Magic Leap Reveals Developer Demo, Confirms NVIDIA TX2 Hardware

130

Today during Magic Leap’s developer livestream the company offered the first details on the compute hardware that’s built into the device, confirming that an NVIDIA CPU/GPU module, and also showed a brief developer sample demonstrating world meshing and gesture input.

Speaking during a developer livestream today, Magic Leap’s Alan Kimball, part of the company’s developer relations team, offered the first details on what developers can expect in terms of the power and performance on the AR headset.

The headset connects to a belt-clipped module which contains the battery and compute hardware. Inside the box is NVIDIA’s TX2, a powerful module which contains three processors: a quad-core ARM A57 CPU, a dual-core Denver 2 CPU, and an NVIDIA Pascal-based GPU with 256 CUDA cores.

The hardware resources are divided between the system and the developer. The headset’s underlying ‘Lumin OS’ is relegated to two of the A57 cores and one of the Denver 2 cores, leaving the other half of the cores for developers to use without fear that system processes will interrupt content. Kimball said that the Unity and Unreal engine integrations for Magic Leap do much of the core balancing optimizations (between the available A57 cores and Denver 2 core) for developers already.

SEE ALSO
Echoing Apple's iPhone Launch, Magic Leap Taps AT&T for Exclusive US Distribution

Magic Leap’s Chief Game Designer, Graeme Devine, said of the device, “It’s a console,” referring to the way that a portion of hardware resources are exclusively reserved for developers, similar to game consoles, as opposed to a platform like the PC where unrelated processes and background apps can easily impact the performance of active applications.

It isn’t clear at this time how the GPU resources will be distributed between the system and developer content, but Kimball said that the system supports a wide variety of graphics APIs including OpenGL, OpenGL ES 3.1, and Vulkan.

The company also offered a brief glimpse of a new developer sample application called ‘Dodge’. While the footage wasn’t shot through the headset’s lens, it was recorded with the device’s capture function, which records the real world from the headset’s camera and then composites the digital content into the view in real-time. So what we’re seeing is the same graphics and interactions that you’d see through the headset, but the field of view and any artifacts imposed by the display/lenses aren’t represented.

In the footage above we can see how the device uses a pinching gesture, similar to HoloLens, to act as a button press. At one point the hand is used to smack a boulder out of the way, showing that the hand-tracking system can do more than just detect gestures. The headset also has a 6DOF controller, but it isn’t shown in this demo.

The blue grid in the demo shows the headset’s perception of the user’s room, and Magic Leap said during the session that this geometry mesh is updated continuously.

During the livestream the company also affirmed that they are on track to ship the Magic Leap One this Summer (which ends September 22nd).

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. See here for more information.


  • Daven Bigelow

    Hmm, surprising the hype around this when this video is almost a year old now: https://www.youtube.com/watch?v=FdDLbCRyZ60

    • LowRezSkyline

      Yeah I keep thinking what has Apple got in their skunkworks compared to Magic Leap when you know they are throwing a ton of cash/energy/devs at it but of course don’t talk about it because, well, it’s better to surprise everyone with exceeded expectations than promising the moon and failing hard (for some reason No Man’s Sky comes to mind…).

      • Apple in my opinion is just as bad. They have a ton of money and I can’t say anything lately has been that spectacular, especially since they had structured light technology for five years to see it being used only for facial recognition, but glad to see other developers using for face scanning and facial feature motion capture.

      • Joe Strout

        Yeah, Apple is clearly working on the skunks, but I don’t think we’ll see anything of that this year or next. ML1 and (maybe) HoloLens will be the only games in town till then. (Unless Lenovo gets off their duffs and finally releases a Mirage dev kit…)

    • dk

      ml1 is a slightly better hololens ……. https://youtu.be/qADZRu1dUMM https://youtu.be/avp5skRo-EA 2 years old hololens demos
      it’s actually putting objects in your vision not like useless phones that r fundamentally limited by using one camera

      • Daven Bigelow

        Look up the Aryzon or Zapbox and you’ll see that “useless phones” are doing much better than Magic Leap (which will suffer the same limited FOV issue as Hololens).

        ARKit and ARCore doing better than this Magic Leap tech demo is further proof that the food chain of AR is still Hololens at the top, followed by premium iPhone and Android devices 2017 and newer.

        • dk

          noooo it’s still a useless demo ….if there was a headset u can plug into a phone …now that wouldn’t be useless

  • Raphael

    Tracking doesn’t look very accurate. Model sliding around due to tracking issues. Not very impressive.

