UPDATE: After our article on Dynamic Range yesterday we received several requests to perform the tests in RAW mode on cameras which support it. We'd deliberately stayed away from RAW mode because it throws up lots of variables, however we have now re-shot the test chart and re-calculated the results for the Nikon D1 and Canon EOS-D30 in RAW mode. (Burnt four hours of my day ;)
It's worth understanding that the Dynamic Range we're measuring here is not the dynamic range of a SINGLE image but rather the difference between the TWO extremes (almost complete black and almost complete white) where the sensor system is about to fail, we want to measure the difference between these two extremes and do so with TWO images.
The best case scenario for every day shooting with default settings
These figures should be seen as a"best case scenario for every day shooting with default settings", you can expect anything within our results (potentially slightly better with less in-camera sharpening).
It's really a method for COMPARING different cameras, because we now have a standardised test we can fairly comfortably say "this camera exhibits 20% better dynamic range than this other camera". In the past all we could say was "looks like there's some more dynamic range here, maybe".
So again, the results below were calculated by shooting our test target (patent pending ;) at the two extremes of exposure (this actually consists of several shots which are analysed later and the two shots just above and just below sensor system failure are used). These two images are then analysed using our proprietary program which does several things, analyses noise levels and calculates the dynamic range of the sensor system based on the difference between the two extremes.
* Nikon Capture, No curve applied, Unsharp mask Intensity 10%, Halo 5%
** Canon TWAIN, All parameters "AsShot (Normal)"
(ISO Sensitivity on above graph is linear)