Digital Foundry Article Technical Discussion [2023]

Status
Not open for further replies.
4090 costs a little bit more than 3x of the PS5. You can play Alan Wake with >30 FPS in native 4K with Pathtracing... So this is basically a much better "option" than a console from price/performance perspective.

So, no, these consoles do not look "very nice in comparision".
The differences are only obvious to those who know what to look for.

PS5 looks incredible and is enough for 98% of gamers. That is the reality.
 
4090 costs a little bit more than 3x of the PS5. You can play Alan Wake with >30 FPS in native 4K with Pathtracing... So this is basically a much better "option" than a console from price/performance perspective.

So, no, these consoles do not look "very nice in comparision".

I am not dening it, the PC version is still much better option. But even without path tracing this is still very good looking game, and yes the PS5 version still looks very very good even against PC max settings.
Its still 3x the price and 10x the power ....
 
No, still images can look very good. But the problem comes with dynamic effects like AO, shadows, reflections etc. Look at these shadows between a 3070 with "high" and DLSS quality in 1440p and the PS5 in performance:

The image quality will fall apart when you move around or have moving parts.
 
Waiting on the PC video from Alex, but I expected much bigger differences in the lighting than in the footage Oliver showed.
 
Last edited:
No, still images can look very good. But the problem comes with dynamic effects like AO, shadows, reflections etc. Look at these shadows between a 3070 with "high" and DLSS quality in 1440p and the PS5 in performance:

The image quality will fall apart when you move around or have moving parts.
PS5 still looks pretty darn fine for hardware released 3 years ago. The noticeable parts are isolated and some are less noticeable than others. The biggest issue for me is the noisy shimmering on some lines and speculars in the distance.
 
Pictures don't show how shit it looks in motion due to using FSR.
I agree this is a problem, the picture doesn't look very stable in motion after rewatching the video a few times i think it will be very annoying on consoles. But as far as i know, the same problem exists in PC version to some degree.
 
And yet it doesnt performs better than a 4 years old Turing card released for $100 less.
Which one are you referring to?
Talking about the price of just a GPU vs a console which is a complete system isn't an accurate comparison.
 
To be precise 3070 relased 3 years ago with 499$ price, ps5 digital released 3 years ago with 399$ price
You ll need a lot more to run a game than a GPU though. Plus the hardware (edit: I mean for a console) is usually finalized for at least around a year before release.
 
Last edited:
I totally agree with Oliver that PS5 version looks very nice in comparison to PC max settings. To be honest its ridiculous how much more powerfull 4090 is (almost 100TF) vs PS5 (10) ! And for the most part the game looks pretty much the same. RT adds some extra fine details but the cost of it is gigantic. "Hey look darker shadows under table! Nice 90TF well spend! "
It is what it is, at least AW2 is one of those games that has great art style and great tech so the final product comes very nicely together.

That's the nature of diminishing returns. Arguably the difference here is larger than the difference between say Horizon Forbidden West on PS4 and PS5. Or any game on Series S vs the Series X version. Its up to the individual to decide whether the extra they are getting is worth the additional money they would spend for it.

The differences are only obvious to those who know what to look for.

PS5 looks incredible and is enough for 98% of gamers.

I think significantly more than 98% of gamers know to look for decent image quality or else developers would all be releasing sub 1080p games.

Additionally, you don't need to have an understanding of different graphics effects or the subtleties of light propagation to simply understand that one version of a scene looks more realistic and/or better.

98% of people probably couldn't tell you if a game was missing ambient occlusion but they could certainly tell you the game looks better with it turned on.
 
That's the nature of diminishing returns. Arguably the difference here is larger than the difference between say Horizon Forbidden West on PS4 and PS5. Or any game on Series S vs the Series X version. Its up to the individual to decide whether the extra they are getting is worth the additional money they would spend for it.



I think significantly more than 98% of gamers know to look for decent image quality or else developers would all be releasing sub 1080p games.

Additionally, you don't need to have an understanding of different graphics effects or the subtleties of light propagation to simply understand that one version of a scene looks more realistic and/or better.

