- Pronouns
- He/Him
* Hidden text: cannot be quoted. *
Hidden content is only available for registered users. Sharing it outside of Famiboards is subject to moderation.
* Hidden text: cannot be quoted. *
Appreciate the thougtful response. A lot of that went over my head, but my takeaway is that 2 1080p cameras using the Switch 2 SoC currently rumored (T239) would be far more capable than phones, and with Tensor cores and even at 1080p, the SoC would produce far superior results to an Apple device that is more powerful on paper but driving 10x the pixels.A dedicated device like the Switch would be way better for AR than phones. High end phones may have better cameras, but they're overkill for AR (you don't need a 50 Mpx camera to do AR on a ~2 Mpx screen), but most importantly software can't be built with those cameras in mind, because not everyone owns a high-end phone. Phone camera set-ups have a huge amount of variability, and software like Pokemon Go has to be built for the lowest common denominator.
Depth estimation is a particularly important factor here. Even on iOS, where developers can optimise around a much smaller set of hardware configurations, Apple only includes its Lidar depth sensors (actually ToF sensors, but whatever) on the Pro models, so software obviously can't be built to require them. Other phones use stereo depth estimation, but with mismatched cameras positioned closely together, the quality isn't great, and it's mostly geared towards fake bokeh portrait photos rather than AR.
A pair of two cheap 1080p cameras on the back of Switch 2 would provide better depth estimation than almost any phone, at much lower cost. Part of the reason would simply be the use of matched cameras much farther apart than is feasible on phones. The more important part, though, is the tensor core performance of T239 and Nvidia's expertise in depth estimation from the automotive side of their business. Most stereo disparity estimation uses a relatively simple mathematical approach which is easy to implement but tends to be noisy and get caught up on small details, which is actually implemented in hardware on the OFA on T239 and other Nvidia chips.
If you've got lots of ML performance, though, which T239 has, you can do much better by using a neural network to perform stereo estimation. Nvidia have been researching this for years, with this paper from 2018 describing a technique that's usable on the Jetson TX2 (even in handheld mode, T239 should have at least 10x the ML performance of the TX2 thanks to its tensor cores). Nintendo could do much better, though.
For one thing, every use case of stereo disparity estimation I've come across (including this one from Nvidia) takes in a full RGB images. This seems sensible, but cameras don't actually take full RGB images. Each pixel on a camera has either a red, green or blue filter in front of it, arranged in grids of four pixels typically with two green, one red and one blue (known as a Bayer pattern) so if you have, let's say a 20 Mpx camera, it actually has 10 million green pixels, 5 million red pixels and 5 million blue pixels. To get a full colour image with RGB for each pixel, software (or hardware) interpolates between the different colours, with a technique called debayering or demosaicing. Further software then applies noise reduction before displaying the picture or sending it on elsewhere.
The issue is that the process of debayering and noise reduction just makes the job of a neural network worse. The debayering step is just simple interpolation, so it's adding data without any of it being actually useful to the neural network. The neural network is getting 3x as much data, but no more usable information. The noise reduction step then potentially destroys information which might be useful to the neural network, so a debayered and denoised image is requiring a neural network to process much more data, but get less usable information out of it, than just consuming the raw sensor data. ML based solutions are very good at extracting detail from raw camera data. I use software called DxO PureRaw for my photography, and it produces far better results than any conventional debayering and denoising software I've ever used.
By single-sourcing a pair of camera modules, Nintendo and Nvidia could create a neural stereo depth estimator based on raw sensor data from cameras with known noise characteristics, which puts them in a better place than almost any other AR product. Furthermore, I would expect there are significant further benefits to be found from moving from a purely spatial approach to depth estimation to a spatio-temporal one. Just like DLSS 2 produces much better and more stable results than purely spatial upscalers by incorporating data from previous frames, depth estimation would benefit significantly by incorporating temporal data. I've seen papers which use optical flow as part of depth estimation, but no ML-based implementations that use a temporal approach such as a recurrent neural net. It seems like a pretty obvious win to me; feed in data from previous frame(s) along with gyroscope and accelerometer data to determine movement (potentially with optical flow as well) and you should get a more stable output that's better at resolving small details.
