Magic Leap One Creator's Edition First Look
cnbc.com
cnbc.com
What are they expecting now? They honestly think developers are going to flock to this device with no audience for $2,250 and produce groundbreaking applications and games that will sell millions of Magic Leaps to consumers to recover their investment?
Why don't they take that 2.5 billion dollars, and prove a single application or game can be developed for this product that people will line up to buy? If they can't do it with unlimited money, almost a decade of time, and the future of their business depending on it, then why do they expect some random developers are going to pay money to do it for them?
I know this post sounds negative, but I'm 100% convinced this product will be dead on arrival and nothing more than a 2.5 billion dollar mistake to show a few hundred people floating jellyfish in their living room.
I would see Magic Leap falls to same category as hololens - which is used for real work in the construction industry - but only on a more affordable price point.
The key question is how easy it is to develop applications for it. Hololens basically consumes unity apps which makes it a pretty easy for development and deployment.
You say that, but we built the leading product for this in 2014 (Visidraft) and AEC still isn't ready in 2018 for this for a host of reasons.
Forgetting for a moment that 99% of these applications are on tablets and phones - which is still too difficult to manage for most AEC firms - the problems inherent in the lack of infrastructure to solve the 3D content management, object localization, user collaboration and markup management are still unsolved. Not technically mind you, but in business processes and workflow.
Add HMD to that and you cut the market deployment by 1/1000th. So a demo here or there from DAQRI or Hololens does not prove the value of it being a serious and concrete product.
Beyond that though, the biggest problem was that it ends up providing the most value as a sales/collab tool between clients and developers, and we constantly heard that architects didn't want to use it because it put too much power into the client's hand. Imagine that.
So I'm not sure where you're getting the "provide concrete added value." Show me the successful product - not a demo, or a pilot product, but one that is generating profit at scale and is integrated into an entire business process at a firm like Gensler (one of our clients previously).
It's a theoretically great use case, but in practice, it requires too much from the user to make it work as a stable product.
[1] https://www.architectmagazine.com/videos/visidraft-augmented...
3D content creation is one of, if not the biggest bottleneck for AR.
I can see a world where retailers on Amazon.com are incentivized or forced to submit 3D models of their products along when creating a for sale posting.
Consumers will put on their AR glasses, or use WebVR to to visualize the item before they purchase.
Amazon will have the gatekeeping power to force vendors to use a certain standard 3D modeling process, or have an in-house team receive the products and use photogrammetry/CAD modeling.
How much value/productivity will this generate - separate question but one worth asking.
However Amazon (et al) have trouble just getting good 2D images from Vendors and is not making much of a push at all to get vendors to build 3D content.
This is a long standing problem: Should retailers do the modeling or vendors? It's expensive and hard on infrastructure to do it, and CAD standards aren't consistent enough to use that avenue. Vendors aren't tech savvy and have questionable practices for investing in marketing/sales materials, including photography because it doesn't scale.
I could go all day on this, but I won't. Bottom line is, none of the major platforms are stepping up and being leaders in AR on the content side. It's pretty marginal efforts so the results are marginal. I think everyone is waiting around for the Apple glasses now.
https://www.recode.net/2017/11/16/16666662/williams-sonoma-o...
I've been really annoyed that all food sellers aren't required to provide a list of ingredients and nutritional info to Amazon when selling on the platform (the US government also doesn't make them available via the FDA). That data would let Amazon build a food graph so people could really search for products that fit them.
For example: I want a cookie that is like an Oreo, but has less sugar.
Today, they just can't answer these questions without making some big assumptions, and can't expose the ingredients via an API so we can build cool products with the data.
http://mixedreality.trimble.com
You are quite right of the challenges but the technical competence of the largest engineering companies is quite sufficient to implement new solutions if they actually provide value and fit into their pipeline (err... that means most of the time that they can display dwg and revit models).
Hololens 2 and Trimble (and Andromeda) ... If they get it right ...
I feel like this is the start of explaining the why: "...architects didn't want to use it because it put too much power into the client's hand."
That's likely true for many more traditional Architectural professionals considering early AEC success case studies with AR/VR. I have to imagine the average Architect in a leadership role is battle-hardened by the changes that have displaced much of their fee, influence and involvement in the existing building construction framework. Giving up more of any of those likely feels like a loss for those individuals. Just as early IT professionals felt the loss of their wizard powers as computers and mobile became consumerized, Architects are feeling the loss of their wizard powers as clients become more informed and engaged.
