As an analog CMOS circuit designer, the only way to get lower noise is
larger area. A larger sensor with the same number of pixels will have
bigger pixel cells, allowing each cell to capture more photons and will
always be lower noise than the equivalent smaller sensor. The only way
I can see that smaller could be better is that smaller may be faster
(lower capacitance) and perhaps allow it to multi-sample and integrate
out noise, or possibly lower power which reduces thermal noise.
Everyone is preoccupied with pixel count these days, but what about the
z-axis? How many pixels does it take to make an image? Early 35mm
photographers made great photos with very grainy low res film. My
co-worker is using a casio digital watch with 120x120 pixels. Not for
everyone, sure, but it has a whopping 2-4 month battery life! Film is
scaleable, need more, get a 4x5.
The limit to using existing SLR lenses has to do with the bucket (pixel
cell) depth and the angle of incidence at the edges with wider angle
lenses. (telescopes are generally not wide angle instruments.) I've
also noticed that the minimum aperture opening is around f11, is this
due to the need to have a wide enough opening to maintain this near
perpendicular angle of incidence to the sensor? For example, even a
pinhole lens will be blurry with an imaging device that has pixel
depth. Could someone make an optical adapter to parallel up light rays,
similar to retro-focus in existing lenses? This adapter, if built into
the body, could also double as seal for the sensor from dust.
Wayne
At 07:52 PM 12/19/2001 -0800, you wrote:
>...One thing though
>that kinda worries me is that with a smaller CCD there
>will be a level where they just can fit any more
>megapixels onto it right?
>
>Mark Lloyd
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|