-
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| (Read the staff posts before commenting!)
- Thread starter Dakhil
- Start date
It's not the biggest red flag. Could just be a devkit vs retail thing, with the extra 4GB being for debug. The amount of RAM that will be available on retail devices would be something devs would be aware of, even if it's potentially subject to change.
Not saying I think the post is necessarily trustworthy, just that those numbers aren't inherently suspicious.
- Pronouns
- He/Him
Well, it all comes down to cost, really. Assuming [REDACTED] is backwards-compatible, then I fully expect it to fall back on a software-based solution at some point. Whether that's at launch, or years after. I assume that hardware-based BC will become expensive to keep including in every new unit as the years go on, whereas software emulation is basically a one-time cost.
I'd be happy enough if they just emulated Switch games to the best of the Switch's ability - i.e. 1080p 60fps with good post-process anti-alising - and then for select, specific games, they release a patch to add DLSS and perhaps a few other enhancements. Hardware or software, they'd need to update them anyway, even if they just want to increase resolution or framerate targets.
I'd say that the most likely scenario is that [REDACTED] runs most Switch games at 1080p with a few games recieving a boost to 60fps, and even fewer recieving a full next-gen update that could be anything from as little as a resolution increase, to something more expansive like updated textures and assets.
D36
#WOHOYHASAWAKENED
- Pronouns
- He/Him
God damnit, I hardly even read the part about other hybrid consoles!
My fault, I am TERRIBLE at reading sarcasm, especially through text.
Goodtwin
Like Like
The video capture was simply not good to begin with and YouTube made it even worse. When you compare the video capture quality compared to other trailers for TotK on YouTube, its noticeably worse. Nintendo deserves some grief for this. They didn't put this 10 minute gameplay video in position to really shine. Even the video quality on the Switch News Page is sub par. Maybe a bit better than on YouTube, but still bad.
D
Deleted member 887
Guest
Huffman coding is a common principle in information theory and in compression, and is a predecessor to the run length encoding scheme of Lempel-Ziv. In bioinformatics there are both Huffman-Lempel-Ziv variants and Hybrid-Lempel-Ziv which are referred to as HLZ. My understanding is that both are bespoke variants designed to make random access across compressed data (absolutely necessary when you're talking petabytes of genetic data) faster.
Thraktor
"[✄]. [✄]. [✄]. [✄]." -Microsoft
- Pronouns
- He/Him
While it's probably technically possible, my guess is that the overhead of such a scheme might exceed the benefit. I'd also say that, when developers are concerned about working within a fixed frame time they'd probably prefer a consistent solution over one that may have a lower average execution time, but more variability.
The BCM43013 would be pretty surprising, as it's a significant downgrade. It only supports one stream of 802.11n at 72Mbps (so not even WiFi 5, let alone 6), whereas the old chip supported dual-stream 802.11ac (WiFi 5, by modern branding). Nintendo aren't getting the most out of the chip, but a quick check on my router shows that it is using a dual-stream connection, with speeds in my case topping out at 173Mbps. A speed check on the Switch itself shows 82Mbps downloads, which would mean slower downloads on the new system even if Nintendo's CDN doesn't improve.
The BCM43013 appears to be aimed at wearables, where data rates are unlikely to go beyond the low single-digit Mbps and batteries are tiny. I'd guess Nintendo will use whatever the baseline dual-band option is, which appears to be the BCM4375 at the moment.
Lol I'm stupid sorry
ShaunSwitch
Moblin
It's to keep 100+ stack overflow tabs open on chrome.
ItWasMeantToBe19
Manakete
Releasing after November 2024 enters into "ha ha, what the fuck?" territory.
I wouldn't really understand it at all as the Switch is clearly declining.
Erika
The New sWitch of Truth
- Pronouns
- They/them
Holiday 2024 is too late. Given their software release schedule and how bad the original switch hardware is in 2024.
- Pronouns
- He/Him
Nintendo and letting an opportunity slip away, name a more iconic duo
Zedark
The guy with the ToV avatar
I don't lol in real life often, but this got me good haha
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
2x6GB LPDDR5
Three words:
Availability, availability, availability.
