After roughly four years of anticipation we finally have our first glimpse of the mysterious headset from Magic Leap, including its unique form factor and various specs. And while it feels like we’ve barely scratched the surface, there’s actually a lot to soak in. For now we only have pictures of the headset and brief impressions of the user experience from Pitchfork and Rolling Stone. Though it's mildly frustrating that Magic Leap haven’t completely lifted the floodgates on their device yet, it appears that they’re following a similar pattern to Microsoft’s initial reveal of the Hololens in January of 2015. Like Magic Leap, Hololens’s creator Alex Kipman spent seven years crafting the technology behind his device. And when it came time to demonstrate it to the press it was done so in locked down environments, meaning no photographs were allowed and it was demoed with pre-production hardware away from the public eye. As Sean Holllister described in his article for Gizmodo, the device he tried was still tethered to a PC via cabling over the rafters above him. What’s missing for Magic Leap though is any sort of public event. When the Hololens was unveiled, the private demos with the press were published alongside Microsoft’s own livestream of their team demoing it for the public in real-time. Perhaps this simply has to do with cost. While two billion dollars in funding seems like a lot, it’s nowhere near the $241 billion in assets that Microsoft has to burn on multiple public-facing events each year. So for now we have a tease of the headset, and the likelihood of Magic Leap’s own public event later down the road. As AR/VR reporter Brian Crecente wrote after his time with the Magic Leap One, “It does sound like they plan on having a full preview press day with hands on demos of full games and apps prior to that”.
Let’s start with what may be the most important factor of Magic leap One: The optics. How we will see the virtual images created by Magic Leap is the bedrock of any AR display and the company has fallen under heavy criticism ever since The Information published a tell-all in late 2016 where it was revealed that they would not be using a fiber scanning display for their first product. Personally, I have always felt that this was a red herring. As optics guru Karl Guttag wrote, FSD is notorious for low resolution, and therefore basically a dead end. It’s also important to note that while FSD was mentioned often in Magic Leap’s vast array of patents, the technology was never mentioned by the company in any of its marketing. So is there a danger to following such patents down the rabbit hole? Brian Crecente had this to say on the subject: “An interested thing I haven't yet reported is that apparently Magic Leap regularly files essentially fake patents. That is they use a method of over filing for patents to obscure the real tech they're using. This creates a smokescreen for their real tech.”
So how do Magic Leap One’s optics stack up? According to Crecente, “the images felt like they had more substance than with the HoloLens”. He also noted in his extensive article for Rolling Stone that when viewing a virtual object, “The world around it still existed, but I couldn’t see through it. It was as if it had substance, volume – not at all a flat image. I was surprised to find that the closer I got to the robot, to an extent, the more detailed it became. Getting close to the floating object didn’t expose pixels; it highlighted details I wasn’t able to see from afar."
Another interesting prospect of Magic Leap One’s optics, but yet still unknown, is how it will function in bright light. To date, the demos have been conducted in dimly lit rooms and this seems to be a standard for other headsets like the Hololens. But Magic Leap CEO Rony Abovitz made an interesting comment regarding usability outdoors in an interview for Time a few weeks ago. “Abovitz speculates that ‘experiments’ in using the eyewear outdoors will likely surface when the product ships. When that happens, Magic Leap will be watching closely to see how people react.” If Magic leap One is indeed usable outdoors at all it would mark another stark difference in how the optics are handled versus the Hololens which becomes completely unusable in outdoor conditions, and is therefore not even mentioned by Microsoft as an option. But who knows, perhaps Abovitz was just referring to using it outdoors at night. Either way, we are obviously many years away from a Magic leap device that turns billboards into personalized advertisements. Maybe that’s not a bad thing, especially in the climate our society is currently in.
Also related to the optics - and a much desired feature for current VR devices - is eye tracking. Crecente noted in his Rolling Stone report that Magic Leap’s device was capable of this while describing rudimentary AI that would follow his gaze and maintain eye contact. This isn’t really the most compelling use case of eye tracking, in my opinion, because this effect can already be accomplished with head tracking in VR devices like the Rift. The bigger question is whether the eye tracking here can be used for foveated rendering, which would result in lower CPU and battery consumption, and in turn produce high fidelity graphics. Lastly, there’s supposedly multiple focal planes (compared to Hololens’s single) though the jury is out on how well that is implemented. Avegant’s own lightfield headset garnered widespread acclaim around the tech-sphere for its usage of multiple focal planes, so this should be an exciting addition for Magic Leap.
