kinect point cloud

hello everybody,

i'm trying to generate a point cloud using max and jit.freenect.grab, as seen in many kinect examples round the web (although most of them seem to be done in c or processing).
with the help of the internet and this forum i achieved the first step by feeding the depth matrix as z plane onto a jit.mesh.

now i'm searching for advice on distortion of the kinect image.

this is what i got so far -
(not much)

first of all:
a very dumb (geometrical) question:
how big does the mesh have to be in x and y dimesions in order to match the centimeter output of the jit.freenect.grab object?

and furthermore:
i found some instructions on the web for correcting camera distortion of the kinect - all of them in c code, and don't really understand all this - has anybody done this already? if so, please help, i would be happy for any advice!

i only have my kinect for 3 days now, and did some videotracking before, but never worked with 3d geometries in jitter..

kind regards