As I said I can give you my code. But it depends on changes that I have made in SDK files. Because I can’t give you those files (with changes), you will have to make changes in your files yourself.

well if you have time and will to explain to me how to make those changes…

First of all you have to know how to disassemble and assmble il code. You need to disassemble SurfaceInput.exe and add InternalsVisibleTo attribute that targets your assembly and then recompile SurfaceInput assembly…

ok I think I’ll read some basic of .net coding before trying anything… keep us in touch with such great work

New video on youtube.

Domino tags are working!!!

Hey Daniel,
Could you explain a bit more on your achievements? Like the first time you showed us that you were able to interact with the launcher only through the simulator, and then you could use your firefly cam, but you had some performance issues because of the upscaling of the simple input and were not able to use bytetags. So could you please explain (without getting into much details) how you are making it better and better?
Thanks a lot for sharing your success and keep up!
Greg

It is not a simulator on the first video. On all videos I’m using single FireFly camera with Surface’s vision system. On the first video I didn’t had a display cable long enough to connect a projector with PC. That is why you see Surface Shell (not a simulator) on the monitor. Then I bought extension cable and could connect projector. That is why you see Shell picture direct on the device.
Surface vision system requires input from 5 cameras. To be exactly it expects 5 parts of the same picture. This can be emulated with single camera. Divide the picture in 5 parts and scale each to the size of single Surface camera. Here is the first performance and quality issue. Performance issue can be solved with parallel algorithm. But quality is still an issue.
I have changed how I process the image. Now I get single image from camera and make 5 copies of it and process them parallel. This gave me 25-30fps (10 before).
Bytetags was not a problem actualy. I have assumed that they will not be recognized because of the low resolution. But yesterday I have tried to print bytetags with other sizes. First I thought I have make them bigger, but it didn’t worked. So I made them smaller. Tag on the video is 70% of the original size. Today I will try identity tag

I have tried identity tag with different sizes. But it didn’t worked :(
But I have found out that bytetags are better recognized if they are 40% smaller than original size (60% of the original size).

no body have surface SDK download link ?
i’ll kill myself from jealous sense to daniel

Hi Daniel

Looks awesome. Got the SDK myself, and was wondering why you had to modify the SurfaceInput file?
Because in the SurfaceInput you have the possibility to load an external (filter) assembly, directly from the menu (File -> Include Assembly).
Or was it just for convenience so you didn’t had to include it manually every time?

You can’t load any external filter because some attributes that are needed are internal. You have to modify SurfaceInput to add InternalsVisibleTo attribute and make internal classes available to your assembly.

how did you get it full screen without the simulator bars?

It is not a simulator. It is a real Surface shell. Look for SurfaceShell.exe in Program Files\Microsoft Surface\…

i think daniel has full sdk because microsoft give it to professional partners.
if you take a look at surface partners site there is a diagram that show the more partnership -----> more freedom to have full sdk