• Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
Status
Not open for further replies.

Mr_F_Snowman

Member
Oct 27, 2017
3,879
Comparing headphones to AR is some wild shit. Wheres the alternative to headphones / earphones? There is no other way to do what headphones do and get sound into your ears without it being blurted out to everyone around you hence they are extremely widespread. Phones / laptops or whatever can all do 99% of the stuff AR would do and will do it better in most parts. The comparison is ludicrously daft.

Also what happens when I'm driving down the motorway in my AR glasses when the overlay glitches out, completely fucking up my field of view and I career and smash into a truck? Or an add pops up doing the same thing?

How do they block out other sources of light? Do they form a complete shield around my face like Geordi La Forge which everyone will absolutely despise and never wear? Or are they magic and can stop light getting into my eyes some other way? If not they will be garbage for any video / gaming use.

How does taking a phonecall work? Phones are designed so there is a microphone near your mouth and speaker near your ear - how does AR replicate this? Do I need to wear headphones and a mic all day along with the glasses?

There a huge list of reasons why AR, like VR is only ever going to be niche
 
Dec 21, 2020
5,066
Can we go to Nintendo has planned to leave the home console space way before the Switch? =P

And switch was their only way of doing that, more specifically the Wii U needed to flop for them to have a reason to do that. They've tried to enter the portable only scene since the GC days, as I said before.

or rather, the end of the GC, to be exact
 
Last edited:

UltraMagnus

Banned
Oct 27, 2017
15,670
Comparing headphones to AR is some wild shit. Wheres the alternative to headphones / earphones? There is no other way to do what headphones do and get sound into your ears without it being blurted out to everyone around you hence they are extremely widespread. Phones / laptops or whatever can all do 99% of the stuff AR would do and will do it better in most parts. The comparison is ludicrously daft.

Also what happens when I'm driving down the motorway in my AR glasses when the overlay glitches out, completely fucking up my field of view and I career and smash into a truck? Or an add pops up doing the same thing?

How do they block out other sources of light? Do they form a complete shield around my face like Geordi La Forge which everyone will absolutely despise and never wear? Or are they magic and can stop light getting into my eyes some other way? If not they will be garbage for any video / gaming use.

How does taking a phonecall work? Phones are designed so there is a microphone near your mouth and speaker near your ear - how does AR replicate this? Do I need to wear headphones and a mic all day along with the glasses?

There a huge list of reasons why AR, like VR is only ever going to be niche

"Ever" can be a long time. In 20 years AR/VR head sets may well be far more commonplace than a standard big box home console in homes globally.
 

UltraMagnus

Banned
Oct 27, 2017
15,670
In 20 years we may be gaming SAO style, in our sleep.

I mean who knows, lol, but I think the "console" as a concept as we know it is probably gone. A standard smartphone will be able to natively generate graphics that are probably close enough to some level of photorealism that its "good enough" for most people and then beam that signal to any display around the house and/or streaming tech will be commonplace.

I can totally see more homes having a thin VR/AR headset, maybe multiples moreso than a home console.

But back to Nintendo now I think it's a good time to try VR for them. An Occulus Rift competior that could also run Switch games on the TV (or in VR view on a 20 foot screen) would likely do a lot better than Occulus Rift.
 
Last edited:

ShadowFox08

Banned
Nov 25, 2017
3,524
p9g5vqdBAoGb0wIc.jpg


don't think we had a die shot of the M1 yet. it seems to be as large as the TX1 at 119mm2
2.6 TFLOPs on a 5nm chip huh.

supposedly he GPU alone uses 10 watts. The benchmarks with these games aren't that impressive. https://www.notebookcheck.net/Apple...t= Manufacturer Apple , no 2 more rows
 

oneroom

Member
Dec 26, 2020
288
You have to take into account that while Apple is spending a lot of money to develop new socks with the latest manufacturing processes, nvidia/Nintendo can't do the same.
It seems that nvidia has been able to use mainly Samsung's manufacturing lines, most of which are used to produce the geforceRTX series.
Currently, there is not an abundance of RTX3000 series in the market.
It is likely that nvidia will continue to focus on geforce for Samsung's production line in 2021.
If the Nintendo Switch (2021) is released, I feel it will most likely be on an older manufacturing process (12nm?).
 

ILikeFeet

DF Deet Master
Banned
Oct 25, 2017
61,987
You have to take into account that while Apple is spending a lot of money to develop new socks with the latest manufacturing processes, nvidia/Nintendo can't do the same.
It seems that nvidia has been able to use mainly Samsung's manufacturing lines, most of which are used to produce the geforceRTX series.
Currently, there is not an abundance of RTX3000 series in the market.
It is likely that nvidia will continue to focus on geforce for Samsung's production line in 2021.
If the Nintendo Switch (2021) is released, I feel it will most likely be on an older manufacturing process (12nm?).
The switch is already on 12nm, so they'd just get a bigger, hotter chip than they have now
 

Deguello

Banned
Jan 14, 2019
269
The last couple of pages of this thread are reminding me of that scene in Bojack Horseman where a film project loses so much focus due to the creatives behind it trying to outsmart the medium that the end result goes from a movie, to an art installation, to a series or inter-related fortune cookies, to a bi-monthly curated box of snacks.

