-
Hey Guest. Check out your NeoGAF Wrapped 2025 results here!
KeplerL2: PlayStation handheld reportedly with 24GB memory, PS6 console with 30GB
- Thread starter IbizaPocholo
- Start date
- Rumor Community Hardware
Panajev2001a
GAF's Pleasant Genius
These were rhetorical paper specs, in practice did it turn out worse or better than that? This is not to talk about XSX vs XSS per-se but to think about PS6 and the Handheld.
With PSSR2 out, a case could be made that PS5 Pro upgrades are doing better than PS4 Pro ones despite a smaller percentage improvement in paper specs.
Last edited:
kevboard
Member
it turned out exactly like that.
the issue was however the RAM pool size, which made it sometimes impossible to have RT enabled, and it also lead to some texture issues.
but in terms of raster performance it's exactly what you'd expect.
it often runs the Series X performance mode at half the framerate and some additional downgrades. so the ⅓ GPU power does play out exactly as expected.
Panajev2001a
GAF's Pleasant Genius
Ok, so the results in practice were a bit worse than the paper specs diff.
kevboard
Member
in some games. in well optimised ones not really.
FingerBang
Member
Where is that assumption coming from?
1) The point of better AI hardware is to avoid that kind of compromise. We already see how turning RT features on strains new hardware less than old one
2) RT is scalable, no one thinks a portable device should run games with the same settings
3) A great upscaler on par with DLSS would make 720p -> 1080p look amazing
vaibhavpisal
Member
Thats how nvidia hardware handles path tracing. 5080 doesn't do native 4k in path tracing.
Its safe to assume PS6 will do these internal resolutions with those settings.
Handheld will probably do same settings at like 120p. Not sure how that will look like upon upscaling.
FingerBang
Member
OK, but here's the thing:
The RT jump for next generation should be really big, for both Nvidia and AMD.
PS6 therefore should be able to do higher resolution than current hardware with PT.
Games will also be written with PT in mind, which would encourage developers to optimize them properly.
PT will probably offer some kind of fallback for weaker hardware. I don't think it's going to be as extreme as going below 720p with upscaling.
But all of this is speculation. I think the PS6P hardware seems pretty solid. Things can be scaled back as long as we don't get any major bottleneck. The problem with current handheld PC with AMD hardware is that the GPU side of the equation is still stuck in 2022.
vaibhavpisal
Member
I was replying to steveholt, who wanted same settings as PS6 on this one.
I think overall experience should be solid.
Am sceptical of any type of RT on it though. Maybe it gets the basic one, like software lumen. Or what id Soft put in their games.
Look at budget nvidia gpus. Like rtx 5050 or 4050. Heavy RT isn't viable on them, they have the upscaling and good RT performance. And they are twice as powerful. Not restrained in power budget.
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
Even Valve doesn't guarantee this with the SteamDeck and your games you buy on Steam. So why should Sony with the PS6 and the PSP3?
Felessan
Member
Even?
Valve guarantee literally nothing, it's just a storefront, they are not ecosystem with fixed spec and curated access
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
You're right. My point is the Valve did a good job communicating what the Steamdeck does and doesn't do. And what to expect. Being that Sony more than likely won't mandate all PS6 games to run on this PSHH, they'll need to properly communicate that to everywhere early.
First order or business, is to NOT call it a PS6 handheld. The name of this thing matters!
Hudo
Member
Because Sony controls the hardware the developers target.
Because all of them are still crutches and don't provide better resolution than native. I use them frequently but all of them have issues. 4k native is 4k native. By same argument you can say that why 1080p when you can upscale from 720p but somehow you can't find anyone arguing for that.
I have 5090 and 4k is still something i can't play natively most of the time. C77 at 4k with pathtracing with no dlss is like 30fps on it which is to me unplayable framerate.
Last edited:
Lagspike_exe
Member
1440p -> 4K with DLSS or other deep-learning algorithms has frequently better image quality than native 4K.
It does not. DF is high as fuck and people are parroting those idiots too much.
To see it in practice go to C77 mox club from game opening and look at LED walls. At native 4k they are sharp and when you move camera nothing happens. But when you use DLSS those same LEDs are getting smeared. Even with latest DLSS. DLSS is good at static images which look great in screenshots but the moment you start to move is when all that sharpness goes away as all of those upscales rely on previous frames.
I think people put too much stock into 'what PS is this'. If I buy 'Game A' that gives me PS4/5/6 license - it only needs to run one of them (or some version of the configuration in between).
But Sony - unlike Valve - does need to mandate 'something' about the library, because if the handheld is a 'handheld library' they're just building another Vita and noone wants that.
And Valve's audience is the 1% so far that is willing not only to put up with PC UX eccentrics but also isn't too fussed about fiddling with making games work when they don't (out of the box). Maybe that grows to 100M+ users some day but they don't exist today, so concessions here could be a target for Microsoft's niche market to chase, but not Sony.
