I know the trace recorder records the camera with camera position, focal point, view up and scale. Would it be possible, and would it be clearer, if we gave users the current roll/ putch/ yaw/ azimuth/ altitude in a trace?
An example of what we have is:
renderView1.CameraPosition = [8.578210035538005, 49.900678928178365, -27.47846414137043]
renderView1.CameraFocalPoint = [0.21706008911132804, 4.0, -5.110947132110596]
renderView1.CameraViewUp = [-0.3150182020483825, -0.3687693998695668, -0.8745128141417048]
renderView1.CameraParallelScale = 13.391445890217907
An example of what I want is:
renderView1.CameraFocalPoint = center of view
renderView1.CameraAzimuth = [123.4567]
renderView1.CameraAltitude = [-15]
It’s true that the currently recorded camera view up vector is not necessarily intuitive, and that the degree of roll/pitch/yaw may be easier to understand. However, if we did report that, it would have to be with respect to a canonical camera position (how many degrees from the original camera position is the roll, for example), and those angles might change depending on the order you record them. That’s all fine, but my point is there is some implicit info you need to make the rotations meaningful.
Note that you can specify rotations by hand currently with
So if you initialize the camera to be centered and look down -Z, you can in your script achieve these rotations. I’m not entirely sold on the idea that that is how we should record them, however.