Another big factor is just the performance of Switch 2 relative to its screen resolution. Even with a bump to 1080p, the performance of Switch 2 per pixel would be way higher than, say the Apple Vision Pro, which is more powerful but has to push displays with 10 times as many pixels, and at a higher frame rate to boot. This allows for more impressive graphics, but also makes it easier to do things like estimate lighting conditions within the scene so that rendered elements feel like they're actually in the environment and not crudely photoshopped on top. This was one of the things that journalists were impressed by when demoing the Apple Vision Pro, and it's also a problem that's well suited to machine learning.
Finally, and most importantly, a Switch 2 with AR capabilities would have Nintendo developing games for it. They've shown how interested in AR they still are with Labo, not to mention how they can bring new things to the table. I'd be exited to see what Nintendo's developers could do with class-leading AR hardware.
On the other hand, if we look at the CPU, the story inverts. The gap between NG and current gen is larger than it was between Switch and last gen.
In terms of hardware, I think it's just a different type of hype. We know the next console will follow the Switch's footsteps, but we don't know how exacly, any new addons, etc. In terms of software, you have to factor in that there are more people interested now than in the NX era. The Switch no doubt has created new nintendo "follow every news type" fans.Is this the most exciting period for Nintendo speculation in recent memory? Hardware hype, Direct hype, Metroid Prime 4 hype, Mario Wonder hype, Smash update hype, and possible TotK DLC hype? In the Switch's 7th year, after they already released a new FE, Zelda, Pikmin, Metroid Prime Remaster, Kirby Deluxe, Advance Wars, Sea of Stars, and even Mario RPG still to come?
Maybe not reddit or youtube, but...Hell, I was a lurker who registered to see hidden content. But that's because I'm an old school fan and there's a wealth of knowledge here, not because I want to spaff it all over Reddit or YouTube.
The Switch does have at least one AR game - Mario Kart Live. They may wish to create more experiences like that without bundling cameras as external peripherals, so more developers have access and more players are likely to try it.Why will Nintendo want to put cameras in a new console? I thought they have given up on that idea long ago. The original Switch doesn't have a camera and no AR games, so what makes everyone think that they might tackle it again?
I mean, Doug knows way more than we do in his position at NoA, I wouldn't worry about anything leaking to him.Maybe not reddit or youtube, but...
*side-eyes your username*
..I bet you'd jump at the chance to feed info back to your dad
Appreciate the thougtful response. A lot of that went over my head, but my takeaway is that 2 1080p cameras using the Switch 2 SoC currently rumored (T239) would be far more capable than phones, and with Tensor cores and even at 1080p, the SoC would produce far superior results to an Apple device that is more powerful on paper but driving 10x the pixels.
I'll reserve judgement on the final design, but this does sound promising.
* Hidden text: cannot be quoted. *
FF7R on PS5 adds extra visual features like volumetric fog and emissive lighting, but that's about it* Hidden text: cannot be quoted. *
the A57 is on part with Jaguar, clock for clock. having half as many cores and lower clock keeps it from matching the PS4/XBO. the A78 in Drake buries the Jaguar six feet under easily. no contest thereHello, I rarely expect to be posting on here, but I had a few questions regarding some discussion from a bit ago, and I thought asking here directly instead of continuing to lurk would be the best way to get answers.
I noted there was some discussion on CPU performance a while back, comparing not the upcoming Switch, but the current Switch to PS4/Xbox One. The claim was that the Switch was quite close to the other two systems, and I'd be interested in hearing how that was concluded. I was under the assumption that the Switch's bottlenecks included the CPU, using Geekbench (which isn't of course a perfect tool for measuring console performance, though pure specs dont describe actual software performance perfectly either), there seems to be a pretty notable gap (2-3 times in performance) between the Switch and PS4, so I'd like to know if Im missing something, I was always under the assumption that even in 2017, the Cortex-A57s, despite being far more efficient as ARM cores, were not able to overcome the clock-rate gap on architecture alone, not even considering the Tegra X1's having half as many cores than the other consoles, so not even considering the massive gap in properly optimized parallel performance.
I actually don't think there's much point in [ HIDE ]. It's not at all difficult to create an account to lurk. If the purpose of hiding things is to keep them within the community, limiting to those who have posted at least one actual thought here seems more sensible to meI’d suggest posters using the [ HIDE ] tag instead of [ REPLY ]. The latter requires the user to reply in order to view the post, leading to an influx of hello messages.
Edit: You’reallmostly lovely people, and it’s good to have more people posting. The suggestion is only to reduce forced participation.
