A reality check for virtual headsets
economist.com
economist.com
To look at today's initial products and call the industry dead, or disappointing, is like looking at an MVP app release and deciding that an entire market does not exist because the initial solution is not perfect.
This type of contrarianism is seductive, and often results in many clicks for the media and the reinforcement of curmudgeonly opinions by those with unrealistic expectations for short-term change (and often who underestimate long-term impacts).
But it's important to evaluate today's VR with the same eyes as when looking at an MVP: does it meet its goals? Does it provide value? Does it show potential if developed further? Is there a product-market fit?
The answer to all of the above is a resounding "yes." Try it if you haven't. VR and AR have the potential to be truly revolutionary technologies. Yes, there is some distance to the goal, but it's eminently achievable - if the tech is not killed off by mis-set expectations, tribalism, paranoia, pessimism, and negativity.
As for investors, it's important to remember that when you're swinging at a pitch, it's no more productive to swing too early than to swing too late. They're both strikes.
VR may take 4 year or 40 years to mature, or it may never take off at all, it's not something we can judge either way.
I think we can conclude it's not going to catch on once they have delivered on the promise of VR and then the market goes meh.
The Rift and Vive (the only VR worth anything) are both very far towards the early adopter side of the spectrum and I think everything follows from that. They are expensive, difficult to set up and keep working, and have expensive dependencies. Prototype hardware already exists that shows all those problems are tractable.
So to me the question is only does VR run out of money before it develops a viable product. Sure there is some small question of what the market will think of what VR has to offer, but when you offer it to people for "free" the reaction is uniformly positive. The only barrier is cost and convenience and whether those lines intersect with the cost to produce VR in time (or ever).
I think those lines will intersect as you can do it on hardware that is O(mobile phone) in cost. Just not actual mobile phones. Those are trash for VR.
I think it's really unlikely that the in four years the market will be so small as to support zero VR. I think it's comparable to the first iPhone which was really not that great.
I know I will use and keep buying VR displays forever even if they freeze at the current level of technology. Even when someone is offering me 16k OLED 2d displays with 240hz refresh rates. There is no substitute for a fully tracked POV and proper depth perception.
So your expectation of what existing VR hardware costs is probably a little off. Let's say that you have a BOM of 500$ to make a Rift. That's 100-300$ to spend on a battery, mobile GPU/CPU and memory, cameras, and associated cooling and packaging. It's within the realm of reason at the 800$ price point.
I think the reality is that BOM will shift around quite a bit. VR might stall at the current level of fidelity for a while (which is fine IMO), but ease of use and price can be improved quite a bit.
Even since launch the cost of the full Rift setup has decreased quite a bit. 1/2 to 1/3rd although I would have to really think harder about what a realistic number is.
And we're discussing when to call it a failure. On a site called Hacker News. OK. But that's kind of my point - people can be so focused on "calling failure" that they miss the incredibly transformative development that's right in front of them.
Regarding a "timeline for calling failure" - who knows? Sometimes success comes to the sole believer who labors on long after everyone else abandoned hope. But how often in tech is something truly a failure? Was AI a failure in the 60's? It brought us LISP and defined the space. The 80's? It brought important foundational work to CNNs. The 90's? Would we be here without Deep Blue? Was mobile a failure with WAP? Palm? Newton? Was Internet 1.0 a failure when the bubble burst in 2000?
All were important steps that got us to our ubiquitously interconnected, AI-enabled, device-driven present.
So why the focus on identifying something as a failure, or tagging the moment that "they're not catching on"? How are you defining success? Adoption? If so, based on what metric - and at what ramp - and based on whose expectations? Investment success? For whom?
Even if this generation of VR fails to gain traction, I truly believe it will be viewed as just one more step towards the inevitable. It's that good.
But my point is that such failure will be due more to social issues than technical ones.
Flip-flops had tons of adoption (literally billions of people used them) when they were the state of the art.
I would say VR consumes about 50% to 75% of my free entertainment time.
If the resolution was a bit higher, I would probably do a few hours of work per day in VR as well. I find I'm fairly focused when working in VR as I have no environmental distractions yet can be in very peaceful surroundings (Mars, space, a balcony overlooking city lights, etc.).
I use an app called Big Screen to bring up my desktop in VR. Interestingly, it's multi-user so others could co-work in the same environment, look over my shoulder at my screen, talk, etc. even from across the world - if I had colleagues with an HMD and wanted to collaborate with them, of course (so much for the focus and serenity part of it!)
[1] https://vrcover.com/product/htc-vive-disposable-hygiene-cove...
