Most of you are probably familiar with the concept of using a camera to pan and zoom around still images. Popularized by Ken Burns in his documentary on the Civil War, it is a great trick for any occasion when you don't have moving video for a scene. You can simulate this by simply animating the position of a still image in virtually any compositing or video editing program. However, there are a number of refinements that can make your life easier, and the end result more realistic.
How big of an image you need to use this technique? Big enough that you never have to scale it larger than 100% when performing a zoom - otherwise, you'll sacrifice some image quality. Figure what is the closest you want to zoom in on a still, and make sure you have enough pixels in that image area to more than fill your video frame.
If you are scanning photos, take the number of pixels you need to fill your frame (such as 480 pixels high for DV), and divide it by the size of the area you want to zoom in on (say, 1.5 inches). This tells you the minimum number of "dpi" (dots per inch) you will need to scan at (in this case, 480 ÷ 1.5 = 320 dpi). It is best to err on high side to give yourself a safety margin later.
(By the way, despite the belief of many editors, there is no "dpi" or "ppi" in video. Your only concern is in the final number of pixels; pixels or dots per inch are a translation only needed for scanning on the way in or printing to paper on the way out.)
You may be tempted to animate the Position and Scale properties of your image to perform your motion control move. However, life will be easier if you animate the image's Anchor Point instead of its Position. This is because scaling takes place around the Anchor Point, and if this anchor is not in the center of final video frame, the image will appear to slide around the screen when you change its Scale. Center the layer's Position in your final frame, and treat the Anchor Point as the crosshairs of your camera as you move it around the image.
Remember that real cameras have mass and inertia; this means they don't start and stop instantly. To perform a smooth move, use the After Effects Easy Ease Keyframe Assistants or their equivalent to add nice end points to your animation. Bend the Anchor Point's animation path to trace the path you want your imaginary camera to take as it moves around the image. These two techniques alone will lift your camera move above looking like a bad game of Pong. Also consider using a bit of interest, as shown in the sequence below.
To perform a camera move on a still image, animate its Anchor Point and Scale values to perform a pan and zoom, respectively. Image from Digital Vision/Beauty.
Follow the basic steps outlined above, and you will have a perfect motion control camera move. Although this works great when you're going for a calming or otherwise unobtrusive effect, it can look eerily unreal when mixed with real camera moves, or when the style you are trying to invoke - such as a documentary - might have a bit more of an edge. You need to reduce the precision of your overall move: not to the point of making the cameraperson seem drunk, but at least to the point of making them seem human.
The first trick is to change the ease in and ease out curves for the Anchor Point and Scale to be different from each other. This will make it seem like the cameraperson was hunting for the exact framing of the shot they want. For example, have the Anchor Point animation settle more quickly than the Scale animation does. To make the cameraperson seem even less perfect, bend the animation velocity curve so that it overshoots its target value for the second keyframe, and then backs up to the final value - like zooming in too far and then zooming back out ever so slightly. To alter these ease amounts in After Effects, open the Graph Editor and tug on the keyframe handles there, or Option+double-click on Mac (Alt+double-click on Windows) on the keyframe and edit its Influence values.