Digital Foundry Face-Off @ Performance Analysis (Latest: DriveClub Preview 2)

[Dr. Pezus]

Wow the 30fps shadow there looks crystal clear

[ethomaz]

The funny thing I find is that the blurry shadows in the PS3 version actually look more realistic than the other 2.

My shadows behind sun looks more like the 30fps... I'm on street right now and tested

[ethomaz]

Eurogamer revisited Titanfall.

http://www.eurogamer.net/articles/digitalfoundry-2014-performance-analysis-titanfall-revisited

Nothing changed.

Well, it's visible now, but what's clear from an initial playthrough is that resolution remains at the 1408x792 pixel count of the launch code with the same 2x MSAA (multi-sampling anti-aliasing). Running key clips through our performance analysis tools, our take on the Update Five is that there is a clear improvement in frame-rates, but we do wonder if the average gamer is likely to notice - baseline performance is still mostly in the 40-60fps range, and there are still dips into the 30s when the game is really struggling under the load.

[Raven]

The fact that Titanfall still dips down into the 30's with a 792p native resolution using an old engine and not attempting anything spectacular visually is either a failure of the X1's hardware, a failure of Respawn to optimize, or both.

[BananaKing]

The fact that Titanfall still dips down into the 30's with a 792p native resolution using an old engine and not attempting anything spectacular visually is either a failure of the X1's hardware, a failure of Respawn to optimize, or both.

most likely a failure on respawns part. Forza 5 has a rock solid 60fps if i remember correctly, and that was a launch tittle.

[Dr. Pezus]

Well TF is unoptimized on PC as well so yeah.

@Banana: Forza has less going on than TF though, but still it is more impressive since it's 1080p too.

[ethomaz]

most likely a failure on respawns part. Forza 5 has a rock solid 60fps if i remember correctly, and that was a launch tittle.

To be fair Forza 5 was not impressive in any technical term... it was like a remaster with low content.

[ethomaz]

Digital Foundry vs Metro Redux (Thanks Kayant@GAF)

For this piece, we want to give you a taste of the work 4A has done, so we'll be stacking up the original Xbox 360 version of Metro 2033 with the Reduxed Xbox One game - as there was no PS3 release there, it's the natural choice. For Metro Last Light, we're breaking out the PlayStation 4 edition of the Redux and comparing it directly with PlayStation 3.

The result is fascinating. Comparing the original Xbox 360 and Redux Xbox One, we strongly suspect that where Last Light assets could help out, 4A used them where they could as direct like-for-like replacements: the mutated wild-life - showcased effectively in 2033's initial level - looks much better and character models are routinely switched out for far superior versions. Outdoor environments share the same basic geography (although we note some subtle tweaks) but the detail level is ramped up significantly - debris and other new geometry elements pepper the landscapes and additional blasted dead trees and foliage are commonplace.

The further you play into the Redux version of Metro 2033, the more the differences stack up. The subterranean sequences still hold up rather well on Xbox 360 - especially in terms of lighting - but the Redux improves things with small environmental detail boosts backed up by far superior precision in effects work, much higher detail artwork and what we suspect is a completely revised take on the lighting.

The original Xbox 360 version holds up surprisingly well considering it's now over four years old - but it's interesting to note that the areas that have aged the most aren't really so much different in the Redux. Most notable is character definition and in the way they move. Not all of the character models are swapped out with higher detail alternatives and those do stick out somewhat. Facial and body movement is a step behind Metro Last Light, which in turn falls short of the more advanced performance capture seen in latter-day last-gen titles.

Comparing the Redux PS4 version of Last Light to its PS3 predecessor is an interesting experience. It's the same game but the level of refinement the new version brings is obviously a highly worthwhile upgrade. The raw mathematics explain much of the differential: 4A opted for a 1152x640 framebuffer on PS3 (1280x672 on 360), while the PlayStation 4 offers up the full monty at native 1080p. The additional pixel-count gives Last Light's intensely detailed texture work the real estate to shine - this more than anything offers up a good proportion of the top-end PC experience on your new console.