Now I'm become Gaming, knower of everythingThis new leaker seems to know too many things in too many directions which gives me a weird feeling. If it's true they had real info in the past it's definitely notable but we're talking Sega projects, SE projects, Nintendo reveals, hardware... That's a lot at once.
Got you, so with 60% the clock speed and half the cores, that lines up far more with what I had been thinking. Gives me back the hope I had before as well that the octo core A78 would give the new system the boost it needs, and also some of my sanity, I thought I was going crazy for a second there XDFF7R on PS5 adds extra visual features like volumetric fog and emissive lighting, but that's about it
the A57 is on part with Jaguar, clock for clock. having half as many cores and lower clock keeps it from matching the PS4/XBO. the A78 in Drake buries the Jaguar six feet under easily. no contest there
There is literally no point in making it exclusive in JP when software sales heavily favor Switch in the region; unless an absolute miracle of miracles happen with PS5 software in the near future which is very doubtful, even then I very much doubt it.Nintendo should make a deal with Square Enix to make Dragon Quest XII: The Flames of Fate a Drake exclusive, given how much the Switch dominates the console market in Japan and that Dragon Quest is the most popular JRPG in Japan its a match made in heaven. Give Square an offer they can´t refuse and Nintendo will have gained their first third party exclusive to launch Drake hugely in Japan.
* Hidden text: cannot be quoted. *
FF7R on PS5 adds extra visual features like volumetric fog and emissive lighting, but that's about it
the A57 is on part with Jaguar, clock for clock. having half as many cores and lower clock keeps it from matching the PS4/XBO. the A78 in Drake buries the Jaguar six feet under easily. no contest there
outside of heavy simulation games, I don't expect Drake's cpu to be overburdened by many games. even Starfield is showing some good scaling on lower end cpusGot you, so with 60% the clock speed and half the cores, that lines up far more with what I had been thinking. Gives me back the hope I had before as well that the octo core A78 would give the new system the boost it needs, and also some of my sanity, I thought I was going crazy for a second there XD
I am wondering now how Geekbench handles their multicore scores. It seems, based on my reading, to be using all the cores for its tests, but not necessarily to score how those cores work for highly parallel tests, but without seeing what code is actually running in those tests, its quite hard for me to tell. Either way actual performance is far more important, if the new system can run good games well, and the current Switch couldn't run most CPU intensive PS4 games well, thats all that matters, but I would be curious to see. Barely curious.
I mean, Doug knows way more than we do in his position at NoA, I wouldn't worry about anything leaking to him.
I’d suggest posters using the [ HIDE ] tag instead of [ REPLY ]. The latter requires the user to reply in order to view the post, leading to an influx of hello messages.
Edit: You’reallmostly lovely people, and it’s good to have more people posting. The suggestion is only to reduce forced participation.
HIDE is actually pretty powerful in terms of flexibility, there's a slew of options you can throw at it. See https://famiboards.com/help/bb-codes/I actually don't think there's much point in [ HIDE ]. It's not at all difficult to create an account to lurk. If the purpose of hiding things is to keep them within the community, limiting to those who have posted at least one actual thought here seems more sensible to me
When they can sell it at 1000$ +.Why will Nintendo want to put cameras in a new console? I thought they have given up on that idea long ago. The original Switch doesn't have a camera and no AR games, so what makes everyone think that they might tackle it again?
A binned Xbox One S APU was briefly available, with the GPU disabled. And of course, folks have installed Linux on Switches. So while both machines are overclocked, we can directly compare themI question if this will be the case. With Switch, developers only have 3 CPU cores compared to the 6-7 core available on on PS4. The A57 cores have better performance per clock compared to the Jaguar cores, but the Jaguar cores are clocked 50% faster, so best case scenario the A57 cores perform on par with the Jaguar cores but there is 3-4 less cores available. With SNG, there will be core parity, 8 cores on PS5 and eight cores on SNG. The Ryzen cores are clocked significantly higher than the A78 cores so there will still be a hefty performance deficit, but seeing as how games often times do not distribute the work load evenly across all the cores. Games that load the first three cores heavy while leaving the remaining cores idling along at 25% would present a bigger challenge because the single threaded performance of the Ryzen CPU will trounce the A78, but that does at least leave the door open for the porting developer to optimize the game to better utilize additional cores on SNG. I know the Ryzen CPU is capable of 16 threads, but I am skeptical that many games make significant use of all 16 threads. Judging from games on PC, most games still tend to load up the first few threads heavy with a significant drop off in utilization after that.
