Hi,
Thanks for your observation. A few observations of my own:
1) The reasoning I made essentially ignores ISO, it is about capturing the best image and for that the sensor needs to be fully utilized.
2) It seems that MF backs have much lower quantum efficiency (QE) than DSLRs, so they need a lot more light.
3) So when I compare curves from DxO I only look at the maximum values ignoring the fact that MF curves are shifted to the left, because of low QE.
4) I actually think that significant progress may have been made on full well capacity on DSLR sensors. The figures from sensorgen point in that direction, but I'm not sure those figures are correct, partly because they give different FWC for same sensor sitting in different cameras.
Anyway:
Sensor | MP | FWC |
Nikon D3X | 24 | 48975 |
Sony Alpha 900 | 24 | 26843 | (note that the sensor is basically the same as in D3X) |
Nikon D600 | 24 | 76231 |
5) The fact that D800 and D600 are quite similar in 18% SNR is that they probably use the same technology. The D800 will capture the same amount of photons per surface area as the D600. FWC is probably proportional to pixel area. So the D800 simply gives better resolution at similar noise levels.
6) The comparison I made was between the Phase One IQ180 and as I said I compared noise at lowest ISO where FWC is utilized fully.
In this case what I see is
Camera/Back | ISO | SNR 18% |
Hasselblad H3DII | 39 | 45.3 |
Phase One IQ180 | 29 | 48.7 |
Nikon D800 | 73 | 46 |
I guess that the IQ180 has a sensor with better FWC than the one Hasselblad uses.
Another factor slightly favoring MFD may be that the sensor don't have OLP filters. The lack of OLP filtering is beneficial to sharpness, so MFD images need less sharpening. Sharpening tends to increase noise. The downside of not having OLP filtering is that it is easy to get fake detail. The D800E has removed OLP filtering, so it is a bit more similar to MFD.
Best regards
Erik
From the article
However practical tests don't seem to confirm this at least when it comes to grey card testing. (upper mid tones)
It's also interesting that there is virtually no difference between the d600 and the d800 despite the difference in pixel size
Here is a comparisson between the D4 and the D800. Same generation cameras.
One with significantly larger photosites. The noise is practically identical.
The big pixel theories don't hold up.....