DeepEnigma
Gold Member
That makes sense. I stopped using cockpit for flatscreen since it's used in VR where I get that less restrictive feel of it. Been doing mostly bonnet now for flat, where RT on with PSSR off is gorgeous.
Last edited:
-
Hey Guest. Check out your NeoGAF Wrapped 2025 results here!
PS5 Pro is getting PSSR 2.0 between January and March 2026
- Thread starter IbizaPocholo
- Start date
- Rumor Hardware
Senua
Member
You mean roof view? Dumbest fucking view ever, at least GT4 spec 2 modded in a proper bonnet view
Also I love how you say bonnet and not hood
viveks86
Member
I can see myself turning it back on if 2.0 does better upscaling so the jaggies go away and the boiling stops. If it gives better performance so they could restore LoD as well, that would be icing on the cake. But I might be asking too much. Not sure why they are messing with LoD at all. May be they are hitting a CPU or memory bottleneck...
Yeah I would think that view is even more restrictive...
Last edited:
DeepEnigma
Gold Member
I think LoD is also a side-effect of the still cross-gen engine woes even if it's current-gen feature set. I notice far less LoD issues with games that are no longer designed around that smaller footprint.
It's a little higher than my tastes, yeah, but it's also literally the same head height as cockpit, with cockpit interiors removed. Some of the nice liveries look great with the RT and lighting, and I can't stand chase cams in racing games any longer. Used to use just the FPS view, but that also feels too low to the ground when passing cars.
What's wrong with bonnet? Isn't that what it's called, bonnet, boot, hood, trunk, frunk, engine bay, etc..
Last edited:
Wait wut? The hell you are talking about? Lol at the people who said DLSS2/3 is better than PSSR
maybe scaling from 1080p but lower than that is far from superior. The only better thing is AA in still BUT in motion and even at low resolution is not that better of the PSSR.
Last edited:
Senua
Member
Roof view feels too high, like you're in a truck. I mean sure, keep that view too but add a bonnet view for the rest of us!
I wasn't being sarcastic, I love that you use bonnet which is the british english expression, usually Muricans say hood
Last edited:
PaintTinJr
Member
You are missing the fact that Cerny showed how TOPS usage was mostly north bridge bandwidth and memory bandwidth bound on the Pro above tensor sizes, requiring extensive tiling, whereas PC DLSS relies on the much bigger L1, L2/L3/LLC GPU caches, and that is the real constraint on consoles including the Switch 2- not withstanding the 27 TOPs and 33.3ms frame time slice - going by RTX 2050 mobile the L2 is 2MB.
And the ~1ms calculation was covered in this post by me.
And your DF image was posted in that thread too, but just like you are confusing inference modelling for the entire solution in DLSS 3.7 using a 2.x model - from the other thread - the inferencing in a game switching out reconstruction in FSR or in TSR for PSSR, is using just ~1ms for the PSSR reconstruction, and the other costs is for taking those results and generating the final image - the pre-processing motion vector calcs - is similar across all the techniques, FSR, TSR, TAA, PSSR so isn't an additional cost of PSSR.
DeepEnigma
Gold Member
Ah, yeah, that's true. I'm also a car enthusiast and track race in and out of season IRL. Watching Top Gear since day 1, then Grand Tour, you tend to adopt the lingo, lol. Also having mid-engine cars, I always hated calling it a "frunk."
I would like if we can adjust the camera height in views. They only allow for you to tweak slightly cockpit and chase.
Last edited:
Zathalus
Gold Member
I'd consider the cost the entire thing. End to end cost for implementing PSSR or DLSS. Of course it's not an additional 2ms over TAA for example. It's less than 1ms over TAA.You are missing the fact that Cerny showed how TOPS usage was mostly north bridge bandwidth and memory bandwidth bound on the Pro above tensor sizes, requiring extensive tiling, whereas PC DLSS relies on the much bigger L1, L2/L3/LLC GPU caches, and that is the real constraint on consoles including the Switch 2- not withstanding the 27 TOPs and 33.3ms frame time slice - going by RTX 2050 mobile the L2 is 2MB.
And the ~1ms calculation was covered in this post by me.
