Digital Foundry Article Technical Discussion Archive [2013]

Status
Not open for further replies.
Yeah, I'm a little surprised to see PC initially exceeding essentially a 8GB HD 7850 that is PS4.

I guess I thought unified 8GB would provide more benefits. Most enthusiast video cards on PC probably are still around 2GB VRAM.

But then you remember the next gen consoles are "really" only 5GB machines, which has to be divided between CPU and GPU, for one thing.

The PS4, the console that shot "high" this gen, is still really quite low, as evidenced by it's 140 watt max power draw.

I figure once the 1st party wares from PS4/XB1 start to show-up next year and beyond, we should start to see some amazing graphics, very comparable to the PC. I just don't see multi-platform developers doing that, at this point and time.
 
Even apart from the obvious differences with respect to effects, PQ differences are pretty big in those XBO vs. PS4 NFS comparison pics.

Certainly NOT the same native resolution, too.

Just look at these 200% crops:

2dqvdav.jpg

What is the source of those caps? If it's a forum poster on neogaf it's probably doctored to make the X1 version look worse. I wouldn't trust anything from that forum. They routinely alter/edit gifs to make X1 look worse/PS4 better I know.

Looking over the pics DF provides there's nothing like the difference in your pic *shrug*.

I think I kind of hate DF's new comparison system though. Not sure. Maybe just give me two full resolution pics instead of the cumbersome magnifying glass tool? Not sure.

Here's my attempt at a unaltered 200% zoom of DF's pics trying to find a similar angle to your cap

uzYAZdb.jpg

Edit: apparently it's some sort of blur effect to replace the lack of bokeh in those pre race scenes on X1.

The blur filter (that substitutes for the lack of bokeh) is low resolution and gives the appearance of geometry rendered at its res.

Actual gameplay where this no bokeh or blur on either platform is the same res.
 
What is the source of those caps?
I just used the pics posted by Ketto a few posts above.

Checked the original DF gallery - and the pictures aren't "doctored":

PS4

XBO

Having a closer look at some of the other examples in their gallery, the differences don't seem as big in most of the other comparison pics, though (especially the gameplay pics are way closer with respect to PQ).
 
Yet another game that comes sporting higher quality effects in the PC version. Despite the power difference in the platforms I'm honestly surprised that so many games are doing this. I expected parity with the new consoles for at least a year or two before PC versions started to add better graphics. And 60fps on a 270X (a midrange GPU)... nice!
Has a console ever had 1:1 graphics with PC at launch? I think the PS4 will hold its own for 2-3 years.

What is the source of those caps? If it's a forum poster on neogaf it's probably doctored to make the X1 version look worse. I wouldn't trust anything from that forum. They routinely alter/edit gifs to make X1 look worse/PS4 better I know.

Looking over the pics DF provides there's nothing like the difference in your pic *shrug*.

I think I kind of hate DF's new comparison system though. Not sure. Maybe just give me two full resolution pics instead of the cumbersome magnifying glass tool? Not sure.
The screen was taken from DF.

The blur filter on the X1 version seems to hurt IQ on some of the screens. These in particular show a pretty big difference.

PS4

X1

PS4

X1
 
Last edited by a moderator:
Check my latest edit guys. Evidently it's some blur filter to substitute for lack of bokeh on X1, but only on pre race scene or whatever not normal gameplay. And the full pics confirm that with bokeh in the PS4 version of the full pic.
 
Has a console ever had 1:1 graphics with PC at launch? I think the PS4 will hold its own for 2-3 years.

X360 in theory could, RAM and CPU aside aside, Xenos was close to top PC cards of late 2005 bar none, which is far from the situation today (melancholy sigh).

In practice I seem to recall PC showing an advantage early though, Oblivion comes to mind?

There's enough diminishing returns PS4/X1 should be ok. Often PS3/360 can hold their own with PC (minus 720P), how much more PS4/X1? NFS Rivals is a good indication. There are differences in favor of PC, but they are slight and reserved for the nitpickers.
 
Check my latest edit guys. Evidently it's some blur filter to substitute for lack of bokeh on X1, but only on pre race scene or whatever not normal gameplay. And the full pics confirm that with bokeh in the PS4 version of the full pic.
Unfortunately that's not the case. The two screens I posted above in my edit are during gameplay.
 
Has a console ever had 1:1 graphics with PC at launch? I think the PS4 will hold its own for 2-3 years.


The screen was taken from DF.

The blur filter on the X1 version seems to hurt IQ on some of the screens. These in particular show a pretty big difference.

