John Sheehy wrote in post #17950815
With JPEGs of "normal" contrast setting, and extreme white-balances like those for incandescent, there can be 3 stops of blue in the RAW, clipped in the JPEG.
I was aware that the histogram is based on the JPEG data; thus with an extreme(ly contrasty) picture profile you could really be fooling yourself. I wasn't aware that under a normal/sensible circumstance you could get that level of misleading data though.
Seems bizarre that a RAW histogram isn't offered as standard. In fact, it would also be useful to have it indicate not just clipped areas, but areas that would fall below some accepted definition of a minimum signal to noise ratio.
Even better, it should surely be possible to code up an auto-bracket option; that would take a series of shots, whereby the darkest exposure held for some detail in your brightest highlight, then the remaining exposures ensured you got something down to a configured number of darker stops. Lots of "ifs" and "buts" there granted, but there would be some practical solution with a bit of thought.
thenextguy wrote in post #17950878
Wouldn't it be something if Canon finally took care of low ISO DR and then they got criticized for worse high ISO?
Well, we've got to have something to whinge about haven't we?