AR remains from what I can see, the fodder of AE start-ups and larger firms placing their bets on the technology to address project design and coordination challenges. VR has already inundated every firm I work with - big and small - on most projects they undertake.
That's not really a problem, rather it proves there is a valuable use case for it.
Depending on the process architects have varying power over the overall project process. This can vary a lot between regions, companies and domains (i.e residential, offshore, bridges, etc etc)
More Simply: It's not solving user problems better/faster/cheaper than current solutions.
It's an order of magnitude harder to create 3D content than it is 2D content. Tying 3D content to the real world in a way that is useful and persistent is incredibly difficult. The AR form factor (phone or HMD) is still wonky and nonstandard. Too many other factors to consider still.
I'd assume that with a BMI model and AR capabilities (even from a phone) where you set a square (or some other shape) with known dimensions on the ground to calibrate and use GPS (or other positioning systems) for determining position/heading, doesn't seem like software for visualisation of models in-site should be that hard. Even on a current phone it should be doable.
What am I missing? Because it feels like there is some complexity I'm not considering.
I don't mean that in a mean way, but for example using GPS for AR anchoring is not viable and will never be. So we have to build computer vision systems for localizing content, and those are complex and expensive and have hardware, bandwidth and other constraints.
BIM is ubiquitous, but not easily integrated with other hardware, and the hardware best suited for it (HMD) are expensive and not yet everywhere - or even widely adopted.
You can get centimeter level accuracy from GNSS with off the shelf custom antennae and software, see e.g. Trimble Catalyst https://catalyst.trimble.com
You can combine that with AR devices to get good enough positioning onsite for review purposes.
It's not precise enough for attaching nails and such, for which there are other location based solutions.
We actually have most of the pieces for a 'smart construction site' in existence as off the shelf components. But the construction industry is such that startups are not very well positioned to lead the change - hence I understand you disenchanted tone.
No, GNSS is not ready to go. Look at the submeter coverage area for GNSS, it's TINY. Go look at work that Todd Humphreys has been doing at UAustin for over a decade trying to do cm specific GNSS localiztion with wider area deployments. It's just not tractable, scalable or anything near ready for a mass market. Like there is no reason for GNSS to grow at all when we're now building cloud vision based localization systems.
You keep referencing Trimble, I get it, they are doing good work in a handful of these areas. But the reality is that the uptake of hardware by firms to support AR at a scale that would support a company, is nowhere close to ready.
I'm telling you, I've skinned this cat from every conceivable angle.
I'm telling you, I've skinned this cat from every conceivable angle."
Ok, that angle I can agree on. I don't think there is any way a small company can provide an AR component and hit it big.
However, if an AR solution is bundled next to other computing infrastucture, with the import of files etc. handled, I've not seen anyone dismissing this concept.
The problem is that you need the entire infrastucture for project management and engineering to be digitized, and then you can add up the AR component to that.
Then AR is just one more "check box" in the feature lists of application suites offered by the incumbents. I think it will become an important checkbox, but like you said, there won't be an "Uber of AR" like there is no "Uber of DWG import".
One thing that I think MagicLeap have got wrong for a business market is not working over glasses (it seems like you need to fit prescription lenses to the device)
That means you can't easily share the device with others, which I would think would be a common use case in business (i.e. most places won't buy one device per user, but a number of devices for a department)
Microsoft can also sell the Hololens at cost to get people building on their cloud platform for that sweet, sweet, reoccurring revenue.
In comparison, MagicLeap has a lot of investors they need to make happy.
Agreed. 10 years from now, AR will still be the future of modern business.
I'd be cautious about that.
I got to try a laser keyboard[1] and it hurts to use it pretty quickly and isn't that accurate. I suppose we could just have a double-wide version of a wrist-wrest for the haptic feedback, but that doesn't necessarily help the accuracy issues.
1 - https://www.amazon.com/Brookstone-FBA_mp-796246-Projection-V...
The most awful thing ever.
One major application that we know about is in the comic book space. That's not a huge market, but there are definitely some hardcore fans that would be willing to pay the premium in order to wallpaper their homes in their favorite giant sized comic books.
