This week I’ve cleaned up the texture atlasing a little, and spend some time getting Scanner ready to show to a slightly wider audience.

Cleaning up the texture atlas

After last week’s investigations into why the atlas triangle edges looked so ropey, I spent a bit of time at the start of the week putting in a new way of blitting the source texture into the atlas. I’ve also changed the vertex order of quads such that they’re pretty consistently aligned, which definitely helps.

Here’s the latest garage scan:

If you get in close to the raleigh box, you can see the texture isn’t looking anything like as distorted as previous versions.

I think this is good enough (given a decent mesh, and a decent source camera texture) to do for now, so I’m going to stop fiddling with it.

Making a product

As I noted last week, I’ve been “doing stuff” for nine months now. I’ve been pretty productive and have had a lot of fun, but there’s nothing external to show for that.

I spent Thursday at Develop:VR, which was great in a lot of ways (some exciting things being shown, with my favourite being an app that localises your phone using photogrammetry in a similar way to how Tango localises - very cool!) I also had the same conversation with a few folks that I’d had in July - namely, yes, I’m having fun, and no, I don’t have a product yet. So, it’s time to change that.

I’m going to start wrapping up Scanner and try and get it into a state where I can put it on the Play Store. While performance and quality on the Tango Devkit are not ideal, there’s still nothing quite like it on the Play store yet for Tango devices. I’m hoping that I can get some feedback from folks who have a Phab 2 Pro, before I try and chase one down myself (they’re still not available in the UK - I’ve been looking at a page saying “coming soon…!” for weeks now).

More details coming shortly, but if you have a Tango Devkit or (even better) a Phab 2 Pro, and would like to help test it, drop me a mail. I’ll be starting with a small beta run first, but should hopefully get something out for release before the end of the year.

I’m not expecting to get any traction with this, but it’s a good experiment and should ensure I have a proper focus for the remainder of the year.