On Sun, 16 Mar 2003 20:19:13 GMT, "David McCall"
<davidmccall@attbi.com> wrote:
>"ralford" <ralford@bigfoot.com> wrote in
message news:1t4da.29200$r26.15459@news1.central.cox.net...
>> Thought I would refresh this thread since it is
becoming a bit unfocused -
>> unusual for usnet ;)
>>
>> Let me attempt to clarify my understanding as
stimulated by Saman's comment.
>>
>> Give the same incoming light, lens diameter, focal
length, f-stop etc, the
>> incoming light is in an angular cone that is then
bent (focused) into a cone
>> that focuses the incoming light over the ccd-plane
(the film plane in
>> analog).
The amount that the incoming cone is bent is dependant upon the
>> focal length, "wide angle" lens have
wider cones. They are all bent into
>> the same cone from the lens to the ccd-plane. We all know this if only
>> intuitively.
This also leads to the depth of field effects, but that is
>> another topic.
>>
>> The interesting point is when the ccd-plane changes
size. A larger ccd
>> would be appropriately higher in the cone (further
from the apex), and a
>> miniscule ccd would be placed at the apex of the
cone. The amount light in
>> any cross-section of the cone (normal to the
central axis) is the same. If
>> ccd's with the same pixel count but varying size
and resulting pixel size
>> are place at the appropriate location in the cone
the light on each
>> individual pixel is the same. In this case, one would expect to find the
>> same total light for a pixel.
>>
>> All of this is of more academic interest than
practical interest. The fact
>> is that larger front elements (i.e the 52mm on the
VX2000 vs the 30mm on the
>> TRV25), the more efficient lenses (f 1.5? v f
2.5?), and the confusion
>> concerning how the pixels are actually
counted: Is R, G, B recorded at each
>> pixel?, how much is used for numerical
stabilization for the single, not to
>> mention how the sensitivity of the three individual
CCD on a 3-CCD can vary
>> from the sensitivity of the cheaper models, etc.
>>
>> The fly in the ointment, thanks to Saman, is that
the light varies along the
>> cone, offsetting the changing size of the CCD
placed on the appropriate
>> location in the cone. The real answer to improved low light capabilities of
>> larger ccd's is probably the "non-linear"
effects that can be controlled
>> with larger and more expensive ccd construction.
>>
>> Cheers - and qualify my comments knowing I have
just return from a bike ride
>> and had a beer!
>No position along the "cone" matters except at
the focal plane. If I understand
>what you are calling a "cone". The CCD will
always be placed at this point no
>matter the size of a CCD. I think what is being missed
here is that smaller chips
>wind up getting shorter optics. If you had a
"normal" lens (20mm??) on a camera
>with a 2/3" CCD, and then removed the 2/3" CCD
and replaced it with a 1/5" CCD
>the shot would
go from a "normal shot" to a "close-up". It would be as
though you
>had zoomed in, but without changing the lens. That is
why they can get away with
>such tiny optics.
I think "RA's" intuitive lens concept is wrong -
and
as I pointed out in the "unfocused" thread, once
all
the reasonable specifications of lighting and lens
characteristics are made, for "apples and apples"
comparisons, the lens and subject lighting can be
dropped from "the equation", and a single even
diffuse source could be used to illuminate the CCDs
for comparison of sensitivity - and then it is obvious
that the larger area of the equal-numbered sensors
of the larger CCD will give it greater sensitivity...;-)
>At the rate they are going with chip technology, you may
soon se consumer cameras
>with nothing but digital zoom. I could see a gigapixel
CCD with a wide angle lens,
>and you just zoom (or pan for distortion control :-)
around the chip(s) to change the shot.
It's called "digital zoom", available now for the
tele end (should one really want to use it! ;-),
but a nifty concept for full-range zooming,
perspective "correction", and format
proportion-shifting (oh, wait - that's available
now, too! ;-).