    • Doctor Bambi

      Yeah, and I thought ML1 was supposed to have proper object occlusion, but when he puts his hand up to block the rock, we still see the full rock super imposed.

      • Good catch. But not surprising, and tells me occlusion is not dynamic, but based on scan of the world much like Occipital Bridge. But in saying that if it has actual sensors for hand tracking, a simple model could be added for occlusion.

    • Good point Raphael. I wasn’t looking for it (trying to do screen caps with my phone), but no surprise since they made sure not to spend a lot of time moving around, which is where most anchoring breaks down. Good catch!

      PS I am going to hold out for the Hololens 2 if it is less or the same in pricing. Loved to see more Unreal support, but they forked the Engine regularly.

      • Who’s this?

        Yep, at least with Hololens the thing wasn’t flying everywhere. Hololens 2 isn’t a thing internally, it skips straight to Hololens 3 since they’ve made such progress.

        • I agree internally, but I think even Microsoft called the next version Hololens 2, but it all blurs together. I think they should put the year after the name when they release it.

        • dk

          man this is getting annoying….every time someone mentions the next hololens ….yes they were thinking about updating the hololens and the articles were saying that they will just skip that and “go for the next version” …….no one cares ….when it comes out it will be the second hololens

    • Jan Ciger

      Yeah, I have noticed that too. Hololens can have similar issues but usually it is less visible than this.

  • Sébastien

    Tracking seems really disappointing and virtual objects seems to be flickering. IMHO, Hololens seems to be a lot better for that…

    • I agree and we still have Hololens 2 to look forward to soon.

      • Sandy Wich

        Wait.. Hololense 2? There was a Hololense 1? Like I heard they launched a dev kit you could buy or something like that, but there was a legit commercial product?

        • Yep for about a year if we are talking about the Microsoft Hololens. If you were part of an educational system or medical center, they had even better pricing. Sadly, at $3000-$3500 was too expensive for my needs.

          • Sandy Wich

            Ahh yea i heard about that version, but maybe it’s for the best, not releasing a product to the mainstream when it’s not ready and give it more time to iron out some kinks

        • dk

          lol …if u google it …..it’s 2 years old now they have sold 50 000 units anyone can buy it for 3k or 1.5k and up on ebay ….but it’s mostly meant for devs/companies ….there r thousands of developers uploading demos daily ……and it’ pretty excellent for something that was developed more than 5 years ago
          2016 real life demo https://youtu.be/qADZRu1dUMM https://youtu.be/avp5skRo-EA not promotional clips
          there r thousands of demos if u look for them
          and the winmr platform is in every win 10 updated pc and when u develop an app for win mr it can be used with their vr headsets in the virtual home and it could be used in your real home with the hololens
          and ml1 is a slightly better hololens……and the hololens2 is coming in q1 2019

          • Sandy Wich

            Yea I didn’t know they were letting consumers buy the dev kit. I wonder why they sold it to consumers though? At that price/quality not enough people would ever have bought it in order to support content development.

          • dk

            at that price it’s really for software companies ….but there is no point to limiting it ….it’s just open to everyone ….it will be for consumers when some version is around 1000 and that won’t be soon

          • Sandy Wich

            You might be right bud, but it’ll come eventually. Hopefully by then they work out all the quirks

          • Jan Ciger

            That won’t happen any time soon. Even though MS has a large margin on the Hololens, the diffractive optics alone costs like $100 per one waveguide, you have 3 per eye (each color has its own), so that’s $600 just in the glass. That’s why I am wondering how much is ML One going to cost because they have 6 per eye (two focus planes, whereas Hololens has only one).

            So unless a different tech is found or these things can be produced much cheaper, the glasses aren’t going to go down in price any time soon.

            (the diffractive waveguides are used because they permit very compact display, otherwise you would need prisms, mirrors or have the image generating element (e.g. DLP/LcoS microprojector) in front of your eyes, resulting in a very bulky eyewear like was common in the past)

          • dk

            my guess would be at least 2500 they want it to be cheaper than the hololens…but maybe at a loss of 1500 per unit ….or smth like that…..they want investors money not really consumers’ money

          • Jan Ciger

            around $2k sounds like a safe bet. I don’t think they can afford to sell it at loss/cost – they don’t have Facebook behind them like Oculus did and they are living off investor money which will run out sooner or later. And if the device will end up underwhelming, they will not get another investment.

