Samsung and AMD = Quest type product?

[funkydmunky]

Well known that Samsung has Radeon GFX partnership. Being that Samsung is the king of displays and has been more then quiet on the VR front for ever, could a Quest 2/3 killer be in the wings? I don't see super gaming phones being the strategy here.

 

[StandartBlitz]

I wouldn't say the silence has to mean much since Samsung slept for years on the gaming monitor market as well. But it wouldn't surprise me either - they had a quick glimpse into VR with the "successful" Gear VR so why shouldn't they try it themselves?
I would be really nice to finally see some competition in this price range.

 

[funkydmunky]

I wouldn't say the silence has to mean much since Samsung slept for years on the gaming monitor market as well. But it wouldn't surprise me either - they had a quick glimpse into VR with the "successful" Gear VR so why shouldn't they try it themselves?
I would be really nice to finally see some competition in this price range.

Uh? Are you up on this or just throwing out?
They did. Odyssey VR/+ It was OLED and had the first No-Screen-Door for VR which some still struggle with to this day (Index)
I feel something is up.
Maybe PSVR-2?

 

[StandartBlitz]

I was but nowadays I don't keep up with most of the VR related news - so just throwing out xD

 

[MichaelSulz]

They did. Odyssey VR/+ It was OLED and had the first No-Screen-Door for VR which some still struggle with to this day (Index)
I feel something is up.
Maybe PSVR-2?

and it's not bad at all)))))))

 

[Chief Blur Buster]

Buzz suggests Samsung won’t directly make it sooner than 2-3 competitors. Currently, vendors capable of quickly doing a profitable Quest 2 competitor right out of the gate are as follows:

- Apple VR with the M1 chip (ETA 2023?), using their massive war chest of funds.
- Valve Index 2 with built-in GPU (ETA 2022-23?), using ready-made storezilla Steam.
- HTC may release a low price consumer version of their existing new enterprise standalone VR headset, once economically feasible to compete with Quest 2.

Honorary mentions of formidable names that won’t hit/return to market as quickly:

- Google Daydream 3 (if not for their gold-trophy winner in # of project cancels)
- Samsung (but they cancelled GearVR & let GearVR app store go fallow)
- Microsoft (Skunkworks possibility, laggard as they may be, they have HoloLens and increasing VR experience, like improved Flight Simulator VR mode, and they are a gaming giant in Xbox).

Apple & Valve has the economic eagerness, hardware, software, app infrastructure & knowhow to launch a Quest 2 competitor hard & successfully — at the right time (after fall in parts costs, fab shortage, pandemic, etc). All three has evidence of already worked (or commercialized in different market) on prototyping Quest-like standalones.

Samsung likely has as well, but it’s clear their economic focuses are currently elsewhere, after the Samsung chairman recently exited incarceration (so they’re for now, being conservative on focussing on established successful markets), and they are clearly focussing on ramping up other product lines for the forseeable future.

VR is an expensive-to-launch niche that requires strong launch-time commitment by an existing gamezilla, storezilla, and/or from well tested currently-active product lines (such VR headsets, consoles, etc), and Samsung has clearly shied away from all of these (for now). They have superior engineers with in-house parts such as OLED, chips, and sensors but insufficient fresh/refreshed app store foundation to bootstrap popularity necessary to compete with established 800 pound app store gorillas (Oculus, Apple & Steam) right out of the gate at beginning. Unless Samsung fully yields to near-perfect SteamVR integration, or publishes a new very open, easier, cheaper app store that creates a mad rush of VR ports, which is unlikely given Samsung’s current priorities. Samsung created standalones already but let all of them essentially go fallow, so they’re probably unprepared to launch as quickly. Actions speak louder than words, like a Google product cancellation, unfortunately

HTC is wisely incubating their brand new standalone first in the high-paying enterprise market - gaining experience for launching a potential future consumer VR headset once (and if) market conditions aligns for them. But a strike exists, HTC sold their smartphone team to Microsoft a few years ago, and VR headsets are full of smartphone tech (cameras, sensors, GPUs, repurposed screens, etc), so they might be short in tech to launch big. However, they still built a standalone which they sell to the enterprise market, so it’s not a large stone as they are more willing to truly happily marry SteamVR than Samsung is.

