Charlietus
Regular
@davis.anthony @Shifty Geezer I haven't even posted the video myself, but even if I had, you would be making a mountain out of a molehill. But that's just me.
Sony PlayStation 5 Pro
- Thread starter chris1515
- Start date
davis.anthony
Veteran
TAA on the PC version doesn't have any sharpening applied at all, which is why it looks so soft.
It's not really difficult to beat it using a different upscaler, it's also another game that doesn't have a brilliant DLSS implementation as it has issues with ghosting and some surfaces, when it shouldn't.
It's not really difficult to beat it using a different upscaler, it's also another game that doesn't have a brilliant DLSS implementation as it has issues with ghosting and some surfaces, when it shouldn't.
Look at the vegetation and how PSSR resolves more details, like if running at native 4K. I don't think you know what a strong sharpening filter would do to those leaves... And there are no visible sharpening artefacts with PSSR here. Besides we could already see PSSR resolving more textures and vegetation details (a sharpen pass cannot do that) compared to native 4K running on PS5 at 30fps. A sharpening filter would enhances aliasing, yet we see the contrary here and same story in Ratchet (in stills there is a lot less aliasing on PSSR vs DLSS).
davis.anthony
Veteran
I do, because I played the game to death on PC and spent quite a lot of time messing around with the sharpening.
There isn't on the PC version either when using DLSS with reasonable levels of sharpening.
Where did I say the extra detail was purely as a result of using sharpening? Hint: I didn't.
The TAA in TLOU is bad, it's bad on PC and using sharpening with TAA is quite common and does offer some benefit to image detail, so it's strange this game doesn't use any sharpening at all with TAA and makes any comparisons to other solutions seems larger than they should be.
No, too high of a sharpening filter does that.
Another game with a poor DLSS implementation......and 'a lot less' is a gross exaggeration on your part.
The image labeled PSSR does look much nicer, but part of the comparison that's missing is that the Pro appears to be using either higher quality textures or a lower LOD bias. High quality reconstruction techniques often require negative LOD bias to maintain texture quality in distant objects, and it would be very noticeable in foliage. But they may also be using higher quality assets in the pro release as well. Hell, it could be both! So, yes, the grass looks better in the PSSR shots compared to the TAA ones. But so far, I'm not convinced that the machine learning algorithm alone is producing such a difference without an adjustment to the assets or settings.
ErneX
Newcomer
Of all the games shown so far this is probably the most interesting one. No idea about the internal resolution other than what DF could pixel peep from the really short clips on the console reveal…
But I’m more interested in the final IQ than the internal res to be honest, and the settings they changed.
snc
Veteran
impressive pro performance comparing to rx6800 with similar cpu (over 30%) !
Last edited:
On consoles with very limited hardware, bottlenecks are completely different than on PC using over the top parts. This is about what I was expecting based on previous benchmarks using modest PC components.
On console every component should be fighting for bandwidth, even in highly CPU limited scenes. And we know RT is far less CPU demanding on Playstation (based on developers input and past games like the Insomniac games or even some UE4 games using hardware RT). But the most impressive in DD2 is the PSSR IQ improvements in stills and in-motion versus FSR2. Night and day difference and probably the biggest improvement ever noted. PSSR is going to be game changing in all those FSR-spoiled games.
Still no word on a Black Myth Pro patch, which is annoying cause I was going to wait until it comes out, but now I haven't been able to stop playing it and I'm a little over half way through the game.
Improvement in what ever noted in what?
From all the games running on Pro I have seen so far. Sorry for my bad english.
Tkumpathenurple
Veteran
It's not your English that's the issue, it's the fact that you repeatedly speak like a sales rep in reference to anything PlayStation and dismiss anything XBox.
PSSR is exciting, but it's no reason to lose objectivity in a technical forum.
Please simmer down, because you're not daft, and you're enjoyable to talk to when you're not whipped up into a PlayStation frenzy.
PSSR is exciting, but it's no reason to lose objectivity in a technical forum.
Please simmer down, because you're not daft, and you're enjoyable to talk to when you're not whipped up into a PlayStation frenzy.
snc
Veteran
Over 30% fps advantage in cpu heavy town section, pc was rx6800 with 5600x
Thanks. So that sounded pretty fishy, so despite my better judgement I decided to watch the video. First thing I want to say is NXG video's are absolutely all over the place in terms of information delivery. He throws stats around one after another without any detailed qualification of what he's referring to in such away that he can often sound like he's contradicting himself mid sentence. Very hard to follow.
Nevertheless I think I managed to piece together most of what he was trying to say after a few re-watches and my initial suspicions are I think are justified.
Two key points stood out for me. Firstly he was assuming a settings match between the Pro and the PS5. I don't think he said what PS5 mode he was comparing to (an example of his disjointed information delivery above) but I am assuming the PS5 high quality/favour graphics mode). He was also unsure whether DRS was being used but still gave performance metrics vs the PC which could clearly be way off if DRS is in use. So that's the first problem.
The second problem is he claims his run was "increasingly CPU bound" throughout the run but didn't provided any specific metrics or video clips corresponding to that, but also claimed the PS5 Pro is running around 60% faster than the PS5 and around 30% faster than the PC. I don't believe there is any way at all the PS5 Pro could be delivering 60% more CPU performance than the PS5 with the same CPU running at up to 10% faster clock speed. So even if we do assume a settings match with the PS5, that 60% improvement is going to be primarily coming from the GPU side. Which means vs the PC the comparison is also on the GPU front, not on the CPU front.
That makes a lot more sense as expecting the Pro to outperform a 5600x by 30% is extremely unrealistic unless there's something very wrong with the PC implementation. However 30% faster than a vanilla 6800 in an RT enabled title sounds fairly realistic given it's improved RT capabilities. Assuming of course the settings and resolution are actually matched.
Charlietus
Regular
When the console is out and conditions and settings can be matched, then some accurate testing will be done. Until then, this isn't accurate enough.
Edit: @pjbliverpool put it very well.
Remedy dropped a big post on their site, showing off Alan Wake 2's PS5 Pro enhancements and going into detail on how they implemented RT. Check it out, it's more in-depth than I would have expected.
www.remedygames.com
Comparison video:
Alan Wake 2 on PlayStation 5 Pro – Behind The Scenes
Take a deep dive into how Remedy’s Northlight engine team updated the technology behind Alan Wake 2 to support the new features of PlayStation 5 Pro.
www.remedygames.com
Comparison video:
Similar threads
- Locked
