How to build a holodeck, week 26

Half a year down, and where are we?

Take a breath, relax.

The last month has been very much spent enjoying the summer. I’ve made very little implementation progress, but spent a fair amount of time thinking about what I’m building, where the market currently is, and where the opportunities lie.

I’ve also found the last few months really frustrating, trying to keep multiple devices all working together. Tango still doesn’t work with Unity 5.4. The camera setup in Vive is a mess. Photon has a voice release that’s 3 months out of date with their main networking code. I know it’s too much to expect for everything to play nicely together, but one could hope.

So, lesson learned - when something stops working, I’m just going to back off from developing on it until it does work properly, which means Tango takes a back seat for now. Once hardware comes out with a decent camera and Unity 5.4 support, I’m back on board, but until then, I’m putting the scanning stuff on hold.

I’ve got a few outside environments up and running in Portals, and a friend who’s getting a Vive next week, so that should help things along considerably.

State of the market, August 2016

Vive is slowly hitting mainstream, but it’s still low 6 figures of devices. I’m guessing (but can’t find actual numbers) that Rift is similar or smaller, so being charitable we’re looking at 250K wired positional tracking HMDs out there.

Hololens gets more exciting to me every time I hear someone who’s actually tried it, but throwing another platform into the mix is going to confuse things for me, so I’m going to hold off on that too.

Oculus touch ain’t here yet, and I’ll not be surprised if it’s a November/December release, which means Vive is the place to be right now.

Intel’s Project Alloy is looking pretty interesting, so I’ll be keeping a vert sharp eye on that in the future, but like all things this year, it’s going to take a while before I can get my hands on it.

Given that the Vive is (imho) the clear winner as far as devices go, the next interesting thing to look at is going to be content.

What next?

I really want to get point cloud aggregation from multiple devices working, and texture mapping on top of that, so I’m going to get cracking with the Kinect as the first data source for this, and then once Tango finally get their fingers out and get it working on U5.4, I’ll bring in that data set too. I’ve got a bit of work to do getting a viable meshing algo working, as I don’t want to fall back on OpenCV, so I’m going to bring my C# port of OpenChisel back on the table and try and get that working in the next month or so.

I’m also going to spend a big chunk of time next month looking at the games and social experiences that are out there on the Vive - I think the time has arrived where there’s actually a few bits of pretty compelling content out there to learn from. Rec Room is proving a lot of fun, and I spent a most enjoyable evening in Brookhaven yesterday with some friends sharing headsets. I think once there’s some multiplayer games of the same calibre as Brookhaven (and hopefully with a bit more interactivity) we’re going to see a bit more interest from the wider public on this stuff.

Next week - probably more holiday time, but hopefully a day or two work on kinect depth data.

Written on August 21, 2016