Digital Foundry Article Technical Discussion [2018]

[BRiT]

The Evil Within 2: Xbox One X vs PS4 Pro Graphics Comparison + Frame-Rate test

No written article yet.

Tango Gameworks adds a fascinating vote of support for this generation's two enhanced machines. With the patch arriving at the tail-end of 2017, Xbox One X's native 3200x1800 update gives it a clear lead in image quality over PS4 Pro.

Resolution is a curious point; while the developer's 1.02 patch notes state 2240x1260 is the native res on PS4 Pro, multiple tests with the EU and US versions right now show a straight 1920x1080 on patch 1.03. We have contacted publisher Bethesda for more information, and will update the video as soon as we know more. In either case, Xbox one X gives a much more satisfying result with 4K selected - and while 30fps target isn't as rock-solid there, it's a better experience overall for those on ultra HDTVs.

Notes and images to pixel count from Dark1X @ResetEra https://www.resetera.com/posts/3236515/

This time, I helped pass judgement on this.

We tried both the EU and US version of the game (had to buy a second copy), we used multiple PS4 Pros, we captured on multiple capture devices, we had four people (including myself) pixel count shots, we spoke with the developer and basically did everything possible to double and triple check it.

We also checked across 3-4 different chapters to make sure dynamic res isn't in use. It's not. It's 1080p.

My guess? They specifically talk about getting faster performance by setting your PS4 Pro to 1080p - I bet there's a bug that is causing it to render out at 1080p in all modes. I have a feeling this was introduced AFTER the Pro patch since there was another patch after initial Pro support.

I remember trying this right after the Pro patch and thinking it looked sharper but never pixel counting but, in its current state, it's clearly just 1080p.

Some 4K shots showing the 1080p count.

Dark1X @RE https://www.resetera.com/posts/3237001/

Ah ha, yes, 1.02 also looked like 1260p to me.

So I'm starting to think it was 1.03 that broke Pro support. Wish we could roll it back to 1.02.


Also, NXGamer did some analysis on a previous version 1.02 of it and their pixie count came to 1260p, which seems to support DF's theory that the latest patch somehow broke something on 4Pro.

 

[BRiT]

Forza Horizon 3: Xbox One X vs PC Graphics Comparison + Frame-Rate Test

Playground Games' premiere open-world racing title gets enhanced Xbox One X support - adding a true native 4K with many more bells and whistles. Here we compare it to PC at max settings, where the contest is surprisingly close - and with a few major plus points on X hardware including the use of 4x MSAA.

 

[Silent_Buddha]

Forza Horizon 3: Xbox One X vs PC Graphics Comparison + Frame-Rate Test

Playground Games' premiere open-world racing title gets enhanced Xbox One X support - adding a true native 4K with many more bells and whistles. Here we compare it to PC at max settings, where the contest is surprisingly close - and with a few major plus points on X hardware including the use of 4x MSAA.

It's unfortunately they weren't able to go for 60 FPS, but outside of that, it's a very impressive upgrade.

I wonder if the CPU is so weak that they weren't able to even try to do a 60 FPS 1080p version of the game on XBO-X.

Regards,
SB

 

[Jay]

I wonder if the CPU is so weak that they weren't able to even try to do a 60 FPS 1080p version of the game on XBO-X.

although i assume that it's simply cpu overhead limitations, can we rule out the 30 base 60 x split not being liked by ms/studios?
It is a multiplayer game even though you can play it single player.

 

[Tkumpathenurpahl]

Possibly, but does a higher framerate benefit you much in a racing game?

I understand that there's gains to be had in the realm of twitch shooters or something like Street Fighter, but I don't know if there's much to be had in other genres. I don't do much racing, but my time with racing games has always indicated that sticking to certain lines is more important than frame-sensitive reaction times.

 

[AlNom]

When you're travelling >88mph, you probably would want twitch-reaction times.

 

[see colon]

Cudos to the devs for this article. Nice!

