I’ve been watching many videos and reading many articles about the Magic Leap One mixed reality headset. Aside from that headset boasting the most impressive set of technologies we’ve seen in years, it does have its bad sides, or maybe it’s better to say, some missing features and imperfections that don’t yet stand in line with our high expectations as a product that can succeed in the consumer market but even for other markets, there are things that can definitely be improved.
Let’s go over some of those things and see where things can be improved, changes or added in future versions of this headset.
Very expensive – The Magic Leap One is a very expensive platform in a consumer market standpoint. Its price starts at $2295 and although it is $700 cheaper than the Microsoft HoloLens Development Edition ($3000), it’s still out of reach of many, like VR was when it first emerged, even more, limited considering its country availability. This, of course, limits its market reach and might discourage some developers to invest in this direction in the first place.
Virtual content doesn’t look super realistic – I’ve talked about this topic in this article. After watching “Tested” video review, we now know that virtual content is never rendered fully opaque and always look semi-transparent to some degree. Some believe realism isn’t that important but for some applications, a high degree of realism can definitely increase immersion. This is, in fact, a limitation of the optical technology used in the Magic Leap One headset.
As explained beautifully by @Nikolaus who answers my question on Twitter:
The light coming from real objects located behind virtual objects, cannot be blocked by transparent lenses. Hence the transparency (real image + overlaid virtual image). Full opacity would require real-time selective light blocking of pixels, which isn’t possible yet.
Limited hand-tracking capability – based on my current analysis, hand tracking seems to be very limited compared to other hand tracking technologies out there (e,g, Leap Motion). It made me wonder why Magic Leap didn’t invest more in this technology. However, they might be working on improving this technology and I’m pretty sure that will see that part of the system getting a big boost in future versions of the headset. Just so you know, companies like Leap Motion worked years on perfecting its hand tracking technology for AR/VR applications (new acquisition on the horizon?)
A narrow field of view – Although you are able to see all the surroundings with the ML1 like with regular glasses, only a relatively small portion of it carries the virtual content display functionality. We are talking about 50 degrees diagonal FOV (40 horizontal / 30 vertical). Although it’s larger than the Microsoft HoloLens offering, it’s still can impair the immersions as you need to rotate your head to reveal content that appears outside that field of view. Some of the content will appear to be cut off as you rotate your head.
Content near you is clipped – The Magic Leap One implements what Magic Leap Calls “Near Clipping Plane”. According to them, this is done to minimize occurrences of discomfort (I’m just following the documentation here). This means that virtual content that is nearer than 37cm (14.57 inches) from the device, won’t be visible to the user. There is no option to turn this thing off. Now, obviously, the problem is where you want to interact with virtual objects in a close range, inspect them in details. You won’t be able to due to this restriction. This is yet another disadvantage of the optical system that can impair the immersion. Just imagine looking close at a real object and it suddenly disappears from sight, not a great experience is it? same goes with virtual content.
In AR, I was very excited about being able to get close to objects and inspect them up close, in some apps, this is even very much encouraged. This is definitely something that I want to be solved or improved in future iterations of this advanced optical technology, there are just no excuses to limiting this range and I’m sure this will be improved in the future.
Observers are disconnected from your experience – this is something that exists also in Virtual Reality (VR). People who don’t have the Magic Leap One can’t see what you are seeing. Yes, I know, this some kind of stupid, because obviously, they can’t. However, it will be great giving observers the option to view what you see (without all the interaction and other capabilities). For example, giving users the option to see the virtual content via a mobile phone or tablet device, using a dedicated app. This way, others can see what you see or interact with. Apart from that, there might be a way to add some sort of interaction between AR and Magic Leap One, similar to hybrid PC/Vr applications, it can be really cool.
Imagine having a social party at your house and inviting friends. It will be great enabling them to see what you see, rather than diving into your own world by yourself and kind of awkward. I have a good feeling that this will come, but not so soon.
Wasn’t designed for outdoor use – The Magic Leap One Creator edition wasn’t designed for outdoor use. Magic Leap even mentioned it on their website on magicleap.care: “We’ve designed Magic Leap One for indoor use in normal lighting conditions.”. This obviously put a huge limit on what developers can deliver for this headset. The outdoor potential is just enormous and in fact, when I envision the future of mixed reality, a big chunk of it is happening outdoors. AR Cloud is going to take this outdoor experiences to a whole new level and for Magic Leap One to have a strong appeal in the consumer market, Magic Leap need to design their headset for outdoor use as well. This is why the first headset feels very experimental. How can the company talk about creative possibilities while putting such a huge limit on the device outdoor usage capabilities?
