1/3 stop less of so-called "PDR" at high ISO is a very theoretical loss as opposed to the practical gain of shooting telephoto lenses with a minimum of vibrations when tripod mounted. Mirror up + EFCS will give you the lowest amount of vibrations to fight at the outset, when you use a DSLR to do tripod mounted tele photography, particularly with light lenses like a 300/4 or a 400/5.6.
I agree and the D810 is my main camera and the one I'd choose for such work. However, my point is that there may be engineering reasons / tradeoffs behind Nikon's reluctance to adopt EFCS across the board in the sensors that they use in their cameras. They have a broad user base whose needs they need to consider. Nikon says that the D810 EFCS may cause uneven exposure at faster shutter speeds with PC-E lenses, for example. So at present it isn't the default mode of operation in the camera but only invoked as a custom function and in a specific drive mode (M-UP).
I assume that the term "PDR" stems from tests done by Bill Claff, tests which I view with much sceptisism.
Right, I was referring to Bill Claff's site. I think the concept of photographic dynamic range is solid (he uses a higher threshold for the noise floor where the SNR for an image segment is supposed to be at least 20 instead of 1 as in DXOMark's tests to be included in the dynamic range; SNR = 1 is just too low to be considered for imaging purposes and I'm in agreement with Mr. Claff on this subject). What I don't like about his methodology is that he analyzes images sent by others. This does not seem like a truly controlled test methology and could be subject to errors due to test condition and tester variability. However, AFAIK DXOMark do use a testing procedure that is in controlled conditions (of course every procedure has some variability and the camera and photon noise themselves, as well as camera sample variability contribute to it). I'm not suggesting one should believe the results of one site blindly but when they point to the same direction there is probably something to it.
DXOMark's dynamic range measurements show that the D810 has worse DR than the D800E and D750 from ISO 100 to 1600. From 3200 to 12800 the D750 is above the D800E and D810. So while there are subtle differences, both sites show (using independent data from each other) that the D810 DR has some loss compared to the D800(E) and is not the best of the Nikons, when compared at equal ISO settings (though at ISO 64 it
is the best). In other areas the D810 does shine, e.g. the dxomark color sensitivity measurements show an advantage to the D810 over the D750 from ISO 100 to 400. One thing that I noticed when using the D750 is that I preferred the D810 colours and resolution at low to moderate ISO, but the D750 allowed some ISO 6400 shots that were underexposed (shooting into the light) to be recovered that could not have produced acceptable results using the D810 or D800. The Df / D4 family / D5 likely show a further advantage at ultra high ISO over the D750 but I haven't used them enough to say what its real world significance is. I do know that at ISO 6400 the light and shooting conditions are often such that it is critical to get the best image quality possible to make the results presentable. I find the D810's ISO 6400 image quality problematic especially if I need to do vignetting correction (which results in a marked blue shift in the corners with this camera, also with the D800 it was worse as there were lines in the noise which are absent in the D810 images).
My overall preference is the D810 but this has just as much to do with the body ergonomics and features as image quality. I like EFCS for tripod based landscape and macro shooting with a telephoto. However, I also do event photography and quite a lot of it is at around ISO 6400 and thus one sensor is not optimal across the board. Nikon now offers at least two cameras with EFCS (in raw mode; the D4s AFAIK can do EFCS only JPG if I understood correctly) and so the users with specific needs should choose these cameras to alleviate their vibration issues, if economically feasible. The D750's lack of EFCS was disappointing to me when the camera was introduced but my main complaint about that camera is that it doesn't fit into my hands well, and in the winter when I must wear gloves it is just uncomfortable to use. However, many people love the D750's shape so it is a subjective thing, depends on the individual's hand and finger size. It does present a very attractive second camera due to its compactness, and I intend to take the D810 + D750 pair to Iceland this summer, as I need a second camera for the trip to use in case of camera failure. Also I use it nowadays in preference to the D810 for some of my high ISO work.
Perhaps Nikon can improve EFCS in the future so that there won't be any need not to use it in regular shooting, and it could also be used in hand held shooting with the viewfinder (I'm not sure why there is the limitation to M-UP mode unless it is simply that Nikon wants us to use it only in carefully controlled situations with cable release, so as to make it less likely that the artifacts of rolling shutter and uneven exposure show up more often). I would hope also that future implementations of VR can operate at higher frequencies and better compensate for vibration caused by the mechanical shutter. As of now, both technologies have limitations and we as photographers have to work around those limitations.
