Pointvision.py and vortexstream.py

[swehcheeh]

Hey james!

hope all is good! i got the LEDs to work with vortex. i wanted to use pointvision.py and vortexstream.py to send OSC protocol to my realtime data but when i use record3d and run after compiling, it immediately crashes and closes the connection! i also tried just using the wirebug.obj as point cloud data to have both vortexstream and pointvision open them up. any ideas what im doing wrong? for the few seconds that both terminals are running and connected, nothing shows up on my displays!

would love to know how you sent real time data to your display if you could please share :)! thank you as always!

on my rpi ssh session:

and then also on my window terminal:

[AncientJames]

vortexstream.py's only function is to capture RGBD data from a connected iPhone using record3d, arrange it in a 128x128x64 voxel array, gzip it, and send it to another machine. pointvision.py's only function is to receive gzipped voxel buffers and copy them to the shared buffer for display by vortex. The protocol is stupidly lazy - it sends an 8 byte header consisting of 4 bytes of 0xff, followed by a 4 byte length for the following data, and then the binary data that has been gzipped. The address of the receiving device is hardcoded as vortex.local, and the voxel buffer is a fixed size. It's really the most minimal setup that let me stream live points.

If I understand correctly, you want stream mesh data using OSC, and I don't think they're a very useful starting point for that. Once you strip out the record3d handling, the raw voxel buffer and the streaming protocol, there's not really anything there.

[swehcheeh]

@AncientJames ahh got it! I was going to write my own anyway but those protocols close session immediately upon launch when I use my iPhone with lidar to try and record stuff from the record3d app so I was wondering if you had any idea what it could be! Other than that, it sounds like it might be because the sizing cap is being exceeded or am I totally off?

Just curious if you know if I would be able to simply put a 3D vocalized mesh that is pre-rendered, and alter it with osc protocol or if that would be too heavy on the rpi