Not a fan of these options. Why does something like this exist but I still can't do a countdown/time zone stamp? Lol.HIDE is actually pretty powerful in terms of flexibility, there's a slew of options you can throw at it. See https://famiboards.com/help/bb-codes/
For example, I can do this for a days registered gate:
* Hidden text: cannot be quoted. *
Or this for a total posts (across all threads) gate:
* Hidden text: cannot be quoted. *
Or even combine things, like this:
* Hidden text: cannot be quoted. *
Perhaps it'd be sensible to move towards something like that? Or perhaps that's too gated, I'm not sure. The interesting thing here is that it's up to the poster how hidden the information should be.
Approving slurp noisesa post with my thoughts?! well I think [garbled noises and visceral screams]
basically I’m very intrigued about what it will be. I’m very interested in keeping tabs on speculation, but not enough to get too worked up about it, meaning… I sometimes get tired and dodge a little
however, I know this for sure:
the launch title is King’s Quest VIII: Mask of Eternity
Flops numbers yeah. But in actual performance, GPU wise, the Switch is a lot closer than 1/5. Closer to 1/3 in some scenarios if you count TX1's more efficient hardware and mixed precision being taken advantage of. This is assuming ports aren't being bottlenecked in CPU and RAM.There is some nuance there, but yeah. To vastly oversimplify you can think of the Switch as like 1/5th of a PS4 in terms of GPU performance, and the NG as, like, 1/3rd of a PS5. That's a very silly way of thinking of things, but it does help you sort of see that the gap is closing, but the gap isn't tiny.
Is it really? Switch A57s vs PS4 Jaguars has a ~3.5x speed gap.On the other hand, if we look at the CPU, the story inverts. The gap between NG and current gen is larger than it was between Switch and last gen.
maybe there is no DLCI've got a feeling the Tears of the Kingdom DLC may be a significant expansion that comes with the Switch 2 and will be a huge highlight of the showcase event, a little unorthodox perhaps (or not really?) but the radio silence is odd considering we're about to enter month five post-release while Breath of the Wild's DLC was confirmed weeks before its launch
Um please don't utter such blasphemy in this sanctuary again please and thank youmaybe there is no DLC
A binned Xbox One S APU was briefly available, with the GPU disabled. And of course, folks have installed Linux on Switches. So while both machines are overclocked, we can directly compare them
Switch vs Xbox One S: The Switch outperforms in single threading so handily, that it is likely that for most tasks, even at base clocks, it outperforms last gen. And as for multithreading, while the 8 cores of the Jaguar get ahead, they don't double the performance. Last gen games were dominated by single threaded code, this is an ideal scenario from a game's perspective.
Similarly, the PS5 APU is available to purchase with the GPU disabled, and while we don't have Switch NG, we do have Orin NX, which is a similar configuration of similar cores. Here are the two, back to back, at clock speeds that looks like the real world.
Orin vs PS5: The single threaded score has widened, exactly as you'd expect. But so has the multicore. Now, you might be right. Practically speaking, this extra horsepower might rarely be a barrier to ports. But there isn't a question that the gap is wider than last gen on both fronts.
Was just thinking the same thing! The next 2-3 months are gonna be very fun. December might slow down a bit, but all of 2024 should be constant anticipation and payoffIs this the most exciting period for Nintendo speculation in recent memory? Hardware hype, Direct hype, Metroid Prime 4 hype, Mario Wonder hype, Smash update hype, and possible TotK DLC hype? In the Switch's 7th year, after they already released a new FE, Zelda, Pikmin, Metroid Prime Remaster, Kirby Deluxe, Advance Wars, Sea of Stars, and even Mario RPG still to come?
Would it be possible for someone to post a summary of what all has been happening the past week?Thank FUCK I caught up! This thread has been on fire this past week.
the most important stuff is that there might have been some presentations at gamescom to developers regarding Drake. after that, it's all shitposting and unverifiable stuffWould it be possible for someone to post a summary of what all has been happening the past week?
It’s so hard to keep up.
Okay so.......Would it be possible for someone to post a summary of what all has been happening the past week?
It’s so hard to keep up.
Would it be possible for someone to post a summary of what all has been happening the past week?
It’s so hard to keep up.
-Switch 2 devkits have been here a while, I can verify it is backwards compatible with a few games tested. New cartridge. New camera feature.