98% of people probably couldn't tell you if a game was missing ambient occlusion but they could certainly tell you the game looks better with it turned on.
There’s a lot of discussion to branch from this post. If you consider why some folks are caught up in resolution, frame rate, and ultra settings, is mainly because those are the easiest things for most people to see, with ultra being a preset for resolution and frame rate.

Very few people can fine tooth comb through a settings menu and see a difference - very few PC gamers would know how to optimize between resolution, frame rate and individual settings to maximize their experience.
 
The more developers push the graphics on these RDNA2 consoles the less compute there is for RT.

Yep, but I was hoping for a "RT quality" mode, where resolution was sacrificed at 30fps for RT reflections. Personally, I find it jarring when a game looks this good but SSR occlusion / disocclusion artefacts kick in, when things that your brain knows are reflective aren't, or when a mirror has to be grainy and blurry to hide the stretchy smeary none reflection that's there instead.

I think the absence of RT reflections - at least as a lower res option - is a bit of a loss even if it's understandable.

And blimey am I glad I don't have to use FSR2.

And yet it doesnt performs better than a 4 years old Turing card released for $100 less.

It launched a year earlier for $100 less, but without a mobo, main memory, CPU, SSD, PSU, case, mouse, keyboard, controller, and a windows license. In other words, if you just bought a 2060 you could play absolutely nothing. Same as if you just bought a 4090 or any other GPU.

Comparisons between the price of a GPU and an entire fucking system are one of the most asinine things that PCMR warriors bring in to shit up multiplatform forum threads.

PC is a strong enough platform in its own right not to need that.
 
If people want to do a proper price comparison between PC and consoles, can they please list both prices and performance numbers within the same posts. eg. 2060 cost $399, total system cost $n hundred, framerate at PS5 settings is x fps. Otherwise we just get lots of one-liners pointing out the issues with every partial one-liner and no obvious way to just look at compare costs and benefits!
Very few people can fine tooth comb through a settings menu and see a difference - very few PC gamers would know how to optimize between resolution, frame rate and individual settings to maximize their experience.
Which is part of DF's reason d'etre, to create 'optimised settings' for people to plug in, those people being moderate enthusiasts who care to research but don't care to fiddle around themselves. On the bottom end of the scale, most will push a button and run with it. TBH I doubt we have any measure of how typical PC gamers operate; best we have is anecdotal evidence from enthusiasts who post on fora.
 
Main problem of Alan Wake 2 on consoles is that it doesnt have 40fps mode. Sad that only Sony bother with this awsome mode. I saw in GamingTech video that 60fps mode is aliasing mess.
 
That's the nature of diminishing returns. Arguably the difference here is larger than the difference between say Horizon Forbidden West on PS4 and PS5. Or any game on Series S vs the Series X version. Its up to the individual to decide whether the extra they are getting is worth the additional money they would spend for it.



I think significantly more than 98% of gamers know to look for decent image quality or else developers would all be releasing sub 1080p games.

Additionally, you don't need to have an understanding of different graphics effects or the subtleties of light propagation to simply understand that one version of a scene looks more realistic and/or better.

98% of people probably couldn't tell you if a game was missing ambient occlusion but they could certainly tell you the game looks better with it turned on.
Yes i fully agree with you on the effect of diminishing returns. I think this gen in general is a great example of this.

"98% of people probably couldn't tell you if a game was missing ambient occlusion but they could certainly tell you the game looks better with it turned on."

This is also true, however when it comes to RT often is the situation of having the same effect but better implementation. I don't think this is something that has a 0 1 effect here (on/off). We have SSR and whatnot baked lighting can look almost as good as RT in some scenarios.
IMO AW2 has a very good RT implementation as the game is very realistic and the RT adds to the overall presentation, it produces a very consistent image. I am sold. However, in some titles, I feel like gluing RT on top of everything doesn't have the same effect. Sometimes it even produces a kind of weird presentation, when you have those super realistic reflections in the game where objects and npcs are not that realistic. It kinda stands out and looks out of place.
 
Status
Not open for further replies.
Back
Top