6GB low to mid range phones and 12GB mid to high end phones are so, so common because 6GB modules of LPDDR5 are so available (and have been for a few years).
Thraktor
"[✄]. [✄]. [✄]. [✄]." -Microsoft
- Pronouns
- He/Him
I had considered Huffman, but I don't think Huffman Lempel-Ziv would be a particularly good name in this case, as many alternative general-purpose compression algorithms (eg DEFLATE) already combine Lempel-Ziv with a Huffman-based entropy coding. My mind just went to "hierarchical" as texture data is hierarchical in a sense (texture -> block -> shape/colour data), and rearranging or working within that hierarchy may lead to more efficient encoding, for example encoding colour data for the entire texture separately from shape data. As I said, though, I'm probably just over-analysing a random letter someone picked, so it's probably not worth putting that much thought into.
Kise Ryota
Bob-omb
Well, like I said, worked amazingly well for me. Servers were stable and the rollback worked as it should. I had a great experience. On the other hand, Smash was garbage.
- Pronouns
- he / him
The most likely scenario is that any game without a specific next-gen patch will run at whatever resolution/frame rate limits it was programmed to for base Switch. If they were (or at some point in the future) doing software emulation that allowed them to ignore those resolution settings, 1080p wouldn't be a limit either.
ItWasMeantToBe19
Manakete
Unless every Switch game needs to be patched for the Switch 2 for different shader compilations and so maybe you could just do a couple adjustments anyway.
- Pronouns
- He/Him
Software emulation is taxing, though. Assuming it'll be about on the level of the Steam Deck, I doubt even Nintendo themselves could write a Switch emulator optimised enough to push beyond 1080p at full speed.
Goodtwin
Like Like
It really would be a historically bad decision. We have seen before how a very successful Nintendo console go from selling very well to selling very little, and there is no momentum left to carry into the next console. The impossible Switch ports are about dead and over with. Any big AAA games that haven't already been ported to Switch probably wont be, and third parties will just wait for Nintendo to put out more capable hardware. Software support outside of the Indies and simpler JRPG's are drying up. Even if Nintendo can maintain monthly releases of first party titles, I don't know how long that can keep bolstering sales. Consumer fatigue sets in as well. People who bought a Switch in 2017 are more and more likely to start looking for an upgrade elsewhere if Nintendo isn't going to offer one. For every new Switch Nintendo sells in 2023, there is probably a user from years past who grew tired of it and moved on to another platform. Nintendo is right at the critical period where their current generation console is starting to decline, and in all likelihood will decline pretty hard, and they have opportunity to reinvigorate appeal of the Switch with new more powerful hardware that gets AAA third party titles flowing to the platform again.
D
Deleted member 887
Guest
My mistake, I thought 4356 was single channel, from reading spec sheets. You're correct on BCM4375 likely being the better option
ShaunSwitch
Moblin
Haha anyone who has spent time coding gets this, glad to spread a little joy.
Sol
Moblin
It has to be 12 GB of RAM, right? 8 GB of RAM means the OS gets 1 and developers for the next 7+ years get around 7 GB to work with.
It reminds me of that speech Mark Cerny gave when he was talking about the development of the PS4. He said he was going around and talking to devs and gathering input, and was told multiple times that if the PS4 launched with 4 GB of RAM, it would be "DOA". So apparently Sony scrambled at the last minute and put 8 GB of RAM in the system.
I feel like that's the case here. A system coming out with 7 GB of usable RAM in 2024 is just a non-starter. And I know the Series S gives devs access to about 8.5 GB of RAM, but the Series S also has a 2.4 GB/s NVME drive to more quickly move data in and out of RAM, along with other techniques to reduce RAM usage. And even then it's still a problem on Series S.
The Switch 2 will be working with slower RAM and a slower I/O solution than the Series S. It absolutely needs 2x6 GB of RAM - 2 GB for the OS, 10 GB available for developers.
