3 ms·

Currently "AR" and "VR" are used to differentiate to separate (but mostly decoupled) distinctions: - Hardware: VR means full eye buffer coverage of both eyes w

by gfodor 8y ago

Currently "AR" and "VR" are used to differentiate to separate (but mostly decoupled) distinctions: - Hardware: VR means full eye buffer coverage of both eyes with no transparent glass, AR means partial eye coverage with all pixels sharing visual field with incoming world photons. - UX: VR typically means full environmental replacement with basic room and body tracking and AR means real world augmentation with virtual entities. In the long run these are separate concerns. We are either wearing headsets with transparent surfaces or we aren't in the long run. This seems clear will have a 0/1 resolution. For the UX side, it seems unrelated -- presumably you should be able to move up and down the spectrum of full environmental override at any moment. In the hardware sense, I expect VR style opaque hardware to win (with passthrough cameras to enable them to simulate transparent glass perfectly.) In the software sense the distinction becomes irrelevant. Through the lens of hardware full eye-covering visors seem to be likely to beat glass on the capability/cost curve readily over the next 5 years so my bet is on Jordi-like visors not glasses for daily drivers.

wuliwong 8y ago
I"ve thought the same thing. I can imagine AR/VR melding into a single device.
neolefty 8y ago
True, he seems to be addressing high-resolution immersive AR vs VR. Is that because he's looking for something to invest in? If so, it's definitely putting the cart before the horse for the average software dev, since AR can be very low fidelity and still be effective. It just has to add some value to reality to be a pretty big win. Mapping software is the main example to me. But maybe mapping is not even AR anymore — it's its own thing? What about Pokemon Go? Is that AR in this context?
mncharity 8y ago
Yes, there's so much confusion about AR vs VR Imagine wearing an "AR" HMD, a nice HMD with an LCD occlusion layer so it can render black. A black dinosaur is perched on your bookshelf. Great AR. Now you turn around. Half your room is gone, replaced by the rolling hills of a Cretaceous lowland. Great VR. But your laptop seems unchanged, sitting on a Cretaceous rock shelf where your desk was, next to a tree stump where your chair was. Great...err, hmm. You turn halfway back. On your left is hills; on your right is a wall and a bookshelf with a dinosaur. Are you doing VR or AR? Both? Half and half? One eye each? Sigh. We so need to disrupt the technology society uses for thoughtful analysis and its dissemination. Or say I'm wearing my VR HMD. But it's showing only my laptop's desktop. Is that still VR? My custom environment has a spatial geometry, and a response to head motion, that is variously aphysical. I joke that VR is in its user-onboarding skeuomorphic UI phase, and I'm much more interested in expert UIs for XR... without the "R". Resemblance to reality as UI design smell. Is that still Virtual and/or Augmented Reality? There's a camera ducktaped to the front of my HMD, providing a video passthrough background. So is that AR? What if I cut side windows in the headset blinders so I can use the real world for balance (old laptop with integrated graphics). Is that AR yet? Unfortunately, VR HMDs still have the unblurry resolution of a 1980's VGA monitor, so I take off the HMD and look at my laptop screen. Is that AR? No? But I'm wearing anaglyph or LCD shutter glasses, and I'm headtracking, so there's fixed-in-realspace 3D content. Is that AR? But what if my desk has several displays, all doing synced high-resolution (but gappy) 3D, across much of my field of view. Is that AR yet? The XR+ design space seems quite broad and rich, but much underappreciated.
- gfodor 8y ago
  Well put. I think the moment of "?R takeoff" is when the first person manages to wear a consumer grade standalone VR headset for most of their day using passthrough. It will probably be an experience for nobody but those with the most hardened stomachs, but it will be an existence proof that you can then build software that can assume you have a raster buffer between your retinas and the incident world photons throughout your day that is 'full-cover' (ie every incident photon to your eyes is software proxied and modulated optionally by world photon sensor information via code, not physical properties like transparent glass.) That is basically the "final visual platform" for compute imho. Transparent AR glasses, if anything, are clearly a transitionary technology since they do not provide full software control over photon delivery to the eyes. (Also, I expect this event to happen quietly and unrecognized in late 2019 probably buried in a forum thread somewhere.) From there, it will just take the hacker community to build apps that provide some type of 10x augmentation by a fully software-proxied visual system and a little bit of cultural normalization of it (probably within circles of teens/kids) to incentivize the acceleration of development of passthrough visor hardware that works better, is smaller, and has longer lasting batteries. But these will be mostly linear engineering problems, except perhaps the frame prediction algorithms needed to deal with camera sensor latency. Whereas the path to a set of AR glasses that mask out the world with full FoV seems a much more theoretical future that stretches the limits of known physics, this seems to be imminently becoming the problem of arranging bits not atoms with the arrival of fairly capable consumer standalone VR headsets with workable passthrough cameras. If you think about it hard enough, if you woke up in 5 years and there was a market with good AR glasses and good VR goggles, but nobody tied the software together to enable all-day VR goggle wear via passthrough, clearly someone would do it and suddenly those devices are insanely more capable than the AR glass equivalent since they can take over your entire visual field and deliver all the same applications of the glasses. So it seems pretty inductive to me, if you buy into a few fairly simple assumptions, that this is our future, assuming that a software-proxied visual system can deliver value to daily life. The only counter argument (and in fact, one of the admitted theses for many people working on AR glasses) is that wearing a VR visor all day will make you look silly -- sometimes that's enough to tip things over for quite a while, but in the limit people are going to converge on what improves their life the most on net. (See: the segway has finally had it's day with the arrival of cheap electric scooters.) Full-cover, rasterized vision seems to be the end game for how our visual system will interface with software systems.

