According to the "measurements" and results graphics:
- a Sony A7R3 should provide the exact same dynamic range at ISO 100 as a K-1 II.
No, it should provide the same 'photographic' dynamic range (PDR), which it does. PDR uses a higher threshold for 'acceptable noise', and if you look at the less dark patches (remember, this is already after a 6 EV push), like the third grey patch from the right, the cameras are essentially even in performance. Any advantage to the K-1 II in this patch is either because of additional 'baseline' NR performed by ACR (ACR performs varying levels of NR from camera to camera, even when turned off), or additional Raw filtering done by Pentax.
(If you're wondering why I might even suspect / suggest the K-1 II might be doing some low level Raw filtering at ISO 100, it's because
if you compare it to the K-1, the K-1 has a more random noise pattern, while the K-1 II has vertical / horizontal 'rice grain' patterns reminiscent of noise reduction).
- there is mentioning of loss of details for the K-1 II but not for the Sony A7R3
You're pointing to a section of the studio scene we know is rendered softly due to the FE 85/1.8 lens. Look more centrally, and the detail in the a7R III is comparable and/or higher contrast.
But then we are not primarily theoreticians, but photographers, so the
proof lies in the image:
https://www.dpreview.com/reviews/im...8&x=-0.6873051525279211&y=0.11229493274677838
Raising shadows is the main usage for a lot of sensor dynamic range.
Maybe I am blind, but the actual photo directly contradicts the bclaff's theory curves:
- the Sony A7R3 shows substantial (!) extra noise compared to a Pentax K-1II, so there is no way it has even a comparable dynamic range.
Some actual SNR measurements of our ISO 100 +6 EV Raw files using RawDigger (a7R III normalized to 36MP):
3rd from right grey patch:
9.2 (K-1 II) |
9.3 (a7R III)
2nd from right grey patch:
6.7 (K-1 II) |
5.1 (a7R III)
Darkest grey patch:
4.3 (K-1 II) |
2.9 (a7R III)
The Pentax and Sony are similar (9.2 vs 9.3 SNR) for 'reasonable' threshold cut-offs. That's why Bill's PDR values show the cameras to be the same. By the time you get to SNR = 5 or so (which I'd prefer not to include in my images, but YMMV), the K-1 II is a bit cleaner, and at SNR of 3 the K-1 II fares even better. That's why Bill's 'Engineering Dynamic Range' (EDR, which uses a threshold of SNR=1 on the lower end) is
better for the K-1 II (13.8 vs 13.6 EV).
But by the time you're at SNR = 5 or below, for an ISO 100 image, you're not likely to want to include those tones in your image, and would probably clip them to black. That's why EDR has limited utility and PDR tends to be more reflective of 'usable' dynamic range (Bill can chime in here on what the 36MP pixel-level SNR threshold would be for his PDR metric).
- the Sony A7R3 has massive loss of detail when using the dynamic range to a large extent, much more than a Pentax K-1II, but somehow I only read the narrative of how the Pentax has some detail rendering issues.
Again, that's not correct - you're conflating lens issues with detail retention due to dynamic range.
-Rishi
