One main issue I am noticing now is that the depth information comes into TD in clear distinct layers which is not what we would generally want from depth information. Luxonis tells me this is called full-pixel layering.
In the TouchDesigner OAK examples, in the OAK oakStereoDepthConfig Base, if I enable subpixel fractional bits and the initialize the camera I do not see any difference in the resulting point cloud.
we have to rework these examples.
Until then, can you give this a try and see if it leads to better results for you? It’s much more directly mirroring the code used in the Luxonis examples.
I have uses for an AI camera like the Oak, but right now im more concerned with good depth info.
Need to find a good ToF cam that can work with mac or switch to a windows laptop