DPReview recently released the Nikon Z5 lab samples, and given all the back-and-forth about full frame vs. APS-C, the industry push for full frame, and my own interest in low light performance, I thought it worthwhile to reproduce some select comparisons from that publication.
Keep in mind that the Nikons are using a brand new lens, while the Pentax and Sony were tested using older lenses, especially the Pentax, where DPReview thought a 12-year-old optical design could be a reference lens. (Unlike Imaging-Resource, DPReview has to my knowledge never attempted to use a consistent reference lens.) I suspect the difference in lens age accounts for the difference in sharpness seen. For both systems, better reference lenses are currently available at those same focal lengths.
Also keep in mind that the recently announced Sony A7C uses the same sensor as the A7 III, and those who’ve had their hands on one have said the image quality is the same.
So let’s look at the image noise, then:
I still think that what the accelerator unit is doing in the Pentax KP is remarkable, even if baking in the noise reduction with no way to turn it off is a little unorthodox (but done by many, if not most, camera makers that apply noise reduction to raw files). My impression has always been that in the Pentax KP, this all really works very well, whereas it hasn’t convinced me as much in the Pentax K-1 II.
For a 2017 APS-C camera to match the image quality of 2020 full-framers is really quite a feat, especially given you can buy two KPs for the price of one Z5 right now (KP, 730 Eurodollars; Z5, 1400 USD or 1560 EUR).