          • dk

            they r basically selling a dev kit ….they can’t sell more than 40-75k in the next 2 years …and the next 2 generations will also not be cheap or that great or small/convenient ……their only hope is investments ….they will get nowhere relaying on consumers…….if they have to subsidise a generation the 1st gen needs it the most

          • Joe Strout

            Have you SEEN how much investment they have?! It’s over $2 billion. They could give the things away for years without making a serious dent in their capital. (And if they’re smart, they’ll do something very close to that, so as to capture as much market as possible before Apple moves in.)

          • brian solomon

            2b is a lot of capital? Not really when considering the product and the infrastructure behind developing it.

          • Jan Ciger

            They don’t really. Also the license explicitly forbids use of the dev kit for commercial purposes – e.g. we aren’t allowed to sell it to a client or use it for e.g. training. For that you are supposed to buy the $5k commercial edition (which is sold too now).

          • Sandy Wich

            You sure? They specifically call it the, “development edition”, on their website, but w/e I guess they can name it w/e they want. Fancy wording for marketing purposes and all that.

            I still wonder what the purpose is selling it to consumers though, very few people would ever pay 3000-5000 dollars, “like 4000-6500 Canadian”, for what is more or less an a pre consumer concept.

            It’s not like it’s a game changer like a flying car or something, right?

          • Jan Ciger

            Yes, I am absolutely sure. The development edition also has no support, no warranty, if it breaks you are screwed.

            That you can physically do something with the development edition doesn’t mean you aren’t breaking their EULA and they won’t sue you for it.

            They aren’t selling it retail, the target for the commercial edition are businesses and you need to contact their sales to get it.

          • Sandy Wich

            Holy crap, no warranty or support?.. They literally pulled an Apple IMac Pro!

            Ahh, tbh I just say them selling it on the store with the ability to purchase it upfront, but I never went further than that, didn’t know that in order to close the deal you’d have to actually contact them privately.

            Damn I just don’t know why they called it a finished product then? Why now release a HoloLense 2 when 1 didn’t really happen?

    • dk

      yep it wasn’t really smooth but who knows how old is this clip they played 50 times ….and if meta nac improve their terrible tracking ml could posssibly do it too

  • gothicvillas

    Before we have a good AR device, there will be VR gen 3 already.

    • Sandy Wich

      Honestly I don’t think there will even BE a real AR device until gen 3 VR, much less good or even decent. Like.. I guess you can call Google Glass AR.. So technically that’s not true, but it certainly didn’t do much or go anywhere.

      Don’t get me wrong, recently I’ve come to see the ~fantasy potential~ for AR, but imo we won’t see AR reach that level, at least something that can be shipped for mainstream consumers, for many years to come.

      But when it does? Sure, I’m on board.

      I imagine someday in 20-30 years we may have the power of current bleeding edge enthusiast pc’s inside a wireless AR/VR HMD

    • Jan Ciger

      And is that a problem? It is not like the two are competing for the same market and the same applications. Furthermore, AR is a **much harder** problem than VR.

      Of course, that is not an excuse for a crude demo by a company that literally lived on hype for years.

      • Lucidfeuer

        You’re getting it slightly off: AR devices are NOT a thing, the science is not even here, so the point is that all these AR “glasses” are fake products, that have no reason to be if not to launder techwashing money.

        • Joe Strout

          I think you need to adjust your tin foil hat. The device is real. whether it will live up to the hype or not remains to be seen, but to claim it’s all a fake is just silly.

          • Lucidfeuer

            Never claimed it was fake. I said AR glasses are as much a thing as self-driving cars…because there are few prototypes and a Tesla car that can slightly steer on linear road doesn’t mean the product they’re dubbed as are actually what they are for the simple reason that the underlaying fundamental tech doesn’t exist yet.

            “Wether it will leave up to the hype” go be a fucking hypocrite somewhere else, there’s not even the slightest hype that is even considered whether from the HoloLens or this…

          • dk

            looool 1st of all in 2018 Google’s company Waymo is starting their public completely driverless service in Phoenix ….and other cities in the following years
            ….what tesla has is good driver assist features …that could some day be an autopilot
            2nd there r ar headsets …but they r basically expensive dev kits at the moment …..they don’t really exist for the consumer….but the technology is definitely somewhat close to that happening ….the 2 years old hololens is proof that it can happen ……and now ML will have a slightly better hololens and in q1 2019 the next hololens version is coming …..and google had a small announcement that they r looking into making a similar device and the samsung standalone that will apparently do ar….and obviously many other companies r researching it

          • Joe Strout

            Sorry, I guess I misunderstood what you meant by “all these AR ‘glasses’ are fake products.” Apparently you mean “not-shipped-yet products,” and of course I agree.