Cool way of doing the scaler...I wonder if they could have used machine learning to train the offline scaling?

If I understand the article right, only 2% of the time the resolution is 1080p?!?!

What the heck DF!!!! What did you measure??

Meh, DF needs better QA if this debacle is really true...this would also explain why the devs went mad.

I'm speculation here, but I would have tested frames during a period of high framerate to try to find the maximum resolution if I were testing. A traditional dynamic scaler would drop resolution if rendering load was too high, so framerate dips *should* be followed by a drop in resolution. If you are able to hit your framerate target, 60FPS in this case, it would be assumed that your rendering load would be light enough to hit the maximum resolution.

Maybe they only hit 60FPS 2% of the time.

 

[function]

I can't help feeling there are some Shenanigans going on in that "dev technical anaysis" regarding the issue of resolution, something on which they were not upfront.

Computing the screen percentage by a single axis (specifically, the vertical one) is a standard practice in the industry because it simplifies every operation related to aspect ratio and screen manipulation. We use the same “r.screen percentage” method as everyone else, properly documented by Unreal Engine devs in the source code (you need to be logged in and granted access by Epic Games to see it). Also, think 4k, 1080p, 720p, etc… They are all one-dimensional measurements of the vertical axis.

The average customer that knows anything about resolution would know that 4K is four times FullHD. You wouldn't talk to customers in promo material (as they did) about going to half of 4K, and not tell them that actually meant 1/4 res, unless you wished to mislead.

Your customers don't read Unreal Engine dev docs, they aren't a texture artist ,and they don't sometimes refer to mip maps and alpha buffers being "half res" as shorthand. They might know about megapixels though, and they probably know about 4K.

Secondly, even programmers are careful not to use terms like "50% resolution" without context when talking about dynamic framebuffers because there is no industry standard for dynamic scaling. Some games scale both axes equally, some scale unequally, and some only scale horizontally, and only a fucker would talk about their horizontally scaling game as "always running at 100% resolution".

 

[Jay]

Possibly, but does a higher framerate benefit you much in a racing game?

I understand that there's gains to be had in the realm of twitch shooters or something like Street Fighter, but I don't know if there's much to be had in other genres. I don't do much racing, but my time with racing games has always indicated that sticking to certain lines is more important than frame-sensitive reaction times.

I don't know if they particularly care if it's a big benefit or not when it comes to multiplayer games.

 

[BRiT]

DF Article on Forza Horizon 3 enhancements is up: http://www.eurogamer.net/articles/digitalfoundry-2018-forza-horizon-3-gets-4k-update-on-xbox-one-x

 

[DrJay24]

30fps really hurts this game, IQ was never the issue.

 

[Reynaldo]

I disagree. I mean, would it be better at 60fps? Of course, but 30fps doesn't hinder it in any way, much less "hurt" it.

 

[function]

30fps really hurts this game, IQ was never the issue.

There is literally not a console on earth that could run this at a solid 60 fps.

... but that leads into a different discussion about CPUs.

 

[DrJay24]

I disagree. I mean, would it be better at 60fps? Of course, but 30fps doesn't hinder it in any way, much less "hurt" it.

Don't get me wrong it's a great game and I've bought the DLC and play it a bunch. That being said, the controls are sluggish and constantly remind me it's not 60fps. Going from mainline FM to H3 takes a while to get used to. Many games are GPU limited and I was thinking this was one and the X would have a 60fps option or at least an unlocked setting.

 

[BRiT]

Two articles over as many days, first up is ... http://www.eurogamer.net/articles/d...ait-was-worth-it-for-ethan-carter-on-xbox-one

The wait was worth it for Ethan Carter on Xbox One
Solid on standard hardware with strong X enhancements.

While the mystery buried within The Vanishing of Ethan Carter is certainly intriguing, from our perspective, it's the journey of its developer that is most fascinating. First released in 2014 on the PC, The Vanishing of Ethan Carter has undergone several surgeries on its way to this new release on Xbox One. In porting the game to PS4 years ago, its developer, Warsaw-based The Astronauts, completely overhauled the game by porting it from Unreal Engine 3 to Unreal 4.