For me, if the quality is high enough that I can read comics without serious eye strain, than I will be buying it.
My reaction to most d&d players I know, is "cool, but no thanks". We are developers and play d&d to get AWAY from technology. We use it where it makes sense (I use a digital 2d map).. but the fun of d&d is in the mind, and having 3d monsters isn't helping the core game.
The chance, to be there, have a foot in the door, when this thing explodes and wraps the whole world in a new layer of devices.
And these devices have the chance to replace everything- you can see and must not touch to function. Want a comprehensive list - look around you.
PC-Screens, TV-Screens, Room-Decorations, Pets, Inflated-Ego-Expression - and all of this, in just one room. This can reform public spaces, city's will have themes on a daily basis- yesterday was Venetian carnival, today its inception town.
And the best of it all- none of these goods will require a environmental footprint to manufacture- trade will have a huge chunk taken out- and moved into a steam-like ecosystem. Who's entrance fee- will look like peanuts to the revenue gained.
What about HoloLens, maybe Google Glass 2, and a few other less well funded competitors like Meta 2?
Whether they can find compelling enough use cases to sustain them in the meantime is the real question (and I suppose that's what you're asking)
For me if I can choose between looking at AR for less than one hour with a big headache because my eyes are playing pinball with my mind and looking for more than one hour to AR or mixed reality and “not notice anything because the resolution is too low” I would obviously choose the second option. Not noticing anything means that magic leap reached their objective. Your experience is seamless, you won’t even notice the different focal planes.
1. There's a neat hardware size progression for ML that was posted on Twitter: MagicLeap's gone from a device like you stare through at the optometrists office down to something comparable to a scuba mask.
https://twitter.com/karanganesan/status/1024970520205307904?...
2. There are a few different ways to achieve AR style displays and what Magic Leap has chosen is the hard way -> lightfields and eye tracking (two technologies that haven't really been attempted to ship at scale).
3. This is still a developers' release - and not truly a "paradigm shift" or whatever overhyped term you want to use.
Separate from all that, I still think the end game of AR (in the next 10ish years) is to be the thing that comes "next" after smartphones. The proper way to view this ML Dev kit is that it's the equivalent to a PalmPilot or other barely functional PDA that had some good ideas, but didn't quite hit the right inflection point of portability, screen, battery life, form factor, etc. but that paves the way to get there.
In HN discussions of it, it's been "over-dismissed". In my recollection, the majority of the comments viewed it as smoke and mirrors. Many were basically or even outright saying it was fraudulent.
See for yourself: https://hn.algolia.com/?query=magic%20leap&sort=byPopularity...
This version of Magic Leap's display has nothing to do with light fields. It is a wave guide optic, just like Epson Moverio, Hololens (the term holographic also implies nothing special in terms of optics, it's a wave guide display combined with perspective rendering), and almost all other see-through HMDs. Magic Leap made headlines when they teased their oscillating fiber display, which had potential to render light fields, but that has not been pursued with this release. That's not to say they won't pursue it in the future, it's just not relevant to this product.
Next, eye tracking actually doesn't affect what is rendered to the display. The relationship of the headset display's field of view to the user is rigidly fixed even if your eyes move; different objects should not be rendered when moving only your eyes.
Rather, eye tracking is used for contextual information and user interaction. Gaze direction could be an input to foviated rendering, or you could use your eyes to select and manipulate digital objects.
Nothing I've said is meant to detract from ML's product; it's a very exciting development and a huge accomplishment for ML and the XR industry. I just think we should keep these points in context.
I disagree. Hololens is used for plenty of custom applications in the construction industry. You just don't hear about them because they are pretty much inhouse projects, likely implemented on top of unity (hence low cost and simple to deploy).
Here is an example https://www.youtube.com/watch?v=gF8H_GAm8mM
No consumer killer apps yet, that is of course true. But I wouldn't see AR in any major way in consumer space - just like you don't have HUD displays in a large category of vehicles, even though they are critical to military cockpits.
Personally, I'd put money on the number of people using HoloLens outside of dev teams in applications like this being in the low hundreds/thousands, but would love to be shown data to the contrary. Cool youtube videos unfortunately do not really indicate anything.
But it brought a ton of clients in the door, so mission accomplished.
According to the article in The Verge [0], CEO Rony Abovitz says it is a “full-blown, working consumer-grade product.”
