Having some odd camera translation behavior I was hoping someone could help me figure out. When the sensor_fit is in auto mode (or the viewport perspective camera in rendered view) I can set the horizontal film size to the sensor_width, set the aspect ratio of the camera image to the aspect ratio of the render resolution (or 3D window) and everything matches up. But if the aspect ratio goes under 1, all of a sudden this method doesn’t work, and the rendered image looks significantly different from the viewport.
It only happens in auto sensor fit or the rendered view. If I change the sensor fit to horizontal or vertical then everything still matches no matter what the aspect ratio is (at least for final rendering).
This wackiness even affects orthographic mode, which isn’t even reliant on the sensor size.
I can see how Lux and Pro Render deal with this issue but their solutions don’t appear to be working for us so I’d like to understand what the camera is doing in this situation.