I may have missed others comments ,if so I apologize for repeat coverage
Jon wrote a while back :
>>>
If a 1/2" CCD has the same no. of pixels as a 1/3" CCD then each pixel will
be effectively "bigger" on the 1/2" CCD resulting in an image which is not
so "smooth". In many situations, if the no. of pixels is the same, it is
better to have the smaller CCD.
<<
=> For identical image print size/magnification, the blockiness,aliasing etc
is absolutely identical, but the larger sensor has some advantages: The
amount of light collected per pixel is greater in the larger CCD, so for
identical technology the signal to noise ratio is theoretically better but it
is a law of diminishing returns. If you double the area the signal to noise
ratio (and film speed) only improves by 1.41. In fact small sensors waste
more area per pixel in between the active sensing area than do large,so the
signal to noise ratio improvement can be greater than the theoretical
sqrt(2). The smaller sensors also suffer from light falloff somewhat like
vigneting caused by the fact the sensor is a 3d not a 2d structure so each
pixel sits at the bottom of a well. The well is relatively much deeper for
the smaller sensor causing added light drop off with wide angle lenses.
Faster lenses should be easier to make for small sensors but it is also a law
of diminishing returns since the manufacturing tolerances become impractical
if the sensor is too small. (Kodak has a long paper on this trade-off) Small
sensors suffer more from dust.
>>
Anyone care to comment ? I'm open to being persuaded from this argument !!
Also anyone got any info on how the sensors differ between TV and stills
cameras ?
<<
=>The video resolutions and readout rates are higher and this can compromise
A/D convertor performance,color bleed through from adjacent pixels and low
light performance.
In fact more and more digi still cameras now offer high speed burst mode. In
particular non-CCD designs using CMOS sensors can trade off speed versus
resolution so eventually video and still cameras will probably become one.
Technically the higher resolutions in still cameras makes data rates very
much higher. Digi cameras based on Nucore's chip set can achieve video data
rates and improved color seperation due to separation of the three electrical
color signals before digitizing unlike most older cameras. (see Nucore
website for more information)
>>
On a final point, the thing that really seperates "pro" TV cameras from
"non-pro" cameras is very often the number of sensors. "Pro" ones have the
light coming through the lens onto a prism which splits the light up into
the 3 colours Red, Green, & Blue, with a separate sensor for each. I know
one of the reasons this is better is only really relevant to capturing a
moving image rather than a still image, but does anyone who is particularly
in "the know" regarding digi stills cameras have any ideas whether this
would improve the quality of the image ?
<<
=>Foveon makes a pro level 25k$ cameras using beam splitters. (Check their
website.)
A beam splitter design achieves better s/n ratio by a factor of approx
sqrt(3). but the real advantage is reduced moire(aliasing) and also a large
reduction in the effects of light fallof with wide angle lenses. Apart from
cost and weight there are some manufacturing tolerance issues in CCD
alignment.with beamsplitters. An advantage of beam splitter designs is also
the color channels do not share the same gain so the dynamic range and bleed
through from color to color are largely eliminated. (Nucore's chip set
adresses this at least partially for single CCD designs).
Regards,
Tim Hughes
>>Hi100@xxxxxxx<<
< 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 >
|