Leica 28 mm lens review (3.1), march 15, 2017
Analysis based on film emulsions
First a few basics.
The main difference between digital and analogue recording is the fact that the digital recording is based on two values, zero and one, or black and white. If we assume for the sake of the argument that we are looking at an edge of a black and white boundary, the pixel in the black part gets a value of zero assigned and the pixel in the white area gets a pixel of one assigned. Whatever the manipulation, these values are not altered when there is some manipulation. In a black and white film emulsion the situation is much different. Here we have black silver on the white side and no silver on the black side. When light is transmitted through the negative the scattering will reduce the black and and white and the result is a darker grey and a lighter grey to the point where both elements are indistinguishable. No more detail will be visible. In the digital processing this scattering does not occur and even when the real values (0 and 1) are close to 0.4 and 0,6, the software will re-establish the values to 0 and 1, so there will be no loss of information. In the more complex case that the analogue version records 0.4 and 0.6 the scattering will reduce this to 0.5 and 0.5. In the digital situation the 0.4 and 0.6 values will be converted to 0 and 1 and no loss of information will occur.
This is the main reason why a high quality negative will produce a medium quality print and a medium quality image file will produce a high quality print.
The camera used was the Leica M7 because the internal exposure meter and the accurate shutter speeds produce reliable results without much thinking when using exposure series with different apertures. The camera was on tripod and loaded with Ilford XP2. Why this film: it is almost grainless, but quite soft and so will capture a wide contrast range (not needed for the resolution test) but for the comparison with the dynamic range of the MM2. The softness of the negative must be taken into account when discussing the micro contrast (contrast of adjacent areas).
The negatives were analyzed under a microscope with a 40 times magnification. A lower magnification does not reveal the extremely fine detail that the lens/film combo can record. A 40 times magnification implies a printing size of 1.44 meter!
To get the equivalent screen resolution for the image files of the MM2 we need a 400% screen resolution. The image file is 5976 x 3992 pixels. The standard screen is 1440 x 900 pixels----> 5976 -/- 1440 = 4.15. The effective screen width is 33 cm: 4 x 33 cm = 1.32 meter.
The actual print size for a high quality print would be 6000 (rounded!) divided by 300 multiplied by 2.54 cm = 50.8 cm.
To produce a print size of 1.44 meter with 300 dpi/inch (best print quality) would require a file width of 1440 cm divided by 50.8 cm multiplied by 300 pixels = 8510 pixels.
No one would print such a large print, but is indicates how much enlargement is needed to discuss the fine points of a Leica lens.
The upshot is simple: anyone who derives conclusions from a visual analysis of a screen resolution of 100% will miss a lot of information.
The important conclusion after comparing the results on screen for the image files of the MM2 and the results after visual inspection of the XP2 negatives is that there is no difference in resolution or definition. The XP2 negatives are much softer than the interpolated image files (in Iridient Developer, Photoshop CC 2017 and Capture One 10), but one can cope with this difference when looking specifically at the fine detail definition. This is an amazing conclusion: a classical 400 ISO film emulsion is as good as the best performing 24 Mp sensor in the Leica stable.
Of course it will be logical that when the image file and the negative are both printed on a print size of width 40 cm (example Epson 3000 print and Focomat/Splitgrade/Schneider lens), both prints will not reveal the subtle differences of both media. This is significant in itself. Disregarding the hyperbole of many members of the digital community, film is not yet outclassed. It is however to be honest much more difficult to get results that are equally impressive.
The general review
Overall the results of the three lenses under review are comparable to the results obtained with the MM2 files. We divide the image area in three segments: center, edge (12 mm radius from center) and corner (21 mm radius from center).This corresponds to the center, the 24 mm length of the vertical side of the negative and the 36 mm horizontal length of the negative.
The classical parameters are used: vignetting, detail definition and overall/micro contrast at three positions and close up performance.
The Elmarit-M 2.8/28 Asph is the most sensible choice. But the Summaron-M at its modest aperture is not much lagging behind. It is in itself interesting to note that the Summaron-M (a very vintage design) performs very well even today with much improved recording media and lenses. It is logical that the old version did get good comments, given the quality of the recording media in the 1950s.
The Summilux-M 28mm is like a turbocharged version of the Elmarit-M 28 mm.
Detailed performance (next part)