Samsung and Google may try again - but probably not after there is already two successful consumer standalones on the market.

Apple may beat Valve to market for a true Quest 2 competitor. but they want to launch a nice high-cost AR-priority unit first before a Quest 2 priced unit. Oculus will still be cheaper than both, but, reportedly many people will pay at least a little extra to avoid Facebook, even if superior/cheaper.

Valve has a giant storezilla for better or for worse, but it’s a lesser of evil for a lot of people who prefer Steam over Oculus, despite Oculus Store generally superior in VR app quality in perfect stutter free framepacing. Valve can simply fork a slightly performance-restrictive version of app store for their standlone, with excellent cross-purchase support, and excellent dev tools with in-VR WINE emulator for easy ports of PCVR games to standalone.

Even Quest 2 already likely has the computing power to run Half Life Alyx at low detail if you see the visuals of Star Wars: Tales from Galaxy’s Edge that has similar open-world visuals of Half Life Alyx at low detail. Some of these standalone VR games on Quest 2 can equal-render some GTX 1070 PCVR games now — it is downright impressive how much graphics now can be done on some mobile GPUs with good VR programmers. Now imagine upgrading it one major mobile GPU generation better.

Although SteamBoxes were not directly successful, the Valve Index VR is successful and consistently sells fairly quickly despite much higher price than Quest 2. Plus their Steam for Linux incubated Windows games for Linux to the point where 70% of existing Steam games runs on Linux one way or another. And Quest 2 runs Linux (as an AOSP fork). So with one generation more powerful GPU (like an M1-competing Snapdragon) can easily run WINE for ARM, and thus allow easy ports of SteamVR PCVR games to Steam standalone. Windows ARM will also (indirectly) become more popular, thanks to M1 FOMO effects, but the big winner will be computerlesss PCVR since it will be a (semi)simple recompile of most PCVR games to run on Valve’s standalone!

They got the hardware tech, software tech, app store, porting skills to launch very hard straight into Quest 2 competitor territory. Support both ports of .apk based (Android AOSP port like Quest) and .exe based (PCVR port) games on their future VR standalone. Both via Steam, their gorilla. This is a double barreled super shotgun in populating their standalone VR app ecosystem.

TL;DR: My personal bet is Valve realistically becomes the first formidable Quest 2 competitor by year 2023 with a good in-VR version of Steam, and a pack-in port of Half Life Alyx Standalone.

Apple and Oculus may still be better quality (higher rez) or cheaper (Facebook discount), but Valve can easily grab a large standalone VR gaming market share just by merely being the world’s first Windows-compatible PC built in headset, literally.

Valve already has the software technology (recycled from Linux steam boxes married with theoretical Index 2) to literally build a complete Windows PC straight into their standalone, using Qualcomm’s upcoming M1 competitor combined with the successful work on ARM WINE (Windows Emulator). Follow the Money shows major spending by Valve on multiple Linux fronts, and the tea leaves tells me Valve will be the one to fire the first big broadside into the now-impressive Quest standalone near-monopoly.

 

[bobzdar]

Problem with that is useable emulation on a lower powered platform has never been successfully done. Apple has done really well with the M1, but their software ecosystem is tiny compared to win/steam and even then not all apps work well - they've hand tailored the emulation code for their most popular apps and those are close to x86 performance, but that's not possible on a win/steam app scale. A mobile steam store would be a better solution and other than Oculus devs that have signed an exclusivity deal, all of the games that run on Quest would certainly be on any mobile steam store. I think they'd be better off shrinking their steamdeck in 2 years and then sticking that into a headset for standalone pcvr. It won't be Quest 2 (or 3) price level, though.