- Pronouns
- He/Him
I'm not praying for Nintendo's downfall, but I'd be lying if I said I would be upset if this happened again with [REDACTED]. Nintendo are at their most creative and most consumer friendly when they have to actually try, and can't just coast on free money.
- Pronouns
- He/Him
It’ll likely be 2x6GB of RAM of LPDDR5 as it would be “cheaper” in the long run considering that in the mobile space, 6GB of RAM as a simple Module isn’t uncommon and will be long lived especially by Apple product standards.
4GB will likely be short lived.
Also, people focus more on LPDDr5X and not LPDDR5T
I still think it’s 5, not 5X. But the most ideal scenario is of course 5T for I think >150GB/s memory bandwidth. I think, I could be mistaken of course and read it wrong.
- Pronouns
- He/Him
D
Deleted member 887
Guest
It would not be hard at all. Running an emulator one piece of specific hardware that only runs on a one other specific piece of hardware with an almost identical architecture is a very different problem from writing an emulator that runs on lots of hardware that is a very different architecture.
Yuzu running on Steam Deck needs to emulate 4 ARM CPUs on top of 4 x86 CPUs. ARM code needs to be captured, compiled to new x86 code, and executed at the same speed that, on Switch, the ARM code natively runs. On REDACTED, you're already on an ARM machine, so the code just... runs.
On Yuzu, this x86 code need to be patched whenever it attempts to call OS services on Switch, with references to new, emulated OS services on SteamOS. Except it's not even emulated OS services on SteamOS, its generic emulated OS services designed to run anywhere Yuzu might run. On REDACTED, they just... call the OS service in the real OS. Maybe a middleware layer for translation, but no need to emulate the OS.
On Yuzu, all Nvidia Switch driver calls need to be captured and translated into AMD driver calls on SteamOS, but again, it needs to be done in a generic way that supports every possible graphics driver Yuzu might run on. REDACTED just needs to capture one set of Nvidia Switch driver calls, lightly edit them to run on the new Nvidia Switch driver, and pass them along.
On Yuzu, every shader that runs on the GPU needs to be captured, decompiled into something called an Intermediate Representation, the Intermediate Representation is compiled in a second Shader Programming language, which depends on which GPU you are using at the time, then recompile that generated shader code. On REDACTED, the shader code can skip those middle two steps and use an ultra faster translator that just updated Nvidia shadercode from version 7.5 to 8.8.
Yuzu is solving a totally different problem from a Switch emulator running on REDACTED, and is doing it while reverse engineering the Switch environment. Nintendo/Nvidia writing a Switch emulator for T239 that could outperform the original Switch by 100-200% isn't trivial exactly, but it's not rocket science.
- Pronouns
- He/Him
It depends on which of the Big N's makes it, though. NVIDIA? Obviously. Nintendo? Not so obviously. Nintendo are yet to write an emulator that outclasses PC emulators from the 90's, so I wouldn't count on them being able to achieve better-than-native performance.
I didn't really consider the fact that [REDACTED] could be so architecturally similar to the Switch that it would be closer to simulation than emulation, though, so in that context, yeah, it seems much more achieveable to go beyond 1080p.
D
Deleted member 1141
Guest
Ouch... You're not wrong, but ouch....
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
My baseline for the backwards compatibility is just that it functions without flaws or shader compilation stutters.
The higher end of my hopes would probably be for unpatched games to have a boost mode for steadier framerates and better dynamic resolutions, and for games to use their TV Mode resolution in handheld mode. While patched games could increase the resolution, maybe significantly.
TomNookYankees
Koopa
- Pronouns
- He/Him
Summer 2024 or later would be pessimistic. Most realistically optimistic is that nintendo announces in June/July 2023 and releases in Nov 2023. I'm leaning towards October 2023 announcement and March/April 2024 release similar to what they did with the switch.
- Pronouns
- He/Him
TonyBaduy
Moblin
- Pronouns
- He/Him
We are getting 8GB LPDDR4X and we are going to have to like it! HmphIt’ll likely be 2x6GB of RAM of LPDDR5 as it would be “cheaper” in the long run considering that in the mobile space, 6GB of RAM as a simple Module isn’t uncommon and will be long lived especially by Apple product standards.