And your DF image was posted in that thread too, but just like you are confusing inference modelling for the entire solution in DLSS 3.7 using a 2.x model - from the other thread - the inferencing in a game switching out reconstruction in FSR or in TSR for PSSR, is using just ~1ms for the PSSR reconstruction, and the other costs is for taking those results and generating the final image - the pre-processing motion vector calcs - is similar across all the techniques, FSR, TSR, TAA, PSSR so isn't an additional cost of PSSR.
Topher
Identifies as young
TrebleShot
Member
XESS 2, DLSS 3+4 and FSR4 are better than it.Wait wut? The hell you are talking about? Lol at the people who said DLSS2/3 is better than PSSR
Based on what? Your feeling? Well FSR4 and DLSS4 I could agree but DLSS3 and XESS 2 eeeh.
Last edited:
Corporal.Hicks
Banned
Even when running at high resolution, PS2 games weren't nearly as detailed as the UE5 Silent Hill: F. In fact even games with prerendered backgrounds on a sixth-generation consoles such us RE1 remake on GC never looked as good.Looks like a high resolution PS2 Pro game.
Not sure what Konami was
SH2 with mods
RE1 Remake
Sillent Hill F
I wouldn't be surprised if the Hinako model in Silent Hill: F had a higher polygon budget than the entire Silent Hill 2 scene on the PlayStation 2.
My screenshots shows the standard maxed-out settings, but this game has hidden settings that can further improve the lighting quality, making Silent Hill F one of the best-looking PC games currently available.
Silent Hill f Looks Incredible With Its Hidden “True Epic” Settings
There are some hidden "True Epic" settings that you can unlock that can greatly improve the image quality of Silent Hill f.
www.dsogaming.com
Last edited:
TrebleShot
Member
I have used them all.
XESS 2 on the MSI claw
PSSR on my PS5 Pro
DLSS 3/4 on my 5090
FSR4 on my XBOX ally X
How many have you used?
Senua
Member
That's awesome, of course we all grew up with Top Gear, I even watched OG Top Gear on BBC2 every thursday evening with my dad in the 90s, before the reboot in 2001. It was nowhere near as entertaining but it was much more factual and cool for geekary.
Nowadays it's North America that is killing it with car content, especially youtube. Jason Camissa and the Throttle House guys being my personal favourites. I find myself saying "Miata is always the answer" instead of MX5. It's funny how we influence eachother through our media
You are the same who said ps5 pro is not capable of ai upscaler so forgive me if I take your impression as a grain of salt. I could tell you I seen in action DLSS3 and XESS2 and at lower resolution than 1080p are very "eh" and not too far away to the PSSR.
Also you people always forgot AI upscaler are based on models experience, ps5 pro is a lot inexperienced compared the pc environment and all considered doesn't stake that badly compared the "older" PC upscalers iterations.
Last edited:
J Required
Gold Member
There's a bit too much snark in this otherwise beautiful comment. Lol
DeepEnigma
Gold Member
PaintTinJr
Member
But it isn't, you could make that argument some gens back, but at least since PS4/X1 when per‑pixel velocity + history accumulation was done for full final image AA became standard - enabled by the vast unified RAM - that cost has been baked in, even if the pre and post processes differ marginally from technique in cost/quality - the inputs and outputs have remained pretty standardized - for the magnitude of scale we are talking versus the cost of the inferencing they are all pretty much the same overhead cost, especially when comparing within the same closed platform like PlayStation 5 Pro
Zathalus
Gold Member
We can see the total frametime spent in the developer image I linked. No speculation, no calculations, no working back from relative framerates. TAAU is about 1.5ms and PSSR is slightly higher at 2.1ms. That's for reconstructing a 4K image at 60fps. So PSSR is about 0.6ms heavier than TAAU for a total rendering budget for the entire solution of 2.1ms. The comparison is the same in a wide spectrum of games I linked. Call of Duty, Alan Wake 2, AC Shadows, GoW: Ragnarok, and one other I didn't show a screenshot of, Control. PSSR is about 0.6-1ms heavier than the regular temporal AA/upscalers, and is around 2ms in total. Which isn't odd, DLSS4 and FSR4 are also both heavier than TAAU/FSR2.
So you are right, just the computational cost for PSSR is under a 1 ms, but it's also correct to say the end to end solution cost is 2ms.
Last edited:
Alebrije
Member
I find logic that Sony releases 2.0 early 2026 and not later so any developer can work on it but speciallly Rockstar with GTA6 so there is a real PRO versión for the Game.
