Kevin Bjorke said:
I'm really curious about this, what sort of DR loss are we talking about (or is it just because of the digital pull-processing which means that you're losing exactly one stop from one end?)?
My technical analysis sofar (on my 1Ds2) indicates the opposite, namely an increase in overall dynamic range. I think the confusion stems from the fact that at double the exposure time (and less amplifier gain) we can capture 2 stops over medium gray before sensel saturation, compared to 3 stops over medium gray at ISO 100 and higher. However, the double exposure amount also improves shadow exposure, and its Signal to Noise ratio.
To summarize what I've found sofar:
Depending on the color channel, the read noise at ISO 50 is reduced by approx. 10%, caused by lower amplification, and the S/N ratio is improved (by approx. 30%) because of the double exposure level.
Both contribute to lower overall noise levels, assuming one takes care to 'expose to the right', but not clip the highlights. For those using incident light metering (or a gray card), there are not 3 but 2 stops headroom before the highlights clip at ISO 50.
There is 1/3rd of a stop additional headroom in all cases, but that will cause colorshifts in the highlights, due to non-linearity of the sensor response near saturation.
Bart