 
Dec 21, 2020
5,066
The switch is already on 12nm, so they'd just get a bigger, hotter chip than they have now
I wonder how well consoles sell compared to gpus, since NV said they had a good year and switch contributed to that, must mean that they wouldn't really be against NS doing well and providing for both their gpu line and the switch line.

And the switch is smaller, it'll offer better yields I think.
 

ILikeFeet

DF Deet Master
Banned
Oct 25, 2017
61,987
I wonder how well consoles sell compared to gpus, since NV said they had a good year and switch contributed to that, must mean that they wouldn't really be against NS doing well and providing for both their gpu line and the switch line.

And the switch is smaller, it'll offer better yields I think.
For AMD, consoles are way more profitable. But what isn't :P

Nvidia makes most of their money from data centers now anyway. Their GeForce cards have fewer chips per wafer, but when you sell a $1500, $700, and $500 chip, you're doing alright. The 3080 even made it to the steam hardware survey at .23%, which is a lot for Steam's install base
 

fwd-bwd

Member
Jul 14, 2019
726
As many of us probably have heard, Nintendo Japan recently posted that Banjo-Kazooie and Blast Corps would be released for the Wii U Virtual Console on 12/30, before they retracted it:


About 2 weeks ago, Nintendo took another curious action of updating the 3DS app Swapnote (aka Letter Box), even renaming it Swapnote Remastered:


Does anybody else find these intriguing? Although the simplest explanation is probably the correct answer (Occam's razor and all that), seeing that there's no new news, allow me to do some irresponsible speculation based on absolutely nothing... well, not exactly nothing. Someone who heard things commented that the NSO will be updated in April; I interpret it as a step in strengthening the Switch platform to support the upcoming iterative hardware updates. As the Nintendo brass said in September's policy briefing, they want to create "a positive cycle" and "long-term relationships" through the Nintendo Account and online services to retain consumers "across platform generations". Here's the briefing slide that I shared before (ignore my doodles in red for a minute):

J9cTfuq.png


It is within the realm of possibility that Nintendo may attempt to accelerate this "positive cycle" by adopting certain 3DS and Wii U online services for NSO—a quick boost to the Switch ecosystem (see the red dotted lines I added to the slide). While the erroneous Virtual Console listings and the mysterious Swapnote update are not a definitive proof of their revival on NSO, it could be an indication that they were being revised for an N64 library and some sort of content sharing service. Since the Nintendo Network for 3DS and Wii U is already on life support, might as well extract some value out of it?
 

Onix555

Member
Apr 23, 2019
3,380
UK
For AMD, consoles are way more profitable. But what isn't :P

Nvidia makes most of their money from data centers now anyway. Their GeForce cards have fewer chips per wafer, but when you sell a $1500, $700, and $500 chip, you're doing alright. The 3080 even made it to the steam hardware survey at .23%, which is a lot for Steam's install base
I wouldnt call them super profitable, the margins are really low.
 

Gurgelhals

Member
Oct 27, 2017
1,709
I wouldnt call them super profitable, the margins are really low.

The difference between AMD's and Nvidia's console deals is that the former desperately needed the revenue streams coming from them during the "lean" Bulldozer years while for the latter – with their highly profitable GPU and datacenter businesses and all – the entire thing always was and remains a nice extra.

It just keeps me wondering how much Nvidia are really invested in that business relationship for the long term. On the one hand, it's just a small (though not insignificant) part of their total business (approx. between 5-10% of their total revenue) and Nvidia is a company primarily focused on the types of business that yield high margins (and console deals are very much high-volume/low-margin-type deals). On the other hand, the Switch deal is basically the only thing that is keeping their Tegra/SoC product line afloat (they're not really making any significant headway in the automotive space and the rest of possible Tegra applications is super-niche). So maybe they're actually seriously committed to sticking with Nintendo here (that is, beyond just selling them a bunch of X1s for chump change simply because no one else would buy them). Idk...
 

oneroom

Member
Dec 26, 2020
288
It is understandable that maintaining the TX1 production line is futile.
The rumor of the TX1 being discontinued around early 2021 makes sense, but is there a tegra that can replace it?
The existing tegra family is a specialized soc for automotive applications.
The die size is large, and there seems to be a lot of wasted space in games.
It is possible that they will design a new tegra family that we don't know about, but it must cost a lot to make a new one.
 

