I would have thought much of the h3dii-39 noise would be averaged out due to the exposure stacking?
It is - compared to a single frame in the stack. But it still cannot possibly catch up to the signal to noise levels of the H4D-40 sensor which (1) has 5x lower dark noise per second, (2) collects twice as much signal per second, and (3) takes the photo in one readout rather than 4 readouts.
Even if you dropped the 3rd thing, and just got both cameras to take a single 1-minute exposure, the H4D-40 sensor is way ahead in the shadows. Getting the noisier H3Dii-39 to take multiple exposures/readouts, while the H4D-40 does it in a single readout, puts the H3Dii-39 even further behind.
Doug agrees with me - a single long exposure is preferable than a stack of short ones. He doesn't address the S/N theory behind this - but I do and I hope that I have explained it adequately. If it's not clear, ask me for more detail.
Are you saying the h3dii-39 image quality would be unusable for exposures up to 1 minute?
I have no idea if it is unusable. That's a judgement call. It should be ok. I AM saying that for a given total long exposure time, subdivided/stacked or not, it is inferior to the the H4D-40.
Will a donation of $100 million eliminate famine in Darfur? I have no idea. Will a donation of $400 million eliminate it? I still have no idea, but I CAN say that it will go further and feed more people for longer than the $100 million will. (Just trying to get away from the ubiquitous medium-format-as-sports-car analogy here!)
When I tested stacking exposures on my 5D MKII it virtually eliminated the noise. I gave it about 5 to 10 seconds to cool after each exposure. I think there where 14 exposures ranging from 15 seconds to 30 seconds. I also found it considerably easier to meter the scene because changing light levels weren't an issue. Trying to get a 5 minute exposure right can be a total guessing game when the light levels are dropping.
If you are struggling to guesstimate long exposure times because of constantly changing light, then that changes things. No point in having a theoretically lower noise, single, very long exposure, if it turns out saturated because you overcooked the exposure time. In that case, do go for stacking relatively shorter exposures. But again, doing so with a lower noise/higher sensitivity camera is preferable. The H4D-40 beats the H3Dii-39.
I guess what I really need to work out from all this is whether or not the h3dii-39 will be able to produce 30" x 30" prints for me of a good quality, better than my 5D MKII that is, from stacked exposure lasting from 2 to 8 minutes.
Otherwise I may need to bite the bullet and get the h4d-40.
In fact, your 5DII beats both of Hasselblads easily, since Canon CMOS sensors have almost negligible dark noise, and much lower readout noise as well once you pass ISO 200 (I wouldn't use less than ISO 400 on my 5DII - I see you used ISO 100 in your shots above - unless you really need those top 2 stops of DR, that is a needlessly noisy setting to shoot at). But I guess your main concern is not with ultimate S/N but rather ultimate, medium-format levels of detail, and medium-format levels of gorgeous Kodak CCD colour. Then try the H4D-40.
Ray