In its updated state, the team worked hard to translate its gorgeous but static global illumination and photogrammetry-based assets and the results speak for themselves. Ethan Carter retains its stylizsd, pseudo-realistic appearance while introducing new Unreal 4 specific features to the mix. We were impressed with the revised version of the game back in 2015 and in 2018, it's surprising just how beautifully it holds up. With forests and structures stretching out as far as the eye can see, Ethan Carter remains as gorgeous as ever.

 

[BRiT]

Next article up is http://www.eurogamer.net/articles/d...-generations-revisited-runs-beautifully-on-pc

Sonic Generations revisited: better than Forces, runs beautifully on PC
A low-cost, high-thrill 4K 60fps experience.

2017 could so easily have been the Year of the Hedgehog thanks to the arrival of two high-profile Sonic releases. It goes without saying that Sonic Mania is something truly special, a brilliant return to classic 2D platforming for the franchise - but on the 3D side of things, Sega came up short. Sonic Forces was a disappointment, a real shame bearing in mind just how good Sonic Generations was - especially on PC. The game has aged remarkably well, has exceptional modding support and runs well and looks wonderful at full 4K on GTX 970-level hardware.

Sonic Generations was released on PC back in 2011, a few weeks after the console versions. Unlike those 30fps efforts though, the PC version allows for smooth 60fps gameplay and higher resolutions. It's powered by the Hedgehog Engine and offers visuals that still hold up beautifully even in 2018. Indeed, in our opinion, the sheer quality of workmanship in the visuals makes this game stand up better than its successor.

 

[iroboto]

Updated Redout analysis.

Honestly, not sold on this variant of dynamic resolution at all. imo they make a good case if it were holding 60fps at all times. But it fails to do that as well. On top of, it only strengthened their resolve on this one. The results of their second report isn't all that different from their first.

 

[Jay]

giving out a response like that, i hope its not only the resolution that they can show is wrong.
I'm not sure many people will be overly impressed if it's higher res but lower settings of the base ps4, lot more inconsistent frame rate to 4pro.

Updated Redout analysis.

Honestly, not sold on this variant of dynamic resolution at all. imo they make a good case if it were holding 60fps at all times. But it fails to do that as well. On top of, it only strengthened their resolve on this one. The results of their second report isn't all that different from their first.

Well as I suspected, the devs really did themselves a deserves with their reaction.
Inconsistent framerate
Lower quality settings compared to other platforms
Even resolution which DF got wrong wasn't as crazy as first seemed.

It's easy to assume the framerate is poor due to their implementation of region based dynamic resolution, but i doubt that's the problem at all.
If it is, then that's an even worse showing than i thought on the 1X.

 

[Allandor]

Well as I suspected, the devs really did themselves a deserves with their reaction.
Inconsistent framerate
Lower quality settings compared to other platforms
Even resolution which DF got wrong wasn't as crazy as first seemed.

It's easy to assume the framerate is poor due to their implementation of region based dynamic resolution, but i doubt that's the problem at all.
If it is, then that's an even worse showing than i thought on the 1X.

That is still really surprising. The xboX has more power than the ps4 pro, but even at 1080p framerate seems to drop for no reason. Maybe they should at least lower their "dynamic" resolution a bit. But still it is not AAA game. Such a small developer may just not have the resources to really analyse what the problem is.

 

[Jay]

That is still really surprising. The xboX has more power than the ps4 pro, but even at 1080p framerate seems to drop for no reason. Maybe they should at least lower their "dynamic" resolution a bit. But still it is not AAA game. Such a small developer may just not have the resources to really analyse what the problem is.

one of the reasons i don't think it's simply a resolution issue.
With the size of drops it has in framerate, would be pretty easy to lower those regions more, especially after the first DF breakdown.
i think it's a general optimisation issue, not simply resolution, which means it would take longer to do.