On Sat, 16 Mar 2002 17:17:52 GMT, Loren Amelang wrote:

>No video camera I know of outputs raw pixels one-for-one,
>and most cameras with claimed pixel counts over about
>400K are using only a part of them for video - the rest are
>for stills only, or for stabilization. The second part of the
>equation is the software used to do the pixel-squishing, and
>you can't really specify that objectively, either.

You are technically right, but perhaps giving the
wrong impression - the "megapixel" one-chippers do use
more than 340K pixels for the motion-video imaging,
but, as you say, some are reserved for other purposes...
And, as you indicate, it is hard to isolate the
characteristics of one part of the "imaging device"
(lens, CCD, processing) from the other parts in these
integrated systems...

>The only concrete data you have to go by are the lens
>specifications. If you want telephoto, get it in the lens. But
>beware - 20X means the longest setting is 20 times the
>widest, and two cameras could be starting from different
>wide-angle capabilities. Try to get "35mm camera
>equivalent" numbers for the long and wide extremes, and
>compare those.

Good advice, though, surprisingly, most consumer video
cameras' zoom ranges "start" at the 35mm-equivalent
of 43-48mm in FL, a very small range of variability,
and high-quality lens converters allow one to modify
the zoom lens range at both ends...

>And after all that, even at 10X zoom your usable resolution
>will be limited by how steady you can hold the camera.

Ummm... "true", but also misleading, perhaps - the
resolution will be there, but the steadiness needed to
appreciate it may not be...;-)

>Optical stabilization should give you better results than
>digital processing, but I think you are in the
>"experimentation required" area here.

One should not make this assumption about OIS vs.
DIS - there are some fine DIS systems that can work
as well as OIS (or even better in some circumstances),
though sometimes with technical liabilities...