I can also visualize something rotating or in motion (or my minds eye in motion) for a few seconds. I often feel a little dizzy when I do this.
Can most people create a mental image and hold it for a long time?
My mind's eye isn't like a framebuffer that I render to. Trying to visualize a direct analog of a still image is difficult, feels unnatural, and requires concentration.
Moving scenes are easier, even if the objects are unchanged. It's easier still when the nature of the objects change too. I think it must exercise the brain's simulation and prediction hardware - it's easier to render and it has a smaller memory footprint. I can "look" at a sphere, but after only a short while it starts to change, move, or rotate as though my camera is changing position. It might spontaneously change color from "red" to "blue", or start to squish into a spheroid. Or change into a scoop of mint chocolate-chip ice cream. Or other elements might come into view.
The changes aren't random - it feels like the scene takes input from a stream of thought just on the edge of consciousness. Maintaining the integrity of the scenegraph - the state and location of the objects requires constant corrective input to return it to its desired configuration. This is a separate and distinct act from the concentration required to "view" the scene in greater detail.
The bigger the scenegraph, the harder it is to stop the motion. I can visualize what starts to feel dreamlike in its detail and complexity, but it becomes impossible to freeze the scene.
There's some work being done in this area under the name Image-Streaming: http://www.winwenger.com/imstream.htm