//DEVIL//
Member
I am not expecting anything from gta 6 on the pro outside of PSSR and higher rez slightly . It will still be 30 fps almost 100% given as a game like this require a better CPU.
And while I don't care about res from 1080 to 1800 or whatever they wanna go for, I am more interested in how PSSR now handle the anti aliasing and the shimmering
Last edited:
It's very easy to reverse the logic of your sentence. The former has a tiny screen the later enormous screen (which is harder miss any kind of flaws), is it normal I noticed more artifacts and instability in the DLSS with a tiny screen (where lower resolution and artifacts should be less visible) than in a bigger screen with PSSR? There wasn't nothing of particular advanced in DLSS2/3 compared to the PSSR outside the more nvidia experience, there are all TAA based and in theory PSSR is more advanced tech wise for the AI. But I know PSSR is not branded Nvidia so it loses the magic. I have tons of friends who always doesn't liked DLSS for the thing people noticed now on PSSR, not saying the later version are not superior or improved a lot, but continue to sell the idea DLSS3 is totally better of the PSSR is not that true outside you are named Bojji.
Last edited:
DenchDeckard
Moderated wildly
Yeah, mustn't be pssr then. It's so strange because I don't have the issue on the OG game on Xbox and pc. It looks strange but not taking too much from my overall experience.
Must be some poor fsr implementation
Zathalus
Gold Member
You're comparing 360p-540p upscaled to 720p-1080p up against games that render at usually over 1080p to usually 4k. Naturally DLSS won't look better. Star Wars Outlaws also uses the tiny version of DLSS, which is worse than PSSR.
Topher
Identifies as young
I've seen worse. Dragons dogma 2 had massive flickering in foliage on PC in specific areas early in the game. It went away elsewhere though. I think a lot of this just comes from the fact that console games are analyzed much more heavily than games on PC.
Ps5 pro PSSR lean more to 720p-800p usually and scaled to 4k; hardly I seen DLSS3 scaled at such range resolution offer so more stable IQ eh, I don't think it was even possible before the DLSS4 transformer models. You have to go to 1440p and 1600p but again I haven't a good memory because the pc wasn't mine so I could be wrong.
Last edited:
Bernardougf
Member
Wait for the tools ? ... wait for the games ... PC
Sony really has come full MS circle
Sony really has come full MS circle
Hyper_hyper
Member
I mean, this is completely normal to update develpoment software which is machine learning based. But your reference is something completely different
Last edited:
Bernardougf
Member
Bojji
Gold Member
DLSS 2/3 has better stability, no "film grain" like noise common for PSSR and doesn't add visual artifacts to games with RT and UE5 games (common to PSSR+UE5 combo).
You can try to prove that this is not the case obviously, but so far you always were just talking about it.
So you are entered in the discussion just to troll. Nice.
Last edited:
Bernardougf
Member
And you just to shill as usual
omegasc
Member
Yeah, maybe there's this "bias" because dlss "has always been perfect" since 2.0.
I only learn of its flaws when they announce new versions lol
I hope PSSR 2.0 delivers. There's nothing to lose since it's shared tech with AMD and will benefit PC and other consoles/handhelds.
Zathalus
Gold Member
Most of the games that look good on PS5 Pro render at above 1080p. Certainly, all the PS5 Pro first party titles are way over that. Games that render below 1080p are thankfully rare on the Pro.
I agree. Not saying PSSR is flawless but I seen a couple of video by Alex about the DLSS4 (there aren't so many as the PSSR) and I have a lot of reminiscences of the PSSR artifacts when he goes to scrutinize some problematic in a sample of games.
Last edited:
DeepEnigma
Gold Member
Depends. RE4R and Village are 800p, AC Shadows also and they are quite good. Unfortunately below 1080p is not that rare especially when 60 fps and raytracing are involved.
Last edited:
Bojji
Gold Member
DLSS was always better than any other reconstruction tech at the time.
DLSS2 better than TSR and TAAU in 2020
DLSS3 better than Xess and FSR2/3 in 2023
DLSS4 better than FSR4 in 2025
DLSS4.5 better than DLSS4 in 2026
Now we wait for competition to catch up...
I get what you're saying but this way of looking at it is flawed because the limit is set only by the frametime cost and a console fps target. Not a "physical" hard cutoff for hardware TOPS that makes it "physically possible" or not.
