

Agreed completely. It's amazing how Driveclub holds a completely solid 30FPS no matter what happens on screen. I'm willing to bet in heavy storms, the frame rate is hovering around 30-40, which is very similar to what PCars runs on PS4 in a storm. No offense to SMS, but DC is a far better looking game running at close to PCars' frame rate IMO.

There is absolutely no point in speculation like this. It has no goal. Especially as we have no idea about what driveclubs unlocked framerate will be or how much more work PCars is doing in simulating its physics.

 

[Brad Grenz]

AMD claim this, NVIDIA claim the opposite. NVIDIA say they do not stipulate how competitors have access to the games builds pre-release on NVIDIA GameWorks (or otherwise) games and that it is entirely up to the developers. I'm taking the stance that we have no evidence either way, but AMD are the ones making those claims that aren't backed up by evidence. Why are you choosing to believe AMD's careful PR statements but not the developers of the game themselves?

Given data we have to actually support the fact that AMD's DX11 drivers introduce more CPU overhead resulting in reduced performance especially on games that are more CPU dependant, it's hardly all that difficult to believe that AMD just simply struggle to keep up without the competitor sabotaging them in matters relating to their drivers.

No, you were taking the stance of blindly repeating what the developers said to defend the state of their game and their relationship with nVidia.

 

[hesido]

Higher res for most people isn't all that noticeable, the frame drops are on both, just less on PS4, they actually mention some shadows being better on Xbox don't they? Tearing? Fair enough.

Having said that, you played a game with constant ghosting? It's easily for me, the worst thing on the list of issues. Halo Reach was really fucking annoying with it.

Shadows seem to be better on PS4. With all due respect, Halo Reach's ghosting problem was not as bad. The problem was it manifested itself more on the gun, which takes up quite a lot of the screen. It wasn't like every single frame being ghosted, so its temporal AA approach was better. Interestingly, this is what DF had to say on the issue back then:

If there's one single complaint we have about the beta, this is it. The effect can look reminiscent of the ages old ghosting LCD problem - a technological era we don't really need reminding about!

This was the old DF. It perfectly summarizes how ghosting effect is perceived (at least for those who do), and this is for a game whose AA solution worked fine most of the time.. The new DF mistook the ghosting as a motion blur artifact. And later when pointed out it's the AA, doesn't even mention how it compares to other temporal AA solutions.

 

[Renekton]

From the Head of Slighty Mad:

Needs more info hmm.

- "Nvidia have not paid us a penny" could be temporal.
- The co-marketing will have eventual positive financial effects on SMS, regardless.
- Does not mention when AMD were given the keys.
- Does not mention if the October mail was pending response from SMS

Everything is conjecture, I guess because SMS's statement isn't 100% clear either (not that they owe us that).

 

[_machine]

No, you were taking the stance of blindly repeating what the developers said to defend the state of their game and their relationship with nVidia.

Well how about those that saw how it was handled throughout the development? Again, we know it's not the case; we've seen how they've handled the improvments and we know that AMD is working on their end because it's a driver issue that's affecting CPU overhead. There were plenty of AMD specific improvments done throughout the development, probably more so than NVidia, yet currently the issues is on the driver side. These conspiracy theories are absolutely ridiculous.

 

[Durante]

No, you were taking the stance of blindly repeating what the developers said

While you are fabricating things out of whole cloth.

What we know from independent evaluations is that AMD's driver overhead in DX11 is more than twice as high as NV's [1], and that they completely fail to effectively support DX11 multithreading.

 

[system11]

Big issue is this:

and this:

and this:

and this:

Some people suggested they can't see it or that you don't see it in motion.

Ummm....... Insert J-Lo 'yeah ok' gif here. It's easily visible during play, looks like some kind of psychadelic video on corners, all we need is for the colours to start cycling through the spectum and some Magic Bus.

 

[Javin98]

There is absolutely no point in speculation like this. It has no goal. Especially as we have no idea about what driveclubs unlocked framerate will be or how much more work PCars is doing in simulating its physics.

It's not really speculation when DC never drops a frame, is it? I mean, at the very least for this to happen, the frame rate during a storm has to be 30-35, which isn't too far off what PCars is running. And I didn't mention physics anywhere...

 

[Durante]

It's not really speculation when DC never drops a frame, is it? I mean, at the very least for this to happen, the frame rate during a storm has to be 30-35, which isn't too far off what PCars is running. And I didn't mention physics anywhere...

Doesn't DC have a limit of like 12 cars?
According to DF, PC only drops below 50 FPS in stress test scenarios with 30+ cars.

 

[Nzyme32]

Big issue is this:

and this:

and this:

and this:

Some people suggested they can't see it or that you don't see it in motion.

Ummm....... Insert J-Lo 'yeah ok' gif here. It's easily visible during play, looks like some kind of psychadelic video on corners, all we need is for the colours to start cycling through the spectum and some Magic Bus.

This is already explained in the OP/article. It's an AA workaround on ps4.

"There's no way consoles can compete with that, but the PS4 edition has an interesting workaround. In addition to EQAA, a 'temporal aliasing' pass is added to the Sony release, blending the previous frame with the current one to reduce flicker on moving objects. During gameplay this works surprisingly well in minimising pixel crawl, but the implementation here is a divisive one. The downside is simple: it creates a ghost image that's very easy to spot in static images, and the effect is noticeable in motion too."

 

[system11]

This is already explained in the OP/article. It's an AA workaround on ps4.

"There's no way consoles can compete with that, but the PS4 edition has an interesting workaround. In addition to EQAA, a 'temporal aliasing' pass is added to the Sony release, blending the previous frame with the current one to reduce flicker on moving objects. During gameplay this works surprisingly well in minimising pixel crawl, but the implementation here is a divisive one. The downside is simple: it creates a ghost image that's very easy to spot in static images, and the effect is noticeable in motion too."

Yes, I know, but it doesn't make it any less awful to look at. It doesn't work 'surprisingly well' as they claim, it just makes the game look horrid. The thing that really amazes me is how hardly any of the reviews mention this. I suppose it's just people used to the trails left by 30fps and image persistence who don't know games shouldn't look like this in motion...

It's easily on a par with chromatic aberration in terms of bad graphical decisions. I eagerly await the patch, I've put the game on the shelf for now, it's just horrible at the moment with this going on.

 

[hesido]

This is already explained in the OP/article. It's an AA workaround on ps4.

"There's no way consoles can compete with that, but the PS4 edition has an interesting workaround. In addition to EQAA, a 'temporal aliasing' pass is added to the Sony release, blending the previous frame with the current one to reduce flicker on moving objects. During gameplay this works surprisingly well in minimising pixel crawl, but the implementation here is a divisive one. The downside is simple: it creates a ghost image that's very easy to spot in static images, and the effect is noticeable in motion too."

Why it is caused is not much important. It's a huge loss of image quality all the same, and no game in future should touch this type of AA with a 10 feet pole.

However, we should all agree that only a few of us can notice it. I've just checked the PS4 part of the public forum and the first two pages did not even contain a thread about this problem. In a game where you are constantly moving, the AF discussions seem funny to me.

 

[farmerboy]

Big issue is this:

Ummm....... Insert J-Lo 'yeah ok' gif here. It's easily visible during play, looks like some kind of psychadelic video on corners, all we need is for the colours to start cycling through the spectum and some Magic Bus.

Serious, not sure if I see this ghosting at all the way its represented in these photos. What I do notice is a stutter, especially as I turn corners. Haven't taken a screen grab, but if I do, is this what I'll see?