No general guidelines regarding max movements can be given unfortunately. All sensors have color cast, which is corrected by the obligatory LCC. To the very least color cast is created through that the sensor color filters change response slightly when light comes in at an angle, so there is currently no sensor technology and never was one that could be free from color cast. However this simple color cast is quite mild and is 100% reversible with LCC.If CMOS sensors apparently don't work well with tech cam movements, what are the alternatives?
I'm asking as someone very interested in tech cams given (i) i have a background in large format film (now sold, but I REALLY miss movements much more than I thought), and (ii) I can see a clear difference in MFD files compared to 35mm FF (the latter is what I currently use), hence why I'm on this medium format part of the site!
Please correct me if my arm-chair analysis is wrong, but it seems the choices are:
(i) CMOS is limited to small movements before running into probs - but has benefits of both LV and with CMOS now being the latest & highest MP backs. An Alpa STC has a maximum movement of approx 20mm, but is there a rule of thumb for a newbie that suggests how much tech cam movement is practically possible with CMOS backs?
(ii) more extensive movements from CCD backs, but no LV and typically older / lower MP. I guess the same question again - is there a rule of thumb for max movements here with CCDs?
(iii) a film back such as the dedicated Alpa roll back. Given the size of 6x7 or 6x9 film, it seems one would need to really use an older Alpa analogue lens (which have very large 160-200mm ICs) to get the full 20mm front rise that is possible with cameras like the Alpa STC. I never saw any probs at all with large format film and extensive movements, but is there any reason why a film back (combined with full use of 20mm movement on a tech cam) would run into any probs??
Why not just go back to large format film, you ask? Well, 120 film is much easier to use (no dust, loading issues, easier to mail for processing, 350mb drum scans render aesthetically nice 40"-50" prints off Acros / Ektar) + I'm attracted to the potential flexibility that a tech cam might offer (ability to use both digital backs and film), + some tech cams look almost handholdable for "casual shooting" but also offer very precise movements for "thoughtful" shooting (all in one package) ...... so there appear to be lots of reasons for why a tech cam might suit me. But the question about the amount of Movements that is realistic / practical (and in what form ... CMOS, CCS, film) is something I'd like advice on. Many thanks!!
Kodak micro-lens free sensors like in the P45+ and Hasselblad H4D-50 etc were the last "issue-free" sensors (that is only predictable fully cancellable color cast). With the Dalsa 6um CCDs (P65+ etc) lightshields disappeared and microlenses were introduced. This required wide angle lenses to be a little retrofocus, and Rodenstock Digarons were designed for this. However there was some miscalculation so there still are issues with microlens ripple, but Capture One is very good at cancelling that out in post. The Dalsa 6um CCDs can still be used on fully symmetric lenses (and some do) like the SK28 and SK35 but then you get issues with significant crosstalk when you start shifting. Crosstalk is mainly shown as desaturation of colors (as the color channels are mixed) and possibly demosaicing failures (mazing).
With the 80MP Dalsa the problems increased further, and yet more with the Sony CMOS. The problem is however that the crosstalk effects are situation dependent. If you have close-to-neutral colors in your scene you don't get much problems despite large amounts of crosstalk. Simply put, there aren't that many situations you get problems in practice. Few notices a reduction in tonality performance.
As there really aren't any options people have learnt to accept using sensors that are not matched with the lenses. The argument goes if you don't really see the performance reduction after proper post-processing why care? I personally care quite a lot because I'm very much against the principle of pushing the sensor out of spec with the lens and getting potentially unpredictable behavior, but others care a lot less and you can't really say that my view is more valid than any other. There are those using the CMOS with Schneider Digitar wides although it's deemed almost impossible. As the CMOS have extremely good DR and tonality it can indeed make up for quite a lot of loss. So a general recommendation cannot be had.
Anyway, Kodak CCDs without microlenses: except for controlled color cast no issues on any Schneider Digitar or Rodenstock Digaron wides. Oh well, you do get pixel vignetting (but no crosstalk) on the SK28 which cuts some of the dynamic range even if center filter is used. Nothing too critical though.
Dalsa 6um CCDs: with a LCC algorithm that can handle microlens ripple (like C1) there are generally no issues on any Rodenstock Digaron lens, shift to the limit if you want. You might get faint tiling with extreme contrast processing, but it's fixable. Schneider wides which are symmetric trigger some issues, even the long SK60XL if you go to the limit of the image circle.
Dalsa 80MP CCD: some issues also on the Rodenstock Digarons but only for large shifts.
CMOS 44x33 and 54x41: similar issues for both, more issues than the 80MP but also more DR which can mask some of the issues. Very much up to personal taste how much shifting you can accept.
Film handles movements extremely well, the cast and crosstalk issues doesn't exist with film as it's built entirely differently. (Or maybe there's is some cast, at least theoretically, but I have never heard of anyone having any problems in practice). Tjv in this forum has shot a lot of film on his Linhof Techno before going digital with a Credo 60 back, and I'm sure there are others. I have myself no experience with film on my tech cam (also a Linhof Techno).
Note that for long lenses all sensors work well.