196

**EOS Bodies / Re: Poll: Most wanted new features for 5D Mark IV**

« **on:**December 23, 2014, 09:09:45 AM »

Interesting that two competing features are chosen, both by about 50% of the respondents. These are higher megapixels and the other is better low light noise.

Those don't compete.

For fixed sensor dimension, higher MP means smaller photo site size which leads to lower signal to noise.

No, it doesn't. Not for the same sensor area.

I like to use this analogy.

Which way do you get more pizza, when you cut the 15" pizza into 8 slices or 12?

A photo site that is larger will have more signal. For equivalent electronic noise, which is doubtful since the weaker signal will need more, statistic very clearly tell us that the S/N in the larger site will be lower. The noise from the signal (number of photons reaching the site) is governed by the Poisson distribution and the variance is equal to the signal. Noise is typically reports as the std dev which is the square root of the variance. For if the signal is 10,000 the variance is 10,00 and the std dev is 100. S/N is 10,000 / 100 = 100. If the area is 1/4 the size the signal will be ~ 2,500 with a variance of 2,500 and a std dev of 50. S/N is 2,500 / 50 = 50.

Which S/N would you rather have, 100 or 50?

That's not the proper comparison.

This is.

Which would you rather have, one pixel with a SnR or 100 or four pixels each with an SnR of 50?