I've created a dense forest of plane image trees using Distribute Points on Faces
and Instance on Points
nodes.
I'm attempting to billboard them to always face the camera.
Plan A was to use object constraints on the original objects. Failed because the instances would not be constrained individually, just mimicking the original asset's rotation.
My current Plan B is to take the camera's position as a vector, subtract it from the position of the instances and use the Align Euler to Vector
node to make each instance directly face the camera.
But two issues are in the way of plan B.
1: I don't know if it's possible to use Geometry Nodes to apply this rotation to all instances individually.
2: I don't know how to access the location of these instances en masse to test if it would work.
I have seen that particle systems would be good to make this work easier on a flat plane, but this is not working on a flat plane.
Position
toSubstract
? Then you would have the direction vector between the camera and the individual instances. ...just a guess. $\endgroup$