Play Switch 6
DF is comparing analytical TAA to DLSS when they say 'native' vs 'upscaled', so indeed it's a false equivalency at best. At minimum they should compare DLAA to DLSS(or appropriate equivalent with other AA tech) if they want to make claims about 'better than native', and we know the upscaled image loses there.
Of course - if they wanted actual accuracy, the test should use a ground-truth with super&multi-sampled AA without the temporal component, but that is no longer accessible in most modern games without jumping through hoops. You can still hack at least some of Unreal games to get there, but not most other engines.
Bojji
Member
Native 4k means 2160p with TAA image reconstruction. And TAA usually have tons of issues compared to DLSS or FSR4, so even if they have lower native resolution they can still produce better results. This is quite objective fact and anyone can test it.
If this was 2160p with msaax4 vs DLSS in older games (when msaa actually worked) - that would be totally a different story.
Yes but it objectively isn't 'native' vs 'reconstructed' - it's just bad science to call it that.
And the only one to benefit from muddying the waters was NVidia, so it only adds fuel to channels like DF becoming PR mouthpiece for bigtech. Even if they did so by accident.
Bojji
Member
Taa was bad long before DLSS was invented. You still have DLAA, no one is taking native resolution from you.
Most people just prefer having almost as good image quality with MUCH better performance.
Last edited:
_oni_
Member
Are you sure you're not using an old version of DLSS?
HerjansEagleFeeder
Member
No disagreement there (it's why TAA was embraced in the first place - which was also a much cheaper alternative to get 'near' SSAA quality).
I just found the concentrated effort to re-frame upscalers into 'better than native' pixels unnecessary and in bad taste. Not unlike DLSS5 reveal which is doing the same but one level up.
It's basically hitting the consumers with a 'you're all idiots, this is what's good for you' stick, repeatedly, and while I accept some marketing PR stunts, that's one treatment that just doesn't sit right with me, no matter what's it in service of.
wayne_joyce
Banned
Totally agreed. These DF folks are nearly as clueless as the MLiD quack. More technobabble than a Star Trek episode and not one of them know anything about graphics or game design or game engines. Just clickbaiters shooting the sh*t. For all the good that Youtube has done for video streaming in the world (think AVGN and the Gaming Historian), it has unfortunately spawned the careers of these know-it-all know-nothing clickbaiter shills parasites making a career out of spewing out garbage for the uneducated masses. These people should be kicked off the platform and replaced by people with credentials, or at least journalists who are objective and don't pretend to know what they are talking about from a technology standpoint. It is sickening and a major turnoff.
J Required
Banned
People are missing a key point: the PS6P isn't comparable to the Series S in the way they think. Unlike a cut-down home console, it's designed first and foremost as a handheld, which already sets it apart from the standard PS6 in both purpose and user experience. Its hardware would likely be built around a large, well-balanced RAM pool targeting 1080p output, meaning developers wouldn't have to wrestle with awkward bottlenecks or fragmented optimization the way they sometimes do with lower-tier consoles.
On top of that, competent machine learning hardware could play a major role in closing the performance gap with its more powerful counterpart, helping deliver visual and performance parity in ways that go beyond raw specs. From a user perspective, that means the experience wouldn't feel like a compromised version of the "real" platform.
Most importantly, a unified library with the PS6 would change everything. Unlike the Vita era, where support eventually tapered off, this approach would ensure consistent game availability and long-term viability. If executed properly, it wouldn't feel like a side device—it would feel like a natural extension of the main platform.
I think Sony has been waiting for the right time to execute this strategy, and now it finally makes sense. Given their developer-friendly approach since the PS4 era, they've built a strong foundation for something like this to work smoothly. If they carry that same philosophy forward, there's a good chance it'll be executed really well and feel seamless for both developers and players.
On top of that, competent machine learning hardware could play a major role in closing the performance gap with its more powerful counterpart, helping deliver visual and performance parity in ways that go beyond raw specs. From a user perspective, that means the experience wouldn't feel like a compromised version of the "real" platform.
Most importantly, a unified library with the PS6 would change everything. Unlike the Vita era, where support eventually tapered off, this approach would ensure consistent game availability and long-term viability. If executed properly, it wouldn't feel like a side device—it would feel like a natural extension of the main platform.
I think Sony has been waiting for the right time to execute this strategy, and now it finally makes sense. Given their developer-friendly approach since the PS4 era, they've built a strong foundation for something like this to work smoothly. If they carry that same philosophy forward, there's a good chance it'll be executed really well and feel seamless for both developers and players.
vaibhavpisal
Member
A different processor from PS6 already means its way more work to port it. Not to mention, much bigger delta in power.
AFAIK, its the only device that uses that processor. Series S, X, PS5 they all used a very popular pc cpu. Devs once they know how to optimise for ram, didn't had any other issues with Series S. it was a one time learning curve worth getting through.
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
They can control that, but they don't. Sony normally doesn't do mandates like that. It's not in their nature to do that at Playstation. They have very little mandates over the decades like that. That's been more a Microsoft Xbox thing, that they do and it's been really bad for them.
