12-11-2011, 01:31 PM
Hi, Hope you don't mind me commenting, but DxO shows also measurement of full SNR. Here one can see the SNR for all possible levels of exposure, at all iso settings. Out of this the dynamic range for each sensitivity with any quality level one like, can be found. Apart from that there is also a graph showing dynamic range as a function of iso. Here one can clearly see how Canon is not so good at base iso but improves as the dynamic range goes up.
Sorry if I have repeated the obvious, but sometimes it seems as if only the final scores are looked at. For me the charm of DxO is that sensor behavior is very well measured and recorded. A shame they do not have any indication of the type of noise. The averaging they do for normalization (print) is only valid is the noise is really random. Things like banding do not gain so much by down sampling to a lower resolution and for that reason is more of a problem.
Sorry if I have repeated the obvious, but sometimes it seems as if only the final scores are looked at. For me the charm of DxO is that sensor behavior is very well measured and recorded. A shame they do not have any indication of the type of noise. The averaging they do for normalization (print) is only valid is the noise is really random. Things like banding do not gain so much by down sampling to a lower resolution and for that reason is more of a problem.