Still an awkward looking design – don’t get me wrong, the Magic Leap One looks way better than any other AR/MR glasses out there. It’s compact, lightweight and was designed to be very comfortable to wear. However, this is not the type of glasses (inc. the external accessories) that you feel confident about going with them outside. Even considering that the ML1 wasn’t built for outdoor use, still, it’s not like the feeling of wearing casual looking glasses. If you walk with the ML1 outside you are definitely going to have some people laughing and staring at you like you were an alien from out of space. Having said all that, people were making laugh at me when I bought a phablet many years back. The thing is that if the product brings a lot of added value and can improve and supplement your daily life experiences, you wouldn’t mind using it, even if it looks a bit awkward.
If Apple adds to come up with AR glasses, I am quite certain that they will add an option to change it looks, like they did with the Apple Watch. A modular design that allows design changes can make a big difference.
Occluding fast-moving objects in real-time – If you look at the Tonandi app, you can see that virtual content position below the user’s hand isn’t masked out. Same goes for if a person would pass through the virtual objects, There is no occlusion of fast-moving objects in the scene. The occlusion of dynamic changes in the environment, whether it’s the user’s hand, another person ow a change in the environment depends on how fast and accurate the meshing system operates. Right now it seems that the performance isn’t there yet to provide the ability to mesh on-the-fly in fast speed. I haven’t talked about the ability to exclude certain elements like a human figure that maybe can be undesired in certain apps. I mean, people come and go and maybe you don’t want them to be included as part of the mesh. In outdoor use when many people scan the environment, the system incorporated with an AR Cloud solution can use average to filter out “noise” real world entities like cars, people, etc, but for single use, from my current analysis and seeing how much time it takes to do the meshing, it’s not possible. I Need to do more work on this topic though.
Update: according to the official documentation: “The world mesh doesn’t update instantly”. It mentions that if there is a change in the physical world, it will take at least a few seconds for that change to be reflected in the world mesh.
Object lighting in MR is quite lacking – When developing for the Magic Leap One, you should know that objects won’t cast shadows (according to the official Art & Animation content guide on the Creator portal on magicleap.com). Shadows give us the ability to have a better distinction of the placement location and distance of objects from other objects (physical or virtual) in the scene. Also, there doesn’t seem to be any light estimation as well (what we have in ARKit and ARCore). I am investigating this topic, stay tuned.
No “true” Black – No just that there is no opaque visualization of colors in Magic Leap One (any pixel is semitransparent), there is no true black. The Magic Leap One uses light-based additive color space (based on the sRGB Color Space). Black is the absence of light, therefore black renders as transparent. To give the illusion of black or darker colors. Based on the official documentation (source) developers will have to surround the dark area with light, which would give the effect of a darker color (that’s how the brain interprets that. When we look at that color area, we perceive its color and shading relatively to other colors in the surrounding area. I guess we will just miss that pure pitch black, well, until a new advanced computer vision technology will find ways to solve it.
Summary
I’m still doing an in-depth research to learn more about the Magic Leap One capabilities and limitations. Even with those disadvantages and others, the Magic Leap One is an extraordinary device that opens up a new world of endless creative capabilities. But like many other similar technology entries, things will take time to cook. The Magic Leap One is making its first baby steps, and although it’s not yet ready for prime time, Magic Leap is making sure that once that moment arrives, it will be among those in the spotlight and leading the Mixed Reality industry forward into the future.
I am indeed quite disappointed because you can see where this is going, higher adaptation in the business/enterprise market with consumers holding their wallet closed until the price drops or waiting even years until that “experimental” stage is over and buying 2nd or 3rd generation products. I’ve seen in VR and most of the signs signal that the same will happen here. Unless another company marches in and speeding up the process. It might be even with a less advanced product, but one that was designed for the mass market with an affordable price tag and good overall set of features. It might happen sooner than we think and than what Magic Leap would do? I don’t want to sit down and wait years until this happens, I know many developers who want to create those ideas now and be able to deliver it to as many people as possible. I mean, the potential, oh the potential (I sometimes hate this word).