3 ms·

I have both a Magic Leap One and a HoloLens 1. I've done work on both. I also have every major VR headset, which I have also written software for each. I've bee

I have both a Magic Leap One and a HoloLens 1. I've done work on both. I also have every major VR headset, which I have also written software for each. I've been working in AR for about 10 years, and VR for almost 6. And honestly, with the latest OS updates, the HoloLens 1, even today, is still the better device. You can compare specs on a sheet of paper all day long, but there are fundamental flaws in the Magic Leap hardware and software that prevent it from being a good user experience. The HoloLens, on the other hand, is just old. Magic Leap's battery/compute pack turned out to not give it an appreciably longer battery life, while also making the device uncomfortable, as the cable constantly tugs on the back of your head. The HoloLens isn't comfortable, either, but that's mostly because the vast majority of people completely ignored the little, rubber strap that came in the case that you attach across your to take the bulk of the weight of the device. The WiFi and Bluetooth an my ML are constantly failing. I've used three different units and have had the same experience on all of them. The controller constantly loses tracking. It uses magnetic tracking and it just doesn't do well. Also, the sensor is on the right side of the headset, and it is legitimately worse to use the controller in your left hand. It's also way too complicated for just-a-pointer, while not being complicated enough for a bespoke user input device. HoloLens has a clicker instead. It's superfluous, but it does the job and doesn't fail. The hand tracking is too broken to actually use. HoloLens' hand tracking isn't anything to write home about, but I've never encountered it doing something blatantly wrong. It's just slow. Magic Leap's is similarly slow and often wrong. This goes hand-in-hand with the ML SDK being poorly documented and rather incomplete. One way HoloLens gets around its rather limited user input capability is by giving you voice recognition out of the box. And in writing software, it's incredibly easy to setup. It's also relatively easy (though takes some thoughful design) to create a speech-based UI that works reliably. Magic Leap, on the other hand, talks about how great it would be to have speech-recognition, but offers nothing in their 1st-party demos to demonstrate it and nothing in their SDK to implement it. The operating system on ML is a hacked-up Android system. I'm sooooo over trying to do 3D graphics on Android systems. All the standalone VR headsets are also a gigantic pain for developing software. If you've ever developed software for the HoloLens, it's a joy in comparison. And Microsoft did a lot to make 2D applications an integral, usable, and useful part of Windows Holographic. While Magic Leap has its own version of that feature, they bury the documentation and don't provide any good samples. Whereas, on HoloLens, you can basically run any pure UWP app, whether it was made for the HoloLens or not.

Zee2 7y ago
Your point about the software ecosystem is very good. I think the advantage you're talking about stems mostly from how Hololens is a first-class Windows 10/10X platform now, serving as one of their flagship new platforms for their new OS. It was clear from early on they wanted Windows Holographic to be clear extension/continuation of their goal for UWP to be the new write-once, run-everywhere platform (even if that goal was a bit aspirational/unrealistic). That positioned Hololens as a more competent general-purpose computing platform. That's a valuable perspective to have on what AR/XR is supposed to look like: new, but not entirely new, and flexible enough to be able to run the old software.

Zee2 7y ago

Your point about the software ecosystem is very good. I think the advantage you're talking about stems mostly from how Hololens is a first-class Windows 10/10X platform now, serving as one of their flagship new platforms for their new OS. It was clear from early on they wanted Windows Holographic to be clear extension/continuation of their goal for UWP to be the new write-once, run-everywhere platform (even if that goal was a bit aspirational/unrealistic). That positioned Hololens as a more competent general-purpose computing platform. That's a valuable perspective to have on what AR/XR is supposed to look like: new, but not entirely new, and flexible enough to be able to run the old software.