Dakhil

Member
Mar 26, 2019
4,459
Orange County, CA
The rumor of the TX1 being discontinued around early 2021 makes sense, but is there a tegra that can replace it?
There has been speculation that the SoC of the next Nintendo Switch model is based on Orin S, except the CPU might be the Cortex-A78C instead of the Cortex-A78AE.
024290f99262aae4bc0ace3f65f37f61.jpg

The die size is large, and there seems to be a lot of wasted space in games.
I'm guessing in the worst case scenario, the die size of the next Nintendo Switch model's SoC is going to be as large as the Tegra X1, whilst being considerably smaller than the Tegra X1+ in the best case scenario.
 

Dekuman

Member
Oct 27, 2017
19,026
The difference between AMD's and Nvidia's console deals is that the former desperately needed the revenue streams coming from them during the "lean" Bulldozer years while for the latter – with their highly profitable GPU and datacenter businesses and all – the entire thing always was and remains a nice extra.

It just keeps me wondering how much Nvidia are really invested in that business relationship for the long term. On the one hand, it's just a small (though not insignificant) part of their total business (approx. between 5-10% of their total revenue) and Nvidia is a company primarily focused on the types of business that yield high margins (and console deals are very much high-volume/low-margin-type deals). On the other hand, the Switch deal is basically the only thing that is keeping their Tegra/SoC product line afloat (they're not really making any significant headway in the automotive space and the rest of possible Tegra applications is super-niche). So maybe they're actually seriously committed to sticking with Nintendo here (that is, beyond just selling them a bunch of X1s for chump change simply because no one else would buy them). Idk...

We don't know, but we have pretty good evidence the Nintendo deal is driving growth for its gaming revenues and more or less kept Tegra from getting axed.
Looked at in those terms, Nintendo is extremely important to them.

I think someone also mentioned previously it was a matter of pride for them to have a foothold in the console market, now they do.
 

DarthBuzzard

Banned
Jul 17, 2018
5,122
Comparing headphones to AR is some wild shit. Wheres the alternative to headphones / earphones? There is no other way to do what headphones do and get sound into your ears without it being blurted out to everyone around you hence they are extremely widespread. Phones / laptops or whatever can all do 99% of the stuff AR would do and will do it better in most parts. The comparison is ludicrously daft.

Also what happens when I'm driving down the motorway in my AR glasses when the overlay glitches out, completely fucking up my field of view and I career and smash into a truck? Or an add pops up doing the same thing?

How do they block out other sources of light? Do they form a complete shield around my face like Geordi La Forge which everyone will absolutely despise and never wear? Or are they magic and can stop light getting into my eyes some other way? If not they will be garbage for any video / gaming use.

How does taking a phonecall work? Phones are designed so there is a microphone near your mouth and speaker near your ear - how does AR replicate this? Do I need to wear headphones and a mic all day along with the glasses?

There a huge list of reasons why AR, like VR is only ever going to be niche
Yikes. Someone really doesn't understand AR it looks like. Saying that laptops and phones can do 99% of what AR can do is misinformed. AR can do far more than you can comprehend; there are plenty of uses that even I won't be able to guess in 2020.

Can laptops work as spatial computers? Can they give you perceptual superpowers? Can they provide presence? Of course not, and even the things they share like 2D computing is inherently better with AR, because AR is more versatile.

There are ways for AR glasses to block light in the field of view, but of course you can't do anything outside the field of view; that doesn't mean it will be garbage for gaming / video use, because we're still talking about a wide field of view with the right architecture.

Phonecalls would use mics built in to the headset which would also be used for audio control of reality, and speakers could be built into the glasses with optional support for earbuds or full headphones.

Do tell me though, how do laptops and phones do holocalls? I'd love to know just how that works.
 

Brofield

Member
Oct 25, 2017
3,896
As many of us probably have heard, Nintendo Japan recently posted that Banjo-Kazooie and Blast Corps would be released for the Wii U Virtual Console on 12/30, before they retracted it:


About 2 weeks ago, Nintendo took another curious action of updating the 3DS app Swapnote (aka Letter Box), even renaming it Swapnote Remastered:


Does anybody else find these intriguing? Although the simplest explanation is probably the correct answer (Occam's razor and all that), seeing that there's no new news, allow me to do some irresponsible speculation based on absolutely nothing... well, not exactly nothing. Someone who heard things commented that the NSO will be updated in April; I interpret it as a step in strengthening the Switch platform to support the upcoming iterative hardware updates. As the Nintendo brass said in September's policy briefing, they want to create "a positive cycle" and "long-term relationships" through the Nintendo Account and online services to retain consumers "across platform generations". Here's the briefing slide that I shared before (ignore my doodles in red for a minute):

J9cTfuq.png


It is within the realm of possibility that Nintendo may attempt to accelerate this "positive cycle" by adopting certain 3DS and Wii U online services for NSO—a quick boost to the Switch ecosystem (see the red dotted lines I added to the slide). While the erroneous Virtual Console listings and the mysterious Swapnote update are not a definitive proof of their revival on NSO, it could be an indication that they were being revised for an N64 library and some sort of content sharing service. Since the Nintendo Network for 3DS and Wii U is already on life support, might as well extract some value out of it?

