Quote:
Originally Posted by Elio
That's true, I know, but we have to consider the whole chip area, I thought... in the same chip area smaller pixels are double than biggest one so the qe lost by one could be gained by next, I guess...right?
|
You are correct Elio in that the same chip area will collect the same signal irrespective of the pixel size, but in the scenario where one camera has pixels double the size of the other camera, the resultant image will then be viewed at half of the size of the other (assuming the same output display device / pixel size is used!). So, the same SNR (neglecting read noise / etc) of the object is being represented at half the output display size. So ... speed has been increased at the expense of detail which has been convolved (lost).
Quote:
Originally Posted by strongmanmike
Hmm?...ok, fair enough I guess  ( all else being equal) 
Mike |
"all else being equal" - yes of course camera read noise, QE, etc will have some impact, and the analogy in the old film days would be film speed, hypering, etc. The difference being that film was a consumable item in those days and when people used f-ratios as a benchmark to compare the relative speed it was understood that film speed also had an impact, as well as optical efficiency (eg. size of central obstruction, etc), etc. Nowadays the CCD is an expensive and integral part of most people's systems (certainly not a consumable like film!), so ... you are right in that I was musing about a simple way to represent the relative "speeds" of systems out there by taking the pixel size into account as well
I probably confused things though with my first clumsy attempt at adjusting the f-ratio (which was incorrect - too late at night for such musings
).
OK, probably enough musing for a Sunday night
