Hi,
Just as a comment, Lloyd has tested several MF cameras, but I got the impression that he found focusing problematic on most of them, with the Hasselblad H4 (?) being a possible exception. He very clearly had focusing issues on both the Pentax 645D and the Leica-S2, AF is just not exact enough and live AF did not work well enough on the backs he tested. The IQ-250 has excellent live view, it seems from the reports.
Lloyd is shooting with a few of the best lenses ever built for 135 format, the Coastal Optics 60/4 and the Zeiss Otus 50/1.4, both diffraction limited at f/5.6 or so and essentially without any colour aberration, so he has a very solid reference ground.
On the other hand, I am pretty sure that a very good lens with an IQ-2xx will outperform even an Otus on a 36 MP 135 format camera.
Something I would mention is that I think that we all do pixel peep a bit to much. Pictures are normally either shown on screen (which really only has 2MP resolution) or in print. Large prints viewed at very close are close to pixel peeping, but looking at pictures at 1:1 view on screen is very critical, because of the large pixel pitch on most monitors (a 4K monitor at 24" inch would be a better match for visual evaluation).
Next issue is that if you look at two systems of different resolution at actual pixels the system with higher resolution is always at disadvantage. Something I do is to up-res both images to print size. Say, if I print at 70x100 (cm) I would up-res to 200 PPI (because that is what my lab uses), than I would compare the pictures side by side at actual pixels. Still looking at that file on screen would correspond to 140x200 cm (as pixel pitch on screens is normally 100PPI or so.
In general, I got the impression that from my experience is that the more controlled the experiment is the more similar the results would be. For instance, shooting a colour checker for WB and adjusting exposure to match can reduce differences between systems quite a lot. Build a colour profile for both and they will become even more similar. Personally, I shoot P45+ and Sony Alpha 99 and I cannot tell the two apart in A2 (16"x23") prints, but I am pretty sure I could see a difference in large prints.
The enclosed screen dumps may illustrate the matter a bit. They show a crop from DPreview's test images at actual pixels in the first image and the second shows the A7r image upsized to IQ-180 short dimension. In the first image the Schneider 80/2.8 LS lens on the IQ-180 shows some weakness while the 85/1.4 ZA-Sonnar on the A7r is essentially sharp. The second picture shows that the up-rezzed image from the A7r falling apart. The third one is a screen dump at 1:2 size, which would correspond more to 200 PPI viewed on print. As a side note, 20/20 vision resolves high contrast detail at around 180 PPI when viewed at
50 cm. So the third image is close to what an observer with 20/20 vision would see in a 145x109 cm (57"x43") print viewed at 0.5 m distance.
Sorry for not using my own images, but I don't have neither an IQ-180 nor an A7r. Would I use my own images there would be a lot of blaming of my technique. The images are available in raw form from DPReview for anyone. I feel this is proper use demonstrating some aspects of testing.
Best regards
Erik
I've always had a hunch that the Hy6 was the best, modern digital MF platform. I've never had the pleasure of seeing one in the flesh, but the system options seem second to none. It's a real shame the platform isn't better supported or hasn't become the standard platform for Phase. This is all old ground to re-cover, I know. I just thought I'd say again anyway, for hope that someone high up is listening!
My impression too is that 80mpx is going to push the lenses on all DMF systems. And you are correct, much of the last few ounces of quality in terms of pixel level sharpness is always going to come down to perfect technique at this kind of resolution. More importantly though, for this kind of money you'd better LOVE the way the lenses draw and the system handles. Last thing anyone most people want is a clinical, aesthetically bland but sharp photo. Unless of course you're shooting charts and such.