I think it works something like this (very simplified) :
Each photosite on the sensor is a little bucket that collects photons. The photons interact with the photosite and generate electrons in proportion to the number of photons. These are shipped off to a ADC converter chip and 'counted'. The result is a numerical value representing the brightness of that photosite recorded in the raw file.
Each photosite has a maximum capacity - if they collect more photons than this capacity, they overflow - this causes clipping or blowout of the highlights. The photosite well capacity puts a limit on how bright a signal can be recorded for any given exposure setting.
The photosites also produce a low level of random signals (noise) even in the complete absence of incoming photons and this can cause the ADC converter to record spurious output.
Fortunately, the noise level is at a reasonably constant level irrespective of the number of photons in the photosite well, so, as long as the photosite captures a large enough number of photons to to drown out the errors, the false signals don't matter too much.
This requirement to capture enough photons in each photosite to overwhelm the random errors means there is however a limit on the sensitivity of the sensor - if too few photons are captured the signal can be overwhelmed by spurious noise.
If we put these two limits together we have:
1. A need to record enough photons to get an acceptable noise level
2. The need to avoid collecting too many thus causing blowout of highlights
To handle both these means defining the base ISO sensitivity of the sensor such that a correct exposure for 'average' scenes records sufficient photons from the shadow areas to overcome noise and (simultaneously) the photons collected from the scene highlights don't exceed the well capacity of the photosites.
A designer could rate the sensor at a lower ISO - this would mean longer exposures which would reduce shadow noise but it would also mean a higher chance of filling the wells and blowing the highlights.
The designer could rate the sensor at a higher ISO and reduce exposure accordingly - this would protect the highlights but mean increased shadow noise as too few photons are collected from these areas.
Corrections to the above welcome!
Why does the Nikon dSLR cameras "only" go down to 200?
Back in high school when I was taking photo workshops, my
instructor loved for me to use Kodachrome ISO 64 slide film. He
said that ISO 64 was even richer than ISO 100.
Many of the "point and shoot" cams even go down to ISO 50, but the
Nikon dSLR cams only go down to 200.
Should this be an issue with me when looking at cameras? I will be
taking general everyday photos of kids, etc., but I do really want
the flexibility of a dSLR.
Chris (newbie)
