Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: What is the status of VR on Linux?
90 points by usernam on Jan 3, 2017 | hide | past | favorite | 40 comments
I'm a programmer with a lot of experience with GL. I'm mostly working with data visualization in general, but I wrote several non-toy 3D engines in the past.

I've been rather oblivious with VR as of date, even though I toyed with 3d glasses on an SGI Octane decades ago. After a couple of experiences with the Oculus though, I'd love to start hacking on VR technology.

As my development platform for the last 10+ years as been exclusively Linux, and again almost exclusively targeting high-end NVidia cards, I have no interest in switching OS whatsoever.

What is the status of VR on linux? It seems that most headsets are platform and vendor-locked. Why is this the case? Are some devices more hackable than others?

If you want to be part of the "platform", then it's not ready yet. Valve demoed SteamVR working on Linux a couple of months ago at Steam Dev Days, but it's not released yet. Given Valve Time, it could be anywhere from weeks to a year from now.

It's possible to write your own OpenVR driver with whatever headset, but it would be quite an undertaking.

If you're just messing around on your own, OSVR is n standalone open source initiative (which also has SteamVR support) that is largely cross-platform but I haven't played with it.

Finally, last month Khronos Group announced [1] that they're going to be developing a joint VR standard with all the big players which should also advance VR towards being more cross-platform.

tl;dr: Unless you're building just for yourself, I'd wait or bite the bullet and install Windows for the duration. (I did the latter, sigh.)

[1] https://www.khronos.org/news/press/khronos-announces-vr-stan...

In theory, http://www.osvr.org/hdk2.html is exactly what you are looking for. In practice... well, it's open-source and the software team would appreciate contributions.

You might be interesting in following along with this: https://github.com/cnlohr/libsurvive

Hacker extraordinaire Charles Lohr is reverse engineering the Vive protocols and has made good headway so far. It also seems that Valve has been giving him the odd hint along the way.

I'm with you! This is one of the last few articles related to 'vr on linux' that I read before the Holidays: http://www.roadtovr.com/steamvr-to-get-linux-and-mac-osx-sup...

AFAIK both Unity and Unreal Engines have 'experimental' linux builds available. I started the year hardware-strapped, but hopefully sometime in 2017 I'll start dabbling with these linux alternatives.

Good luck!

You might want to look at this lightning talk from 33c3:


We really need to standardize VR drivers with a hardware class that needs to support, in a standard fashion, x types of standard i/o.

The various VR platform vendors have only recently come together on a universal standard, which will ideally be implemented on both Windows and Linux: https://www.khronos.org/vr

Serious question, what's the incentive for the current vendors to do this? I feel that given how nascent the space is, everyone is rushing to try to be the next iPod/iTunes store combo, and you don't get that by sharing.

If there were only a single major vendor, then they would have an effective monopoly which would take years to catch up to, such as iPhones/Android. Instead, we had both Vive and Oculus release at around the same time, and both have seen reasonable success. Prospective buyers worry that they'll get locked into one ecosystem, hoping that the decent games on other platforms get ported over eventually. That indecision probably also stops a large number of prospective buyers from purchasing either platform, because they would rather play it safe than to spend $800 on backing the wrong horse. I know a guy that got burned buying an HD-DVD player, it sucks.

With a shared standard, games developers can build to a common target, and consumers can buy hardware based on actual hardware differentiators. Oculus will be able to say "buy our headset, it has better feature X and it supports all the Vive/GearVR/Cardboard games!" Customers will feel more comfortable just picking a horse, knowing that even if it doesn't win it'll at least finish the race.

You also don't get there by dying out because you're missing software. Fragmenting a tiny starting market (like Occulus attempted at the beginning) is what would kill the whole technology for years to come.

VR makes an excellent demo, but has anybody found themselves playing VR games regularly? For me, it seems a little like the 3D situation a few years ago. I bought a TV, got the glasses, and was able to play some PS3 games in 3D. It was a lot of fun... for an afternoon.

"3D" has nothing to do with VR. Period.

In fact, 3d is nothing but a marketing scam, it is not really 3d, but a fixed stereoscopic view. When I tested the "3D" on a TV, it was clear that it was going nowhere.

Why? Because "3d" does not update your view depending on the movement and rotation of your head. As simple as that.

VR has been used successfully in the world for decades in CAD, simulation, the different armies, and they spent tens of thousands of dollars that used to cost the device for a reason.