Yeah, I mean maybe I'm just prone to conspiracy theories, but if they had the release of two N64 games on Wii U in their database, either someone's job is on the line for this little prank or plans got shuffled around and this curious little detail was mistakenly not edited in time.

Certainly having some Rareware representation on N64flix at launch would certainly go a long way in forming confidence that they'll have some good selections in the library to start
 

Mr_F_Snowman

Member
Oct 27, 2017
3,879
Yikes. Someone really doesn't understand AR it looks like. Saying that laptops and phones can do 99% of what AR can do is misinformed. AR can do far more than you can comprehend; there are plenty of uses that even I won't be able to guess in 2020.

Can laptops work as spatial computers? Can they give you perceptual superpowers? Can they provide presence? Of course not, and even the things they share like 2D computing is inherently better with AR, because AR is more versatile.

There are ways for AR glasses to block light in the field of view, but of course you can't do anything outside the field of view; that doesn't mean it will be garbage for gaming / video use, because we're still talking about a wide field of view with the right architecture.

Phonecalls would use mics built in to the headset which would also be used for audio control of reality, and speakers could be built into the glasses with optional support for earbuds or full headphones.

Do tell me though, how do laptops and phones do holocalls? I'd love to know just how that works.

Weird tone to take considering your argument basically boils down to "AR can do far more than you can comprehend; there are plenty of uses that even I won't be able to guess in 2020". You fail to give any weight to your argument besides what is essentially hype for a totally unknown product and ignore completely the fact common sense tells you people don't want to wear chunky, heavy glasses and earphones constantly all day.

When you say things like "2D computing is inherently better with AR, because AR is more versatile" you could say this about anything considering there is no detail or explanation. Surely VR is also better then than traditional 2D computing (you know as its more versatile) but that hasn't taken off? Why is AR so much more likely to blow up?

Is the computing power built in to the glasses? If so how are people meant to wear these things for extended periods given the weight? How do you get them to weigh no more than a normal pair of glasses and look good? If the computing is done on another device it has to be constantly tethered to how is that convenient? And even then you presumably still have to have at least some kind of battery, cameras, mics etc built into the glasses - presumably they aren't magically weightless? Any actual counterpoints to these with anything or substance?

I don't doubt AR will have it uses, but a future where they replace phones is a lifetime away
 

Dekuman

Member
Oct 27, 2017
19,026
AR will be useful but I feel like it will be a logical extension for phones with a glasses wearable

Gaming AR will be more specific
 

oneroom

Member
Dec 26, 2020
288
There has been speculation that the SoC of the next Nintendo Switch model is based on Orin S, except the CPU might be the Cortex-A78C instead of the Cortex-A78AE.

It is hard to say when the specs of the Nintendo Switch (2021) will be finalized, unless you are an insider.
However, we have to take into account that hardware development takes a considerable amount of time from specification determination to completion.
The cortex-a78 was announced in May 2020, so it seems unlikely that it will be used in Switch (2021), both in terms of timing and cost.
 

Skittzo

Member
Oct 25, 2017
41,037
It is hard to say when the specs of the Nintendo Switch (2021) will be finalized, unless you are an insider.
However, we have to take into account that hardware development takes a considerable amount of time from specification determination to completion.
The cortex-a78 was announced in May 2020, so it seems unlikely that it will be used in Switch (2021), both in terms of timing and cost.

Being publicly announced at a certain date doesn't mean it wasn't known behind the scenes. That said, I do agree it's probably a bit too new. There was an article over the summer that mentioned Samsung presented A76's to Nintendo so that's probably the best we can realistically hope for.
 

ILikeFeet

DF Deet Master
Banned
Oct 25, 2017
61,987
It is hard to say when the specs of the Nintendo Switch (2021) will be finalized, unless you are an insider.
However, we have to take into account that hardware development takes a considerable amount of time from specification determination to completion.
The cortex-a78 was announced in May 2020, so it seems unlikely that it will be used in Switch (2021), both in terms of timing and cost.
Orin was announced to use the A78 back in late 2019
 
Dec 21, 2020
5,066
As many of us probably have heard, Nintendo Japan recently posted that Banjo-Kazooie and Blast Corps would be released for the Wii U Virtual Console on 12/30, before they retracted it:


About 2 weeks ago, Nintendo took another curious action of updating the 3DS app Swapnote (aka Letter Box), even renaming it Swapnote Remastered:


