• Hey everyone, staff have documented a list of banned content and subject matter that we feel are not consistent with site values, and don't make sense to host discussion of on Famiboards. This list (and the relevant reasoning per item) is viewable here.

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (New Staff Post, Please read)

Reader mode
ItWasMeantToBe19 said:
I mean, they could just literally use FSR3, lol.

If Xenoblade 4 runs at 20 FPS and using FSR3 can bring it up to 30 without being horrible... Maybe.
Click to expand...

Yeah you're correct, but I think Nvidia probably wouldn't want DLSS upstaged like that. I'm pretty sure they can maybe make a custom solution for the Switch 2.

The reason I think they don't want to do it for their PC cards is because they want to be able to say frame generation is only for 40 series or better Nvidia cards so you better buy the newest card (there's a sales variable at play). But for Switch 2, since that isn't really a concern maybe they can do the whole "oh look, actually we can do a kind of a frame generation using this custom Nvidia chip that's only in Switch 2 consoles!".

The other thing I've wondered is if Nintendo themselves may not be using actual DLSS, but something that's basically almost the same as DLSS just under their own name (for future compatibility issues). So maybe Nintendo isn't bound to some of these rules and can say "hey, well we're using our own version of DLSS as is, lets study what AMD has done with FSR and maybe we can implement some of these frame generation features too". Nintendo did file a patent for their own form of upscaling ala DLSS.
 
Machrider said:
Yeah you're correct, but I think Nvidia probably wouldn't want DLSS upstaged like that. I'm pretty sure they can maybe make a custom solution for the Switch 2.

The reason I think they don't want to do it for their PC cards is because they want to be able to say frame generation is only for 40 series or better Nvidia cards so you better buy the newest card (there's a sales variable at play). But for Switch 2, since that isn't really a concern maybe they can do the whole "oh look, actually we can do a kind of a frame generation using this custom Nvidia chip that's only in Switch 2 consoles!".
Click to expand...

I mean, they've brought a bunch of stuff to RTX 2000+ cards, it's probably just that their machine learning approach gave terrible results without better flow data.
 
0
ItWasMeantToBe19 said:
Sega leaks constantly and may have been the original dev kit leak source.

They have managed to keep zero of their games from leaking in the last several years.
Click to expand...
Sega isn't Nintendo.

I'd be worried about talking about Nintendo more than I would Sega.

If Sega has a leaky ship, perhaps they need to borrow some of their ninjas. Shut that shit right down.
 
Tentacle-tropes said:
The shares increase but there is little reason for Nintendo to go all digital currently. For instance take JP where if the split is 40/60 then they are needlessly cutting mindshare & sales for no reason. Add onto the fact that they haven’t taken steps to be a fully digital company yet; signals they view retail as an important aspect of the company still.
Click to expand...
I agree, which is why I said Switch 3 or 4.
 
TechnicalTip6578 said:
I'd like to ask credibility of this guy
Click to expand...

As Mobius said, could be speculation rather than claiming to know. But regardless, I'll repeat what I've said before about him. I'm not aware of anything he's ever leaked correctly (not to say he hasn't) but I am aware he claimed Star Fox and Metroid were getting a cross over on Wii U years ago, and that seemed to have no truth to it.
 
oldpuck said:
This requires arguing that the SNES, N64, GBA, and 3DS were not successful.
Click to expand...

Two of those still qualify with what I said. Two others are just casualties of war.

SNES was successful, but it did sell less than the NES. And even though it was #1 in Japan, it was permanently 2nd in Europe and 2nd in the US for a while. So perception-wise, maybe it was time to retire the Famicom/NES name.

Similar thing to the N64. It sold even less than the SNES and it was a distant 3rd in Japan and a distant 2nd worldwide. It also lost Nintendo a ton of 3rd party support.

GBA. Well, they didn't plan to retire the GB line, but after the success of their "third pillar" (DS), they were forced to.

3DS. Same thing. The Switch forced the DS line into retirement.
 
0
ItWasMeantToBe19 said:
Would be literally the fastest announcement to release in console history, with only the Xbox One coming close?