For example, the NASA uses it to train astronauts to move in the space station, the army uses it for training pilots or soldiers to navigate using the stars. Engineers use it to collaborate on designs, like a plane or a car.

I was mostly talking about consumer applications and that's why I mentioned television. If the major users are professionals, then it isn't nearly as interesting to me as it would be if it were a big consumer product success.

With VR (and more interestingly, AR), I don't doubt that there are some professional applications. But soldiers and astronauts just aren't a huge market.

I may sound super negative, but I do hope that it takes off. As a demo it's fantastic (aside from the motion sickness). I just haven't seen anything very compelling yet. Have you?

there are plenty of compelling games that VR either enables or augments out there. off the top of my head:

"job simulator" (yes) is enabled by both a VR headset, and VR controllers. you couldn't achieve the same without VR

"elite dangerous" is significantly improved with VR. pretty much anything with a cockpit would be improved, because you can look all around you. this one in particular is an amazing experience

*edit: there's also an abundance of VR storytelling that are exploring new ways of giving experience without the use of words, VR artwork where there no real point other than just to experience a world. there are a lot of REALLY cool, not just "for demo" experiences you can have on VR

I wouldn't want to play Elite Dangerous without VR, same for racing sims.

(I'm currently playing ED without VR since I lent my Oculus to a friend and I really miss it.)

The 3D TV comparison is completely tired at this point. VR is a new form of HCI, not a questionably improved display technology. Besides that point, if you're only looking at VR as "games" you're missing the boat entirely.

I would say that spatial controllers are a new* form of HCI, but not necessarily the VR viewer. Although I would concede that there are a few experiences that track what you're looking at to interact, it's really mostly focused on the controllers.

Many VR gamers who use traditional dual-analog style controllers concede that it's not really different in terms of HCI model.

* = "new" meaning unconventional and/or untapped in the current mindshare of HCI. Obviously this stuff has existed for decades.

I think the usability of spatial controllers is pretty tightly linked to use of a VR viewer, since it's the HMD that constructs the "space" for the controllers to work in. I don't think Tilt Brush would translate well to a regular monitor with spatial controllers. I can see some potential with 3D displays, but I think if spatial control enough was a paradigm shift then the motion controllers we've seen in gaming applications previously would have translated to professional applications the way VR is now.

VR gamers using traditional dual-analog style controllers aren't leveraging the full spatial immersion of VR, so they don't report a spatially immersive experience. When I describe "VR as HCI" I mean everything involved in immersing someone in an altered reality, and the HMD is only part of that. I think it's more apt to say "XR is a new form of HCI," to include tech like the Hololens, but that terminology hasn't really entered the common lexicon yet.

VR and XR as a whole give unprecedented access to the primary I/O pathway of the human: our eyes and our hands.

> I think the usability of spatial controllers is pretty tightly linked to use of a VR viewer

I see what you're getting at, but remember that the first wildly successful spatial controller was released over 10 years ago: the Wiimote.

> The 3D TV comparison is completely tired at this point

Why do you say that? I think the 3D TV comparison is fairly accurate. If VR (or AR) requires anything more bulky than regular eye glasses, I can't see it becoming a huge success. Part of why I think VR makes a neat demo only is because the current crop of VR headsets get uncomfortable very quickly.

> VR is a new form of HCI

New? I first tried VR 25 years ago and I know it's older than that.

> if you're only looking at VR as "games" you're missing the boat entirely

So if gaming isn't the primary user and driver of VR hardware and software, what is? What industry is putting the most investment into VR right now? From the current generation, I really have only looked at consumer setups like Vive, Playstation, and Oculus. I haven't tried Microsoft's product or Magic Leap so maybe they would change my mind?

I mean "new" as in "newly practically accessible." CAVE wasn't exactly accessible to your average hacker hobbyist but anyone with a gaming computer and $800 has access to VR - and going forward, anyone with a solid GPU and $400 will have access, and so on. The 3D TV comparison is accurate in that both things require you to have something inconvenient on your head, but I'm pretty surprised if you genuinely find the Oculus Rift to "get uncomfortable very quickly." And, remember the devices currently on the market are akin to consumer-ready dev kits - particularly the Vive. They are MADE for early adopters who are prepared to be inconvenienced.