4GB will likely be short lived.
Also, people focus more on LPDDr5X and not LPDDR5T
I still think it’s 5, not 5X. But the most ideal scenario is of course 5T for I think >150GB/s memory bandwidth. I think, I could be mistaken of course and read it wrong.
- Pronouns
- He/Him
So true grandpa, don’t worry the nurses are coming to give you your dinnerWe are getting 8GB LPDDR4X and we are going to have to like it! Hmph
lattjeful
Like Like
- Pronouns
- he/him
Can't believe you just spoke this into existence.We are getting 8GB LPDDR4X and we are going to have to like it! Hmph
On the off chance we get 8 GB of LPDDR4X, I know who to blame.
TonyBaduy
Moblin
- Pronouns
- He/Him
In my times we had 4GB of LPDDR4 and that was plenty!So true grandpa, don’t worry the nurses are coming to give you your dinner
Last week I was at the GDC, and I’m now just catching up on things. No, I amn’t in the game industry, but someone hooked me up with a pass. For professional reasons (don’t ask) I mainly went to product monetization talks and missed the fun sessions such as ray tracing and temporal upscaling.
Just skimming through the thread, there were quite some fireworks this past week. Ignoring the discourse around the false prophecy (Samsung 5LPP), I’d like to respond to @Thraktor's excellent post regarding the OLED model and DisplayPort. A while back I went down the same rabbit hole, but emerged with a different—admittedly less exciting—conclusion: The design of Switch OLED is less likely the remnant of a “Pro” model, but a pragmatic pandemic-era product that affords Nintendo the maximum level of adaptability. A hardware that can be enabled to support 4K on TV, HDR on OLED display and TV, and higher performance profiles if/when necessary.
Since I still have a lot to catch up on offline, I’m going to write my reasoning below rather haphazardly. Please pardon anything unclear or imprecise.
[Background]
According to Bloomberg, the mass production of the OLED model started in June 2021, and I’d guess that the product design concluded in 2H 2020 and production lines set up in 1H 2021. Note that the Pfizer-BioNTech vaccine was only emergency-approved by the FDA in December 2020, and the world’s logistics was still in chaos during that time frame. → When the OLED was being finalized, the business risk and logistical uncertainty due to the pandemic was still at an elevated level, to put it lightly.
Thanks to @oldpuck and a Korean leaker, we know that Nintendo tested an overclocked TX1, but it obviously didn’t work out. A Chinese uncle shared that in 2020 there was a Switch prototype that wasn’t the OLED, and a Japanese uncle stated that a Pro model was planned. As @LiC pointed out, the “11 developers” report came out after the OLED announcement, therefore these disgruntled devs probably weren’t complaining about the SWOLED. → Assuming the rumors were credible, the cancellation of Pro doesn’t seem to be a last-minute decision.
[Considerations]
So if the SWOLED isn’t a repurposed “Pro” model, why is it 4K capable? (And possibly HDR ready too; I’ll explain it further down in the post.)
[Reserved capability: better performance]
We all know that the Mariko SoC is capable of better performance than currently allowed. It can be unlocked with higher performance profiles in exchange for a shorter battery life. Some might think that the additional profiles would be a hassle for the developers; if they are planning to build cross-gen titles, however, supporting an in-between performance tier may not need much investment.
[Reserved capability: 4K and HDR TV output]
To @Thraktor's point, four data lanes are wired from SWOLED’s Mariko to the USB matrix. Since both the TX1 and PI3USB30532 support DisplayPort 1.2, the OLED model in theory can transmit 4K/60Hz/HDR10 video to the dock via DP Alt Mode. Note that at this data rate, HDR10 content doesn’t require chroma subsampling or Display Stream Compression (DSC, which wasn’t supported until DP 1.4 anyway).