(Paul Gale has no credibility)
Click to expand...
...152 days is more than the 134 Nintendo Switch had, and in-line with the intended but scrapped roll-out of 3DS, which was pushed back for software reasons, not marketing.

Short reveals to release are simply not unusual, they're not unprecedented, and with the sole exception of Wii U, have been Nintendo's MO for a decade, with revisions or "extensions" of a family getting even less notice.

OLED Model was barely an upgrade and got two months to simmer, and still boiled over and sold through for months.

It's simply not the case that a 152 day reveal to release cycle is unusual, unworkable or even difficult.
 
ItWasMeantToBe19 said:
I mean, they could just literally use FSR3, lol.

If Xenoblade 4 runs at 20 FPS and using FSR3 can bring it up to 30 without being horrible... Maybe.
Click to expand...
The point is that DLSS is generally better than FSR thanks to hardware acceleration. So if the latter is seeing these kinds of gains, imagine the former.
 
The only meaningful challenge introduced by a shorter reveal-to-launch period is logistics - getting the products produced and shipped to stores within that short time frame. Typical launch challenges like coordinating marketing, producing a launch lineup, etc. are problems that can be addressed in advance to reveal with proper planning, and we're already seen signs of some of these things happening for a while now.

A few months between reveal and launch should be more than sufficient to get the products onto stores, especially since COVID taught everyone the value of good supply chain management.
 
0
Machrider said:
Here is FSR3 on the ROG Ally, it is able to take a game that runs at about 32-40 fps only over 80 fps.



I have to wonder what the input lag is like.

To be honest I don't think you need to do something crazy like 30 fps to 100 fps ... that's not really that useful for a device like the Switch 2. 30 fps to even 45 or 60 fps would be a big difference maker.

I wonder if Nintendo or Nvidia can sort of study what FSR3 does and incorporate that into something for the Switch 2 with DLSS.
Click to expand...

Are you trolling? I am asking genuinely, I can't tell if this is a joke.
 
0
I would've thought a quick turn around from reveal to release is pretty much expected at this point?

Let's say it's a June launch - announcing by early March, or in March itself, means Nintendo have the new product announced well ahead of their fiscal briefing in May and their AGM in June itself. Or maybe it's revealed in June for a September or early October launch. Both scenarios put us less than 150 days from first reveal to release.

The form factor is now a known quantity (hybrid) rather than an unproven gamble, and the next system may well share the same brand (Switch); Nintendo don't have that much heavy lifting to do in terms of concept and branding (though surely if it is super conservative conceptually, you'll see criticism online that there's no hook or twist or surprise - personally I won't rule that sort of thing out).

Really, the question would be: do Nintendo need anything other than a short reveal to release window? I don't see any reason why they would. I wouldn't be surprised at all if reveal to release turn around is even shorter than the Switch itself.
 
Neoxon said:
The point is that DLSS is generally better than FSR thanks to hardware acceleration. So if the latter is seeing these kinds of gains, imagine the former.
Click to expand...

I think people are overly fixated on massive FPS improvements. For a portable device even if you can get a boost of like 10 fps ... that can be a considerably performance change in the playability of a game.

28 fps vs 38 fps is a pretty significant difference, if Nintendo and/or Nvidia could figure out how to get some frame generation using DLSS or a DLSS type fork (which I what I suspect Nintendo is using anyway), it would be pretty good.
 
MisterSpo said:
I would've thought a quick turn around from reveal to release is pretty much expected at this point?

Let's say it's a June launch - announcing by early March, or in March itself, means Nintendo have the new product announced well ahead of their fiscal briefing in May and their AGM in June itself. Or maybe it's revealed in June for a September or early October launch. Both scenarios put us less than 150 days from first reveal to release.

The form factor is now a known quantity (hybrid) rather than an unproven gamble, and the next system may well share the same brand (Switch); Nintendo don't have that much heavy lifting to do in terms of concept and branding (though surely if it is super conservative conceptually, you'll see criticism online that there's no hook or twist or surprise - personally I won't rule that sort of thing out).

Really, the question would be: do Nintendo need anything other than a short reveal to release window? I don't see any reason why they would. I wouldn't be surprised at all if reveal to release turn around is even shorter than the Switch itself.
Click to expand...
I'd argue being able to give dev kits to more developers, aside from trusted partners, would be a good reason for Nintendo to announce the hardware early enough so something like March announcement to a September release for example. Not really a deal breaker necessarily but I certainly it's something they'd consider.
 