[0]: https://www.theverge.com/2018/8/8/17662040/magic-leap-one-cr...
I'm with the author, I absolutely think this kind of thing will be the next step in computing. It just seems so obvious. Get rid of screeens and wear a headset that lets you put application windows anywhere in your view.
I'm glad they questioned a lot of Magic Leap's claims and didn't just parrot back the taglines and act like it's magic.
That being said, I don't think Magic Leap is going to be the product that "wins". It seems from this article that it's going to run its own "OS", with it's own special apps, and most likely limited integration with other stuff.
I don't want a new "device", a new os, a new "platform", I want a screen that I strap to my face. One that I can eventually connect to my Linux machine, or my windows desktop, or my MacBook. Maybe in wrong, but I feel like Magic Leap is going to die if it demands that kind of integration, regardless of how technically impressive it is.
Hololens on the other hand is something I'm giving a good amount of consideration to, because they seem to be focusing on "a screen you strap to your face".
Anyone who's tried VR knows what that means - 1920x1080 is not enough useful lines for serious work with text, it's just barely enough for gaming.
I think you are right about higher resolution being needed for real "work", but since this doesn't run a "real" OS either, i'm not sure how much work you could get done on it anyway.
If by computing you mean video/game consumption then yeah sure, but for pretty much every other application (programming, editing, reading, etc) this is a poor medium.
Games and stuff are nice, but everyone I've known that uses a vive or oculus gets tired of needing to move around in VR pretty fast.
But being able to get rid of the 3-4 screens on my desk, and replace them with a headset that I wear? That I can put windows everywhere? Behind me, far left, even throw one up above me to glance at? That's HUGE imo. And the cost can quickly make sense.
I'm in the market for a new set of screens for my machine. I want to go to 4k. I'm looking at realistically $1500 for new screens. If something like Magic Leap or hololens was only 2x more than that and could get even "1080p equivalent" resolution compared to my screens, i'd happily drop the extra cash right now.
Yes, it's still not at that point as far as I can tell, but I don't think needing to increase resolution a bit is going to be the thing that ends up making VR/AR "impossible".
Are you sure about this?
Because that's the idea i'm envisioning. Being able to put papers or applications on a virtual whiteboard that completely encircles you. Move them to where you want, make them "fullscreen" (full-view?), maybe even be able to move them forward and backwards in focus and "position". Hell, with some extra work, we could even put them on surfaces around the room. Throw a monitoring program on the wall next to my window, "project" keybindings on top of my keyboard's keys when i'm pressing parts of a combo, etc...
Hell, even if you hate the idea of an "augmented reality desktop", you could just project "fake" screens and use them the same as you normally do.
I mean it's not like our current tools just go away and I'd start craning my neck to look up and to the left as far as possible for 9 hours a day... You could still have tools like "multiple desktops", and change/move them. But now I have the option of putting my task manager above me at like a 60 degree angle straight ahead and can glance up at it when needed.
(And before you ask: Yes, I use one laptop screen for all my work.)
What in the world do you put on those other seven screens?
> offline documentation (i.e., not in a web browser)
This deeply confuses me. Why does it matter if the documentation is in a browser or not?
man and info come distraction-free out of the box.
I can't think of a situation where I'd need to focus on all the things you list simultaneously – even just having e-mail on the screen while I'm trying to code is distracting – so I'd rather have them hidden. Sure, I'll have a couple code windows and documentation or logs open, but that fits with room to spare on a single monitor.
The only people I know using 4+ screens are monitoring Satellites or industrial processing equipment.
With this, you could just float individual windows around you. Quick splitting and navigating between splits in Vim is already one of the features of this editor I use the most - and if I could do that in 3d, leaving small splits all around me and moving them around, it would be an immense boost in productivity.
Of course you would, that's like saying "if something like a Tesla could operate completely autonomously in all weather and road conditions I'd happily drop the extra cash right now!". Unfortunately, this is a "hard" problem and while I'm not saying it won't happen, we are nowhere near anything that can render 1080p VR monitors... 4k VR monitors are unlikely to be possible (in a consumer product) within the next 20 years, and you can definitely be sure that anything like that will not be wireless.
Think about where this will be in 10-20 years, not where it is now. It will be like putting a pair of glasses on, not a huge headset. Once your workspace can exist in 3 dimensions, new paradigms will emerge for arranging a "desktop" that we can't even envisage right now.