Just like with CAVE 25 years ago, there's a lot of ongoing research efforts to use VR in areas outside gaming - such as replicating experiences over and over that can't dependably be replicated in real life, or leveraging VR as an HCI tool in existing applications. To refer to AR, Boeing has already used a hologram device for aircraft electrical assembly for years. Educationally, VR has serious potential in fields like medical instruction and training.

I agree, gaming and porn will and are driving adoption of VR to the wide consumer market, but its actual applications go way beyond a litmus test of "well the games still aren't fun enough."

Basic interaction design is still being done on VR and AR. There are new forms of HCI currently being made around them. http://elevr.com/some-rather-different-social-vr-experiments... http://elevr.com/all-about-the-context/

> So if gaming isn't the primary user and driver of VR hardware and software, what is?

Perhaps pornography? VR could offer a completely novel set of experiences

The porn industry has already dumped an insane amount of money into VR and is presumably seeing quite a return, since the major players have only accelerated their pace of output. Just like with other technologies, porn companies have already driven the consumer innovations streaming VR video from the internet, capturing 3D 180 footage, and defining best practices for immersive content. I'm sure "novel" experiences are on their way eventually, but right now the money's being made just making 3D180 videos for mobile VR platforms.

It'll get there. It's a new technology and developers are still figuring out best practices as well as getting the hang of how to build quality experiences that will resonate with users. This process is further burdened by the fact that there's currently a bit of fragmentation among device capabilities (positional tracking, motion controls, "room scale") that heavily impact potential game design decisions. Not only that, but VR also has a slightly higher barrier to entry because 2d asset design is not an option (the relative simplicity of 2d game design has been a huge boon to the indi game community on every platform) as well as the fact that upfront hardware purchases are pretty much a must because it's impossible to emulate sophisticated VR features like motion controls etc.

I've had my vive for over six months now and still play a couple times per week. I've lost interest in traditional games at this point.

I was going to say, I am 30 and have been an avid FPS PC master race gamer since I was 16 and in college, and since I have had the HTC Vive, its been only to play with friends I miss that I do any traditional gaming. Other than that, I am either working on my VR game, or playing/testing something in VR.

This is going to change everything from music production and video editing, to gaming and entertainment.

I think VR will be a big success in the coming years, but I doubt it will make many inroads into music production and video editing. I've actually done a lot of music production and video editing work and the practical resolution on these devices is just nowhere close to where it needs to be to be worthwhile vs a multi-monitor setup. Of course the technology will get better, but it's a long way off with many technical hurdles on the horizon before substantially higher resolution systems will be viable.

If you use VR to mimic a multi monitor setup, for sure.

But there are potentially new ways to organise media and interact with clips which VR allows.

I'm really interested to see what we can do in VR for video editing.

I'm more excited about the possibility of creation in VR: sculpting and CAD. Tested had a pretty good demo and discussion recently about how it works. https://www.youtube.com/watch?v=Pf8sKAuzR0k

Absolutely! Check out some of the new games, like Smashbox Arena and Arizona Sunshine.

The whole game changes when you have physical controllers that you hold in your hand.

As a VR dev I'm seeing a lot of users spending significant time in my game. It's, AFAIK, the longest dedicated Vive story-based game, so that might explain some of the longer play times.

(Left-Hand Path - http://store.steampowered.com/app/488760 . Horror RPG with a heavy Dark Souls influence).

But still - it's clear that there are a lot of people playing VR games consistently.

(I also play VR games quite a bit, but I'm just a single data point.)

Yes, I do. And I cannot wait for the games and applications that will come out this year. I am hooked.

My family got PSVR a few months ago. We play VR games about as much as non-VR, if not more. I find it very engaging. It's a fundamentally different and very compelling experience.

I've produced commercial 3D video content (blu-ray extras, commercials, shorts), while not believing in it at the time.

Yet now using VR I think it's incomparable. I'm a believer.

Play Batman: Arkham VR on PS4 and if you're not convinced then you'll never be...

The open-source Vrui VR toolkit[1] has some level of support for the Vive and Rift[2].

[1] http://idav.ucdavis.edu/~okreylos/ResDev/Vrui/

[2] http://doc-ok.org/?p=1508

If you want to do VR dev on linux I think Web VR is probably your best bet right now.

I believe Openvr does work on Linux, with some modifications. Valve is actively developing it.

> What is the status of VR on linux?

It doesn't work, at all. Sorry.

Even OSVR?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact