K, let me try and make that a bit more clear. 1) You evaluate the overall 3D bounding box of the container having all your objects. 2) From this box you extract 8 3D points corresponding to the corners of the bounding box. 3) You project these corners to screen coordinates. 4) Of all the projected coordinates, you extract the max Y and min Y values, and then calculate the distance between them, the yRange. 5) Now calculate a ratio, something like yRange/stage.stageHeight. 6) Use the ratio to calculate how much to move the camera from or towards the target, something like targetCameraDistanceToObject = 1.5*currentCameraDistanceToObject*ratio.
Note that the apparently arbitrary 1.5 value has to do with the camera's zoom and focus values. There might be a simpler way, but this worked like a charm for me. hth!
