f-number equivalent between m43 and FF

Started Mar 25, 2014 | Discussions thread
Signal = Amplitude of the Light Waves Electric Fields
1

The signal is always what we want/need to know. What we need to know is one or more parameter values estimated by the mathematical model applied to the data collected by the measurement apparatus.

In digital photography we need to know only one thing, the electrical charge stored in a capacitor for each sensor site. The amount of electrical charge is the signal.

Forget about photons. Photons can not explain diffraction. Waves are a better model because you don't have to switch between models to describe all the parameters/phenomena relevant to photography.

Light waves have three components that are perfectly oriented to each other by 90 degrees. One is a propagation vector. The second is a magnetic field. The third is an electrical field. The lens changes the propagation vectors' paths. The sensor sites selectively converts the electrical field vectors into electrical charge (the magnetic filed is ignored). The raw-file data contains numerical estimates for the light waves' electric-field amplitudes available at each sensor site when the shutter was open.

Light Waves look at the second graphic from the top on the right.

The light-wave electric-field amplitude is a state of nature. Our estimate of the true, but unknown amplitude contains errors (uncertainties) from at least two sources. One is shot noise that is inherent to the process of converting light energy to electrical charge. The other is read noise which is the error in the amplitude due to thermal fluctuations in the electronics. There are other sources that increase the estimates' uncertainty as well.

Everything in the data, besides the light waves' electrical-field amplitude, after the waves leave the sensor sites' micro lenses is either noise (psuedo-random fluctuations) or artifacts (non-psuedo-random errors). These factors increase the uncertainties in the electric-field amplitudes' estimates .

The read noise has structure (white, red, pink, etc., etc.). In fact modeling the noise itself is commonly used to improve signal parameter estimates in astronomy because the signal-to-noise ratios are oppressively low. Knowledge about the noise improves the probability for the signal parameter estimate. In photography the SNR is sufficient so people have not spent time modeling the noise. And, the calculations require parallel computing architecture to complete in a reasonable amount of time.

Complain
Post ()
Keyboard shortcuts: