Khronos Officially Releases OpenXR 1.0

[starshipeleven]

He doesn't believe in VR, and nothing you say will convince him otherwise. I've been down this road with even more use cases than you listed.

He's just here to hate on VR and any of its adherents.

Why are you talking of VR as if it was a religion?

VR has niche uses, and high cost/complexity. So far none has found a real money-making reason to actually have it, they were all stunts.

It does not help the cause to lie, you are only deceiving people but not solving the problem.

[Spazturtle]

VR adoption is not only growing but the rate of adaoption is growing, so no VR isn't going away and yes it is going somewhere.

VR headsets are getting both better and cheaper, lots of the issues either already have been or are being worked out. Things like SteamVR, OpenXR and VirtualLink ports on graphics cards are making VR headsets easier to adopt.

[starshipeleven]

Which is why MS Hololens was doing it in 2015, with a cell phone SoC and a DSP?

And a tower PC nearby. And a very handpicked setup for reviewers.

And Skyrim-grade glitching when it was shown to more people later in that year.

And Google Tango used a Tegra-based tablet, back in 2014?

It did what Pokemon Go does, more or less. It's not bad but it's kind of limited, you know.

who are you to say it's not going anywhere?

I'm one that knows where VR actually adds value and where it does not.

So far only place is entertainment and some simulation. This isn't enough on its own to propel it forward beyond niche gaming use.


For AR the situation is both better and worse. It's more useful, but it's still basically a fancy 3D HUD, so on its own it's far from the game changer that dreamers think it is until very late in its development (i.e. when it can do the crazy shit you see in movies, which is mostly doable for AR).
It can reuse a lot of the work done by robotics and other fields that need image processing (say, autonomous vehicles, which is far from a dead end even if it's not really around the corner either) though so even if it's not pulling its own weight it will go on.


Now, brain-machine interfaces like the (mostly vaporware) thing hyped by Musk lately are going to have a much bigger impact, and would have much more applications if they are really as imagined.

[kpedersen]

To catch my drift. The biggest issue with VR is not cost, it is not hardware. I can survive the crappy VR UIs. It's not even the driver quality. No, it's the horrendously shitty bundled & forced software.

Hah! I agree. It is actually nice to see someone in the same boat trying to use VR for potentially sensitive data whilst trying to avoid the consumer bullsh*t.

You can see some more of my "journey" here: https://forums.freebsd.org/threads/vr-in-freebsd.71411/

One thing that you might want to look into if you did feel that Valve's VR solution is fairly stable for your uses is: https://partner.steamgames.com/doc/f...s/steamvr/info
Check out the "SteamVR In the Enterprise" section. You might want to also consider a steam emulator (i.e from cs.rin.ru (perfectly legal to use in the UK because you own the hardware)).

But yeah, ultimately OpenHMD and OpenXR will be the future. Will keep with them if we can.

[kpedersen]

Lies? Uh, if you're happy with such graphics, good for you. That's not even DayDream quality.

That isn't quite the point I was making. The listed minimum specs for the Vive are: "Nvidia GeForce GTX 970, AMD Radeon R9 290 equivalent or better"
It basically wont run on anything lower than this. It is artificially restricted in the installer or in the setup tool.

As you can see, DayDream doesn't quite need this (an Integrated Intel chip could even be enough if it supports >=OpenGL 2.1) and it would be a shame to have to buy new GPU hardware (or chuck out a perfectly good card) just to appease the hardcoded restriction.

Check out the slightly misguided information on this forums: https://forum.vive.com/forums/topic/...s-530-on-vive/

"Unfortunately, VR requires a pretty beefy graphics card, bare minimum recommended is a GTX 970 (or AMD equivalent)"

Simply not true. All you are doing is rendering the scene twice at not a particularly high resolution using a moderately expensive distortion shader. I don't even see why this couldn't work on something like Windows 2000. I am sure that with a few patches to OpenHMD, it could be made to work.

[coder]

It would be trolling if I'm wrong in what I said. Am I?

Here's the current Wikipedia definition:

In Internet slang, a troll is a person who starts quarrels or upsets people on the Internet to distract and sow discord by posting inflammatory and digressive, extraneous, or off-topic messages in an online community (such as a newsgroup, forum, chat room, or blog) with the intent of provoking readers into displaying emotional responses and normalizing tangential discussion, whether for the troll's amusement or a specific gain.

So, if you upset people by posting inflammatory messages with the intent of provoking readers into displaying emotional responses, whether for your amusement or a specific gain, that would make you a troll. Factual correctness is actually irrelevant.