Numerous 360 video portals and creators present and produce content including BBC, Disney, National Geographic, New York Times, etc. VR tie-ins to feature films include Spiderman Homecoming, John Wick, Games of Thrones, and many others.
Companies such as Volkswagen use VR for training and development. SIs such as Accenture have built practices to support enterprise adoption of VR and AR technology for manufacturing, training, and advertising.
Though a lot of content includes thin "experiences," there are so many high quality, rich applications and games that a backlog (more content than time) is a common problem for working adults. Oculus alone will be publishing a major game each month through the end of the year. Fallout, Skyrim, and other game titles are getting the VR treatment as well.
Content is not the problem. The outdated perception of a lack of content is a huge problem.
[1] http://store.steampowered.com/vr/#p=0&tab=TopSellers [2] https://www.oculus.com/experiences/rift/ [3] https://www.volkswagen-media-services.com/en/detailpage/-/de...
No it's not. The big differentiator this time is not graphics (though they're vastly improved) but reduced latency and lag and more accurate head tracking. That's the key to making it feel immersive an making it less likely that people will throw up.
VR is a much harder problem than the 80s pioneers expected. Previous display technologies "worked" for VR, but they gave you neckache and/or severe nausea after a few minutes. We needed a breakthrough in lightweight, high-resolution, low-persistence, low-cost displays. That breakthrough arrived when Samsung started putting OLED displays in the Galaxy series smartphones.
We finally have headsets that really work - headsets that provide immersive experiences, headsets that can be worn for more than a few minutes at a time. We've got GPUs that can push pixels fast enough to maintain a stable 90fps. We still have a huge way to go, but I think it's far too early to declare VR as vapourware or a technological dead-end. We're only just getting started.
Television emerged as a broadcast medium in the 1920s, but didn't become mainstream until the 1950s. I wouldn't say that's a particular good example of fast adoption (or at least, not significantly faster than VR).
It's not merely graphics that make today's VR better than VR in the 90's - and actually viable. There has been a confluence of enabling technologies including:
- High bandwidth internet - High DPI OLED screens - Development of VR-specific techniques such as low-persistence pixel strobing, rolling shutters, pre-warped images, asynchronous reprojection, etc. - Computer vision developments for 6 degree of freedom tracking - Submillimeter tracking of room position for the VR user and his/her hands - Fast GPUs to allow for consistent 90 frame per second display - Cheap precision manufacturing for complex systems - Lens technologies - Lightweight materials - Capacitive touch technologies to enable finger tracking - High bandwidth I/O ports (USB3, HDMI) - Fast, cheap CPUs - Widely available Graphics engines and toolkits to accelerate and democratize content development
None of the above were available in 1998.
Yes, one important result is better graphics, but the sum total of the above is an experience where your brain forgets it's inside a simulation. This comes from low "motion to photon" latency such that your movements track 1:1 with the virtual world, "good enough" graphics (but could be better), a wide enough field of view (but could be wider), and so on.
People didn't buy VR in the 80's, or 90's, or 00's, because it sucked. The technology wasn't there. That is no longer the case.
Future iterations in this generation will widen the field of view, further improve pixel density and brightness, lower weight, reduce cost, and increase portability - all important evolutionary steps that incrementally improve the experience.
[1] https://en.wikipedia.org/wiki/The_Sword_of_Damocles_(virtual...
The problem is not in the technology. It's in the content. When the current adopters will start using the systems they already have semi-regularly then and only then I can agree it has a chance.
- It needs to be at a price point akin to a tablet computer
- It needs to be able to run for more than an hour without overheating
- It needs full 6dof head tracking and ideally 6dof controller tracking with motion to photon latency comparable to vive/rift
- It needs to be comfortable to wear for an hour for most people
- It needs to deliver a graphical experience on par with something like a PS2 or PS3
- It needs to be portable and wireless so it can be passed around and shown off to others easily
- Most importantly, it needs to be simple to set up and use. Ie, no external computer, no wires, no mounting of sensors on walls, no need to plug or unplug devices before/after use, etc.
No such device exists on the market today but many of them will exist soon. Choose any one VR device out there today and it will meet a subset of these but not all of them.
If devices meeting these specs don't have some form of mainstream appeal a year of the first release then I'll be revisiting my assumptions about VR. (I have been working in VR for the last 3 years)
it is not like a futuristic product that for now is too big or battery doesn't last, etc. it is by definition something nobody wants because it is exclusive, in the sense that only one person at a time can experience it and be completely cut off from the surroundings. and there is no way to solve that even in infinite interactions.
it migth work for teenager gamers and porn. who knows. but definitely will never see a huge adoption because it is inherently anti-social.