ItWasMeantToBe19 said:
I mean, they could just literally use FSR3, lol.

If Xenoblade 4 runs at 20 FPS and using FSR3 can bring it up to 30 without being horrible... Maybe.
Click to expand...
The point of using hardware-assisted AI upscaling is because it’s better. If FSR3 can bring Xenoblade 4 to 30 fps, then DLSS 3.5 should bring it to 40 fps with much better image quality. In almost every single test, DLSS is better than FSR.

The only reason these devices use FSR is because they have no other option. This circles back to the discussion of Nintendo “having no choice”. They chose DLSS, which is the better solution.
 
Last edited:
Teal'c said:
Now many are starting to call it Nintendo Switch 2, coincidence or something more?
Click to expand...
It's a very descriptive name. We know they aren't going to position it as a pro at this point, it's definely their next gen system. Nobody knows the final name, switch 2 is a good substitute.
 
Teal'c said:
Now many are starting to call it Nintendo Switch 2, coincidence or something more?
Click to expand...
It's just a natural naming convention that works. Apple's managed to get up to iPhone 15 and nobody's getting confused over it.
 
bmfrosty said:
You think someone finally made that argument successfully at Nintendo?
Click to expand...
Nobody needs to make that argument when most of the most successful product lines on the planet (e.g.: smartphones) have been successfully using it for years without any need to change.
 
InsaneZero said:
Nobody needs to make that argument when most of the most successful product lines on the planet (e.g.: smartphones) have been successfully using it for years without any need to change.
Click to expand...
Has Nintendo ever put a 2 in a hardware product name except for the 2DS? Yes I'm being facetious, except at a small level where it appears that just putting a 2 after the name hasn't ever been something that Nintendo has done, even when it would have made an insane amount of sense like with calling it the Wii 2 instead of the Wii U.
 
Machrider said:
Here is FSR3 on the ROG Ally, it is able to take a game that runs at about 32-40 fps only over 80 fps.



I have to wonder what the input lag is like.

To be honest I don't think you need to do something crazy like 30 fps to 100 fps ... that's not really that useful for a device like the Switch 2. 30 fps to even 45 or 60 fps would be a big difference maker.

I wonder if Nintendo or Nvidia can sort of study what FSR3 does and incorporate that into something for the Switch 2 with DLSS.
Click to expand...

I don't think Nintendo devs would be blocked from using FSR3? I mean, there are Switch games that uses FSR2.
 
MisterSpo said:
I would've thought a quick turn around from reveal to release is pretty much expected at this point?

Let's say it's a June launch - announcing by early March, or in March itself, means Nintendo have the new product announced well ahead of their fiscal briefing in May and their AGM in June itself. Or maybe it's revealed in June for a September or early October launch. Both scenarios put us less than 150 days from first reveal to release.

The form factor is now a known quantity (hybrid) rather than an unproven gamble, and the next system may well share the same brand (Switch); Nintendo don't have that much heavy lifting to do in terms of concept and branding (though surely if it is super conservative conceptually, you'll see criticism online that there's no hook or twist or surprise - personally I won't rule that sort of thing out).

Really, the question would be: do Nintendo need anything other than a short reveal to release window? I don't see any reason why they would. I wouldn't be surprised at all if reveal to release turn around is even shorter than the Switch itself.
Click to expand...
Yeah, Nintendo hasn’t done a “long” turnaround since the Wii U, and they’ve launched a few systems since then. Also if it’s a backward compatible system, they probably don’t have to worry too much about a smaller launch lineup than usual.
 
Giancarlo said:
Digital Foundry talk about the possible Switch sucessor techinal/graphical perfomance

Click to expand...

why the heck are the comments contains alot of pessimism?

People seem to be relying on power only for the console but that's not the only thing that should give a definitive idea on the console

As DF pointed out, the console, while as strong as PS4/XONE, will have advantages over those consoles in CPU, RAM, type of storage & DLSS
 
Concernt said:
...no, we literally won't.