On the OG dock, the STDP2550 DP-to-HDMI converter is only capable of HDMI 1.4b output, capping out at 4K/30Hz. (Technically HDMI 1.4b can transmit 4K/60Hz/HDR10 using chroma subsampling, but that isn’t an industry standard and thus not supported by STDP2550.) The OLED dock instead contains an RTD2172, which converts DP 1.4 to HDMI 2.0b, thus permitting 4K/60Hz/no HDR, or 4K/60Hz/HDR10 with chroma subsampling.
So here we have a 4K and HDR capable output chain from the source to the sink, just waiting to be activated. And lest we forget that the OLED dock inexplicably contains an updatable firmware.
[Reserved capability: HDR on OLED display]
The model number of the OLED panel in SWOLED is AMS699VC01. “AMS” denotes a rigid AMOLED, and “V” a panel in Samsung’s V series. Other devices that I could identify using a V series rigid AMOLED include:
To maintain efficiency and prevent burn-in, most OLED panels feature an Auto Brightness Limiter (ABL) that dials down all individual pixels’ luminance as the video content as a whole becomes brighter. You can see the ABL in action here (pay attention to the brightness of the blue Windows background). The Alienware AS3423DW, for instance, is rated at 1013 cd/m2 when 1% of the screen is white (1% APL), but drops to 258 cd/m2 when 100% of the screen is white (100% APL):
To no one’s surprise, there’s no industry standard regarding how “peak brightness” is measured. Alienware advertises the aforementioned monitor’s peak brightness at 1% APL. RTINGS.com, on the other hand, only tests peak brightness down to 2% APL. Likewise, “typical brightness” can be measured at 100% APL or another condition that only the manufacturer knows.
Further muddying the water is the Auto Static Brightness Limiter (ASBL) feature. It automatically dims the whole display if the video content is mostly static. For example, If one plays a visual novel such as the Great Ace Attorney on an OLED panel with ASBL, after a few minutes the brightness will decrease because the video is mostly static. You can see the ASBL in action here.
AFAIK, there’s no ABL or ASBL on the SWOLED. Better yet, the YouTube videos that I’ve seen testing the OLED model’s brightness were all done with a 100% white screen. Under these conditions, a panel exhibiting 350-370 cd/m2 luminance is actually pretty respectable. If AMS699VC01 is similarly spec’d as the other V series panels listed above, Nintendo may be able to activate the ABL to allow for a much higher peak brightness, and in turn HDR.
The VESA DisplayHDR “True Black 600” standard seems to support my hypothesis. The “True Black” tiers are specifically for OLED displays, taking into account their superior contrast ratio and differing peak/typical brightness levels. As the table below demonstrates, an OLED panel capable of sustaining 350 cd/m2 at 100% APL (for which the SWOLED already qualifies) and reaching 600 cd/m2 at 10% APL (for which the SWOLED should be able to qualify after enabling ABL) can be considered—among other requirements—HDR capable.
Note that the True Black 600 is currently the highest tier of DisplayHDR for OLED, and “provides up to 50X greater dynamic range and 4X improvement in rise time compared to DisplayHDR 1000”. The output would be gorgeous if Nintendo can enable it. Even if they drop it down to True Black 400/500 to conserve battery, the result can still be enough to satisfy most gamers.
[So you're telling me there's a chance]
Eh, I don’t want to give anyone false hopes, but can see three potential paths forward.
The Drake Switch is released on time at a reasonable price, and a Drake-based Lite follows shortly: This would be the best case for Nintendo. The OLED model would likely be phased out quickly, therefore no need to activate any of these reserved capabilities.
The Drake Switch is released timely but at a higher price point; the OLED and Lite remain in the market: Some major Nintendo franchises are family friendly/oriented. I’m not sure if the Mario, Pokemon, AC, Kirby, and Yoshi series would ever release a Drake exclusive until Nintendo is able to create a Drake-based Lite. To maintain the marketability of SWOLED (maybe even Lite) and prolong its service life, the higher performance profiles for Mariko could be unleashed. For product segmentation reasons, the 4K output probably won’t be turned on. The HDR capability might go either way.