            I’m still not sure what you mean about there not being any hype. Clearly there’s been quite a lot around this product, and I’m sure some folks will be disappointed with the first version when it ships. Personally, I remain cautiously optimistic.

          • dk

            hmmm where did my comment go ……anyway …..Google’s company Waymo is making their completely driverless service open to the public in 2018 in Phoenix …..and other cities to come in the next years
            what tesla and other companies have r good driver assist features that could become actual autopilot at some point
            ……ar headsets exist the technology does exist ….but r basically expensive dev kits and it will take time to reach the consumer and for the tech to get a bit better

      • Sandy Wich

        Well being a head mountable entertainment device does kind of make them competitors, although AR’s uses can go beyond that by having in-home window displays, car windshield displays, cellphone AR etc.

        I’d say they were… Semi competitors right now considering AR development is focusing on the head mounted display right now, but in the future if things don’t go well they have an easy escape route while VR doesn’t.

        And yea I know what you mean.. Those fake demos were just… A horrible thing to do. That’s what happens when you really wanna complete against VR, but aren’t ready to do real things with your own technology yet so you resort to bullshitting people.

  • Thanks Ben. I skeptical of their approach to CPU processes, but can’t verify this myself. Also, if they are using NVIDIA SLAM libraries, they are CUDA bases so GPU will have to be shared with these tasks as well as the graphics display. I also didn’t catch power capacity since I can’t seen this drawing less than 50 watts.

    As I mentioned in one my tweets, I also curious if their surface mesh is created on the fly, or scanned before hand. Also, it mesh density will play a big part in how well occlusion and anchoring works (Raphael makes a point about sliding in his post). You have not covered Occipital’s Bridge headset much, but much of their SDK does similar things using their Structure sensor. My guess is a similar AR type product is in the works using their Structure Core as well.

  • Something else I should mention. I was quoted pricing of $400 10K volume (Q4 2016) for the TX2. If the pricing is still close to this, there is no way the Magic Leap will sell for less than a $1149, but they may have to with the Hololens 2 announcement coming…

    • LowRezSkyline

      What’s the expected price of Hololens 2 (or 3 or whatever they are going to call it)?

      • Can only speculate, but “less expensive” than the Hololens ‘1’ . But knowing Microsoft they will price it competitively with the MagicLeap. My hope is that is less than $1500, better graphics, FOV, faster performance and battery life. I was okay with the weight, but if it is lighter as well, that will work for me.

        • Jan Ciger

          well, “less expensive than Hololens 1” just means <$5000 that the commercial version (not the devkit that isn't allowed for resale or commercial use) costs.

          • That’s true, I forgot that. But as you have posted elsewhere, we are looking about $1000 in components cost (I think someone actually did a full component breakdown cost, but can’t find it). My guess is the mold tooling for the frame was pretty substantial for a small run add all the engineering & manufacturing cost involved, the $3500 price seems about right. But this is speculative since I have not actually seen an actual parts breakdown or project accounting report from Microsoft.

    • dk

      there is now way it will sell for less than 2k ….and that will be at a loss of at least 2k per unit

    • Jan Ciger

      The diffractive waveguides they (and Hololens) use cost about $100/pop to make too. Hololens has 3 per eye (one for each color), so that’s $600 just in glass. Magic Leap One has 6 because of the two depth levels. So unless they found a low cost way of making these, that’s about $1200 just in glass costs per unit.

  • Dreaded ghost rocks :(

    Also, was there a noticeable delay in response to rock collision?

    https://uploads.disquscdn.com/images/dd6092b2f72aceda8770fe396a502878fc63e7f6ac1ed93e62447f8eb59a362c.png

    • Daemon Hunt

      Love this, I saw it too. Sad.

  • Zachary Scott Dickerson

    Far cry from the CGI demos we saw in the past.

    • Sandy Wich

      Yea I remember those fantasy concepts. They really kind of pissed me off instead of trying to quietly say, “This is AR!!!!”… No. That was a concept.

  • Molmir

    No wonder the footage wasn’t shot through the headsets lens, how else would they hide the plastic covering much of your vision and a field of view even lower than what’s left of the real world. Oh and this headset is supposed to be exclusively for indoor use, yet has darker glass than my ray bans. But they hid that aswell, didn’t they.

    • RJH

      ‘Tragic Weep’ I like that :) very good.

