HoloLens 2 Remote Collaboration is fake?

[GlebBraverman]

Facebook, Twitter and other media are still rumbling on the news from MWC-19 and Microsoft release of its HoloLens 2 with “Spatial” remote collaboration demo. The project was presented by Spatial team and CTO of Mattel (iconic toy producer). The crowd is super-excited about the demo, however we’d love to give some thought on what Microsoft’s recent release really stands for:

1) Turning 2D images into full 3D avatars in few seconds sounds quite challenging. I mean, there’re awesome guys like Wolf3D, but it still doesn’t work just like that even for them. Have a look at Jinha Lee (00:21 on the video) standing with his back turned to the crowd, this is simply impossible to create out of 2D photos – not online or within seconds - ever!

2) Moving on to magical search engine (from 02:10) with predefined 3D objects: there is a 3D model with animation for every search. Looks more like a fantasy to me.

3) There is plenty of black color in the video, while HoloLens holograms are created by adding light to real life light. This means that white appears bright and black renders transparent. To put it simple you cannot see black color in Hololens.

Bottom line, we believe the whole thing was a replay with pre-recorded user actions(best case scenario) or a video. What really confuses is why the presented case was this basic. At some point the Mattel CTO states that the remote collaboration gives a faster way to the market + reduces time for travel. This is so true, BUT: in the remote collaboration demo we’ve spotted just the collaboration part.

[lacion]

1, there are demos of avatars created in real time for VR and AR, even back in the days where all we had was Kinect. all in real time. the avatars show in that presentation were pretty crappy compared to the ones demoed in more recent VR devices that track eyes and lips.

2. its a demo, that was most likely staged so that 3d models would appear, they may be searching inside their own DAM/MAN system that already has the assets.

3. black colors can be reproduced in transparent screens. magic leap does it why not the hololense?

[GlebBraverman]

1. Not a good example. Kinect has a depth sensor - this guys claim to do it from a 2D photo.

2. OK, then what the point of showing it if it is prerecorded? that does not show any real value to users. Basically nothing different from a video.

3. Its impossible to disguise real light the way you say it. This is fundamental technical limit. I am not sure how magic leap is doing it if doing at all.

[GlebBraverman]

https://m.youtube.com/watch?v=3UTYxaXtc0s&t=35s