Re: 1um pixels, anyone?

Alfred Molon wrote:
In article <4412EE19.1040503@xxxxxxxxx>, username@xxxxxxxxx says...
No, (maximum signal)/(noise floor) = dynamic range. Noise in the
signal (the best possible) is the square root of the number of
collected electrons (converted photons).
[...]

Please explain why noise should be the square root of the collected electrons.

Photon arrival and detection is well-modeled as a Poisson process.
The true illumination is the proportional to the expected number
of events in the exposure time, and in the Poisson distribution
the standard deviation is the square root of the expected.

http://en.wikipedia.org/wiki/Poisson_process
http://en.wikipedia.org/wiki/Poisson_distribution


--
--Bryan
.

Relevant Pages

  • Re: 1um pixels, anyone?
    ... signal is the square root of the number of ... collected electrons. ... Note too that a read noise of only 5 electrons in a cheap ... Please explain why noise should be the square root of the collected ...
    (rec.photo.digital)
  • Re: shot noise poisson distribution and central limit theorem
    ... If arriving photons follow a Poisson ... So the mean number arriving goes with the square root ...
    (sci.physics)