 

[Chief Blur Buster]

Problem with that is useable emulation on a lower powered platform has never been successfully done.

Perspective --

It's not emulation of the GPU nor CPU -- it's Wine running native ARM and GPU code without instruction interpretation. The reference is that Valve has already done a lot of work on this via Steam Boxes.

Also Android in Quest 2 "emulates" Java already as part of AOSP's cloning of Oracle's Java APIs. Wine is no different from an "API emulation" perspective. Remember the Oracle lawsuits?

Remember, we're not emulating x86 games. We're talking about developers recompiling Windows x64 games to Windows ARM64 games -- sometimes a (reasonably) simple effort. And getting it to run native CPU + GPU. Then it's mainly just a matter of a HLE API system such as Wine. Theoretically they could actually use x64 like Steam Deck and run full-on PCVR games standalone, as they alluded to Steam Deck being relevant to their future standalone VR plans. A real Windows 10 x64 embedded OS could actually run directly on the headset. But they could just as easily decide to do the ARM route instead.

Fundamentally, API-layer emulations are no different, like Quest 2's existing defacto Java API emulation (as part of the modified AOSP fork -- the operating system of Quest 2), versus Linux Wine existing Windows API "emulation layer. These are HLE's (High Level Emulation), not low level, and have extremely minor performance penalties.

I have a Quest 2, and multiple GPUs including GTX 1080 Ti and RTX 3080, and can make comparisons on the mobile progress lagging behind desktop progress. Current multi-gigahertz multicore CPU/GPUs like those used in Quest 2 -- are fan cooled (with a clever slit exhaust at top edge) -- are NOT exactly "a low powered platform" from a mobile perspective. It is actually far more powerful than some tower PCs from 10 years ago.

In addition, it does not preclude developers from optimizing the VR apps, or buying time to port it from an ARM EXE format to an APK/Android package like those Quest 2 uses. It can support both, in theory.

The goal is to rapidly populate the headset's library, while providing a profit incentive to fund the manufacture of the said headset; currently HTC prices their standalone in the thousands. Imagine a large generation past Snapdragon XR, into something approximating M1 territory.

Even if the HLE API overheads (Wine) adds some penalty, Valve did some amazingly incredible job to the point where frame rates of the emulation equivalent or sometimes higher than native sometimes (for certain cherrypicked games). So M1-like performance may only be bumped back to better-than-XR-like performance, which is sufficiently good enough for initially populating a launch library.

Thanks to their existing code optimizing work on SteamBoxes and the upcoming Steam Deck -- for those who have been paying attention to Steam Boxes. Sure, Linux Steam Boxes were a semi-unsuccessful product, but perspective is needed to understand that HLE of API's can be fairly low overhead, to the point where other system efficiencies can make up. But, this, humbly, is fully relevant to cutting the costs of launching a VR headset.

Weirdly, Valve keeps spending lots of money optimizing Steam for Linux, despite low commercial potential; and they've made some recent source code commits -- Look at https://github.com/ValveSoftware/wine -- and you can see active development. It's very close to being a high-performance Windows HLE API system for ARM64 CPUs already, even though right now it's mainly x86 -- it also now has ARM support. Even though a port of Steam hasn't yet hit ARM, it would not be difficult for Valve to do so.

I think it will probably take a major app store vendor (e.g. Valve and their Steam) to be able to afford to lower the price of the headset by incubating the software library fast enough, to ramp up the manufacture of the headset, in order to lower the prices as much as possible to three figures instead of four figures. Apple can afford to price high to consumers (luxury pricing especially if they bundle the best mobile CPU, GPU and screen together), but Valve wouldn't have the same power, and would need to price three figures, even if higher than Quest 2. One way to do so is use an app store to justify pricing the headset with a lower (or non existent) profit margin, which is why this scores highly.

The AAA developers will optimize their software more to max the number of polygons - but HLE APIs can be a tool to help developers rapidly port their games from other VR platforms.