In an age where marketing often promises 60fps only for the final product to fall short, it's refreshing to see that 4A's approach is brutally uncompromising. There is indeed no 60fps 'target', or even what we like to call a 'perceptual' 60fps - where performance often dips below the ideal, but not to any great, noticeable detriment to the experience. This looked absolutely rock solid.

On PlayStation 4 running Last Light, we see a solid 60fps from start to finish (screen-tear is only evident on FMV cinematics, curiously enough). Across 26,000 sample frames, each is entirely unique. For our Metro 2033 Xbox One captures, just two frames are dropped - one accompanied by a fleeting screen-tear cascade, again proving that adaptive v-sync is in place. What this means is that the Metro Redux offers up a consistency in visual and controller feedback that is transformative in nature, significantly improving the feel of the interface between player and game. In a first-person shooter, that kind of improvement is priceless.

We'll cover like-for-like performance on Xbox One and PlayStation 4 in a forthcoming pre-launch update, but we're going into that testing with the expectation of very close results. Differences kick in at the resolution level: PS4 hits its 60fps target at full 1080p, while Xbox One currently stands at a curious 912p native resolution - that would be something in the region of 1620x912 (assuming square pixels). The original plan for Xbox One was to ship at 900p, but the June XDK update (returning the Kinect GPU resources to developers) has allowed for a tiny resolution boost - our guess here is that 4A opted to bank the additional resource to help lock down that all-important frame-rate rather than really push the pixel-count. If so, that's the right trade.

Summary

PS4

• 1920x1080p @ 60fps
• Rock solid framerate

Xbox One

• 1620x912 @ 60fps
• Rock solid framerate

PS3

• 1152x640 @ 30fps

Xbox 360

• 1280x672 @ 30fps

[ethomaz]

DAT June SDK upgrade.

BTW Redux seems like a easy buy for me on PS4 even more after Pezus saying good things about it on PC... I'm glad the 1080p@60fps experience is perfect on PS4.

[ethomaz]

I will update after but Diablo 3 analyst is out.

Funny because they first said the PS4 have framerate issues (it was a confirmed bug by Blizzard) and Xbox One was running at 900p solid... after Day one patch the PS4 runs solid and Xbox One 1080p with framerate dips.

And we have the first word about resolution from MS:

"We did find it challenging early on to get it to 1080p. That's why we made the decision to drop to 900. That's what we demoed and were showing around E3 time. And Microsoft was just like, 'This is unacceptable. You need to figure out a way to get a better resolution.' So we worked with them directly, they gave us a code update to let us get to full 1080p."

http://www.eurogamer.net/articles/digitalfoundry-2014-diablo-3-performance-analysis
http://www.eurogamer.net/articles/digitalfoundry-2014-diablo-3-xbox-one-1080p-patch-performance-analysis

[Shinobi-san]

Another win for PS4...although i was hoping 4A would push the graphics a bit more in comparison to PC. Surely they can squeeze more out of the hardware of the PS4.

Perhaps we will see a major bump with their next Metro release. Amazing how this franchise has managed to become pretty successful and it was initially done by a very small team.

Sometimes it really pays off to push the graphical boundaries on PC, it automatically gives the game exposure amongst the PC fanbase even if its just a benchmark...this is where Crytek failed horrible i wonder why they dont/never did realise this?

If Crytek had made a Crysis 2 that focused a lot on the strengths of the Crysis franchise and on pushing graphics...things would have been better for them. And seeing how the PC fanbase is moving more towards legally buying software..its a good market to be strong in. They threw all that away...

[ethomaz]

Face-Off: Diablo 3: Ultimate Evil Edition (Thanks Ethelwulf@GAF)

With a full 1080p full HD resolution now in place for Diablo 3 on both PS4 and Xbox One, the time has come to investigate each in their fully updated states. After all, we've seen performance drops on Microsoft's platform since its bump from 900p to 1080p, but does Sony's hardware suffer similar issues at the same stress points? And from a visual perspective, does Blizzard keep all the bells and whistles of a maxed-out PC playthrough for each?

Compared to PC on its highest settings, the fully patched PS4 and Xbox One games run with a no-compromise approach. We get the high quality, smoothed dynamic shadows under each character here - with no pixellation seen on the lower settings. The core texture-work around the Highlands is beautifully preserved too, and explosive spell effects during cut-scenes show little change to alpha buffer resolution, It's all intact.

