Quote: "On the basis of these tests I would judge the A7s to have almost a three stop advantage over the A7r between ISO 6,400 (which is as high as I would use on the A7r) and ISO 51,200 where the A7s reaches its useful limit (for my type of shooting)."
A7r has about 50% quantum efficiency and the sensors are of similar size.
First of all, image quality is subjective
That being out of the way, when I saw Den Lennie's Scotland shooting, I spontaneously said to my partner, the A7S will have somewhat usable (read "printable") files up to ISO 50K. Around ISO 25K they are probably a no brainer. Of course that was an out of my guts first impression.
As for dynamic range and pixel size, shooting at higher ISO, it is obvious that large pixel cameras have significantly better dynamic range than small pixel cameras. Now that is not out of my guts, but a hands on observation I can make any time.
I learned over the years that theoretical sensor performance and hands on exposure/print sometimes differ. Remember the Canon G10 causing a stir? How could it be that a G10 produces nearly identical quality on an A3 print of certain
(say "intimate landscapes") pictures compared to a high spec Hasselblad? Impossible....not!
Just a few days ago, I read on a forum the opinions that Sony's new developed curved sensor will never make it into a camera, and these opinions were flanked with a lot of theoretical background, some factual, some mixing up terms and conditions, and some claiming underlying physics prohibit the use in a camera.
My gut tells me we will see this sensor in a camera in the next 12 month.
Now I quickly leave this field of "thermal noise from dark currents" and shoot some pictures.