That said, it is also highly likely Valve will launch without Wine in their headsets -- they have a large number of routes of competing with Facebook because of their existing excellent in-house VR-hardware knowhow, in-house VR software accomplishments (e.g. Half Life Alyx) and their storezilla. Even if one doesn't want Valve to enter the market.

I am also a software developer and understand what I'm talking about:
HLE API's for Vulkan is very low translation overhead (aka low emulation penalty). Quest 2 uses Vulkan, and this is now also used by some PCVR games too. Half Life Alyx uses Vulkan APIs. While WIne's D3D is slow (it translates to OpenGL), Wine's Vulkan is relatively fast (it stays close to metal). When using games that support Vulkan in Wine, they run with a hell lot less performance penalty. Even if this is just quickly ported to APK or APK-like format instead of kept as an ARM64 EXE for Wine, it would not be too difficult for Valve to do a pack-in of Alyx for any theoretical Index 2 standalone VR -- regardless of which route they do (APK route or x64 (like Steam Deck) or ARM64 EXE or ARM64 Linux binary route)

We need more competition, but if we're speaking of "likelihoods" instead of "wants", it's quite obvious this will be an irresistible "new app store market" for Valve.
The bottom line is Valve clearly appears to have a lot of ammo to fire a massive salvo of an entry into the standalone VR market -- with and without Wine (Windows emulation).

Given Samsung's prowess in good CPU/GPUs/screens, I'd prefer Samsung to launch something. But Samsung has no massively profitable app store nor fast appstore-incubation ability applicable to standalone VR, necessary to sell millions of units at consumer pricing. So if someone wanted to bet a mortgage on who's the next non-Oculus VR vendor to enter the consumer standalone market to sell millions of units, it's Valve or Apple or both before Samsung does.

 

[funkydmunky]

Even Quest 2 already likely has the computing power to run Half Life Alyx at low detail if you see the visuals of Star Wars: Tales from Galaxy’s Edge that has similar open-world visuals of Half Life Alyx at low detail. Some of these standalone VR games on Quest 2 can equal-render some GTX 1070 PCVR games now — it is downright impressive how much graphics now can be done on some mobile GPUs with good VR programmers. Now imagine upgrading it one major mobile GPU generation better.

There is nothing in the mobile space as of now that is even remotely close to a GTX 1070 8GB in capability. A game may look the same on different platforms, but that is not to say the under laying HW is comparable.

 

[Koldur]

There is nothing in the mobile space as of now that is even remotely close to a GTX 1070 8GB in capability. A game may look the same on different platforms, but that is not to say the under laying HW is comparable.

The bold text is the thing what counts. In the end the hardware doesn't mean anything, it is all about how things are perceived in a game. It is downright impressive.
So, in all reality, you are confirming what was said in the text you quoted.

 

[bobzdar]

Theoretically they could actually use x64 like Steam Deck and run full-on PCVR games standalone, as they alluded to Steam Deck being relevant to their future standalone VR plans. A real Windows 10 x64 embedded OS could actually run directly on the headset. But they could just as easily decide to do the ARM route instead.

It wouldn't be just as easy to go the ARM route as they'd have to have all devs recompile their games, reset default settings and/or tune to the hardware etc. where with the steam deck it'd be plug and play with the existing (vast) steamvr library and let users adjust the settings themselves. It'd be way more difficult to go the ARM route to get all steamvr games - they'd most likely have to settle for existing oculus games ported to their store and maybe get a few more converts at best.

 

[Chief Blur Buster]

There is nothing in the mobile space as of now that is even remotely close to a GTX 1070 8GB in capability. A game may look the same on different platforms, but that is not to say the under laying HW is comparable.

That may be, specs sheet wise.

However, the whole product of efficiency (hardware and software) with incredible optimization feats done by VR developers on Quest 2, including foveated tricks (higher detail for screen centre) and the focus on getting more polygons/shaders visible per clock cycle where human vision matters.

