This would wrap the frame and common metadata (eg framenumber and frame time) in a common structure. This would allow downstream operators to be built in a generalized way. All different camera types would then have a custom adapter class that converts CameraFooData into CameraData