Does anybody else find these intriguing? Although the simplest explanation is probably the correct answer (Occam's razor and all that), seeing that there's no new news, allow me to do some irresponsible speculation based on absolutely nothing... well, not exactly nothing. Someone who heard things commented that the NSO will be updated in April; I interpret it as a step in strengthening the Switch platform to support the upcoming iterative hardware updates. As the Nintendo brass said in September's policy briefing, they want to create "a positive cycle" and "long-term relationships" through the Nintendo Account and online services to retain consumers "across platform generations". Here's the briefing slide that I shared before (ignore my doodles in red for a minute):

J9cTfuq.png


It is within the realm of possibility that Nintendo may attempt to accelerate this "positive cycle" by adopting certain 3DS and Wii U online services for NSO—a quick boost to the Switch ecosystem (see the red dotted lines I added to the slide). While the erroneous Virtual Console listings and the mysterious Swapnote update are not a definitive proof of their revival on NSO, it could be an indication that they were being revised for an N64 library and some sort of content sharing service. Since the Nintendo Network for 3DS and Wii U is already on life support, might as well extract some value out of it?
If I remember correctly, what makes this even more peculiar is that the Switch, 3DS and the Wii U run on the same eShop. Switch is basically keeping the 3DS and the Wii U eShops alive and open.

Unless I'm mistaken. Though, 4 Years almost after the death of the Wii U, I hear people comment that they buy a thing and can access it to make purchases on the corpse.
 

Dakhil

Member
Mar 26, 2019
4,459
Orange County, CA
It is hard to say when the specs of the Nintendo Switch (2021) will be finalized, unless you are an insider.
However, we have to take into account that hardware development takes a considerable amount of time from specification determination to completion.
The cortex-a78 was announced in May 2020, so it seems unlikely that it will be used in Switch (2021), both in terms of timing and cost.
Trust me, I'm no insider, nor do I want to become one.

Since Nvidia announced on 17 December 2019 that Orin is going to use Arm Hercules cores, which is the codename for the Cortex-A78, it's very likely Nvidia (and other companies) probably knew about the Cortex-A78 at least a couple of years before it was formally announced by Arm, which skittzo0413 and ILikeFeet have mentioned.

This is speculation on my part, but I think it's very likely Nvidia knew about the Cortex-A78 since 2018, especially since Orin was first announced by Nvidia on 29 March 2018.
 

DarthBuzzard

Banned
Jul 17, 2018
5,122
Weird tone to take considering your argument basically boils down to "AR can do far more than you can comprehend; there are plenty of uses that even I won't be able to guess in 2020". You fail to give any weight to your argument besides what is essentially hype for a totally unknown product and ignore completely the fact common sense tells you people don't want to wear chunky, heavy glasses and earphones constantly all day.
In what world is the end goal of AR chunky heavy glasses? We're talking 30-40 grams, and even earphones will be optional.

I have already laid out plenty of uses of AR in this thread, so I have actually done my part - I'm saying that there will be plenty more uses that I haven't thought of because there is a set of ingredients that I'm aware of that will lead to things no one saw coming, the way no one saw SnapChat being used frequently on a slab of plastic in their pockets.

When you say things like "2D computing is inherently better with AR, because AR is more versatile" you could say this about anything considering there is no detail or explanation. Surely VR is also better then than traditional 2D computing (you know as its more versatile) but that hasn't taken off? Why is AR so much more likely to blow up?
It's simple really. It's too early, for either of them. VR will also be a better computing tool than anything we use today, but it's very clear that the clunkiness and low clarity of today's headsets cannot possibly live up to that. When I say more versatile, I mean that AR can do whatever a laptop does but provide more on top. I can sit in Starbucks and type away either on a flat surface or using a keyboard, but unlike a laptop now I have all the screen real estate I could ask for. As many screens as I want, in whatever position I want, with more intelligent interfaces - thanks to eye-tracking and gesture shortcuts.

I could also have a colleague in a holocall of sorts, where we could share screens, 3D models, and collaborate as if we were both in the same Starbucks.

Is the computing power built in to the glasses? If so how are people meant to wear these things for extended periods given the weight? How do you get them to weigh no more than a normal pair of glasses and look good? If the computing is done on another device it has to be constantly tethered to how is that convenient? And even then you presumably still have to have at least some kind of battery, cameras, mics etc built into the glasses - presumably they aren't magically weightless? Any actual counterpoints to these with anything or substance?
This is clearly a difficult process to get right, and it's going to take quite a while indeed. Compute can be cloud based in the long run, and various optimizations can be made to significantly reduce bandwidth and battery consumption via foveated rendering and smart use of AR maps where the glasses do not re-scan the same area twice, but adhere to a global map and update as that map changes.

Cameras and mics are smaller than you might think. Overall this is a big challenge, but not one that is insurmountable.
 
Last edited:

Hermii

Member
Oct 27, 2017
4,685
Being publicly announced at a certain date doesn't mean it wasn't known behind the scenes. That said, I do agree it's probably a bit too new. There was an article over the summer that mentioned Samsung presented A76's to Nintendo so that's probably the best we can realistically hope for.
The same rumor also said x1.
 

