Darned good point, Iliah, once more you pop up with a "goody"

.
I hadn't even thought about that one. For all we know the "noise"
we are seeing is an artifact of what is opening the NEF.
Towards the shadows, the range information occupies in histogram is
more and more compressed, so the quality of post-processing
algorithms like de-mosaicing, responce curves, color space
conversion becomes here more important IMHO.
As far as I understand from experimenting with implementing RAW
decoding myself, those three are actually responsible for abnormal
noise levels in many cases, and what we see is not noise of the
sensor, but posterisation.
Wouldn't it be a hoot if the delay is getting Nikon View and Capture in
shape?
Improvement of RAW rendering quality between NC 1 and recent
version NC 3.5x is IMHO to a big part due to handling color space
conversion between the sensor color space and editing color space.
Same sensors, same firmware - but what a difference just because of
Nikon Capture/View processing algorithms improved, whatever it was
the improvement (or, rather, improvements).