Huh? I'd love to get rid of the chunky, heavy, power-hungry screen. Read lying in bed? Check! Program on an escalator? Check! Watch a big-screen movie while eating in McDonalds? Check!
I don't think wearing headsets is a practical way to do everyday computing and it can even be bad for your eyes.
With a monitor you can look away, look out the window, etc. but with a headset the screen is always in front of your eyes, so you have to take the whole thing off if you want to give your eyes some rest from the screen.
It may be good for shorter special tasks like gaming, vr assisted surgery and such, but it doesn't seem a good match for programming, for example, which you do all day.
With a good AR headset, that focal point can be adjusted and moved around (possibly even in realtime if you want, move a screen to be "further away from you"). Not to mention that when you aren't looking at something that the headset is "generating", it's just basically a clear piece of plastic that you can see through. They are basically bad sunglasses.
You can look away from a "projected" screen just as easily as you can look away from a physical screen.
Check out a Hololens if you ever get a chance. You can choose if the object you create is always in front of you or 'stuck' in a fixed point in (virtual) space. If I place a 'screen' on a wall it will stay exactly where I put it no matter how I move around.
I don't care if i'm tethered to a $3000 desktop PC, because I have an office anyway and it's where i'm going to work.
Presumably you're still going to need a keyboard.
Unfortunately it doesn’t matter what you (the consumer) wants. Do you want a single search engine that dominates 90+% of the market and serves shitty ads to you all over the internet? Do you want a single walled garden for your social presence online that captures/sells your data and serves you shitty ads? Do you want 2 main mobile OS’s that tightly control their platforms and the apps available to you?
I will say this though...people saying they want a screen strapped to their face, as you did, is an advertisers wet dream and these tech companies can’t wait to sell you that screen and serve you those ads.
It seems obvious to me that there's a progression. Back in Apple II, C64, DOS days computers generally ran a single app at a time.
Windowing systems were invented allowing multiple apps to share the same screen. This is what we've had for 30-40 years.
For AR though we need multiple apps to share the same 3D space all running at the same time just like I have 14 apps running on my laptop right now.
I need my mapping app drawing the path to my destination. I need my transportation app showing me the next bus/train coming. I need my virtual pet app adding my virtual pet around my feet. I need my social networking app to show me my friends info next to them in me view. I need my dating app highlighting people that are potential matches.
All of this has to happen simultaneously and at 90fps. That will require an entirely new OS and possibly new hardware (dual GPUs, ... or pre-emptable GPUs) one for compositing the 3D world at 90fps and one for apps so if app takes too long to generate it's portion of the scene the entire OS doesn't start running low. An old OS that only runs a single 3D app at a time and who's apps don't share the same 3D space will be a huge hindrance to AR being truly revolutionary.
I'm not saying Magic Leap's OS is that OS but that OS is not MacOS/iOS/Windows/Linux/Android
Sure, it might require special hardware and software (even your screen has it's own software and processing that it is doing on the signals coming from your desktop), and at the end of the day the OS would need to change and evolve to handle this new paradigm, but it's not impossible to imagine a system where the "PC" does what it does now, and just tells the "VR" sub-processor/sub-machine where to put the rendered "window".
Already there are some VR apps that basically wrap your normal "windows" desktop in almost a cylinder around you and you can move screens around on it as much as you want with current VR headsets.
One of those is not like the others: “Linux” is not a graphical environment. If you had said “X11” or “Wayland” (or even “Gnome” or “KDE”, etc.), I would have agreed with you, but “Linux” could perfectly well be the base of an entirely new 3D-based application environment.
This is something that I've given some practical thought to. The approach that I considered was to treat each application process as a client to a '3D space' service through which it could add and modify defined geometry.
Failure/slowness of any given application would leave the existing application geometry in-situ within the 3D environment and avoid the most jarring extremes of user experience. I imagined that the service interface itself would mandate the availability of simplified geometries together with meta information. This would allow the service to appropriately degrade the 3D rendered environment to maintain high framerates.
I think this approach could be effective and would not require a new OS.
Apps which crash or otherwise stop rendering trigger a fade out to the white-room VR space, whereas frame-skips are handled by hard drops (so the effect is like a teleport to avoid motion sickness).