Mr_F_Snowman

Member
Oct 27, 2017
3,879
In what world is the end goal of AR chunky heavy glasses? We're talking 30-40 grams, and even earphones will be optional.

I have already laid out plenty of uses of AR in this thread, so I have actually done my part - I'm saying that there will be plenty more uses that I haven't thought of because there is a set of ingredients that I'm aware of that will lead to things no one saw coming, the way no one saw SnapChat being used frequently on a slab of plastic in their pockets.

It's simple really. It's too early, for either of them. VR will also be a better computing tool than anything we use today, but it's very clear that the clunkiness and low clarity of today's headsets cannot possibly live up to that. When I say more versatile, I mean that AR can do whatever a laptop does but provide more on top. I can sit in Starbucks and type away either on a flat surface or using a keyboard, but unlike a laptop now I have all the screen real estate I could ask for. As many screens as I want, in whatever position I want, with more intelligent interfaces - thanks to eye-tracking and gesture shortcuts.

I could also have a colleague in a holocall of sorts, where we could share screens, 3D models, and collaborate as if we were both in the same Starbucks.

This is clearly a difficult process to get right, and it's going to take quite a while indeed. Compute can be cloud based in the long run, and various optimizations can be made to significantly reduce bandwidth and battery consumption via foveated rendering and smart use of AR maps where the glasses do not re-scan the same area twice, but adhere to a global map and update as that map changes.

Cameras and mics are smaller than you might think. Overall this is a big challenge, but not one that is insurmountable.

Well thank you at least for a more measured response. But obviously I'm not saying the end goal is to have chunky, heavy stupid looking glasses. I'm asking how it is physically possible to have a device with computational power as good (or better) than a phone, that has an array of physical sensors and batteries in it that doesn't weigh anymore than a standard pair of glasses (that people already dislike wearing and find uncomfortable)? These aren't issues that are easy to hand wave away because things will improve in the future.

I'm not saying AR won't have uses - I'm saying comparing it to headphones is stupid and saying it will replace phones is ludicrous. At best you will still need another device that actually does the computational stuff so it doesn't replace anything, its going to rely on (presumably phones for the most part) doing the grunt work or otherwise you are 100% going to have big, chunky, heavy stupid looking glasses.

There are also many aspects that are worse - you talk about using any flat surface as a keyboard, gesture inputs and eye tracking - something you can literally do already with available devices but they are not popular because the experience is not actually very good. Why do you think haptics are a thing?

To actually have a decent AR experience you would realistically need not just glasses but a seperate computing device and some gloves with some kind of haptics - which will provide a great experience and have plenty of uses sure. But suggesting they will miraculously circumvent all these issues, totally replace phones in everyday life and be as ubiquitous as headphones is a complete fantasy
 

DarthBuzzard

Banned
Jul 17, 2018
5,122
Well thank you at least for a more measured response. But obviously I'm not saying the end goal is to have chunky, heavy stupid looking glasses. I'm asking how it is physically possible to have a device with computational power as good (or better) than a phone, that has an array of physical sensors and batteries in it that doesn't weigh anymore than a standard pair of glasses (that people already dislike wearing and find uncomfortable)? These aren't issues that are easy to hand wave away because things will improve in the future.

I'm not saying AR won't have uses - I'm saying comparing it to headphones is stupid and saying it will replace phones is ludicrous. At best you will still need another device that actually does the computational stuff so it doesn't replace anything, its going to rely on (presumably phones for the most part) doing the grunt work or otherwise you are 100% going to have big, chunky, heavy stupid looking glasses.
You're going to see many AR glasses tethered to phones this decade because the kind of ultra slim glasses that are being worked on in labs are not going to be ready as there will be plenty of optics problems and manufacturing issues that will be needed to be solved before that happens. That doesn't mean it won't happen though, just that it's going to be a long process. It's precisely the right kind of architecture and smart optimization that gets us there - breakthrough optimization like foveated rendering, resulting in serious power savings.

There are also many aspects that are worse - you talk about using any flat surface as a keyboard, gesture inputs and eye tracking - something you can literally do already with available devices but they are not popular because the experience is not actually very good. Why do you think haptics are a thing?
Gesture tracking today is very limited regardless of what device you're using. As seen in that sample video I showed on a previous page demoing AR gestures, we can get to a much more convenient and ultimately more accurate set of gestures in the future; hand-tracking is still in it's infancy in consumer devices but is near perfect in labs.



Eye-tracking doesn't get used much for 2D displays, so there isn't much of a push to design interfaces around it. Every VR/AR device will have eye-tracking as standard in the next few years, and so the entire medium will be built to leverage this. HoloLens 2 lets you scroll pages by gazing towards the bottom of a page. What I'd like to see is using eye-tracking as a second cursor, where I can map my middle mouse button to act as the eye's click.