In the doomsday scenario of Drake being canceled or severely delayed, or the global supply chain going to s*** yet again: Although Nintendo certainly don’t want this to happen, IMHO they made plans for this during the height of the pandemic. The OLED model would be hard-pressed to expand its product scope, by enabling all its reserved capabilities, to ride out the storm.
P.S.: This is also a long and roundabout way of saying that if the Switch successor reuses the same OLED panel, it probably can support HDR.
Just skimming through the thread, there were quite some fireworks this past week. Ignoring the discourse around the false prophecy (Samsung 5LPP), I’d like to respond to @Thraktor's excellent post regarding the OLED model and DisplayPort. A while back I went down the same rabbit hole, but emerged with a different—admittedly less exciting—conclusion: The design of Switch OLED is less likely the remnant of a “Pro” model, but a pragmatic pandemic-era product that affords Nintendo the maximum level of adaptability. A hardware that can be enabled to support 4K on TV, HDR on OLED display and TV, and higher performance profiles if/when necessary.
Since I still have a lot to catch up on offline, I’m going to write my reasoning below rather haphazardly. Please pardon anything unclear or imprecise.
[Background]
According to Bloomberg, the mass production of the OLED model started in June 2021, and I’d guess that the product design concluded in 2H 2020 and production lines set up in 1H 2021. Note that the Pfizer-BioNTech vaccine was only emergency-approved by the FDA in December 2020, and the world’s logistics was still in chaos during that time frame. → When the OLED was being finalized, the business risk and logistical uncertainty due to the pandemic was still at an elevated level, to put it lightly.
Thanks to @oldpuck and a Korean leaker, we know that Nintendo tested an overclocked TX1, but it obviously didn’t work out. A Chinese uncle shared that in 2020 there was a Switch prototype that wasn’t the OLED, and a Japanese uncle stated that a Pro model was planned. As @LiC pointed out, the “11 developers” report came out after the OLED announcement, therefore these disgruntled devs probably weren’t complaining about the SWOLED. → Assuming the rumors were credible, the cancellation of Pro doesn’t seem to be a last-minute decision.
[Considerations]
So if the SWOLED isn’t a repurposed “Pro” model, why is it 4K capable? (And possibly HDR ready too; I’ll explain it further down in the post.)
- Nintendo had enough time to remove those two pins to the PI3USB30532 crossbar, and could’ve kept the original DP-to-HDMI converter instead of upgrading to the RTD2172. Since every penny saved adds up to a hill of beans at Switch’s volume, this was certainly a deliberate design choice.
- At that moment the supply chain was still in chaos and the efficacy/availability of COVID vaccines was also in question, Nintendo probably couldn’t predict with high confidence when they’d be able to release the Switch successor.
- Hence, the company needed the OLED model to keep the platform’s momentum going for an undetermined amount of time—it could be years before the successor reaches the market. Even when it does, the quantity might be extremely limited (see: PS5 and XSX in 2021), necessitating a prolonged service of the SWOLED in that scenario.
- For these reasons, I believe that the OLED model was therefore spec’d with reserved hardware capabilities to cover all eventualities.
[Reserved capability: better performance]
We all know that the Mariko SoC is capable of better performance than currently allowed. It can be unlocked with higher performance profiles in exchange for a shorter battery life. Some might think that the additional profiles would be a hassle for the developers; if they are planning to build cross-gen titles, however, supporting an in-between performance tier may not need much investment.
[Reserved capability: 4K and HDR TV output]
To @Thraktor's point, four data lanes are wired from SWOLED’s Mariko to the USB matrix. Since both the TX1 and PI3USB30532 support DisplayPort 1.2, the OLED model in theory can transmit 4K/60Hz/HDR10 video to the dock via DP Alt Mode. Note that at this data rate, HDR10 content doesn’t require chroma subsampling or Display Stream Compression (DSC, which wasn’t supported until DP 1.4 anyway).
