Could you explain the difference between per pixel noise/DR and overall sensor noise/DR?
In other words if pixels performance was the same would the higher resolution MKII have better overall noise/DR?
First of all, caution with the term "per pixel".
A single pixel does not have any noise. One can talk about noise only relating to a group of pixels (some area of the image).
Example in the attachment: a smooth (even though it does not look like smooth), evenly lit, unicolored spot is selected from a raw image and the noise is measured as the standard deviation.
a. In case of the 40D, the red channel's average values are at the 7.52nd EV of the dynamic range (from saturation), and the noise is 12% (compared to the normalized output value).
b. With the 50D (a completely different image, but that has nothing to do with it), the selection average is at the DR 7.63 EV, only slightly more in the shadows than the above example, but the noise is 15%.
If these samples are accepted, then they show that the 50D's noise level @ ISO 200 is somewhat higher than that of the 40D, around this depth of shadows.
This is the pixel level noise, here measured on the non-demosaiced raw data.
The
sensor noise is IMO a muddy term. Who on earth cares for the "total sensor noise" or for any average of that?
As Ray and some others are understanding it (please correct me if I am wrong), they are trying to convert images of different sizes (i.e. of different pixel counts) into some common denominator and quantify the noise on the result. This approach ignores not only the problem of image deterrioration by resizing, but the fact, that reducing the image size simply
hides the noise, while hiding the details as well.
So, if you want to use your 5DMkII as a 10 Mpix camera, then yes, you can evaluate the result by resizing the image and comparing it to a Nikon D3.
My point was, that reducing the image size and sweeping the noise under the carpet does not increase the dynamic range, as truely lower noise would do.