To actually have a decent AR experience you would realistically need not just glasses but a seperate computing device and some gloves with some kind of haptics - which will provide a great experience and have plenty of uses sure. But suggesting they will miraculously circumvent all these issues, totally replace phones in everyday life and be as ubiquitous as headphones is a complete fantasy
Realistic force feedback haptic gloves are being built, and do exist in enterprise as bulky contraptions - but will need to be slimmed down for consumers. I never really mentioned these because I don't expect society will choose to wear these outside much for every day AR use. Instead, these seem much more applicable for stationary AR/VR experiences in the home. Providing haptics on the go may just come from a form of a wrist-band which is enough to trick you into believing something touched the tips of your fingers.
 
Last edited:

Dekuman

Member
Oct 27, 2017
19,026
If Nintendo is moving away from TSMC and needs to redo the TX1 SoC in anycase, I wonder if what is going to happen with the Pro is some frankenstein solution where the keep the TX1 GPU but swap out the A57s for a more capable CPU. So the Pro might be something like 4-5X CPU improvement (with 6 or 8 cores) and 1x on GPU with a simple OC.
 

ILikeFeet

DF Deet Master
Banned
Oct 25, 2017
61,987
If Nintendo is moving away from TSMC and needs to redo the TX1 SoC in anycase, I wonder if what is going to happen with the Pro is some frankenstein solution where the keep the TX1 GPU but swap out the A57s for a more capable CPU. So the Pro might be something like 4-5X CPU improvement (with 6 or 8 cores) and 1x on GPU with a simple OC.
but why. there's no "swap out", they have to make something from scratch anyway
 

Dekuman

Member
Oct 27, 2017
19,026
but why. there's no "swap out", they have to make something from scratch anyway
Correct, if they have to make it from scratch due to TSMC IP, but want to keep maximum scalability with their OG Switch line, they may opt to keep the same GPU and just upgrade the CPU. This is close to what they did with the new 3DS except the GPU upgrade could be more significant (iirc, the n3DS GPU wasn't upclocked at all, just the CPU with higher clocks and more cores)

This would allow them to disable cores/clock down ther CPU for the OG Switch that sells alongside their Pro Model using the exact same chip.
 

ILikeFeet

DF Deet Master
Banned
Oct 25, 2017
61,987
Correct, if they have to make it from scratch due to TSMC IP, but want to keep maximum scalability with their OG Switch line, they may opt to keep the same GPU and just upgrade the CPU. This is close to what they did with the new 3DS except the GPU upgrade could be more significant (iirc, the n3DS GPU wasn't upclocked at all, just the CPU with higher clocks and more cores)

This would allow them to disable cores/clock down ther CPU for the OG Switch that sells alongside their Pro Model using the exact same chip.
for what reason do we assume that a newer architecture is incompatible with the current software stack? also, the issue with comparing the pro with the n3DS is that the 3ds was two chips: a separate cpu and gpu. if they really wanted to keep everything the same, just more, why not stick with the A57 and add another cluster of A57s? it'd be more akin to the Wii > Wii U, especially with the turning off cores for the based Switch (though, there's still no point in doing that)
 

Hermii

Member
Oct 27, 2017
4,685
Correct, if they have to make it from scratch due to TSMC IP, but want to keep maximum scalability with their OG Switch line, they may opt to keep the same GPU and just upgrade the CPU. This is close to what they did with the new 3DS except the GPU upgrade could be more significant (iirc, the n3DS GPU wasn't upclocked at all, just the CPU with higher clocks and more cores)

This would allow them to disable cores/clock down ther CPU for the OG Switch that sells alongside their Pro Model using the exact same chip.
This made sense in the age of the 3ds, but the Switch platform was developed in the age of hardware agnostic APIs and dev environments. It would be incredibly short sighted if the current software stack is not compatible at the very least with modern Nvidia architectures.
 

Dekuman

Member
Oct 27, 2017
19,026
for what reason do we assume that a newer architecture is incompatible with the current software stack? also, the issue with comparing the pro with the n3DS is that the 3ds was two chips: a separate cpu and gpu. if they really wanted to keep everything the same, just more, why not stick with the A57 and add another cluster of A57s? it'd be more akin to the Wii > Wii U, especially with the turning off cores for the based Switch (though, there's still no point in doing that)

No reason at all, i'm splitballing ideas, 'out of left field' stuff Nintendo may pull. Doing a frankenstein build would certainly be one and allow some of the contradictory information we're hearing to make sense.