This manage to make some games like “Star Wars: Tales From The Galaxy’s Edge” (have you played it through?) on Quest 2 slightly more detailed than Half Life Alyx (I have played too) on minimum detail.

That is the baseline I’m using for comparision sake. There are subtleties, like reduced shader detail, but you disable that with Alyx at minimum detail anyway.

The bottom line, the mobile performance envelope now is capable of encompassing Half Life Alyx, especially with the next generation mobile GPUs along with similiar VR developer optimizations.

Once you play through both games (on your Quest 2), I’d like your comparison commentary with the builtin (Star Wars) vs linked PCVR (Alyx configured to minimum detail). You’ll see lots of Star Wars:TFTGE more detailed than Alyx min detail setting. They are both massive sized loading-screen-less outdoor open worlds with lots of high resolution textures and polys. Heck, even use a different LCD headset like Index for Alyx, if you wish. (use similiar screen tech and resolution for a fair comparison, e.g. blacks and colors).

 

[bobzdar]

Capable, yes, but as a simple port, no. It will be a complex port with tons of optimization, not a recompile with minimum settings and go. Look at the PC version of Onward vs. Quest 2 version, play them both and you'll see that the mobile version has had further optimization/graphics tweaks - and that's best case scenario at this point as it's now owned by FB. We're still a few years off for mobile chips to be able to hit low end PCVR of 2-3 years ago to where VR games can be recompiled multiplatform releases with mobile running on low settings vs. higher for PC. There's too much geometry, texture etc. work that still happens to get mobile up to snuff for even old games. Mobile VR is at 2016 PC level right now, so 5 years behind.

 

[Chief Blur Buster]

I think we’re closer than you think, considering M1 type chips greatly outperform Snapdragon XR.

Most PCVR games are also configurable to run on 2015-era gaming PCs, especially with very clean optimized Windows installs. Many GTX 980 VR games were able run perfectly smooth on 780 when resolution was configured lower (blurryvision) - a simple change, though sometimes needs override with sub 1.0 scaling values in software such as 3rd party Oculus TrayTool utility to override default VR resolution. Like 1280x720 VR resolution. The 780 was made in year 2013.

Which can then, in a proper implementation, be re-enhanced by DLSS/FidelityFX-like algorithms. 720p back to 1440p. Boom. Big problem solved. Mobile implementations of DLSS/FidelityFX-like algorithms also now exist too — reportedly, Steam Deck supports it without needing the games to support it.

The OS of a standalone headset has far less crud, so you can lop off a whole generation of requirement, especially with Vulkan API games.

All the above combined, plus “easy low lying apple optimizations”, lops about two GPU generations backwards to PCVR-port viability in a standalone. There are multiple engineering paths to get there sooner than you think.

Moreover, frame rate amplification technologies such as FidelityFX/DLSS 2.0 did not exist in 2015. And it will be version 3.0 DLSS / 2.0 FidelityFX by 2023.

My prediction is two years, ala 2023, for a true Quest 2 competitor with Alyx low/medium visuals. This includes minor HLE overheads, if any.

Possibly visuals of standalone VR in 2023 can reach similar to 2019-era Alyx med-high detail setting via a newer >3:1+ framrate amplify of DLSS 3.0 algorithm like those alluded in Blur Busters Area 51 Research section. Especially with minor mobile-specific dev optimizations of a company of Valve’s budget, and avoiding HLE “emulation” overheads via a native port instead, as Alyx was apparently pre-emptively coded in a way to be easily ported to other platforms. 3rd generation frame rate amplification technologies enables about a 4x resolution increase at same framerate with the same fully-loaded GPU shaders with no easily noticed oddnesses. But let’s be pessimistic and just assume Steam Deck like framerate amplification technology performance multipliers only.

Though could slip to 2024 due to parts shortages.

 

[funkydmunky]

I think we’re closer than you think, considering M1 type chips greatly outperform Snapdragon XR.

