Originally posted by Gimbal Thanks.
So if I'm reading this right, when down sampling the 40Mpix picture to approximately 8Mpix it will beat the K-x with 0.2 EV in DR (IF you stay at base iso, otherwise it will not beat it). But if you stay at the original "screen"-resolution (after all, the high resolution is what this camera is all about) you will not beat the K-x. You will still have the high resolution, but the DR is nothing special, is that correctly understood?
Both, dynamic range and noise are properties which depend on spatial frequency (just like audio noise depends on frequency too). E.g., how would you define dynamic range for film which doesn't have pixels at all?
A spatial frequency-independent discussion of the properties is impossible (except in the internet, of course
).
--
Just to illustrate one detail: There is no need to downsample to have a high dynamic range at low spatial frequency. It is in the image anyway. If you printed the image, you would have slightly better shadow (black) detail but still
much better gray and white detail!
If you would draw SNR over spatial frequency, the 645D curve would be above K-x for all frequencies, with K-x having zero SNR at the high end (diue to lack of sensor sampling resolution).