On the OG dock, the STDP2550 DP-to-HDMI converter is only capable of HDMI 1.4b output, capping out at 4K/30Hz. (Technically HDMI 1.4b can transmit 4K/60Hz/HDR10 using chroma subsampling, but that isn’t an industry standard and thus not supported by STDP2550.) The OLED dock instead contains an RTD2172, which converts DP 1.4 to HDMI 2.0b, thus permitting 4K/60Hz/no HDR, or 4K/60Hz/HDR10 with chroma subsampling.
So here we have a 4K and HDR capable output chain from the source to the sink, just waiting to be activated. And lest we forget that the OLED dock inexplicably contains an updatable firmware.
[Reserved capability: HDR on OLED display]
The model number of the OLED panel in SWOLED is AMS699VC01. “AMS” denotes a rigid AMOLED, and “V” a panel in Samsung’s V series. Other devices that I could identify using a V series rigid AMOLED include:
- Realme X50 Pro, 90Hz, DCI-P3, HDR10+, 1,000+ nits (peak)
- Meizu 17 and 17 Pro, 90Hz, DCI-P3, 700 nits (typical), 1100 nits (peak)
- Vivo S9, 90Hz, 98% NTSC, HDR10+
- Vivo V20 Pro and S7, 98.5% NTSC, HDR10
- OnePlus Nord, 90Hz, DCI-P3
- Honor Play5, HDR10
- Google Pixel 4a, HDR
To maintain efficiency and prevent burn-in, most OLED panels feature an Auto Brightness Limiter (ABL) that dials down all individual pixels’ luminance as the video content as a whole becomes brighter. You can see the ABL in action here (pay attention to the brightness of the blue Windows background). The Alienware AS3423DW, for instance, is rated at 1013 cd/m2 when 1% of the screen is white (1% APL), but drops to 258 cd/m2 when 100% of the screen is white (100% APL):
To no one’s surprise, there’s no industry standard regarding how “peak brightness” is measured. Alienware advertises the aforementioned monitor’s peak brightness at 1% APL. RTINGS.com, on the other hand, only tests peak brightness down to 2% APL. Likewise, “typical brightness” can be measured at 100% APL or another condition that only the manufacturer knows.
Further muddying the water is the Auto Static Brightness Limiter (ASBL) feature. It automatically dims the whole display if the video content is mostly static. For example, If one plays a visual novel such as the Great Ace Attorney on an OLED panel with ASBL, after a few minutes the brightness will decrease because the video is mostly static. You can see the ASBL in action here.
AFAIK, there’s no ABL or ASBL on the SWOLED. Better yet, the YouTube videos that I’ve seen testing the OLED model’s brightness were all done with a 100% white screen. Under these conditions, a panel exhibiting 350-370 cd/m2 luminance is actually pretty respectable. If AMS699VC01 is similarly spec’d as the other V series panels listed above, Nintendo may be able to activate the ABL to allow for a much higher peak brightness, and in turn HDR.
The VESA DisplayHDR “True Black 600” standard seems to support my hypothesis. The “True Black” tiers are specifically for OLED displays, taking into account their superior contrast ratio and differing peak/typical brightness levels. As the table below demonstrates, an OLED panel capable of sustaining 350 cd/m2 at 100% APL (for which the SWOLED already qualifies) and reaching 600 cd/m2 at 10% APL (for which the SWOLED should be able to qualify after enabling ABL) can be considered—among other requirements—HDR capable.
Note that the True Black 600 is currently the highest tier of DisplayHDR for OLED, and “provides up to 50X greater dynamic range and 4X improvement in rise time compared to DisplayHDR 1000”. The output would be gorgeous if Nintendo can enable it. Even if they drop it down to True Black 400/500 to conserve battery, the result can still be enough to satisfy most gamers.
[So you're telling me there's a chance]
Eh, I don’t want to give anyone false hopes, but can see three potential paths forward.
The Drake Switch is released on time at a reasonable price, and a Drake-based Lite follows shortly: This would be the best case for Nintendo. The OLED model would likely be phased out quickly, therefore no need to activate any of these reserved capabilities.
