atk@xxxxxxxxxx writes:
<< According to these calculations an 8meg pixel camera should be
on par with typical high (but not optimal) 35mm system. I thought the number
was much higher -- what is wrong with my calculations ?
>>
Alan,
this is a topic that has been discussed a fair bit on the list and
you could probably profitably check the archives.
Very briefly:
1) the digital camera vendors specify the pixel count for the (RGB) three
color sensor pixels, used so that for an effective full color pixel the pixel
count is reduced by a factor of 3. So a "3Megapixel" camera actually is at
most a 1Megapixel resolution. In reality it is less than this for a number of
reasons some of which are discussed in 2. They also try to reduce artifacts
by interpolating pixels by averaging using nearby pixels of the same color
this generally limits the resolution further.
2) The two samples per line of resolution is a theoretical limit ("the
Nyquist limit") for sampling. You actually need to sample more pixels to
resolve a line pair.
Think of this: if your sensor pixels in the "film plane" exactly line up with
the lines you are imaging you get the correct lines resolved. *But* if you
offset the sensor by 1/2 a pixel you get the average of half a black and half
a white line or a medium grey - no image at all! Further if the pixels are at
just a slightly different pitch than the image you get a picture with nothing
but very wide dark and lighter bands. This is termed aliasing or moire. With
seperate color pixels this can lead to even more strange artifacts. see below
urls for an example
Bottom line you need a fair number more pixels than you might at first think.
Check the Foveon Website for some insight on color sampling issues and
moire/aliasing:
http://www.foveon.net/interp.html
http://www.foveon.net/moire.html
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 >
|