Given parity is essentially achieved everywhere else, has the touted Xbox One's resolution boost changed anything in performance testing? With both running at a full 1080p, we see a clear margin when running the New Tristram Gates, during the Skeleton King boss fight, and during packed skirmishes on the South Highlands plains. It reaches as low as 52fps on Xbox One, whereas with the rendering glitch now fully attended to on PS4, Sony's hardware never drops a frame in these areas.

But is this performance a deal-breaker for Xbox One? Given how infrequently we bumped into the issue, even during frenzied sand dune battles with two allied AI players in tow, the issue is surprisingly rare. Most drops tend to be imperceptible, and in the end, we resorted to scanning hours of footage to track most shifts downwards from the 60fps mark. A screen filled with effects-spewing creatures isn't necessarily the ticket to a frame-rate drop - even at its extremes the Xbox One holds up well save for a few choice moments.

In terms of frame-rate metrics, the Xbox One does show flickers of strain in meeting its recent 1080p upgrade - something never elicited by the PS4, which holds at 60fps in likewise tests. These drops, as infrequent as they are, make it harder to call the Xbox One release a purely locked 60fps experience in pragmatic terms, but it's clearly a solid performer. So much so, that it begs the question: could the PS4 release have been pushed any harder to take advantage of its own GPU headroom?[/b]

Summary

PS4

• 1920x1080p @ 60fps
• Rock solid framerate
• Compared to PC on its highest settings
• Post-process anti-aliasing focused on high contrast edges
• Dual Shock 4's features are utilised in subtle ways

Xbox One

• 1920x1080 @ 60fps
• Reaches as low as 52fps (most drops tend to be imperceptible)
• Compared to PC on its highest settings

PS3

• 1120x584 @ 60fps
• 2x MSAA
• Variable framerate targeting 60fps
• Tearing is 32% on average

Xbox 360

• 1120x584 @ 60fps
• 2x MSAA
• Variable framerate targeting 60fps
• Tearing is 24% on average
• Performance advantage over PS3

[u4gReservoirDogs]

Nothing we didn't know already. 

[ethomaz]

DF 4A (Metro) interview... amazing.

http://www.eurogamer.net/articles/digitalfoundry-2014-metro-redux-what-its-really-like-to-make-a-multi-platform-game

They really covered all the issues with gaming development on PS4 and Xbox One.

[ethomaz]

Performance Analysis: inFamous First Light (Thanks Kayant@GAF)

First Light takes place predominately across the first of the two islands available in Second Son while also offering a slew of indoor areas within the DUP containment facility. Prior to testing, we had hoped that, by utilising more confined environments, we might actually see frame-rates improve, perhaps even closing in on that 60fps mark. That turned out to not be the case at all, however, with the average frame-rate during these sections remaining mostly under 40fps. While the visuals in these sections remain quite detailed and impressive in their own right, it does suggest that their engine is not entirely bottlenecked by handling large environments.

Conversely, we were surprised to find that gameplay set within the city itself actually ]seemed to operate at a slightly higher average frame-rate. While exploring the city we found that the frame-rate stuck more closely to a 40fps average

Furthermore, massive particle displays and certain combat scenarios could actually drop performance levels below 30fps for a longer duration resulting in obvious slowdown. What was odd, however, was the fact that some of these dips appeared more often when using the 30fps lock, which kind of eliminates the advantage of using it. This is in contrast to Guerrilla Games' work with Killzone Shadow Fall, where its 30fps option is a totally consistent, completely reliable lock.

First Light offers the 30fps lock feature as well, and it works in very much the same fashion as Second Son. The frame-rate is capped at 30fps but occasional dips are noted. We put it to the test by running as quickly as possible through the environment and examining the results. The dips persist but do seem to appear less frequently resulting in a slightly more stable experience. It's not a flawless 30fps presentation but it's close enough that it makes using the option worthwhile - the action in First Light actually feels more fluid than the main game.

Summary

PS4

• 1920x1080p @ 60fps
• Variable framerate with dips below 40fps (even the 30fps lock have dips)
• Slight better performance and visual than inFamous Second Son