The Drake Switch is released timely but at a higher price point; the OLED and Lite remain in the market: Some major Nintendo franchises are family friendly/oriented. I’m not sure if the Mario, Pokemon, AC, Kirby, and Yoshi series would ever release a Drake exclusive until Nintendo is able to create a Drake-based Lite. To maintain the marketability of SWOLED (maybe even Lite) and prolong its service life, the higher performance profiles for Mariko could be unleashed. For product segmentation reasons, the 4K output probably won’t be turned on. The HDR capability might go either way.
In the doomsday scenario of Drake being canceled or severely delayed, or the global supply chain going to s*** yet again: Although Nintendo certainly don’t want this to happen, IMHO they made plans for this during the height of the pandemic. The OLED model would be hard-pressed to expand its product scope, by enabling all its reserved capabilities, to ride out the storm.
P.S.: This is also a long and roundabout way of saying that if the Switch successor reuses the same OLED panel, it probably can support HDR.
Im not feeling late year announcement and release in q1. Would they really want to lose holiday sales? I think Switch 1 sales would tank hard.
ItWasMeantToBe19
Manakete
If the Switch 2 doesn't release until Q4 2024, would 2x6GB LPDDR5X be cost viable for the Switch 2.
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
We can't know, but more importantly we know the SOC was designed, taped out and manufactured some time ago, so the likelihood of 5X is slim, though possible.
ItWasMeantToBe19
Manakete
"know"
If this doesn't release until November 2024, wouldn't that suggest that perhaps they moved on to something else or what.
D
Deleted member 887
Guest
The GB/GBC/GBA emulators on NSO are absolutely best-in-class, and with maybe one feature that they've clearly intentionally chosen not to implement (super GameBoy support) are absolutely unrivaled. NERD is an entire arm of Nintendo dedicated to R&D including emulation, and hold a number of related patents.
Nintendo has been delivering PC emulators to developers since the late 1980s, when Nintendo delivered a high quality SNES emulator to speed up SNES development before the console even launched. Nvidia has no significant emulation experience.
Open Source emulators, when popular, often receive absolutely huge amounts of technical investment. Yuzu, for example, has more active contributors than the entirety of MonolithSoft, and is over half a billion lines of code. The standard process for estimating software development costs (COCOMO II) suggests that it would take a team of 100 people 2.8 years to write at a cost of about $27 million.
Of course Yuzu is an incredible effort, but it's unsurprising it excels at the exact sorts of edge cases that the fan community cares about, while totally lacking in the turn-key operation that is critical for a commercial software company. This isn't indicative of Nintendo's ability to deliver a quality emulator.
Considering Nintendo will 1) likely keep new hardware in the Switch family, 2) has emphasized a desire for continuity across platforms, 3) has only delivered one non-BC handheld in it's entire lifetime, 4) will likely need some level of emulation for BC, Nintendo has a combination of the skills, the platform, and the market push to make Switch on REDACTED run at high perf and quality.
LiC
Member
That 4chan thread is very blatantly a repackaging of our own rumors and speculation in the format of a (fake) leak post. There are multiple one-to-one mappings between things discussed here and statements in the post (hopefully I don't have to explain why this makes it considerably less credible, despite being more believable). Meanwhile, the only new or different thing in it is the totally out there "CCD" concept. I think someone clearly combined existing speculation with a wild guess at a StreetPass-like gimmick.
It's not a matter of "pushing beyond" 1080p. The games have built-in maximum output resolutions, and they cannot be exceeded without modifying the games themselves. If the resolution is dynamic, then additional performance overhead may allow games to stay higher in their range than they would otherwise, but the maximum will not budge without a patch.
That said, certain games are already rendering at higher resolutions than can be output, and will benefit from higher output resolutions for free, but those are fairly rare.
Concernt
Optimism is non-negotiable
- Pronouns
- She/Her
"Know", as in "we can't know". Because we. Can't know.
While we "know" about the SOC. 2 years is a very short amount of time to design an all-new SOC, and it's extremely unlikely T239 was scrapped due to Linux updates.- Pronouns
- He/Him
Are there devices in mass production using LPDDR5X right now? If not, then no.
- Pronouns
- He/Him
That was SCD
Last edited:
