Robe,
I was searching internet for simplest solution - webcam driver for kinect.
http://www.kinectdevs.com/forums/filebase.php?d=1&id=20&c_old=5&what=c&page=1
now I have all the beauty offered by hierro skeleton plugin, and rgb from the same source thanks to Soroush from kinectdevs.com. It is not elegant, but it works! You can add it as ordinary videoin devs. Thats all.
I’m rewrote Hierro’s plugin and now it able to show both streams. I will publish it sources, when clean whole code. I’m also will remove skeleton support, because i think it’s not stable for this release. Will add it again in future versions. Is anyone will have troubles with GitHub? Or i must publish to sourceforge?
what is unstable about skeleton tracking ?
@rogalag:
thanks. So you can use Hierro plugin for the depth and taht tool for the RGB as a Videoin without syncing problem or performance trouble… I’ll try
@alg:
Add it as Contribution on the VVVV forum. The skeleton is working in the hierro plugin for me, but I’m agree with a “simple” RGB+depht plugin.
Have you tried to align the streams (depth warping) like in the OpenNi example?
Thanks guys
@robe, no syncing problems, quite stable on XP SP3, works like a charm!
Btw, you can choose RGB or Depth or follow-me cam.
@alg, on my XP sp3 hierro’s skeleton is stable…
@robe Ok, i’m planning to put binaries to contributions and sources to GitHub. Yes, i’m tried it with displacement shader.
@rogalang thx for info, i will test it again i think.
@alg: you’re free to put sources anywhere of course. the advantage with putting them on sourceforge with vvvvs other plugins is that then the plugin could be included in our automatic daily build-process. this way you could get daily info if your plugin builds fine and it could be automatically integrated with the next addonpack… you choose.
@joreg With this features i think sourceforge is better place for it. Or i can just make a mirror for autobuild.
After the first release, I was busy with job, but, had time to play and develop also other stuff on kinect technology, like multiple support and multiple texture , aligning, custom recognition, hands generator and so on.
I read your post about how to split or isolate functionality into plugin, as talked in past with vux & sogokugenki, the best solution will be a chain of node depending on device, to achieve better performance and plugin for different purposes.Actually is really a nice talk, but waiting for microsoft SDK to develop this framework, thus, multiple kinect support is actually very tricky and bugged in c# interface.
I have some old plugin with skeleton, and multiple textures, they are not optimized and just for test purposes, I’ll post them on plugin thread.
Not goin to clean code or soemthing like that for exposed reasons, waiting for different framewrok or openni update to make it modular as could be needed :) so, if there isn’t an ordered changelog or documents explaining how to use, please, don’t complain, a contributions is something to be shared and free, and my main activity/interest on that plugin, was coding it.
Hierro, thanks for that contribution, even not “clean” is very interesting!
hi rog,u’re welcome, that version is for demo(i do demo to myself yes); set background off to obtain better performances :)
Here the update sites/all/modules/general/pubdlcnt/pubdlcnt.php?file=/sites/default/files/uploads/OpenNIGenericNode.zip&nid=57736 :D
Link is not working
thank you :D