PS4

X1

PS4

X1


The first set of pics was explained, the second it looks like probably the bokeh effect had faded out on PS4 while the blur was still in effect at that microsecond on X1, misleading comparison. It's clearly the start of a race.
Unfortunately that's not the case. The two screens I posted above in my edit are during gameplay.

Well as you know most of the 70+ pics dont show such a discrepancy so I'm not sure what you're trying to prove, it's clearly not the norm.

PS4 does look slightly better to me than X1 in most of DF's screens, but I wonder how much is power of suggestion.
 
NFS Rivals is a good indication. There are differences in favor of PC, but they are slight and reserved for the nitpickers.

Well, apart from 60fps... ;)

Other than that, the differences are indeed minor but no more so than the average difference between PS4 and XBO versions of titles. The fact that they're there at all though in (most) launch games on day one says a lot for how the next 6 years might look.
 
Well, apart from 60fps... ;)

Other than that, the differences are indeed minor but no more so than the average difference between PS4 and XBO versions of titles. The fact that they're there at all though in (most) launch games on day one says a lot for how the next 6 years might look.

XO should be very happy if it goes like NFS Rivals! Much less difference than COD:Ghosts for example.
 
Yet another game that comes sporting higher quality effects in the PC version. Despite the power difference in the platforms I'm honestly surprised that so many games are doing this. I expected parity with the new consoles for at least a year or two before PC versions started to add better graphics. And 60fps on a 270X (a midrange GPU)... nice!

The advantage of the Pc version is the fact that you can unlock the framerate. I'd quite like to see the difference between the One and the PS4 with unlocked frame rates. Judging by the DF Battlefield 4 framerate tests, the PS4 had a fairly consistent 10fps advantage, despite the fact it was rendering at a 50% higher resolution. I'd bet that Rivals is running at a much higher rate than the videos imply, just not a locked 60fps.
 
Yet another game that comes sporting higher quality effects in the PC version. Despite the power difference in the platforms I'm honestly surprised that so many games are doing this. I expected parity with the new consoles for at least a year or two before PC versions started to add better graphics. And 60fps on a 270X (a midrange GPU)... nice!

Also, let us not forget that COD: Ghosts is running at essentially a consistent 60fps on PS4 at 1920x1080, whereas a Titan struggled to achieve that with only slightly increased graphical effects. So let's not rule out these machines just yet. ;)
 
PS4 is in fact the sharpest of the 3!

Car aside, I much prefer the clarity of distance rendering of the Xbox One version over PS4. Look at the three barricades down the track, they are far too obscured on PS4 compared to One.

In this case, less is more.

This blur effect is just for the start of the course when the camera is turning around the car.

The PS4 version is in fact the sharpest of the 3 cause there is some kind of blur applied in the background for the X1 and PC version.

kztu.png


This blur on the X1 and PC has been missed by the DF analysis. But it is the first noticeable difference I have spoted between the 3 versions. Weird.

I have the impression that the blur on X1 is supposed to mask some missing AA, present on the PS4, but I could be wrong. The PS4 definitely has better edge detect AA than the X1 (even better than PC) I think, looking at the post lamp.
 
Last edited by a moderator:
Also, let us not forget that COD: Ghosts is running at essentially a consistent 60fps on PS4 at 1920x1080, whereas a Titan struggled to achieve that with only slightly increased graphical effects. So let's not rule out these machines just yet. ;)

Comparing performance at different quality settings tells you very little. The highest settings in PC games can often half performance for very little visual gain.

According to Digital Foundry there are quite a few improvements in the PC version of Ghosts (http://www.eurogamer.net/articles/digitalfoundry-call-of-duty-ghosts-next-gen-face-off) so you certainly wouldn't want to be comparing performance at the PC's highest settings. Knock a few settings down a little though and you see the likes of the 270X sitting comfortably above 60fps average:

http://www.tomshardware.com/reviews/call-of-duty-ghosts-pc-performance,3683-7.html

No need for anything close to a Titan there. Whether these settings are equivalent to the console though is anyone's guess. Note DF also states the PS4 version, while usually at 60fps can drop as low as 40fps under heavy load.
 
I have the impression that the blur on X1 is supposed to mask some missing AA, present on the PS4, but I could be wrong. The PS4 definitely has better edge detect AA than the X1 (even better than PC) I think, looking at the post lamp.

X1 even looks a tiny bit sharper than the PC in that shot too. Perhaps down to a difference in the AA method being used? From what I gather it's not selectable on the PC version, you just get whatever the devs baked into the game.
 
Comparing performance at different quality settings tells you very little. The highest settings in PC games can often half performance for very little visual gain.

