I wish to generate synthetic data for analysis of structure from motion scene reconstruction for a rig of cameras that look like that:
As you can see, my two cameras should act like if there were a mirror modifier that uses the rig as origin, but as it's impossible to put modifier on cameras.
I have tried with constraints, but it's hard to get real mirror, it usually gimball lock as it doesn't work like I wish in local space, I don't think using drivers would be any better in this regard (I don't master quaternions algorithm for mirror)
I wanted to know if it was possible using geometry node, I've seen that everything node had started to land on recent versions of blender, I still only see geometry node and no object-node as of now, I wonder if such thing is yet possible without python scripting or not?
Thank you.