Using photon noise to determine unity gain ISO, part 2

A Nikon D800E produces these results at a count of around 500:

The Sony RX-1 at a count of around 4000 looks like this:

And at a count of 500, it’s a little different:

The Leica M9 at Zone VI has a pronounced hump:

The Sony NEX-7 had anomalous behavior at ISO 6400:

We saw this kind of thing in previous noise testing on the NEX-7. I surmised then, and I still beleive, that it occurs because of some noise reduction that you can’t turn off that kicks in at ISO 6400. If we take that point out, we get this:

Again, less than a third of a stop variation.

I think I’m pretty much ready to declare victory. It does appear that there are some things going on with the cameras that are different from the model. I might look into that later.