Like, Valve have really done an incredible job on the user application interface there.
Examples:
In my state we have a govt funded web/native app that provides the fuel prices at all the service stations around you. If there was a mapping extension you could turn on, it would show this directly within the mapping app, and also an indicator "cheapest fuel within X km is at Y location" when you're in navigation mode.
In camera view with an extension turned on, provide object detection similar to how QR codes are recognised.
This is already true with even the oldest and most rudimentary AR/VR systems, eg, Gear VR. The interface and head tracking is 90fps all the time. If the app is slow, the viewport will not freeze. Just it will be stuck on one frame.
It may come from ML 4 generations and a decade from now, it could come from google, apple, microsoft, or some as-yet unknown company... but it will come and it'll change everything.
Magic Leap has been under a lot of pressure to prove that they are not vapor ware. Which, is why I assume that they chose to launch this now. Unfortunately, their vision and their product are obviously still miles apart and the content is a little underwhelming right now.
Probably the best way to look at this would be as SDK for early adopters. The first version of the product that doesn't require a ton of NDAs to be signed to even be allowed in the same room with it. However, the price point is such that this is clearly not intended to be a mass market product.
For a mass market product the price could actually be fine but then the experience needs to match the price. People spend way more on fancy TVs, high end laptops, gaming hardware, etc. For that to happen there needs to be a supply of compelling content.
The UX for AR is going to have to be substantially different from whatever your current OS does. Overlaying a lot of 2D windows in your peripheral vision doesn't really make a lot of sense. Sure, it's a neat trick but probably not that practical. Similarly a lot of mobile UX is inherently 2D currently and actually optimized for tiny screens manipulated by clumsy fingers. Slapping some AR on that is not going to be very compelling.
With VR this was more straightforward since there were a lot of 3D games that with a little work can be awesome in VR. So far gaming is the killer use case for VR. People spend money on the latest VR kit so they can play all the awesome games that are out there and immerse themselves.
On the AR front the most popular thing so far has been pokemon go and a range of me too style clones of that. That was only possible because of a critical mass of cheap, camera equipped mobile phones. With AR we need new content specifically made for AR. Most VR content does not port over well to AR since the whole point of that content is full immersion and the whole point of AR is not replacing reality but enhancing it.
Clearly magic leaps strategy for this is to have their own OS, app store, devices, and platform. This sounds like a risky strategy since clearly all their marketing is about the qualities of the device and how cool the technology powering that is. It's significant how they are not talking about the rest of the ecosystem.
From what I've heard via a friend of a friend on the hololens team, MS is putting a lot of resources into building out that with a lot of cool R&D in the pipeline. MS of course has a long history of cool R&D projects that ultimately never turned into products but this looks like they are pretty serious about it.
MagicLeap is "An OS for VR/AR stuff that you make apps for", and that's not what I'm personally looking for regardless of how many other OSs it will integrate with.
My "perfect" AR system is one where I can run Slack, Chrome, my code editor, a few terminal windows, and more, all unchanged. And MagicLeap looks like it will never be that. HoloLens will, even if only on windows.
Now I might be wrong, and MagicLeap might knock it out of the park and become a new "category" of device like phone apps did, but I still can't use a phone OS to do real work outside of a few choice situations, and I REALLY want a VR/AR device that lets me do real work.
I wish them the best, and while I get annoyed at every article that talks about it handwaving away the tech and calling it "magic", I hope they do well, because more competition is always a good thing, but from what I've seen they aren't making a device for me.
> A Dynamic Digital Lightfield is a binocular display that can project digital objects into the world such that light enters the eye as if it were reflected from a real object. Our lightfield is biomimetic (mimics our biology/physiology). Magic Leap One creates a seamless experience where digital and analog lightfields combine into a single scene. For example, with Magic Leap One, you can render digital flowers in a physical vase so they can appear fresh every day.
Like that. How is that different from what Hololens does?
> Abovitz told me that the light can only be interpreted by a human — not a camera — so part of his problem in launching his product is that he can't film the experience.
How is a camera sufficiently different from the human eye that you physically can't film an image through the headset? Or is this just saying you can't film it in the sense that a flat image on a monitor can't really convey the experience of trying the headset on for yourself? (The latter is a well known problem with VR.) If the latter, again, how is that different from Hololens?
