This is one in a series of posts on the Nikon Z7. You should be able to find all the posts about that camera in the Category List on the right sidebar, below the Articles widget. There’s a drop-down menu there that you can use to get to all the posts in this series; just look for “Nikon Z6/7”.
I have read reports of banding occurring even in dark-field images made with the Z7. I happen to have a bunch of those, so I went looking. I took dark-field images made at all the ISOs between 64 and 25600 inclusive, and averaged all the columns in each row. Then I plotted some histograms.
As you can see, plane 4, the blue plane, has a double peak, as does plane 2. The larger peak looks like the peaks in the other channels. The smaller one has a lower mean value. I calculated the skewness of each of the planes at each of the ISOs. Here’s a graph of the skewness versus ISO setting for the blue and Gr planes:
Negative skewness means the long tail is on the left. I’m going to discuss it in terms of the absolute value. The skewness goes down with increasing ISO as the second, lower peak is gradually drowned out in the read noise. The reason the skewness increases again at ISO 400 is that the read noise drops there because the conversion gain increases.
This is, to me, indicative of a firmware error on Nikon’s part. It’s OK for the standard deviation of the PDAF rows to drop because on interpolation, but the mean values should not.
If we look at the average values for the dark-field images in the PDAF rows and in those rows that have no PDAF pixels, we see this:
It is interesting to me that the offset doesn’t change much with ISO, but the visual effect does because the greater noise at higher ISOs masks the shift.