If quality VR were available at a price point similar to a phone, comfortable to wear for hours at a time, I have no doubt many people would get one. If it also comes with quality AR, most will probably never take them off.
Except many of these people are using their phones to connect with other people. Literally the opposite of "antisocial technology".
That's why I believe VR will become mainstream as soon as the technical challenges are solved, just like phones.
Last night I could have been alone after my wife and kids were in bed. Instead, I was in a 5-on-5 person zero-G battle akin to the Battle Arena in Ender's Game. Communication, coordination, and just trash talking are critical.
The night before I went on Star Trek missions with three friends (only one I had met in person). Despite them being from 3 different parts of the world, I could look around the bridge and see them right there. We may as well have been in the same room, with eye contact, body language, and hand gestures.
Facebook bought Oculus because of the pro-social opportunity to connect people worldwide. Though Spaces is early and a bit rough, it clearly demonstrates how compelling it is to have face-to-face talks with people inside and outside of VR.
Echo Arena: https://venturebeat.com/2017/05/22/lone-echos-vr-multiplayer...
Star Trek Bridge Crew: https://uploadvr.com/star-trek-bridge-crew-review/
Facebook Spaces: https://www.facebook.com/spaces
Maybe, but most MVPs also don't go anywhere, and lots of markets are fads and die. The question is why this is different for VR?
(And I've lived the previous VR fad, when it was going to revolutionize stuff back in early/mid 90s. What we have left from that era is the Lawnmower Man).
Lots of hobbyists (or rather say: gamers) were keen on getting a decent soundcard or 3D acceleration card - until it became very mainstream.
In contrast, it feels like the whole mobile app paradigm was a more emergent trend that rose from the hardware getting better and it being an 'obvious' thing to do with software on that hardware. Like, smartphones weren't designed specifically with apps in mind, but they emerged as a phenomenon naturally, in response to demand. But maybe that's just hindsight talking, maybe it too was ultimately pushed from the top down by companies building apps and promoting them. Hard to say I guess.
If any company can push VR into the mainstream, or at least accelerate the process, Facebook has to be one of the best positioned, and I hope it works. I just wonder if the hardware doesn't need to get dramatically (5 years) better before it becomes obvious that ordinary people will want to incorporate it into their daiy routine (and then, possibly more in AR form than VR form) - indeed, to be fair, at F8 a very similar message was delivered in some talks.
Well they do own Oculus, so they have a vested interest in making sure VR succeeds, whether it is a demand or not.
The form of presentation impacts blocking and scene-changes. Theater in the round would result in different staging than a traditional auditorium, for example.
A 360 degree movie is essentially the inverse of the theater experience. Instead of fixed locations pointing towards the action, the viewer is in the center of the world with potential for action all around.
People will pay attention to what's happening, there's no need to worry too much about someone simply ignoring the action. Besides, even casual YouTubers do positional sound nowadays, which can be used to "lead" the spectator.
Do they react to what the player is looking at or where they're standing? I think what the GP meant was that "you can't have David Attenborough going on about the lovely penguins whilst the user is staring at some polar bears."
You'd think that device manufacturers would then subsidize content development to differentiate, but that didn't seem to pan out either http://variety.com/2017/digital/news/oculus-story-studio-shu...
There's still no killer app. You can play FPS games. Works fine, for people who like shooters. You can run roller coaster sims. Works fine, gets boring fast. You can visit Second Life. Kind of cool, if you like SL. You can visit High Fidelity, which is a higher-res Second Life where nothing is going on. Boring. (Worse, the High Fidelity CEO is now talking about "sovereign identity on the blockchain". If one branch of hype isn't working, pivot to another.)
VR conferencing? Working in High Fidelity and Second Life now. HF even has gaze tracking and facial gesture capture.
In the early days of VR, there was hope that VR would provide a more natural interface to CAD programs. Autodesk put effort into VR to try to make that work. Turned out that trying to do precision work in VR is not only difficult, but really tiring.
It's all been built, and nobody came. Best Buy is removing the Oculus Rift demo stations from stores due to lack of interest.
Personally I think that features like virtual workspaces (e.g. multi-monitor displays / overlays) and HUDs (e.g. vehicle navigation) are going to be game-changing outside of the entertainment industry, when the headset tech gets to where it needs to be; this will probably be better implemented as an AR overlay (a la Hololens) than full VR.
Take for example software development. It has huge potential to improve how efficiently we can work together remotely. It is often difficult to communicate very well with remote team members, and to develop a rapport with them. But imagine if you could be pair programming right beside them, in a VR setting, pointing to some code in a floating IDE or something, and explaining to them how it works. Or you could pull up a virtual "white board" and start drawing maybe.