  • mellott124

    Is this the spec reveal? Disappointed. What is the FOV and resolution?

    • dk

      as usual they didn’t say anything ……from the patents and people’s analysis ——>the fov 45 horizontal 30 vertical 55 diagonal….the res is about 1280 by 960 or something like that …..therefore good angular resolution….and they have only 2 depth planes at 1m and 3m

      • mellott124

        About what I was expecting and somewhat disappointing. Didn’t think they would break any laws of physics but was secretly hoping.

        • dk

          well it’s basically phone hardware powering it ….and the optics for more depth planes will be quite complex and quite expensive …..if everything works as intended ….it’s a pretty nice improvement over the 2 years old hololens ….but nothing “magical” …..also the next hololens is coming in q1 2019 and google r apparently working on a similar device …..so there is some progress in the ar world

  • fklds

    $5B

    • Lucidfeuer

      That’s a nice tech-washing sum, the product always was “fake”…

  • impurekind

    Beyond disappointing–again.

    VR is absolutely brilliant already and is just getting better–AR has yet to impress me. . . .

    • Sandy Wich

      I had a few people bring up to me that the interest in AR isn’t what’s happening right now, it’s the possible fantasy potential of the technology in the years to come.

      OBV AR is trash atm and would fail completely to get off the ground if they ever tried to sell a product, but in 10 years?… Maybe it’ll be different.

      • Merithios

        AR with the Hololens is mostly used in industrial context, where they have the budget, manpower and use cases to get real projects going. I know some quite interesting projects, which really have proven to be usefull like: https://www.youtube.com/watch?v=_PyCBWgJU68

        Thyssenkrupp is one of the innovative companies in this sector, but also BMW, VW etc. are doing quite some research for usefull applications on it.

        To sum it up: It’s not complete trash, but the normal consumers are simply ignored, because the industry pays much better.

        • Jan Ciger

          And there is actually meaningful stuff to do. No consumer will pay $2k for a crummy headset to play Pokemon on or to watch animated ads on their cereal boxes …

          • JJ

            Exactly. There are some pretty cool medical applications that i’ve worked on that are going to be pretty big. They wont be for this gen solely because the hololense isn’t ready for hospital use, like it can’t be properly sanitized so it can’t be used. But the apps are very cool and work excellent.

      • impurekind

        Yeah, it should be pretty cool in the future for what it is, but right now it’s just rubbish for the most part. The tech just isn’t ready to do AR full justice for now.

    • dk
  • CURTROCK

    This is it? Wow. (Not a good wow, either)

  • sfmike

    So unimpressed…….

  • brubble

    HAHA! smh

  • Daemon Hunt

    Well that twitch stream: fail. Utterly boring, nothing new, the viewers are
    gonna be pissed. For all that money, I would spend some of it on better
    presenters to start. Highly energetic, engaging Gen-Y or Millenial kids
    would be way more gimmicky. Pay for good talent and you may be able to
    mask your guarded-ness better and replace it with hype.

    Speaking
    of… You know when you are talking to someone who is clearly being
    really guarded (cagey)? These guys turn guarded and
    harrassed-face-pretending-not-to-hear-or-see-the-chat into an art form. I
    think the word “FOV” was mentioned in the chat at least 763 times.

    • Joe Strout

      Yes, but I blame the teenage chat trolls for that, not the presenters. I thought they did a decent job considering.

  • Samos

    VR HMDs fitted with camera can do decent AR.
    AR HMDs can’t do any VR.

    Proof of an HTC vive can do AR thank to its forward looking camera :

    https://www.youtube.com/watch?time_continue=24&v=shB1ZoJwyms

    • Jan Ciger

      Um, have you actually tried that? I did. Through-the-camera see-through has its own problems, namely the inevitable and very noticeable camera lag. There is no point having high frame rate low latency HMD when you have e.g. 100ms processing delay before the video from the camera arrives, is processed (e.g. for tracking or overlaying the augmentations on it) and displayed. And I am being generous there, the delay if often much more than that – e.g. in the case of Vive we are talking about the video having to arrive to the PC over USB for processing and then back for display. A truly usable system would need to do all the video processing on-board, using some heavy duty FPGAs or something of similar caliber to minimize the latency. And, AFAIK, nobody has built and HMD like that so far.