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
I think people put too much stock into 'what PS is this'. If I buy 'Game A' that gives me PS4/5/6 license - it only needs to run one of them (or some version of the configuration in between).
But Sony - unlike Valve - does need to mandate 'something' about the library, because if the handheld is a 'handheld library' they're just building another Vita and noone wants that.
And Valve's audience is the 1% so far that is willing not only to put up with PC UX eccentrics but also isn't too fussed about fiddling with making games work when they don't (out of the box). Maybe that grows to 100M+ users some day but they don't exist today, so concessions here could be a target for Microsoft's niche market to chase, but not Sony.
Play Switch 6
That Play Switch 6 name is all the Devil's work. No way man
But to your first point, Sony clearly will need to come up with a way to not allow a game to run on previous generation consoles. Whatever that config is or however they make those wrapper work, they have to allow devs to make a game that runs on the PS6, but not the PS4, PS5, or PSP3.
A PSP3 "verified" label works. We know that works. Valve has proved it. Even Microsoft has their "Xbox Anywhere" label for games where that applies. It's not really that hard. And I'm assuming they are working on that policy and engineering that technical ability on the backend, right now as we speak.
adamsapple
Or is it just one of Phil's balls in my throat?
J Required
Banned
Im amazed you didn't like what I said then. OT: I KNOW Sony will do it right.You sound a lot like Jason Ronald talking about Series S
asdasdasdbb
Member
That Play Switch 6 name is all the Devil's work. No way man
But to your first point, Sony clearly will need to come up with a way to not allow a game to run on previous generation consoles. Whatever that config is or however they make those wrapper work, they have to allow devs to make a game that runs on the PS6, but not the PS4, PS5, or PSP3.
They are going to require the handheld's support. That's why LPM exists.
J Required
Banned
What different processor? All is Zen6 and RDNA5. The game package will need minor tweaking to scale down to the handheld.
As I said before, unlike MS Sony will do it right.
kevboard
Member
Microsoft also did it right. the issue is that you can't rely on developers doing proper optimisation.
you have to hope that the handheld is successful, if not, it will end up with awful unoptimised ports just like the Series S.
and it's very easy to see in just the most recent release, Pragmata.
why does the Series S run at 720p with TAA, instead of utilising FSR3 (which the game supports) to reconstuct the image to 1440p or even just 1080p? why not offer a quality mode that targets 30fps and runs at native 1080p, maybe with FSR3 to 1440p?
the very simple answer is that the Series S port has basically zero priority. the hardware is not the issue.
I can guarantee you that if you exposed the full settings menu many users would find better settings than what Capcom used. that wouldn't fix their suboptimal texture pool handling of course.
and if the PS6 handheld ends up as a low priority system, it will have the same fate.
it's in the same position after all. less memory, dramatically cut back GPU, and on top of that also a dramatically cut back CPU.
mckmas8808
Mckmaster uses MasterCard to buy Slave drives
"IF" they do, then that's a HUGE Playstation failure. And would be Mark Cerny's first policy-like failure when it comes to new hardware. The only other minor one was not having a pathway for PSVR1 games to play on PSVR2 headsets. But I understand how that happened.
Forces games that are released in 2035 to be playable on a PS Hand Held with the power of a console from 2013 (but with better RT and ML) is a HORRIBLE idea. You guys will see. Just watch (assuming you are correct btw).
Is the PS6 portable less capable than the Switch 2? Because basically every developer will be supporting that going forward.
kevboard
Member
we'll see if Switch 2 support will continue to be this good once the PS6 and Helix are out.
we already have games being delayed now, because the optimisation work takes that long on it.
Bond was just delayed, Elden Ring is taking ages to release, Borderlands 4 might be cancelled the way they just are completely silent, and there are titles we know wouldn't run on Switch 2 unless completely revamped.
so I wouldn't count on the Switch 2 helping forever.
you should instead hope that the base PS5 will remain a focus for developers for as long as possible. because that is the system that would be closest in performance and is actually plausible to be supported longterm by multiplatform releases.
the Switch 2 is ok for now, but it already struggles with CPU limited games now, imagine what happens once consoles with Zen6 CPUs are on the market to replace the gimped Zen2 CPUs we have now, that are already about twice as powerful as the Switch 2 CPU.
Last edited:
Loxus
Member
I believe it's the other way around. The difference in specs is bigger than with the Series consoles to just scale down.
Games well be made for the Handheld/PS5 first, cross-gen lives forever. It's a shame really.
Magister
Member
I don't give a shit about Playstation handheld. PC handhelds are often unereliable, but they will always have more exclusive games. Even if PS handheld is stable af, it won't replace a handheld PC for me. I'll just get a PS6 and spend less money instead of buying 2 different things. The worst thing about this setup is that diversifying your library often is a pain in the ass as you don't know for which platform you should get the game.
Last edited: