🚀 go-pugleaf

>The first question deals with auto-focus.  It used to be that cam-corders
>that had auto-focus also had a lens that was visibly turning any time
>the distance to the focal object changed.  Now, with the new and even
>smaller cameras, you often don't see this.  Is the lens just smaller and
>internal or is some non-mechanical means used to focus. The manual to
>my new Sony cam-corder says something to the effect that the camera
>focuses by maximizing the difference in light intensity between adjacent
>pixels.  I guess this makes sense - a blury picuture would have the
>light more dispursed between pixels, and a sharp picture would have
>higher contrast between pixels.  If a non-mechanical method is used to
>focus, does anyone have any more details?

Maximizing contrast is how it *decides* when focus is best; it still (as far
as I know) needs to mechanically move lens elements to change focus. 
The moving element is probably just not the outer surface, as you suspected.

>The second CCD camera question deals with the CCD's response to IR.
>What is the response curve of a CCD camera?  How quickly does the CCD's
>sensitivity taper off as the frequency of the light drops below visible
>and into IR?   And how come I can see the IR on my monitor/view-finder?
>(try pointing your TV remote at the cam-corder).  My potential application
>is a pseudo-night-vision device.

It depends on the CCD used, and the optical filters in front of it that
make specific CCD elements sensitive to red, green, or blue. But as you've
observed, they do see some IR. Your monitor displays this signal as
visible image (TV remote IR LED's are actually pretty bright).

I have specs here for a Kodak mono CCD chip; the quantum efficiency (sensitivity)
peaks at 27% around 400nm (bluish), and drops off more or less linearly for
higher wavelengths: 16% at 550nm (green), 10% at 650 (red), but still at 3%
at 850nm (IR). The color version of the same chip has filters that peak around
425, 550, and 650nm; the blue and green ones drop below 1% +/-100nm of center,
but the red filter looks more like a high-pass.

As far as night-vision, all you can do is experiment. Find a bright non-visible
IR light source (no, LED's aren't *that* bright!) and see what the camcorder
can see. Good luck!
----
"How am I typing? Call 1-818-354-6508"               jack@robotics.jpl.nasa.gov
Jack Morrison/Jet Propulsion Lab/MS107-102 4800 Oak Grove Dr, Pasadena CA 91109