All-in-all the product does look promising; it appears to be less bulky than competing headsets and the devkit is less expensive too. I'm just not sure whether the display technology is a competitive advantage or not.
However, I think Magic Leap has reduced the number of layers down to something like three [edit: two], from the many more layers they were showing before.
I call it a "Light field display" - see https://en.wikipedia.org/wiki/Light_field See also the reverse "Light field camera" https://en.wikipedia.org/wiki/Light-field_camera
The "dynamic" part I don't understand yet - something to do with eye tracking allowing a higher resolution than otherwise would be available?
Where's "multifocal display" from as terminology?
As for where I heard the term "multifocal display", I'd recommend checking out Douglas Lanman's talk at SID Display Week, where he goes into detail on a whole bunch of possible solutions to vergence-accommodation conflict in VR headsets: https://www.youtube.com/watch?v=GbRvkS3GDsw
It was nothing like Magic Leap, no 3d, just test 2d images - and yet, these razor-sharp, light images that I was able to see in front of the sun, were better than any computer display that I've ever seen. I don't know how would I convince anyone about the quality of the product aside from telling them to see for themselves.
This is most obvious in something like the Meta2: imagine you started with a ballcap, used double sided tape to attach a smartphone to the bottom of the rim (touching your eyebrows and screen pointed down), lastly attach a reflective piece of plastic from the forward edge of the brim angling down to around your nose.
Or just look at pictures of it here (though they are all cleverly angled to not show the eyebrow screen): https://www.metavision.com/
The resulting image is both digital (from the LCD) and the reality of what you're seeing around you, but by necessity it's dimmer and the digital screen images look more translucent.
The promise of MagicLeap and lightfield projection is that you can get a fundamentally much more solid looking image by directly projecting light into your eye (with no intermediary lenses), though it's unclear exactly how much of this is happening in the current MagicLeap developer release.
It's not a perfect analogy, but it's the difference between looking at a computer screen with and without some tinted sunglasses, or a couple of 2 way mirrors between you and your monitor.
So they basically have the same category of tech as the hololens but with a few additions, tricks and optimizations.
You're correct in that the display tech isn't the differentiator right now. They are hoping the ML One does well enough to support actually getting the FSD to market. It's iterative not revolutionary.
That's not to take anything away from them, this is an exceedingly hard problem and space so it's astonishing they are getting anything out at all.
However, the general concept of a "light field display" is worth knowing.
Normal displays work by creating pixels that glow. Each pixel emits light in all directions. That passes through your eye if it is looking at the object, which focuses it back into one point on the retina.
In a (pure imaginary) light field display, instead individual rays of light are produced, not pixels. The individual rays can be faked so even though emitted very near your eye, they are focused by your eye and your brain thinks they are further away.
Possible advantages of this (off the top of my head based on reasoning from the physics and the way vision works and the way existing displays are used):
- Can create objects at any distance - including infinity. Your brain (it's stereoscopy tricked) will believe they really are at that distance.
- The viewer can change focus, and the objects will correctly blur and come in and out of focus.
- An individual large object could have varying distance and focus across its form as in the real world.
- Because of the improved stereo and focus, in an AR display especially, there will be less contrast between the projected objects and the real world. They will just seem "more real" in comparison to real objects.
- Vision correction. In theory a light field display could project images that correct for any vision defects you have - dynamically. (In practice the talk of special lenses for glasses whereas in the OP article shows that Magic Leap isn't at a level it can do that yet)
- Special effects. So of course you can then do other funky things - play with faking visual defects for special effects. Your character could appear to need glasses, and then get them as a powerup etc. Or you could make someone's vision blur when they're tired or about to die. Or have a boss which is always out of focus in an odd way.
Of course Magic Leap won't be that good. It is only digital. It will be granular both in normal resolution and in effective depth of field resolution. But still... It really does sound like it has some of the above benefits.
From the very weak way ML is using the term light field, anytime there is light, there is a light field. An LCD monitor produces a light field.
I believe this technology will advance, but I think some of the unrealistic expectations do not help.
I am just unsure which demographic they are targeting with a $2500+ headset. The tagline: free your mind. With a poster that even includes a reference to a VW microbus. Obvious roots in psychedelic culture. For which, undoubtedly, the experience would be immensely hedonic and synesthetic. How large is this audience of virtual shamans?