It could also go a long way to making work more comfortable. With your computer being just a VR headset, it would be easy to work in a laid back recliner, standing in your office, or even lying down.
In AR, the graphics don't have to be as good. Text and line drawing alone would be useful. This translates to less-clunky headgear. The Microsoft HoloLens is far less clunky than the HTC Vibe or Oculus Rift. Being able to walk an oil refinery or a server farm or a WalMart and have displays of what everything was and how it was doing would be valuable.
I think VR and AR will be very big in the future. I don't know how long it will take. The current generation might fail for some understanding of the word but I think they are on to something.
Perspective[0] is a fabulous capacity (and Renaissance "invention" for artworks). Our eyes need 3D, but our brain is just fine with a bit of fake perspective.
PS: slightly tangential but, the only personal interest I have for VR is watching movies alone as if I were in an empty theater. I find it very cool.
I don't think the NES would have succeeded as much as it did were it not for the Mario Brothers + Duck Hunt bundle. The light gun turned out to be a gimmick, but at the time it felt like such a huge technological leap over Atari, and Duck Hunt was the way to show it off.
VR needs a Duck Hunt + Mario Brothers.
I think Nintendo gets this. Wii had Wii Sports, the Nintendo Switch has Zelda: Breath of the Wild.
They're more than just games, they're brands that through their ubiquity generate cultural capital for the platform.
I can only use a Vive if I have a high powered gaming computer in a place conducive to gaming.
You highlighted two major barriers of entry for the larger market: 1. Initial Cost 2. Technological knowledge required (to inform the purchase and to set it up)
Another point adding to the cost is that VR games of any quality [or by larger studios with resources for quality development] are new games, and as such usually run around 60 to 100 dollars each. And they don't just come on a disk you pop in and play.
Sidenote: Playstation VR, which is substantially easier to set up -- the headset is more expensive than the console at this stage. If you get the starter bundle it's almost 3x the cost of the console.
To many people, it's outrageous. And the argument would likely follow, "there just aren't that many good games for VR" or "I just bought a new 4k tv, why would I buy that _now_". (At least this is what I've heard so far)
Playstation VR seems like a terrible deal at this point with the Rift price drops and deals. Most of the cost of a VR capable PC with a Rift with vastly inferior experience.
I don't have a Playstation myself, I play on a PC when I do. The rest of what you said sounds about right.
Today, a vr rig, hotas controller setup and elite:dangerous already is a killer combination for vr. Put the price is steep and the appeal is much more narrow than Mario Bros.
[ed: personally I think vr will be more of a paradigm shift, than a game gimmick though. I think creative tools like tilt brush will combine with multi-user world's a la second life or croquet/open cobalt. And I think it'll grow a bit "top down" first: with early adopters and creators / developers.
But one thing that has to go are the wires - probably first by higher performance/watt and a move to "backpack pcs" to run the software.
Before that, Sony has the right idea: lower ambitions and let players sit in a sofa. Works fine for driving/simulators etc. ]
Is this still not an issue?
The system came out in the US in 1985; the Action Set bundle you described came out in 1988. The Japanese release of the system was in 1983. It was a huge leap in technology, as you would expect with 6 years between the release of the 2600 and the Famicom; the NES graphics chip is much more capable, and the system has much more ram and address space, but the cpu is only clocked 50% higher (and is essentially the same, but with a larger address bus, and interrupts in the NES), the sound chip is better too. None of this detracts from your main point -- packing in the right software sells the hardware.
I mean, it's stupid, all the games and apps are stupid, all the videos are stupid, interaction is clumsy, almost everything I try to use shifts steadily to the left until I can't see anything anymore without turning my head, and eventually turning my whole body around to keep up with the shift, and all around it just barely works at all. But, it's still amazing.
That tells me something really big is going to happen; the Internet started out the same way. It was stupid. You could watch a video, if it was 1/10th the screen size, buffered for several minutes, and you didn't need it to look very good. Regular old television was a million times better. I feel like we're at that same point right now with VR. It all sucks right now (probably because I have a cheapo VR headset that uses my phone and my phone isn't well-supported by the VR players, but it was cheap), but we're only a year or so into the "anyone can afford it" phase.
Think how many years it took for the tech to use the full potential of the Internet took to make it into the majority of homes. I think we're still way too early to tell how VR is gonna play out, but when I use it, I feel a bit of that old "OMG, this is gonna be huge in ways I can't predict" feeling the Internet gave me the first time I logged on.
This sums it up perfectly.