I doubt we see specs at announcement. 🫠
Click to expand...
We more than likely won’t get specs but we will get 4k capable and can look at the games being shown. Either doesn’t work either way.
 
0
If Switch OLED reveal-to-release was four months, and that was a revision, I’m expecting the successor to have a five month window like the OG Switch.

I wouldn’t be surprised if Nintendo does a four month window like with the OLED, but since that was a revision, and the Switch 2 is an entirely new console generation, I think Nintendo could be a bit more careful and add a month.

Tho the benefit of a shorter release window is that the hardware is closer to release for when they have solid confirmation if how their crossgen and nextgen software pipeline is gonna look like
 
the only prediction I'd seriously make about the naming convention is that it won't remove the Switch branding but it'll also make it as obvious as possible this is a next generation device,.

And my less than sure guess would be to go with a prefix rather than a suffix to further differentiate it from current Switch refreshes (Nintendo Switch LITE, Nintendo Switch OLED), Super Nintendo Switch or something as cheesy as it sounds.
 
Machrider said:
I think people are overly fixated on massive FPS improvements. For a portable device even if you can get a boost of like 10 fps ... that can be a considerably performance change in the playability of a game.

28 fps vs 38 fps is a pretty significant difference, if Nintendo and/or Nvidia could figure out how to get some frame generation using DLSS or a DLSS type fork (which I what I suspect Nintendo is using anyway), it would be pretty good.
Click to expand...
Thus far all frame generation is about producing one "fake" frame between every two "real" frames. Doing something like Real 1, Real 2, Real 3, Fake seems like way more trouble to figure out than it would be to just ignore it for a few more years. At least as I can see it, there'd be two options.
1) Repeatedly turn frame generation off and on. Which would lead to pretty uneven animation.
2) If you wanted 1/4 of frames to be Fake, but also have proper timing between frames, it would have to be doing something weird like producing Real 2 and Real 3 with imagery from a tiny fraction of a second earlier than they'd normally have shown, while the Fake frame was also being worked on concurrently, starting before it even had those frames to work with. So I don't see how this one would work in real-time.

I think this go around the best bet is to just get extra frames by setting the rendering resolution even lower and taking the GPU gains, and just being stuck with it if it's the CPU maxing out first.
Multiverse said:
The point of using hardware-assisted AI upscaling is because it’s better. If FSR3 can bring Xenoblade 4 to 30 fps, then DLSS 3.5 should bring it to 40 fps with even better image quality. In almost every single test, DLSS is better than FSR.

The only reason these devices use FSR is because they have no other option. This circles back to the discussion of Nintendo “having no choice”. They chose DLSS, which is the better solution.
Click to expand...
Yes. But as far as we know, they did not choose a generation of hardware built to take advantage of DLSS Frame Generation. For people with RTX20 or 30 cards, FSR3 is the better option than DLSS Frame Generation, because it is the only option.
 
darthdiablo said:
I don't think Nintendo devs would be blocked from using FSR3? I mean, there are Switch games that uses FSR2.
Click to expand...

What I mean is I wonder if Nvidia/Nintendo can make a custom implementation like FSR3 which operates with DLSS instead, maybe to better image quality results (perhaps using the Tensor cores).

Not looking for like 35 fps to 80 fps gain here, even like 10-15 fps boost would be plenty.

I know Nvidia says frame generation is only possible with the 40 series cards, but I really have to wonder if that's 100% legit or if there is a lot of sales-influence on that claim (they need a reason to sell people on 40-series cards).
 
bmfrosty said:
Has Nintendo ever put a 2 in a hardware product name except for the 2DS? Yes I'm being facetious, except at a small level where it appears that just putting a 2 after the name hasn't ever been something that Nintendo has done, even when it would have made an insane amount of sense like with calling it the Wii 2 instead of the Wii U.
Click to expand...
The next Switch hardware is by all accounts an iterative upgrade of the original Switch rather than any significant and fundamental change to the core concept. Nearly every product line is using a numeric-based naming convention for differentiating between iterations because a number is simple to understand than word(s). This has practical implications since marketing doesn't need to tear their hair out trying to clarify any confusions between "Nintendo Switch OLED" and "Nintendo Switch 2 OLED" as opposed to "Super Nintendo Switch OLED".

