I am trying to do some abstractions around the pdp_freenect object, to process the depth information and turn it into a 3D scene.

Have played a bit with [pdp2gem] in combination with [imageVert] but this stuff is not really what I want.

First of all the [imageVert] is not able to be textured, or otherwise manipulated, beside [translate] and [scale] etc
Second, it seems to not be really what I want.

I am not clear about the route I specifically want to take, but I feel I want to do something for each pixel in the 640x480 output.

Been looking at the numerous objects to get data out of pix/pdp surfaces, but I look up to the task of running with 640*480 [pix_data] objects hooked to [cube] for example.
Even lowering the resolution to 160*120 makes it a lot of objects ..

[pix_dump] looked fascinating, but crashed pd on the smallest images I feed it ..

There must be some better way that I'm missing?
Please advice me into a proper direction

you forget the link to the actual project, http://www.ludique.cl/labs/simple-kinect-touch/

And a screenshot

I'm so happy this works!

http://www.pdpatchrepo.info/hurleur/check_it_out_with_kinect.png

Oh seems like [repeat] and [separator] are really all that is needed

This is sorta my final post here to this thread, unless something says something cools

http://www.pdpatchrepo.info/hurleur/pix_data_test4.pd

Well, GLSL is still pretty much over my head, and before advancing to the next abstraction I would still like to try to get this working.

Attached my latest attempt (still renders just one cube )

http://www.pdpatchrepo.info/hurleur/pix_data_test2.zip

did you install libfreenect before compiling the newest pidip external?

I grabbed libfreenect from https://github.com/OpenKinect/libfreenect.git

Well, some little productive night here.

Now I'm using pix_video to feed 40*30 [cube] objects into a GEM scene through [pix_data] and [translateXYZ] which is sorta working.

Besides the fact that I'm only seeing one cube

http://www.pdpatchrepo.info/hurleur/pix_data_test1.pd

btw, it looks pretty cool already, but not cool enough

http://www.pdpatchrepo.info/hurleur/kinect-imageVert.png

check out this example we talked about on irc.

in it, i mix together three videos to seperate [pix_texture] and [rectangle] surfaces, and capture it all into a [gemframebuffer] which goes to a fourth [pix_texture] with a [rectangle] that actually gets displayed on the [gemwin]

http://www.pdpatchrepo.info/hurleur/GEMtriple_video_mixer.pd

sorry manut, Windows CE laptops are not the best target for running PD
Better grab a real x86 netbook