In addition to the optics there are two other noteworthy improvements over the now three year-old Hololens: Hand tracking and 6DOF controller input. Hololens has crude hand tracking that limits it to just a few hand gestures, such as the “pinch” maneuver to select and resize objects, but aside from that there isn’t much you can do with your hands. It has yet to be seen how complete Magic leap’s hand tracking will be but Crecente says it will track individual fingers including “velocity and direction”. In one demo centered around the music of Icelandic band Sigur Ros, Pitchfork writer Mark Hogan describes using his hands to reach out and manipulate virtual objects which reacted like organic musical instruments. This is an important distinction when compared to Hololens, because the way that hand tracking works on Microsoft’s device the hand gestures are done only outside the user’s peripheral vision, away from the small field of view the objects are centered inside of. The reason it’s done this way in the Hololens is because you can’t actually use your hand to select a virtual object; rather you use your head gaze to select the object with a cursor, then snap your fingers outside the FOV like you would press a button on an Xbox controller. If you were to try and use your hand to actually select an object in the Hololens the optics would immediately inhibit you because your hand is incapable of occluding the object. In other words, if you were to try to press a virtual button on your wall while donning a Hololens your hand would appear underneath the button, not on top of it as your brain would expect it to. Whether there is proper hand occlusion in Magic Leap One is still up in the air, but it’s hard to imagine the Sigur Ros demo reacting to hand gestures in the way described by both Pitchfork and Rolling Stone without it.
There is a remote control for the Hololens which can be used in lieu of the hand gestures, but in my time owning one I never used it. It might be better referred to as a clicker, very similar to the one that ships with the Oculus Rift. Magic Leap One will also ship with a controller, but it is much more capable as an input device, and appears to be on par with HTC’s Vive wands. It has six degrees of freedom, meaning its position in space is recognized by the headset. This is crucial to immersive gameplay in VR headsets and it will likely be a game-changer for AR headsets as well. When the Vive launched it was the first VR headset with tracked controllers, setting it miles apart from the Rift developer kit that only had head tracking. Ask any Rift owner during the first year of it being on the market, and they will readily admit that the lack of tracked controllers was a serious drawback for gameplay. And there's already a compelling use case for this tracked controller on the horizon: Dr. Grordbort’s Invaders. First seen in Magic Leap’s infamous “Another Day In The Office” concept video, this shooter aims to be the flagship game when the One launches sometime this year. The tracked controller will apparently act not only as the gun pictured in the original video, but be occluded by the optics to appear as if it resembles the blaster as well. Contrary to what was reported in The Information, Grordbort has been in development since 2012 and according to its creator Greg Broadmore last month “It's only a hop, skip and a leap from here to show you all the crazy shit we've been building with Magic Leap these last 5 years.” My personal favorite game on the Hololens has been Fragments, a truly transcendental experience that places life-sized AI in your home who actually sit on your furniture while discussing matters of a kidnapping investigation. Dr. Grordbort’s Invaders will likely have that same intensity, but with a 6DOF blaster in your hand.
In the honorable mention category, another positive is that because the Magic Leap One splits its headset away from the CPU, GPU and battery it’s much lighter on your head. A discomfort of the Hololens was that after about 45 minutes of wearing it it would become unbearable for me, which mostly made movie-watching a chore. Crecente wrote that “The weight felt very evenly distributed around the crown of my head. The pads inside, which are all replaceable, helped a ton with comfort too.” Also worth mentioning is that apparently the device will be capable of interfacing with your PC, which theoretically would mean infinite battery life (at least tethered). And lastly, while Magic Leap’s device has fallen under unique criticism for not catering well enough to women users because of its pocket-able processor, it will in fact ship with a strap which anyone can use to hitch it to their body. The recently released TPCast wireless transmitter for the HTC Vive ships with a similar accessory.
There are two sore spots for the Magic leap One, though the details of both are still undetermined. The first, and in my opinion the most important, is price. Magic Leap’s CEO compared the device to “a premium artisanal computer,” setting the bar pretty high for its price tag. Additionally, the company’s Twitter account said “Magic Leap One will be priced similar to high-end laptops.” This puts the device right around the two thousand dollar mark. A report from Business Insider earlier in the year had it a bit lower at roughly $1500. If it happens to come under $2k it will be a much easier pill to swallow for developers and enthusiasts than Hololens’s $3k price tag. But that doesn’t do much for broad market adoption and the general public. The price is certainly justifiable though. A Vive package, including the PC to run it, will also cost you the same amount, generally speaking. My hope is that like with the Vive (whose product marketing head Jeff Gattis left HTC to join Magic Leap) we see a hard push to make the public aware of the product despite the high price tag. HTC had a tour bus go around the world to give demos of the Vive to people unable to attend tech conferences, and they doled out headsets for free to Youtubers, otherwise known as influencers. As a result, the Vive scored a tremendous lead over the older and better-known Rift until it eventually gained feature parity with the Oculus Touch controllers and roomscale tracking.