The biggest one for me is, (assuming TX1 production ends early 2021) how they will handle the very popular OG Switch performance profile going forward. I just don't see them moving 80+ million users into legacy mode in 2021, the transition will take time. If they don't have at least a couple of years of inventory , then they will need a way to continue making OG Switch at the OG perf profile. In that sense, I can't see them bundling expernsive new SoC into OG Switch configuration downclocked to hell with DLSS disabled. It could just be cheaper, even if redoing the die is included to have 2 chips, or if they go with 1 chip, have a chip that can work on both a Pro and OG.

Of course I could be wrong and they may very well come out and tell people who have an OG Switch that their model is now a legacy one, but I'd be more surprised by that than Nintendo going with a brand new SoC in 2021.
 

Skittzo

Member
Oct 25, 2017
41,037
No reason at all, i'm splitballing ideas, 'out of left field' stuff Nintendo may pull. Doing a frankenstein build would certainly be one and allow some of the contradictory information we're hearing to make sense.

The biggest one for me is, (assuming TX1 production ends early 2021) how they will handle the very popular OG Switch performance profile going forward. I just don't see them moving 80+ million users into legacy mode in 2021, the transition will take time. If they don't have at least a couple of years of inventory , then they will need a way to continue making OG Switch at the OG perf profile.

Of course I could be wrong and they may very well come out and tell people who have an OG Switch that their model is now a legacy one, but I'd be more surprised by that than Nintendo going with a brand new SoC in 2021.

They don't really like to discuss specs in general. It wouldn't shock me if they just released the *new* Switch with some cosmetic differences and said nothing about performance increases (unless it has DLSS for 4k), and we find out it does have a substantial processing upgrade like the *new* 3DS. That way they're not telling anyone that their old Switch is a legacy model.
 

Dekuman

Member
Oct 27, 2017
19,026
They don't really like to discuss specs in general. It wouldn't shock me if they just released the *new* Switch with some cosmetic differences and said nothing about performance increases (unless it has DLSS for 4k), and we find out it does have a substantial processing upgrade like the *new* 3DS. That way they're not telling anyone that their old Switch is a legacy model.
Oh that is possible too
 

ILikeFeet

DF Deet Master
Banned
Oct 25, 2017
61,987
No reason at all, i'm splitballing ideas, 'out of left field' stuff Nintendo may pull. Doing a frankenstein build would certainly be one and allow some of the contradictory information we're hearing to make sense.

The biggest one for me is, (assuming TX1 production ends early 2021) how they will handle the very popular OG Switch performance profile going forward. I just don't see them moving 80+ million users into legacy mode in 2021, the transition will take time. If they don't have at least a couple of years of inventory , then they will need a way to continue making OG Switch at the OG perf profile. In that sense, I can't see them bundling expernsive new SoC into OG Switch configuration downclocked to hell with DLSS disabled. It could just be cheaper, even if redoing the die is included to have 2 chips, or if they go with 1 chip, have a chip that can work on both a Pro and OG.

Of course I could be wrong and they may very well come out and tell people who have an OG Switch that their model is now a legacy one, but I'd be more surprised by that than Nintendo going with a brand new SoC in 2021.
crippling chips is just a waste of money. if they're gonna make all switches use them, then there's no point in spending extra resources to a lower the performance of a chip. I expect them to just take less revenue on OG styled switches, while making more on premium switches. that way, ever new buyer is on the better performance profile and will give incentive for devs to prioritize that
 

Dakhil

Member
Mar 26, 2019
4,459
Orange County, CA
I wonder which is more likely for the next Nintendo Switch model's SoC:
  • The SoC will have roughly the same die size as the Tegra X1 (118 mm²)
  • The SoC will have a considerably smaller die size than the Tegra X1+
 
Jan 10, 2018
7,207
Tokyo
They don't really like to discuss specs in general. It wouldn't shock me if they just released the *new* Switch with some cosmetic differences and said nothing about performance increases (unless it has DLSS for 4k), and we find out it does have a substantial processing upgrade like the *new* 3DS. That way they're not telling anyone that their old Switch is a legacy model.

That would be such a U-turn as compared to the strategy z0mbi3e, myself, my mom and my uncle who works at tsmc heard they were following. It would personally sour me quite a bit to see Nintendo being so conservative in their approach.
 

Skittzo

Member
Oct 25, 2017
41,037
That would be such a U-turn as compared to the strategy z0mbi3e, myself, my mom and my uncle who works at tsmc heard they were following. It would personally sour me quite a bit to see Nintendo being so conservative in their approach.

Hmm? I'm saying it would still be a substantial upgrade, they just wouldn't actively discuss the specs publicly. I don't see how that would be any different from what you've heard.
 

fwd-bwd

Member
Jul 14, 2019
726
I got around to skim through the annual Famitsu survey of Japanese devs, and encountered this irrefutable proof of a much more powerful Switch model in 2021:

Tomio Kanazawa, Toybox Inc.: "In 2021, I'd like to sublimate Deadly Premonition a bit more"

There you have it folks. Happy new year!
 
Status
Not open for further replies.