      Optical see-through HMDs have the same camera lag but you notice it less because it affects only the virtual augmentations, not your view of the world itself. I.e. it doesn’t make you motion sick …

      • JJ

        Yupp same, I as fortunate enough to get to pursue vr/ar for my job and have worked with a lot of these pass through ar solutions. All of which have the lag and pretty much all have warping and distortion that immensely disassociate yourself from what you’re seeing. Your hands don’t feel like your hands, your vision feels like you’re super drunk. The tech just isnt there yet but it is cool watching it develop

      • Samos

        Thanks for all the valid points and explanations. On the short term I believe that the latency problems will be settled. An onboard chip just for that purpose seems in order…. just like Smartphones that apply Snapshat filters on the go with little latency.

        When VR HMDs will have such capabilities, and that may be soon, most AR HMDs may be less enticing.

      • brandon9271

        I think it would be fairly trivial to design a headset that displays the ourside world without to much delay. The viewfinders on old camcorders were capable of this. The AR function of the Vive seems to be an after thought. Still, i think they could alleviate some of the perceived latency with some kind of reprojection.

    • dk

      technically …at the moment pass through ar is giving u a really shitty and delayed view of the world not to mention big headset…….and u can cover up an ar headset and they do have vr apps that don’t use the external world for nothing but tracking ….but it’s a small fov and the picture is not great……. but great ar and vr in a small form factor will be possible one day http://lightfield-forum.com/wordpress/wp-content/uploads/2013/07/nvidia-near-eye-light-field-displays.jpg

      • Joe Strout

        I agree. Using a camera for pass-through AR is a stupid way to do AR. That will never catch on beyond academic exercises, because looking at the world through cameras will never be as good as receiving photons directly with your eyes. (Well, not for a very long time, anyway!)

        Conversely, doing VR with an AR headset is trivial: cover up the lenses with some black construction paper. Boom, done.

        • dk

          well if pass through actually looked good and the delay was almost nothing ….it is a great addition to any vr headset …..but a headset mean to be used for that doesn’t make too much sense mainly because it’s big

          • Joe Strout

            True enough. And it could be an important feature, if nothing else, to show a bit of the environment when you get too close to it so that you don’t bump into or trip over things.

  • Trekkie

    Disappointing…. Also remember ML is indoors only.

    • JJ

      yupp mine as well just get a hololense

    • Ben

      I thought you had limited access to a development model and you were thoroughly impressed with it’s capability and SDK?

      • Trekkie

        Would like to point out that the context of this discussion is “THE DEMO”.

        • Ben

          What makes the difference between this demo and their previous demos? The early demos were impressive I thought.

          • Trekkie

            Since there was no difference I am pissed. However the specs are very impressive – very good scene meshing (as seen in the wireframe overlay), very good hardware spec (top end Tegra X2, the best in the mobile hardware business), 6DOF controller etc. All impressive but this demo sucked in contrast with other demos by, say, project Northstar, which are awesome to say the least.

  • RJH

    So how much and when? (not that I will be purchasing).

    • Joe Strout

      When = this summer, supposedly (which surprised me — and I’ll still be a bit surprised if they make that). How much, nobody knows. If it’s over $5k, I’ll have to think about it. :)

  • MarquisDeSang

    AR games suck, please make AR useful for production or office work. I would love to see AR XCode with an infinity of space for my code.

    • Jan Ciger

      VR/AR is used for ‘production’ work since ages ago already.

      However, what you really want is Hollywood-style monitor replacement and there we are far off due to low resolution, low FOV displays and even basic ergonomics issues such as keyboard input or eye/neck strain (craning one’s neck to the side while typing on a keyboard is something you will regret super fast). If you want more space for your code, get a 4k TV, works great.

      • JJ

        Yeah marquis/jean is far separated from the industry. Even though i agree that I can’t wait for AR tech to improve, its going as fast as it can and there are some interesting products already but they’re limited by both hardware and software at this point. I work with AR image recognition and spatial mapping and its tough, the tech is still so infant that its working but not really good enough yet. So people outside of the industry are going to complain its not what they want but they dont understand its a process thats taking time to develop.

        the magic leap is crap btw. After this demo its clear to see it can only do what the hololense could do years ago.

  • Sponge Bob

    6DoF Controller demo pleeeeeeaze !!!

    Without good controller the whole thing is useless to me (for productivity apps like CAD etc)

    • Daven Bigelow

      Windows MR headsets do quite well on desktop, something like their controllers for Hololens V2 would absolutely dominate the AR industry.

      Either that, or Leap Motion’s hand tracking finally coming to mobile / standalone.