A laptop 3050 is also only 57 TOPS yet it runs the same path tracing and DLSS models as a 3352 TOPS 5090. Just not well, right? Same is true with Switch it was designed to run DLSS but the difference is that while you may put your 3050 on ray/path tracing and a given res and hit 20fps the Switch version has to hit a standardised performance target framerate so a specific version exists for it for those targets alone not because somewhere between 27 and 57 it becomes "physically impossible" to do otherwise. The switch 2 is physically capable of it and how much of that 57 TOPS is taken by pathtracing absent on switch too yet it runs standard DLSS like the 3050. The difference lies in the fact that you can't release a console game hitting 10fps and call it a day so it has bespoke low frametime cost DLSS simply to hit fps targets for given hardware. Its actually an added effort over 3050 laptop DLSS support. Why would this idea not apply to PS5 Pro and PSSR to make something bespoke and lightweight requiring as little silicon as possible for PS?
On PS5 Pro the use of PSSR is the same tradeoff just at a higher res/settings/raytracing which uses its hardware as best it can at the time. So when they created PSSR they still tried to create "tiny" DLSS much like switch 2. As tiny a frametime cost as they can while trying to maintain raytracing and 60fps or 120hz modes in games, hoping to boost IQ to near 30fps quality resolution with minimal frametime cost. PSSR is also "limited" from the beginning by the same constraints and they try to use the full capability of the hardware.
I agree some early model artifacts absolutely can be improved with minimal performance cost and those advancements in quality/efficiency naturally come as time goes by but I was debating the idea that one hardware has unused headroom while another does not based on their absolute theoretical TOPS. This is simply wrong to me. They both have bespoke upscalers based on their framerate/settings targets and naturally higher fps/res/raytracing/settings require more computing power but this hardware is based on the framerate targets and settings and not so much vice versa especially for Pro. They dont really have unused headroom in TOPS, it's simply because those initial performance targets required that hardware.
DLSS just worked backwards, it was developed before for other hardware then it targeted weaker hardware with bespoke DLSS for their perf targets, whereas PSSR was developed for PS5 Pro from the beginning still taking that given hardware and 60fps framerate/higher settings targets into account. The PS5 Pro has the disadvantage that it is tied and always compared to PS5/XSX so anything lower in framerate or settings or much higher in price is considered a failure of the system too, so they have to keep PSSR frametime cost very low with a high res hence why it has a given TOPS not because of some "maximum model size" that they're not hitting yet and leaving TOPS on the table or something. It's all about frametime targets for a given res/settings and the methods and efficiency will improve on both.
Had standard DLSS not existed outside of Switch 2 I doubt standard CNN DLSS would have even came to it at launch to be honest. They would have concentrated on the bespoke tiny DLSS/"NSS" even with those 33ms frametimes just as Sony seem to have only concentrated on one PSSR version performance target for now.
They will release PSSR2 and will no doubt improve things but the performance cost is unknown. It may even be more costly in terms of frametime cost. That would really prove the idea that there is unused hardware headroom wrong. I hope it is in fact more efficient in addition to quality improvements though. We might get several profiles for use. Assasins Creed Shadows PSSR turned out well when they worked with PS. So maybe they push specifc ones for different engines/games based on feedback from devs and we might even get a tiny PSSR for 120hz modes or something. We know very little about it for now.
Sorry I've ranted for so long.
Hyper_hyper
Member
No one ever said that PSSR is flawless, but for its 1st iteration it is tremendously better than it was DLSS 1.
winjer
Gold Member
Although that is true, there is a big caveat with that statement. DLSS1 was a spatial upscaler from Feb. 2019.
You are comparing abandoned tech from 7 years ago with PSSR.
I have to repeat it every single time because I know the usual suspect when you try to praise the PSSR they start with their rethoric upscaler war with the DLSS2 is far better and blablablah which is not even true at all.
Just look to the reaction
Last edited:
winjer
Gold Member
And be honest, FSR4 is a high standard. It only loses to DLSS4 and even that is not by a huge margin.
omegasc
Member
Yup, it's a deserved "best of all" title on every iteration so far. I just turn it on and that's it.
What I mean is media rarely pointed flaws besides "it may get a little blurry in perf settings". Until the next version is announced and then they suddenly notice blur, trailing, artifacts. I'm not trashing, I think it's kinda natural when you don't know anything better and is as good as dlss has been since 2.x
another genius in the list.
Last edited:
Similar threads
- Locked