And here's a couple of actual targetting pods:
In this one, the polygonal cover would be stationary, and the camera would move inside around it.
And even in those where it seems the dome has some turning capability of its own, the actual camera will still be some distance inside it and have some independant movement. (in, ie, doing active image stabilization of its own to reduce vibrations from the rest of the vehicle)
In fact, in targetting pods where the entire cover rotates as well, these would significantly reduce extreme movements by the smear.
Since in one with a static cover, if the vehicle turns a significant amount the smear would shoot away some direction as the camera makes a "big" movement inside the cover.
Whilst if both the cover and the camera track, then this wouldn't happen, but the dome's movement and camera's own stabilization inside it would still differ a bit, which would create some back and forth movement.
(until the vehicle turned enough that the dome runs out of motion range. Which ofc wouldn't happen if the pilot itself is trying to keep the smear within the pod's field of view)