The other downside to the Magic Leap One is the FOV, though we don’t have a specific number on that either. Basically the FOV (field of view) is the window in which the virtual objects appear. In AR devices there is a bit more leeway here than with VR devices, because with those your peripheral vision is actually cut off, resulting in binocular-vision. With AR, the edges of the FOV only cut off the parameters of the virtual images, not your entire vision. However, low FOV -- at least in my experience with the Hololens – means frustration. At roughly 33 degrees diagonally, the Hololens’s FOV is more than three times as small as the Vive’s 110 degrees and is an exercise in futility, meaning you can never truly achieve a level of comfort which permits you to use the device as naturally as you can a VR headset. Watch a movie on Netflix in the Hololens and you’ll quickly find that you have to hold your head completely still so as not to cut off the edges (unless you shrink the virtual screen down to the size of a laptop). Play Fragments and you’ll find yourself having to stand some 15 feet away from the NPC’s in order to see their whole bodies. Walk any closer and they’re lower torsos get clipped and disappear. There is a silver lining to Magic Leap’s FOV, however. According to Crecente, it is “much larger” than the Hololens. While his description of holding a VHS tape in front of your eyes to simulate how wide it is subject to wild interpretation, I found a separate description of his to be more helpful. With the Hololens the FOV limitations are immediately apparent. You don’t have to look for it, in other words. But Crecente writes the following in regards to the One’s FOV: “I first noticed it when I walked up to something and looked down. There was a stark horizontal line at the top of my vision where the image simply vanished.” To me, that sounds like it is not immediately noticeable, and in my case, it didn’t take walking up to something and looking down. Hopefully the FOV is around 50 degrees, and from Crecente’s description it may be close to that. Ultimately, what you want is an FOV that doesn’t constantly call your attention.
Probably the most subjective and controversial aspect of the Magic Leap One is its look. Smaller and much lighter than VR headsets, but much bigger than Google Glass or even ODG’s yet-to-ship R-8’s, the One is an in-betweener. It’s likely that Magic Leap had aimed early on to have the size of the headset be much smaller, but it's limited by today’s technology. No doubt, future generations will achieve that sleek form factor; the question is how many generations down the road. The subject was awkwardly brought up at Fortune’s Brainstorm conference in July of 2016 when an interviewer asked Abovitz if their system would “require glasses”. He responded “It’s a lightweight headset”, likely trying to manage expectations. So, even though Magic Leap’s headset isn’t the small pair of glasses we dream of, it’s also not meant to be worn outdoors in public; at least not yet. I’ve long believed that while Abovitz and company have had lofty ambitions of users strolling into a local Starbucks and it magically turning into a Star Wars cantina, the infrastructure just isn’t there this early on. This version is a baby step towards that interconnected mixed reality world, but first it needs to work well in the home and office, and then it needs way, way more partnerships and a very large user base to warrant building that sort of “Oasis”, if you will. For now, how the device's appearance is fine for "at home" with friends (as most VR devices are), and maybe in another 8 years or so you’ll be able to drive down Sunset boulevard wearing what look like RayBans while Disney characters dance in the streets like it’s just another Saturday. Maybe by the time Magic Leap Three is released.
Barring any terrible surprises down the road, I personally I can't wait to pre-order the Magic leap One. I thoroughly enjoyed the several months I had of owning a Hololens, but it's time for the next level. I would have killed for an extra ten degrees of FOV, and the Walkman form factor is something that I look forward to trying out, considering how well the TPCast works at delivering similar satisfaction with the Vive (and soon for the Rift). Also, what will developers do with eye tracking? The FOVE headset has gotten strong praise for its implementation. And how will the multiple focal planes enhance the realism? Avegant's lightfield prototype wowed the tech community with its depth of field. What will they do with hand tracking? Leap Motion was a blast on my Rift DK2, but always felt limiting by not having any tactile feedback, something AR can actually deliver. And how many variations are there with a tracked controller? Both the Vive and Rift have already cracked opened that universe, so how will it feel to actually look down and see a Portal gun in your hand... in your living room? I think what's most promising about Magic Leap's device is that the possibilities are limitless due to its Swiss army knife configuration. And that's just what we know about the hardware. With the promise of a future launch event, where we'll likely see what developers and partners have been working on for the past few years, there's a whole lot more to learn. Up next: Magic Leap's developer portal, opening soon to the public...