I can think of good applications for this technology for example in shipyard inspection, factory management, city utilities workers, police, even farming operations.
So far the gaming industry was driver behind tech, but with augmented reality being so useful for industry (navigating warehouses, for example), this might change. The prices will drop, but I don't expect ML will be priced at <$1000 anytime soon.
Psychedelic application would be very interesting. Imagine experiencing something like Waking Life or Scanner Darkly wherever you look. A gimmick, but good demonstration of tech.
They haven't done that.
> Tracking was generally good, but objects occasionally shifted or jittered. A few times, animated objects seized up altogether, which might have been an issue with tracking, Lightpack performance, or something else.
I tried a HoloLens last year[0] and, whatever its other fauls, the tracking was perfect and objects moved smoothly like they were actually real. If Magic Leap is behind Microsoft's two-year-old tech, that's a bad sign, especially when they've pitched themselves as being categorically better than all other AR.
[0] My thoughts here: https://news.ycombinator.com/item?id=14106028#14107939
Interesting deployment strategy.
Even worse, are their pricing tiers [0] - you have to pay an extra $500 for "Early access to beta software and developer tools", which is just the height of absurdity. You then have to pay ANOTHER $500 for a "Priority Service Plan" and a "shoulder strap". I figured the company was shady but this is just hilarious.
This isn't always the case. The best Centaur Chess players (Human + AI) still lose to the best AI: https://www.chess.com/news/view/stockfish-outlasts-nakamura-...
I also find it interesting that instead of making it stand alone, they require a belt/pocket clip computer unlike Hololens. The headset seems much more compact, makes me wonder what Microsoft will be able to squeeze out in v2 of Hololens. Either way, competition is good.
Personally, the only "convincing" I need is price, weight/convenience, and software. All will come in time, I'm certain. Eventually, these have a real chance of being way better than phones/laptops, e.g. they'll likely use much less battery (as the "screen" is smaller).
So, have they really accomplished anything other than creating a new HMD, one with less of an ecosystem than existing products?
I still remain curious what do they really mean by the lightfield. Does anyone have more information here?
The best I can infer is they don’t actually have a “screen” with pixels per say, but they somehow attenuate the light as it goes through the “rectangle”. Why I say this is different than pixels is that it almost seems like they are doing something more “analog”. By attenuating the light as it comes in, rather than adding its own light, I see how that you could get a different “type” of light that may be better from a human processing perspective. The “screen” would effectively consist of a high resolution array of “windows” that selectively attenuate the light going through them. It would also explain why they can’t take “video” of the experience.
Do others agree with my shot in the dark? Or is there some other way they could be realizing this technology?
Disclosure: I’m not buying a magic leap, I have no financial interest on it, I just want some working environment that I would really really enjoy.
Weight concerns are valid - I wish they included actual figures for the weight of the Magic Leap for comparison.
What will be very interesting is to see what Microsoft have come up with when they release their next version (which is looking like Q1 2019).
If Magicleap is still well ahead at that point, then it shoudl go well for them, but if MS can leapfrog them, that's going to dampen their prospects.
[1] https://www.kguttag.com/2018/06/10/magic-leap-livestream-02-...
The Magic Leap One is
sort of like looking through
a window within your
field of view
AR glasses won't be a thing until they stop trying to put a tiny TV screen in front of your face. It needs to fill your visual field.
They also won't be a mass adopted thing until they don't make you look like an 80s cartoon character.
The question remains whether VR will still have it's time before AR becomes reasonable.
Until that is answered, VC money will continue to be sacrificed in a giant fire on the alter of getting there first.
The problem at the moment is we are technologically well behind where we need to be to make practical AR in terms of screen resolutions.
I think the screen resolution isn't that big of a deal for many applications (mostly games and stuff like it), and it's not like it's impossible right now, just expensive.
The bigger hurdle from where I'm sitting is getting framerates and latency in a good position. Everyone I've seen use a current gen VR headset gets horribly motion sick in a lot of cases. That NEEDS to be fixed, and I think that comes down to crazy high refresh rates, and super low latency, both of which really crank up the processing power needed.
But regardless, I don't think resolution is going to be what makes or breaks VR/AR. We worked with computers just fine when they were ~500 "pixels" tall, and it's not like screen resolution technology has hit a wall that it can't get past yet, things are still improving.