Most PCVR games are also configurable to run on 2015-era gaming PCs, especially with very clean optimized Windows installs. Many GTX 980 VR games were able run perfectly smooth on 780 when resolution was configured lower (blurryvision) - a simple change, though sometimes needs override with sub 1.0 scaling values in software such as 3rd party Oculus TrayTool utility to override default VR resolution. Like 1280x720 VR resolution. The 780 was made in year 2013.

Which can then, in a proper implementation, be re-enhanced by DLSS/FidelityFX-like algorithms. 720p back to 1440p. Boom. Big problem solved. Mobile implementations of DLSS/FidelityFX-like algorithms also now exist too — reportedly, Steam Deck supports it without needing the games to support it.

The OS of a standalone headset has far less crud, so you can lop off a whole generation of requirement, especially with Vulkan API games.

All the above combined, plus “easy low lying apple optimizations”, lops about two GPU generations backwards to PCVR-port viability in a standalone. There are multiple engineering paths to get there sooner than you think.

Moreover, frame rate amplification technologies such as FidelityFX/DLSS 2.0 did not exist in 2015. And it will be version 3.0 DLSS / 2.0 FidelityFX by 2023.

My prediction is two years, ala 2023, for a true Quest 2 competitor with Alyx low/medium visuals. This includes minor HLE overheads, if any.

Possibly visuals of standalone VR in 2023 can reach similar to 2019-era Alyx med-high detail setting via a newer >3:1+ framrate amplify of DLSS 3.0 algorithm like those alluded in Blur Busters Area 51 Research section. Especially with minor mobile-specific dev optimizations of a company of Valve’s budget, and avoiding HLE “emulation” overheads via a native port instead, as Alyx was apparently pre-emptively coded in a way to be easily ported to other platforms. 3rd generation frame rate amplification technologies enables about a 4x resolution increase at same framerate with the same fully-loaded GPU shaders with no easily noticed oddnesses. But let’s be pessimistic and just assume Steam Deck like framerate amplification technology performance multipliers only.

Though could slip to 2024 due to parts shortages.

I think you are confussing the chips CPU capability with a GPU or the intergrated GPU of an ARM chip. It is a vast amount of difference between a GTX 1070 and ANY ARM gfx.

 

[Chief Blur Buster]

(my original intent in this thread -- was to explain why, regardless of GPU performance, I believe Valve is the one most likely to become the first multimillion-unit-selling competitor standalone VR to Quest 2)

I think you are confussing the chips CPU capability with a GPU or the intergrated GPU of an ARM chip.

No, I am not.

(Perhaps you might not be already aware who I am)

As inventor of TestUFO and founder of Blur Busters -- I own several VR headsets, I own several GPUs, several PCs, several Apple products (applicable to my extensive ARM experience), I have made some contributions to the VR industry, and the progress in VR has been nothing short of incredible.

It is a vast amount of difference between a GTX 1070 and ANY ARM gfx.

Yes, but talking real world results rather than spec sheets. Do you have a Quest 2?

Also, check this post above for some multiplier factors that contributes to the somewhat smaller real-world-experience chasm.

I'm a member of the PC Master Race, and have multiple GPUs -- the point I am making, however, is that mobile GPUs have been very rapidly multiplying in power very rapidly, combined with some optimizations, which helps the two ends converge closer together, as explained in the post above.

In addition to VR, PC GPUs, I also have some of the best ARM GPUs too -- including iPads, experience with Mac M1, own an RTX 3080, some 5 GHz overclocked gaming PCs, and dozens of gaming monitors.

I am commenting based on actual real world visuals, instead of specifications sheets

The best PCVR games at the highest resolutions definitely can blow away the standalone VR, but this isn't Atari 2600 style graphics we are talking about here. We're comparing somewhat older PCVR-capable desktop GPUs to the Quest 2 here instead, and the gap is smaller than most people thought when they try both out.