Keep in mind that these are all 1st generation products. When an immersive VR system comes with a wireless headset, costs less than $500 and doesn't require 30 square feet of space, then it will begin to take off.
Remember the first iPod? We are going to be in this early adopter phase until the technology matures.
VR is not 3D TV. People who use immersive VR actually like it. 3D TV was annoying to most people after the novelty wore off.
Current generation video game consoles are also incomparable to Pong. But hey, that's the difference between 1st and 8th generation consoles.
A lot of things have improved but most of the problems are things that were recognized before and still aren't perfect.
They still need to get smaller, lighter, higher resolution, faster refresh rate, 4-8X GPU power, and wireless, but in another decade they should be awesome.
I kid. It's a movie from the 90's about VR turning groundskeepers into geniuses. I'm over simplifying. Watch it. I dare you. I might watch it again.
The question is whether they are enough, not whether they are merely better.
Autodesk had a VR ping-pong demo, with the paddle position reporting to the computers. It took a small rack of 1980s PCs and, I think, Matrox graphics cards. Matrox had one of the few graphics cards of the 1980s with hardware geometry support, the 4x4 multiplier unit. Nobody bought it because it was so expensive. There wasn't enough compute power in the 1980s for general collision detection, let alone game physics, so you couldn't align things by pushing them against each other. Various kinds of "snap" were tried, but now there was a mode problem. In VR, you don't have a keyboard, and doing keyboard-like things is hard. Mode controls are a problem.
The '90s brought all sorts of 3D input devices, from robot arm like devices for input to magnetically levitated trackballs. All that stuff has disappeared. There have been real improvements in 3D GUIs driven by a mouse with a thumbwheel. (Try Autodesk Inventor or Autodesk Fusion, both of which have free trials.)
Newer VR systems don't use gloves; they just give you a sort of game controller for each hand. OK for gaming, but not for fine manipulation. Good for shooting, not for putting a nut on a bolt.
This stuff has come a long way. It tracks every slightest movement of your head and controllers. The graphics and sound will immerse you in a world you will believe is real.
Very nice gaming is possible, as well as recreation. People could visit fantastic virtual worlds just to be somewhere else and feel good; a kind of therapy, if you will.
I don't understand what these insipid nay-saying wankers are blabbing about.
IMHO walking around is the real thing in VR; games not directly designed with this in mind are not exploiting the unique possibilities of the medium.
I wonder if anyone has tried using this tech to somehow trick the body into activating what happens when one actually goes running? Can I get a workout by just watching myself run in this VR world???
"The mind makes it real"
Conversely, apparently elevated heart rate due to gaming does not improve aerobic fitness.
So I'm thinking that, broadly, there will be no such thing as a free lunch here (as per usual).
I know lowering the lag reduces this issue, but it does not completely solve it.
I have a friend who has an interest in VR so he tried the various models being demoed in shopping malls when he could find them in the US and the EU. Always the same problem: motion sickness. So he would ask the people managing the demo kits and would usually get the same answer, a bit more than 50% of the people who try have that same problem. Then a few months ago I came across an article that was essentially raising the same point (can't source it, sorry. The number they gave was closer to 40%).
So basically it seems about half of the target audience asked to shell out a small fortune will have "puking on the carpet" as their most memorable VR moment... Not really a recipe for success.
Having tracked hands, moving around (instead of turning with a controller) and using teleport (or just room scale 1:1 movement) reduces the percent of dizzy people to zero. If you use slow artificial locomotion with comfort features (e.g. FoV reduction), that number is about 5%.
We have a design meeting and visualisation tool for engineers and desginers to discuss projects with their management, clients and other domain experts. This has significant benefit for our customers and it will not go away anytime soon.
Perhaps the hype cycle will cause the current HMD manufacturers to exit the market and VR will retreat to these niches but it's definitely not going away.
I think it will happen though for the one reason that Bethesda is delivering Doom, Fallout 4 and Skyrim for Sony PS4 VR for this Christmas. That will give Sony a big reason to keep yelling about PS4 VR and will restore the faith of people like me who bought a VR headset early that is now just sitting there waiting for good software to use it with.
The next challenge facing vR beyond that point is whether the three titles mentioned above deliver on the promise of VR - if they do not then VR is finished, at least for this round of enthusiasm about the idea.
FPS in VR only works from a static standing point, otherwise the VR sickness kicks in really fast.
VR still doesn't have the three variables. But it's getting really close!
IMHO, the future is the replacement of the screen of smartphones by glasses with AR. The phone will be in my pocket with my hand and the glasses will be near my eye, ear and mouth.
*Or anyone "authoritative".