That's why I asked you to state your interest. If you're just here to provoke people, then you need to quit yer trollin'. You don't have to like VR or believe in it, but please leave us alone.

[coder]

Why are you talking of VR as if it was a religion?

Because that's how you seem to treat it - like you're Richard Dawkins, here to show us the light and cure us of it.

It does not help the cause to lie, you are only deceiving people but not solving the problem.

Lying about what? What problem are you solving?

[coder]

And a tower PC nearby. And a very handpicked setup for reviewers.

You don't know what you're talking about. Hololens was self-contained, as was Google's Tango.

Today, you can download and run AR apps on regular iPhones and many Android phones.

And Skyrim-grade glitching when it was shown to more people later in that year.

Guess what? It didn't even ship in 2015 (nor was it planned to), so pre-release software issues are understandable. But the hardware of the original Hololens was finalized back then. That was my point - you seemed to imply there was a gap in hardware capabilities needed for AR, but this is just a demonstration of your ignorance. ...if I'm being charitable. My cynical side says you're just engaging in nit-picking, in an attempt to "distract and sow discord by posting inflammatory and digressive, extraneous, or off-topic messages".

It did what Pokemon Go does, more or less. It's not bad but it's kind of limited, you know.

Again, you clearly have no clue what you're talking about. I'm not even going to correct you, because if you cared about accuracy, you'd actually bother to educate yourself about the subject, before posting this nonsense.

I'm one that knows where VR actually adds value and where it does not.

By what authority? All you've so far demonstrated is ignorance and trolling. You have no credibility left, on this subject matter.

[coder]

That isn't quite the point I was making. The listed minimum specs for the Vive are: "Nvidia GeForce GTX 970, AMD Radeon R9 290 equivalent or better"
It basically wont run on anything lower than this. It is artificially restricted in the installer or in the setup tool.

Okay, I understand now.

I think they were just trying to establish a minimum hardware spec that developers could target, so that they could ensure an acceptable experience for users. The problem they were trying to avoid is someone with a below-spec PC trying a bunch of apps and quitting in frustration (or nausea), after finding that none of them run at a usable speed.

So, it's not that the HMD required a GTX 970, but rather the Steam VR apps had that as a baseline expectation.

As you can see, DayDream doesn't quite need this (an Integrated Intel chip could even be enough if it supports >=OpenGL 2.1) and it would be a shame to have to buy new GPU hardware (or chuck out a perfectly good card) just to appease the hardcoded restriction.

Microsoft also showed this, as the minimum spec for their HMDs was actually integrated graphics. However, Microsoft took the approach of creating two tiers. So, the lower-tier was more like phone-grade VR, while the upper tier was closer to the graphical quality people were experiencing with Vive and Rift.

Simply not true. All you are doing is rendering the scene twice at not a particularly high resolution using a moderately expensive distortion shader. I don't even see why this couldn't work on something like Windows 2000. I am sure that with a few patches to OpenHMD, it could be made to work.

It just depends on the app's graphical sophistication. It's the same reason you need a decent GPU to run many recent games at good settings and framerate.

A careful look at the Pi's screenshot shows that it has extremely low polygon count and seems to use just simple texture mapping.

[coder]

The limitation is the human user, you can aggregate all you want, but he is going to be able to focus only on a few of these at a time. We are already at near-peak capacity with a 4k screen, and you really don't want to distract people in the field with bullshit feeds, they need to look around 360° already to avoid getting shot from some random asshole. The people on the field must receive directions and commands aggregated by their support intel analysts and chain of command back at the base that together can process much more information than a single human alone.

The benefit that AR can provide is contextual awareness. It can provide you with relevant and important information for your specific context.

To use the example of troops on patrol, it could warn you of potential snipers or hostiles approaching, even showing you the direction (if outside of FoV) and potentially highlighting them. Perhaps this data is generated by drones, but perhaps it can even be provided from neural-network based image recognition that's running locally, on the device. You could really have eyes in the back of your head.

Again, using the patrol example, it could also highlight subtle changes that occurred since previous patrols (possibly even by different soldiers), so you can see where potential IEDs might be hidden or where foes might be lying in ambush.

And imagine being able to see where an incoming airstrike or artillery is going to hit, so you can see if it's on-target or if you're too close & need to take cover. Imagine being able to see the location of "friendlies", so you can better coordinate with them, as well as avoiding friendly-fire.

Plus, with soldiers kitted out with multiple cameras and full 3D tracking, they can relay the data back to a command post, so their commander can see exactly where they are, what they see, and what's around them. All in an integrated, multi-asset battlefield display, that he might even be viewing in VR to see in 3D.