Also, the fan-cooling of a Quest 2's Snapdragon XR also prevents its thermal throttling, and keeps it performance envelope surprisingly close to some desktop GPUs that were in use at the beginning of PCVR's introduction, especially when accounting for the additional factors described above. With desktop GPUs, it is easier to slack off on optimizations, but with mobile GPUs it's almost a life and death matter to get those pixels, triangles and textures rendered on time to avoid a stutter, so some really clever optimizations came in to slightly shrink the chasm between desktop-and-mobile GPUs.

Also, helping the standalone case here, is Moore's Law is currently far faster (at the moment) on speeding up performance-per-watt, aka mobile GPUs, which shrinks the time interval between state-of-art desktop GPUs and state-of-art mobile GPUs, to the point where a mobile GPU is now batting less than a 10-year-old desktop GPU, but with various standalone-specific optimizations thrown in, mobile GPU can now battle 5-year-old midrange desktop GPUs (when combining all the current tricks crammed into Quest 2). The 2016 PCVR entry level system versus the Quest 2 today, for example. Specs sheets are not everything -- visuals are.

If you read some other Quest 2 threads, other people have already confirmed in other threads, what I'm actually seeing with an actual Quest 2, actual desktop PCVR (Rift original, Valve Index), while also owning multiple desktop GPUs -- ever since I purchased my first 3dfx Voodoo card in 1997, I've been keeping up with GPUs all the way to the current GTX 3080. Back in the day, I was actually a multiplatform software developer, so I have been exposed to both PC and Mac (so real world experience on both sides of the fence). The optimizations that have been done on the ARM side has often been greater than the optimizations done on the PC side of things -- especially when we're talking about the best mobile GPUs running on a well performance-optimized system.

Not armchair spec sheet reader stuff -- actual real world visuals stuff. It's sort of like how incredibly optimized some console games became because of console limitations, etc. Visually, in the real world (not spec sheet stuff), some of the more detailed Quest 2 games do actually match PCVR quality for several games (especially when you define "PCVR quality" of a ~2016 caliber PCVR system, roughly similar poly/shader/texture detail but lower resolution than what you can do with newer headsets and 2000 or 3000 series RTX). That a standalone is batting straight into the ballpark of early modern PCVR (2016-era specs), is still an astounding accomplishment visually, despite only running

My question #1 is: Have you ever played some cherrypicked Quest 2 games such as Star Wars: Tales from the Galaxy's Edge too?

If yes to question #1, then my question #2 is: Have you made any comparisons between Quest 2 and an approximately 2015-era PC? Or Half Life Alyx at minimum detail (to simulate)?

If you reread my posts in this context, and also have actual experience visuals-vs-visuals with a Quest 2, then you would already be agreeing with me.

We're not talking about spec sheet stuff in this thread -- we're talking about actual real world performance and visuals. We're all focused on proving our respective points, and we're both right in different ways. So to refocus back on the ball -- my focus in this thread was talking about actual real world visuals achieved & who will be able to compete with Quest 2 -- and that's what I replied to.

 

[funkydmunky]

The only time real-world visuals would be even remotely close between a 1070 and and ARM gpu's is when the game is in no way pushing the limits of the PC hardware. Many VR titles are optimized for a broad spectrum of hardware with many titles going very low poly/texture. Granted the Quest2 has a little bit of last gen console magic going for it where far lesser hardware is pushed to its limits and highly optimized for. That said a 1070 would be considered very anemic at this moment for PC-VR so the reality of the gap is even larger. The fear of course is the Quest will market dominate putting PC titles that do push said limits of tech into a non profit situation and are simply no longer produced.
I think Valve is very important in the future of PC-VR and I believe their next product is going to be a form of Index "Lite" possibly with Steam Deck integration?

 

[ChosenUno]

With the kinds of visual fidelity you get in more AAA mobile games, I don't see a problem on the GPU side of things, especially since these days HFR gaming on mobiles are a thing now, plus with my experience in VR, visual fidelity doesn't actually matter as much as interactivity, smoothness (FPS) and tracking.