by gfodor 8y ago

wuliwong 8y ago

I"ve thought the same thing. I can imagine AR/VR melding into a single device.

neolefty 8y ago

True, he seems to be addressing high-resolution immersive AR vs VR. Is that because he's looking for something to invest in? If so, it's definitely putting the cart before the horse for the average software dev, since AR can be very low fidelity and still be effective. It just has to add some value to reality to be a pretty big win. Mapping software is the main example to me. But maybe mapping is not even AR anymore — it's its own thing? What about Pokemon Go? Is that AR in this context?

mncharity 8y ago

Yes, there's so much confusion about AR vs VR Imagine wearing an "AR" HMD, a nice HMD with an LCD occlusion layer so it can render black. A black dinosaur is perched on your bookshelf. Great AR. Now you turn around. Half your room is gone, replaced by the rolling hills of a Cretaceous lowland. Great VR. But your laptop seems unchanged, sitting on a Cretaceous rock shelf where your desk was, next to a tree stump where your chair was. Great...err, hmm. You turn halfway back. On your left is hills; on your right is a wall and a bookshelf with a dinosaur. Are you doing VR or AR? Both? Half and half? One eye each? Sigh. We so need to disrupt the technology society uses for thoughtful analysis and its dissemination. Or say I'm wearing my VR HMD. But it's showing only my laptop's desktop. Is that still VR? My custom environment has a spatial geometry, and a response to head motion, that is variously aphysical. I joke that VR is in its user-onboarding skeuomorphic UI phase, and I'm much more interested in expert UIs for XR... without the "R". Resemblance to reality as UI design smell. Is that still Virtual and/or Augmented Reality? There's a camera ducktaped to the front of my HMD, providing a video passthrough background. So is that AR? What if I cut side windows in the headset blinders so I can use the real world for balance (old laptop with integrated graphics). Is that AR yet? Unfortunately, VR HMDs still have the unblurry resolution of a 1980's VGA monitor, so I take off the HMD and look at my laptop screen. Is that AR? No? But I'm wearing anaglyph or LCD shutter glasses, and I'm headtracking, so there's fixed-in-realspace 3D content. Is that AR? But what if my desk has several displays, all doing synced high-resolution (but gappy) 3D, across much of my field of view. Is that AR yet? The XR+ design space seems quite broad and rich, but much underappreciated.

gfodor 8y ago
Well put. I think the moment of "?R takeoff" is when the first person manages to wear a consumer grade standalone VR headset for most of their day using passthrough. It will probably be an experience for nobody but those with the most hardened stomachs, but it will be an existence proof that you can then build software that can assume you have a raster buffer between your retinas and the incident world photons throughout your day that is 'full-cover' (ie every incident photon to your eyes is software proxied and modulated optionally by world photon sensor information via code, not physical properties like transparent glass.) That is basically the "final visual platform" for compute imho. Transparent AR glasses, if anything, are clearly a transitionary technology since they do not provide full software control over photon delivery to the eyes. (Also, I expect this event to happen quietly and unrecognized in late 2019 probably buried in a forum thread somewhere.) From there, it will just take the hacker community to build apps that provide some type of 10x augmentation by a fully software-proxied visual system and a little bit of cultural normalization of it (probably within circles of teens/kids) to incentivize the acceleration of development of passthrough visor hardware that works better, is smaller, and has longer lasting batteries. But these will be mostly linear engineering problems, except perhaps the frame prediction algorithms needed to deal with camera sensor latency. Whereas the path to a set of AR glasses that mask out the world with full FoV seems a much more theoretical future that stretches the limits of known physics, this seems to be imminently becoming the problem of arranging bits not atoms with the arrival of fairly capable consumer standalone VR headsets with workable passthrough cameras. If you think about it hard enough, if you woke up in 5 years and there was a market with good AR glasses and good VR goggles, but nobody tied the software together to enable all-day VR goggle wear via passthrough, clearly someone would do it and suddenly those devices are insanely more capable than the AR glass equivalent since they can take over your entire visual field and deliver all the same applications of the glasses. So it seems pretty inductive to me, if you buy into a few fairly simple assumptions, that this is our future, assuming that a software-proxied visual system can deliver value to daily life. The only counter argument (and in fact, one of the admitted theses for many people working on AR glasses) is that wearing a VR visor all day will make you look silly -- sometimes that's enough to tip things over for quite a while, but in the limit people are going to converge on what improves their life the most on net. (See: the segway has finally had it's day with the arrival of cheap electric scooters.) Full-cover, rasterized vision seems to be the end game for how our visual system will interface with software systems.