You put a great deal of weight on stitching a series of smaller frames versus using an MF frame (or two or three). Stitching of course is an option, but it deals only with the total pixel count in the image. As we all know, there is much more to sensor quality and image quality than pixel count. But even at the level of pixel count, my immediate reaction to these options was exactly that of Nick Rains: how much premeditation, time and effort do you want to put into a stitching approach pre and post capture, versus getting one or several MF captures containing all the pixels you'll ever need? Yes, the latter costs much more money as an investment, and the former costs a lot more time - perhaps even at the expense of getting the image, depending on conditions - and there is, for many people, a real time value to money. Over time, it all adds-up. I don't think the value of MF photography using the better of today's equipment (2010) is at all nullified by the stitching alternative, although stitching remains a viable and worthwhile option in many circumstances.
Mark,
Fair analysis indeed.
I mostly agree with you but would have agreed 100% if you had talked about "a perfect MF body". Indeed, I agree 100% that a
perfect MF body that would be available at reasonable prices would indeed be a better base landscape camera than a DSLR + stitching.
Now, how close are existing MF bodies from that perfect camera and how many images does a top landscaper shoot a year that would justify the cost?
A key point that was not even mentioned by Mark is the real resolution achieved over 100 frames. That is one of the main gaps between current MF and the perfect camera I was talking about. FX bodies will be within a few percent of the optimal every single time thanks to live view. I believe that many MF landscape shooters have been suffering from mis-focusing on many frames, the lower the light level the worst the issue. Our friends at Diglloyd have been reporting once more in the 645D review about how difficult it is to focus the body accurately on a plane subject (granted real world applications will be less impacting). The real world resolution of bodies without live view should probably considered to be 10 to 20% lower than the pixel count indicates. Beyond that, it might just be me, but I always was
very frustrated when coming back from a shoot with my MF body to notice that half of the images shot with my 10.000+ US$ camera were not correctly focused. For me that frustration was a major problem.
I didn't intend to enter the MF vs DSLR debate (my point was only about stitching), but others have take the discussion there, so let's assume for a second that most stitching is done with DSLRs.
Regardless of the time t relevance of DxO, there is little denying that the improvement of pixel quality of FX is faster than that of MF. I don't believe having ever felt limited by the DR of my D3x while I did feel limited by both my former DSLRs and my Mamiya ZD. Leading photo publications like Chasseur d'Image and Photo in France have recently reviewed the 645D and found its DR to be in the same ballpark (or a little bit lower) than
their current reference, the D3x. For what it is worth, these guys combined sell over 500.000 copies a month. Diglloyd similarly found the DR of both S2 and 645D to be inferior or similar to that of the D3x. If a gap remains in favour of MF it is small and closing. Considering that the MF bodies use sensors with photosites of the same size as DSLRs it only makes sense that DR is similar by the way (knowing that DSLRs use CMOS that are typically less noisy - noise defines DR). The only rationale for MF bodies to still have better DR is their higher price.
Together with the reduction of the gap between FX's resolution and the needs of most real world applications, I feel that many
landscape photographers (few images, high quality focus) today are in fact best served by a high end FX body plus the stitching option that will enable them to reach whatever resolution is needed for those images that deserve a large print. I wouldn't have been shocked had Mark, after presenting in a fair way these different options, had reached the conclusion that he feels that the balance still leans towards the MF being the better option.
But a fair comparison there was not.
So, considering the overlook of all these factors, it shouldn't come as a surprise that the initial article is perceived as a plug for MF manufacturers disguised into a opinionated write up. For what it is worth, the only place I know where the relevance of DxO results is widely questioned is this very community with a strong presence of MF owners and sellers. You and I appear to belong to this category of people feeling that we owe to ourselve to own and broadcast a balanced view of things, but LL as a whole seems to be leaning more and more to a one sided view of things aimed at counterbalancing the rest of the world.
LL vs DPreview. LL says MFBDs have 6 stops more DR, DPreview says they have the same... so it must be around 3 stops, right?...
Cheers,
Bernard