Page 1 of 1

Accurate sensor dynamic range measurements

Posted: Tue Nov 17, 2020 12:40 am
by sdaconsulting
I'm hearing a lot of discussion that with the R5 sensor, Canon's dynamic range has finally caught up to Sony. But that is not correct. Canon is using mandatory, baked-in noise reduction to get higher numbers in testing by blurring away noise. But you could of course do the same thing with Sony and modified-Sony sensors in Nikons, Fujis and Sony mirrorless cameras and get even higher DR numbers with less loss of detail. And the noise reduction algorithms possible in powerful computers without time constraints are far superior to the firmware noise reduction algos that must process an image on low powered processors in ~1/100th of a second.

I think people want Canon to have "arrived" therefore they have decided Canon has accomplished the feat of catching up with Sony. But the unavoidable noise reduction simply proves that they haven't.

Posted: Tue Nov 17, 2020 7:12 pm
by E.J. Peiker
If you want accurate, real world Dmax comparisons based on photographic dynamic range (S/N=20) rather than theoretical (S/N=1), which is what the manufacturers use, this is the place to go:
https://www.photonstophotos.net/Charts/PDR.htm

You can select different cameras and compare them. We had quite a bit of discussion on this in an earlier thread - go down a couple of posts for the beginning after clicking on this link which takes you to the right part of the thread:
viewtopic.php?f=57&t=289494&start=80

Posted: Thu Nov 19, 2020 10:28 am
by sdaconsulting
Pretty sure none of the DR charts, including Bill Claff's, can do anything to account for on-sensor RAW noise reduction except put an asterisk on the results.

Bill should do this for the Canon R5 and R6 cameras since they are filtering out noise which produces higher SNR measurements.