Numbers are also convenient for naming future iterations since it is unambiguous as to the lineage of a product line. You know exactly that a "Switch 3" is an upgrade of "Switch 2", but the same can't be said for "Super Switch" and "Ultra Switch". Nintendo's recent history (i.e. the Switch era) has demonstrated that they are entirely willing to abandon long standing practices and if they offer no practical utility, and they are also willing to adopt new practices that have a demonstrably proven track record.
 
Last edited by a moderator:
Machrider said:
I think people are overly fixated on massive FPS improvements. For a portable device even if you can get a boost of like 10 fps ... that can be a considerably performance change in the playability of a game.

28 fps vs 38 fps is a pretty significant difference, if Nintendo and/or Nvidia could figure out how to get some frame generation using DLSS or a DLSS type fork (which I what I suspect Nintendo is using anyway), it would be pretty good.
Click to expand...
Looking at your posts, frame generation doesn't do the thing that you're thinking it does, and FSR doesn't have anything to offer Nvidia they do not already have. Not trying to argue with you, just saying, from a technical perspective, what you're hoping for isn't viable.

Frame generation isn't designed to make low frame rates into half-way decent frame rates. It produces a very bad result when you try. AMD says their minimum is 60fps before you add frame gen for good results.

If you're starting with 28fps, as in your example, the minimum latency added is 35ms. The generated frames are of poor quality, because of how far apart the native frames are. And the frame pacing is very bad, which is why you see so much microstutter in that video, when the frame rate drops out of the Rog Ally's VRR window. Although it is hard to judge the quality of the frames just because FSR upscaling is so generally bad at that low a resolution - I can't tell what are frame gen artifacts and what are upscaling artifacts.

You say "10-15 fps boost would be plenty", but that's a huge boost if your game is already only 25 fps. That's a more than 50% increase. That requires a lot of spare performance, and without hardware acceleration, I'm not sure it would even be doable in the majority of games.
 
0
InsaneZero said:
The next Switch hardware is by all accounts an iterative upgrade of the original Switch rather than any significant and fundamental change to the core concept. Nearly every product line is using a numeric-based naming convention for differentiating between iterations because a number is simple to understand than word(s). This has practical implications since marketing doesn't need to tear their hair out trying to clarify any confusions between "Nintendo Switch OLED" and "Nintendo Switch 2 OLED" as opposed to "Super Nintendo Switch OLED".

Numbers are also convenient for naming future iterations since it is unambiguous as to the lineage of a product line. You know exactly that a "Switch 3" is an upgrade of "Switch 2", but the same can't be said for "Super Switch" and "Ultra Switch". Nintendo's recent history (i.e. the Switch era) has demonstrated that they are entirely willing to abandon long standing practices if they offer no practical utility.
Click to expand...
I'm arguing that it should be Switch 2. I'm also arguing that Nintendo doesn't really do that for inexplicable reasons.
 
TechnicalTip6578 said:
why the heck are the comments contains alot of pessimism?

People seem to be relying on power only for the console but that's not the only thing that should give a definitive idea on the console

As DF pointed out, the console, while as strong as PS4/XONE, will have advantages over those consoles in CPU, RAM, type of storage & DLSS
Click to expand...
"nintendo" and "tech" inherently breeds bad faith discussion

Machrider said:
What I mean is I wonder if Nvidia/Nintendo can make a custom implementation like FSR3 which operates with DLSS instead, maybe to better image quality results (perhaps using the Tensor cores).

Not looking for like 35 fps to 80 fps gain here, even like 10-15 fps boost would be plenty.

I know Nvidia says frame generation is only possible with the 40 series cards, but I really have to wonder if that's 100% legit or if there is a lot of sales-influence on that claim (they need a reason to sell people on 40-series cards).
Click to expand...
it's legit. DLSS FG is reliant on the OFA improvements to generate new frames. from what I've seen, FSR FG is much lower quality. it produces frames faster, but you're paying for it
 
Reader mode
You must log in or register to reply here.
Please read this new, consolidated staff post before posting.

Furthermore, according to this follow-up post, all off-topic chat will be moderated.
Last edited by a moderator:


Back
Top Bottom