Originally posted by BB_Zone28 That said, variations in DR between FF and cropped captures from the same sensor does seem both surprising and perplexing. Also, you may be on to something about how cropped and FF sensors/capture modes ultimately get evaluated for the benchmarks of the testing site.
What is the variable that people are missing is how we determine the floor at which the DR is limited and this is noise. Now you have to remember that noise is bandwidth dependent what this means is how closely we are going to view it.
The very same can be said about how much noise is visible as you increase the iso in your camera depending on how much we are going to zoom in as look at the photo will determine how much noise your are going to see.
When we are using only the cropped portion of the sensor your are basically zooming in more to makeup for what was cropped away. This will then appear as more noise just the same as iso 1600 on fullframe looks as noisy as 800 on cropped.
A very simple test than can be done is to take an image of a white piece of paper underexpose it by 10 stop and then lift the shadows by 10stop. Now take that image and print it in 8-10, then take the very same image and crop it by a factor of 1.5 and then print that also in a 8-10 and view them side by side.
This should also allow you to see that is more that the total amount of light that is captured ( via larger use of the sensor) that determines the amount of noise and DR you can capture.