On Wed, Jan 04, 2006 at 01:35:50PM -0500, Bob Shell wrote:
>
> On Jan 4, 2006, at 9:11 AM, Steve Jolly wrote:
>
> >A 1mm shift of the camera will only produce a 1mm shift of the
> >image on the sensor if the lens is at 1:1 magnification. That's
> >not true in the general case. Lenses magnify. Different focal
> >lengths magnify by different amounts. This applies to image shifts
> >just as it does to the images themselves.
> >
>
> I still do not think this is correct. A point projected on the
> sensor by any lens will move 1mm if the camera body moves 1mm.
> (Equivalent to moving the sensor 1mm).
>
> Bob
No, Bob - it won't, if you consider rotation rather than translation.
Here's the thought experiment.
Consider the camera set up (on a tripod) with a 100mm lens, focussed on
a brick wall a little distance away.
Now apply a small rotation (pan) to the camera, just enough to move the
centre of the image by one brick. The sensor needs to move by the size
of the image of a brick to compensate.
Replace the 100mm lens with a 50mm lens, and repeat the experiment. The
angular rotation to shift the center of the image remains the same, but
the amount of sensor movement needed to compensate is reduced by half.