I don't think you'll see a Quest 2 competitor from anyone outside of Google, Microsoft or Sony, considering price, Valve is a distant possibility too.

Valve can possibly release a VR headset using identical specs to the mid-tier Steam Deck, running Windows, for $499. I doubt they can make the $399 pricing, but we'll have to see. Valve hasn't demonstrated inside-out tracking yet, no idea how much that'll drive up cost. And Valve isn't in a position to be able to eat a loss on every unit sold, since they don't have any tie-in products.

Google, Microsoft and Sony can subsidize and build competitive headsets. Google would be the worst off, since I don't remember them having done any VR tech research yet, while both MS and Sony have had plenty of experience building VR products, but would most likely to build a standalone headset, using their Play Store, while both MS and Sony would probably tie in to the consoles and maybe PC for MS.

A pipedream would be for Google to make a platform akin to Android One for VR, and then existing phone mfgs can jump in and build VR headsets off of that.

You know what, a dark horse would be Xiaomi. They just might be crazy enough to do it, and do it well.

 

[Chief Blur Buster]

Valve can possibly release a VR headset using identical specs to the mid-tier Steam Deck, running Windows, for $499. I doubt they can make the $399 pricing, but we'll have to see. Valve hasn't demonstrated inside-out tracking yet, no idea how much that'll drive up cost. And Valve isn't in a position to be able to eat a loss on every unit sold, since they don't have any tie-in products.

Valve now earns billions per year revenue from Steam -- Valve has the Steam application store, which would be a side revenue stream that allows them to stay close to the approximate $499 level. Though they have a mixed record of hardware, their app store is wildly successful, and their Index VR product is reasonably successful.

Even at $699, many will still buy a Valve standalone. It'd still be a legitimate competitor, as long as the specs could outperform a Quest 2. Mind you, that would compete with the Quest 3, but many don't want Facebook.

Google may be best off releasing a reference standalone-VR platform as you suggest. Google did some VR called the Google Daydream, a standalone headset with a built-in screen and smartphone GPU -- operating roughly similar to Oculus Go (before Quest). But it's one of the many cancelled Google projects, so I'd be surprised if Google gives VR another go after floundering.

Microsoft does have Windows Mixed Reality, which is rather mixed and they'd have to give it another concerted go at it while involving their XBox division, it is far more likely they'd add a wireless headset requiring the XBox GPU. Perhaps via a WiFi 6E technology dongle dedicated to ultra high-speed peer to peer connection to a cordless XBox VR headset. It would not be a true standalone, but it would be far more powerful freedom than PSVR. Sony could also do the same.

The necessary AI for inside-out tracking is becoming more widely available, possibly with some upgrades to some open source projects. It may take time to get as good as Oculus and add bare-hands tracking, but inside out tracking without hand-tracking can still be begun by Valve. There now currently isn't too many missing technology pieces when it comes to Valve, with their experience with Steam Deck and Index VR.

 

[noko]

Well known that Samsung has Radeon GFX partnership. Being that Samsung is the king of displays and has been more then quiet on the VR front for ever, could a Quest 2/3 killer be in the wings? I don't see super gaming phones being the strategy here.

I see potential in Apple and/or Samsung making a super sleek, lightweight headset and driving it with one of their phones. If the person already carries a phone on them at a more coinvent place vice weighing down the headset with the needed processing power, I think that would be a more affordable/cheaper route with the flexibility of a smartphone interface, Wifi, 5G, bluetooth already built in capability, app stores etc. I would say Facebook would be the one trying to catch up.

 

[Chief Blur Buster]

To follow up on my bet that Valve (rather than Samsung) is probably first to market a Quest competitor — there is some new Valve standalone hints in library files from reverse engineering the latest build of SteamVR:

https://arstechnica.com/gaming/2021...-vr-headset-teased-by-deep-steamvr-file-dive/