According to Digital Foundry there are quite a few improvements in the PC version of Ghosts (http://www.eurogamer.net/articles/digitalfoundry-call-of-duty-ghosts-next-gen-face-off) so you certainly wouldn't want to be comparing performance at the PC's highest settings. Knock a few settings down a little though and you see the likes of the 270X sitting comfortably above 60fps average:

http://www.tomshardware.com/reviews/call-of-duty-ghosts-pc-performance,3683-7.html

No need for anything close to a Titan there. Whether these settings are equivalent to the console though is anyone's guess. Note DF also states the PS4 version, while usually at 60fps can drop as low as 40fps under heavy load.

It's very rare. I've played that game to death already and it's certainly not consistently dropping frames. I'd also suggest that the PS4 is running at higher quality than 'High' from that Tom's Hardware article. It's listing SSAO on low, AF on normal, and blur / terrain detail as 'off'. I'll take a look later today to see whether motion blur is on, but I'm pretty sure that terrain details must be on. It's a shame we can't directly measure them really, but my experience the PS4 is not as underpowered as a lot of people that only play PC games would have you believe.

Again, the PS4 version of that game (and indeed almost any game) run at locked framerates. This means that the averages framerates will also be lower because you can only measure the troughs and not the peaks.

FYI - I'm not saying the PS4 or Xbox One are near to a Titan, only that they shouldn't be underestimated on a multiformat game with a locked 30fps.
 
It's very rare. I've played that game to death already and it's certainly not consistently dropping frames. I'd also suggest that the PS4 is running at higher quality than 'High' from that Tom's Hardware article. It's listing SSAO on low, AF on normal, and blur / terrain detail as 'off'. I'll take a look later today to see whether motion blur is on, but I'm pretty sure that terrain details must be on. It's a shame we can't directly measure them really, but my experience the PS4 is not as underpowered as a lot of people that only play PC games would have you believe.

Again, the PS4 version of that game (and indeed almost any game) run at locked framerates. This means that the averages framerates will also be lower because you can only measure the troughs and not the peaks.

FYI - I'm not saying the PS4 or Xbox One are near to a Titan, only that they shouldn't be underestimated on a multiformat game with a locked 30fps.

While some settings from the Toms comparison are undoubtedly lower, it's possible that others are higher, for example the shadow advantages DF mentions (there's no mention of what settings shadows are set at in the Toms comparison). And while I agree you're correct about the fixed frame rate making average framerates difficult to compare, you can get a good idea from the minimum. If DG is saying the PS4 version can sometimes drop to 40 and the Toms comparison shows the low (in this scene) as being 49 (more like 51 if you look at the line graph and exclude the momentary blip) then you've got to conclude that the average on both must be pretty close.

Unfortunately without an exact match in quality settings, or at least an attempted match, further comparisons using Ghosts as a benchmark aren't going to offer much useful information.
 
Yet another game that comes sporting higher quality effects in the PC version. Despite the power difference in the platforms I'm honestly surprised that so many games are doing this. I expected parity with the new consoles for at least a year or two before PC versions started to add better graphics. And 60fps on a 270X (a midrange GPU)... nice!

I'm sure the PS4 drivers aren't in a completely finished state. I don't expect this to happen for some time, so why are people acting like this is the maximum performance?
 
I'm sure the PS4 drivers aren't in a completely finished state. I don't expect this to happen for some time, so why are people acting like this is the maximum performance?

Who said anything about maximum performance? And do you really want to get into which system is able to make more optimal use of it's peak performance?

The point is that were seeing launch games across the board sporting higher settings in the PC versions. That didn't happen at the start of last generation so it could be a sign of how this generation will play out.
 
If DF is saying the PS4 version can sometimes drop to 40 and the Toms comparison shows the low (in this scene) as being 49 (more like 51 if you look at the line graph and exclude the momentary blip) then you've got to conclude that the average on both must be pretty close.

I certainly wouldn't exclude a 'momentary blip' on the PC, but retain the same blip from the PS4. Nor would I suggest that the shadows on High are definitely of a higher quality than those on the PS4 and consider the comparisons the same. Blur is an effect that is always there and in my experience is fairly expensive in terms of rendering budget (if it's even present on PS4 - I need to check). I honestly don't believe that you can use the Tom's Hardware article as a direct comparison.

Rivals is directly comparable because the settings would appear essentially the same across PC and PS4. If the framerate were unlocked, we'd actually have a very good idea of exactly which PC GPU the consoles are the equivalent to. We know what they're equivalent to in terms of raw numbers, but the OSs would play a part.
 
Status
Not open for further replies.
Back
Top