      • JJ

        yeah man I work with leap motion for my job and its not perfect but working with it is clear that theyre on the verge of something that were going to be suing for a long long time. There are some use cases that cause errors, like crossing hands confuses it same with crossed fingers but aside from those types of obvious issues you can get some pretty unbelievable results.

        • JJ

          obviously i meant using not suing haha

        • Sponge Bob

          hand/finger tracking is useless for thinks like CAD and precise drawing/sculpting/building

          Controller is a must

          touchscreen never replaced computer mouse

          likewise, finger/hand tracking will never replace good precise controllers

      • I have read in some MS developer threads that this has already been done with the current Hololens, unknown how well it worked, but there is no reason LED point constellations would be easier to detect, map & transform over detecting the contrast of your hands, assigning a bone structure and then determine when fingers are arrange in a position for a matched action. As far as processing, I have no trouble moving around and tracking my controllers in the “Cliff House” with my “Belt Nuc” an Intel NUC NUC7i5BNK ($349 on Amazon), an off-the-shelf Li-ION 6 AH battery pack and 5″ LCD screen w. capacitive touch for 2-3 hours of VR use.

        • Daven Bigelow

          Does the WMR work like Oculus Touch controllers where it’s LED constellation + the frequencies of each LED are different?

          The ability for the camera to know not just the distribution of LEDs but which specific unique ID they are constantly makes the tracking much better than optical non-IR/non-depth sensor or non-time-of-flight hand tracking.

          • Sponge Bob

            those controllers are way too big, not ergonomic and pretty much suck

            besides, the don’t work outside in direct sunlight

            for AR on the go – those are useless

            Magic Leap magnetic controller ?

            I dunno

            DEMO pleeeeeaze !!!!!

            I personally think they are full of sh1t just like Sixense guys

          • I don’t think so, since they are all the same color, but that is not to say the change the duty cycle of each LED at a rate faster than the eye can see. I do know that the LEDs do change brightness, but unsure if the controlled individually. I have asked for more details on the controller since I want to use the process for a project I am working on, similar to how the Vive Trackers are used, but I have not heard back from Microsoft.

        • Daven Bigelow

          Oh forgot to say, awesome little setup there! I can’t wait until Hololens level tech is as cheap and easy to set up as your mini VR kit there

          • Thanks. I wish Microsoft would create just the visor without the processor, like their WMR headsets that you can control with your PC module (like mine :) reducing the cost even further as well as improving the render quality and frame-rate. If Magic Leap is successful with separate battery and CPU/GPU pack, than Microsoft should do so as well, but based on the Intel Core i7-8809G variant. If they can get this down to TDP of 35W. I think it would be an amazing device, especially if they continue to handle sensor integration with a discrete chip.

      • Sponge Bob

        its not “OR” but “AND”

        both controller AND finger/hand tracking are a must for AR

  • Ivan Si

    Everyone is pissed and dissapointed with the demo. The CEO write in Twitter some lame excuse about how the video affect the lightfields. If they know that why they show the video instead of the live demo?

    Hololens show more impressive lives demos 3 years ago!!!!!!! I remember the first time I saw a LIVE demo of Fragments and Roboraid, wow that was impressive. That Fragment game it was so incredible how the characters interact with real chairs and sofas.

    This demo is nothing new. You can do the same demo with arkit, arcore or google tango.

    For me the Magic Leap is not ready and they are years away for a working prototype and they do the video using other device like a PC, Mac, ARKit or ARCore.

    • JJ

      exactly, lol so much money wasted here

    • Joe Strout

      I think you’re projecting. I was pleased with the demo, and a number of others I’ve talked to were as well. We got some real specs, we got a real demo — I think there’s just no pleasing some people.

      • Joe Strout

        I was also intrigued to see that the API has a built-in capture function (and what that looks like). As a developer, I was wondering how I would ever post YouTube videos of my software — that question is now answered.

      • NooYawker

        I guess when Leap Magic is promising to change the world as we know it, then this is the first actual demo using the product… it’s shit.
        ESPECIALLY when hololens is doing so much more than this for years and they haven’t bragged about their superior product at all.

        • brian solomon

          Hololens is doing more? Hmm. How so?

          • NooYawker

            Uh.. it’s an actual product that has been for sale for some time now. All you have to do is google or youtube hololens demo. This demo from leap magic is embarrassing

          • brian solomon

            Yes hololens is a product on sale, a pretty shotty one in terms of what it actually gives you for the money. Yet it was a first go. So I cut them slack.