-FF7 Remake looks and runs like a PS5 game on Switch 2 devkit. Port took no time I'm told. Could be a launch game, not sure yet.
I'm expecting/hoping TotK DLC news any time.I've got a feeling the Tears of the Kingdom DLC may be a significant expansion that comes with the Switch 2 and will be a huge highlight of the showcase event, a little unorthodox perhaps (or not really?) but the radio silence is odd considering we're about to enter month five post-release while Breath of the Wild's DLC was confirmed weeks before its launch
BLASPHEMY!maybe there is no DLC
Would it be possible for someone to post a summary of what all has been happening the past week?
It’s so hard to keep up.
To put it simply :Would it be possible for someone to post a summary of what all has been happening the past week?
It’s so hard to keep up.
Yeah, with talk of CPU no longer being the bottle neck for the current gen systems, I think there are far greater concerns than Drake for the new system. Especially bandwidth.outside of heavy simulation games, I don't expect Drake's cpu to be overburdened by many games. even Starfield is showing some good scaling on lower end cpus
Got you, so I think that's where the confusion is coming from. Im not quite sure how to come to terms with such different results with Geekbench seeming to claim PS4 has slightly better single core performance than Switch, but again, there are so many variables involved with these tests (Im going off 2019 geekbench scores, but no clue if those are 20nm or 12 nm, and if they are overclocked or not. Also geekbench might not compare to these other tests). As for multi-core stuff, yeah, I think that's the main point of contention, given how much of last gen was focused on single threaded code, that would be more than enough reason to believe that the Switch wasn't in practice that much less performant than the PS4 and Xbox One, and how it performed in practice is far more interesting than the results from benchmarks and tests.A binned Xbox One S APU was briefly available, with the GPU disabled. And of course, folks have installed Linux on Switches. So while both machines are overclocked, we can directly compare them
Switch vs Xbox One S: The Switch outperforms in single threading so handily, that it is likely that for most tasks, even at base clocks, it outperforms last gen. And as for multithreading, while the 8 cores of the Jaguar get ahead, they don't double the performance. Last gen games were dominated by single threaded code, this is an ideal scenario from a game's perspective.
Similarly, the PS5 APU is available to purchase with the GPU disabled, and while we don't have Switch NG, we do have Orin NX, which is a similar configuration of similar cores. Here are the two, back to back, at clock speeds that looks like the real world.
Orin vs PS5: The single threaded score has widened, exactly as you'd expect. But so has the multicore. Now, you might be right. Practically speaking, this extra horsepower might rarely be a barrier to ports. But there isn't a question that the gap is wider than last gen on both fronts.
I passed all those activity checks with more than 99% of all my activity being in this thread alone.HIDE is actually pretty powerful in terms of flexibility, there's a slew of options you can throw at it. See https://famiboards.com/help/bb-codes/
For example, I can do this for a days registered gate:
* Hidden text: cannot be quoted. *
Or this for a total posts (across all threads) gate:
* Hidden text: cannot be quoted. *
Or even combine things, like this:
* Hidden text: cannot be quoted. *
Perhaps it'd be sensible to move towards something like that? Or perhaps that's too gated, I'm not sure. The interesting thing here is that it's up to the poster how hidden the information should be.
Thick with two c’sDouble thick in the top half to facilitate double-stacked ROM chips, yes please!
My god, were the Jaguar really that bad? Admittedly, even the A78s hold up better than expected against the Zen 2s (in this benchmark, of course).A binned Xbox One S APU was briefly available, with the GPU disabled. And of course, folks have installed Linux on Switches. So while both machines are overclocked, we can directly compare them
Switch vs Xbox One S: The Switch outperforms in single threading so handily, that it is likely that for most tasks, even at base clocks, it outperforms last gen. And as for multithreading, while the 8 cores of the Jaguar get ahead, they don't double the performance. Last gen games were dominated by single threaded code, this is an ideal scenario from a game's perspective.
Similarly, the PS5 APU is available to purchase with the GPU disabled, and while we don't have Switch NG, we do have Orin NX, which is a similar configuration of similar cores. Here are the two, back to back, at clock speeds that looks like the real world.
Orin vs PS5: The single threaded score has widened, exactly as you'd expect. But so has the multicore. Now, you might be right. Practically speaking, this extra horsepower might rarely be a barrier to ports. But there isn't a question that the gap is wider than last gen on both fronts.