            As a developer and a creator the hololens was pretty useless. Thats why almost nobody used it outside of enterprise. It was also so fragile and expensive.

            ML will be out shortly as they said. The demo was an example for developers. It’s not trying to shatter the earth, and is showing some features. I agree it would have been nice to see a better presentation. Apple is very good at presenting even the least exciting ideas in an exciting way, offen their live demos hiccup though. Look back at apples history.

            I wouldn’t be so negative and harsh especially when you’ve never tried the actual hardware. You might look foolish when it launches.

          • NooYawker

            So you think this demo is good? Interesting.

          • brian solomon

            No. Not necessarily. You like to infer a lot from very little. Not always the most productive use of your time and others.

            I personally would have preferred a demo showing non-game content as an example, but it isn’t a bad demo.

            Microsoft didn’t really show off much prior to official launch, ya know? You seem to like using that as a comparison. Most companies don’t show off anything prior to actual launch so I think it’s nice they’re starting to.

          • brian solomon

            Also just so you’re fully aware. The hololens never actually launched as a consumer product. They only released a dev kit and that is the “product” you’re speaking about.

            So if a company like Microsoft, who’s been around forever put all their effort into their tech and only managed to burp out a dev kit, maybe you should analyze magic leap in that context. Where as here’s a company that just invented something, built a company around that from scratch and off a very small sum mind you.

            How much do you think it cost Microsoft to develop and build the hololens dev kit? Even with all their previous infrastructure already in place?

            Just some questions to ponder while you develop your own views. Seeing things outside of your own narrow perspective is helpful sometimes. That also wasn’t a pun around Hololens nearly unusable FOV. ;)

          • NooYawker

            So as a consumer I should take into account the cost and company size when I buy an inferior product? Gee, this jacket sucks but they put all their effort into it so I shouldn’t buy the much nicer jacket? And let’s not forget this company promised me they were going to put out a life changing jacket. Sometimes thinking you’re so clever makes you look silly.

          • NooYawker

            It’s Magic Leap that built up all this hype, then they present this terrible demo. How else are people supposed to react? Either way my original statement stands correct.

    • dk

      nah it’s ready …..if the jerky tracking is just a side effect of it recording ….they r going to sell the units pretty soon
      …and it’s just a basic demo …it has some neat improvements over the hololens….nothing that revolutionary

  • Remember when this is what Magic Leap promised in 2016? Sigh…

    I am going to take sometime to break down the Twitch Stream video tomorrow on “3D in Review” (2:00 pm PST) and compare it to some of the other demos from other front-runners in ‘Mixed Reality’ like MS Hololens, Occipital, Meta, Dreamworld, Leap Motion and even AntVR. Along with some phone kits (one as low as $15 that even allows monocular SLAM IO tracking) https://uploads.disquscdn.com/images/4e0b0c46f4d33e8d109857f35d55f49e1f09b40946b134cbb2cc7b1bb741fdf4.gif

    • brubble

      Yeah, even as a fraudulent demo this is pushing the boundaries of its FOV.

  • Here is static frame comparison of yesterday’s live-stream and their promo in 2016.

    Why am I so mad? Investors poured millions of dollars into a promise (Including Alphabet) that should have gone to less ambitious but equally groundbreaking ideas that may have not promised the world (through a HUGE lab device demo) a revolutionary change, but an affordable stepping stone we can all enjoy right now.

    There are reports that Apple has scaled back its iPhone X because it was too expensive for the general public. If this is the case, how will Magic Leap expect to sell this at its price (or higher) to that same public? For those who believe this is not the market they are going after, just take a look at the image below. Both cater to fantasy & gaming, and with the money they have thrown at comic book creators, FX companies and gaming developers, what other markets were they going after?

    I am working with OMSI right now to put a VAMR weekend learning camp for years 10-16 (early September) and it is a struggle to talk about funding for 20 $150 graphics cards to support a $200 MS WMR headset, let alone a product like this that will cost $1000+ that can be done with a $15 cardboard AR box and smartphone that for kids will be just as good of an “eye-brain experience” as seen in this demo. https://uploads.disquscdn.com/images/4f34b2b3646962609312025f2bc5ab7176dc1a7f7cfe12fa8d7c4775c421a918.png

  • nipple_pinchy

    It has a long way to go.

  • VR4EVER
  • At least Magic Leap is good to make us laugh

  • After all the hype and that is the best demo they could put together? Jeez. Graphics are straight out of the 80’s. Tracking clearly needs work as the monster floats about on surfaces. I am just very underwhelmed by that.