I believe that the assertion that if a print is viewed from far enough away, resolution doesn't matter is commonly used to oversimplify or simply dodge thinking about print requirements. Further, who has the right to someone else what the correct viewing distance is to his own prints?! Taking the argument to the extreme, if you stand a mile from a print, the print resolution will indeed not matter. :)
Originally Posted by pen s
Its simple to quantify print resolution requirements, just using geometry. This is a simple experiment I have done: print a pattern of line pairs/mm up on a sheet of paper, alongside a smooth grey tone, and walk away from it. At the point the texture of the line pairs disappears, the two pictures will start to appear the same. Measure this distance. From here you can calculate the given number of line pairs a print requires for a given viewing distance, for you. You might pick a pattern with a frequency of 1.0 lp/mm. If the viewing distance at which this appears to become a smooth tone without texture is 4m, and one views prints at 50cm, then you have a print resolution requirement of 8lp/mm.
Yeah, and the additional irony is that I also use a Bronica ETRSi!
Originally Posted by wblynch
I mention print sizes as I think they are easier to interpret than raw lp/mm on the negative. When I shoot a batch test of lenses, I am in fact just calculating the resolution they are capable of laying down on a negative (this is a convolution of developer, agitation, tripod, focus accuracy etc etc). Its important to shoot a batch as then the relative performance can be seen easily and as resolution testing is a bit subjective in that the limit of the resolution is determined by the definition of acceptable minimum visible resolution. This is inherently dependent on the operator, or "nut behind the camera". By doing several lenses at once, it ensures some consistency though, and I can end up with a relative measurement. Whole thing takes about half an hour, I note it down in a book where I have previous results, and when I get a spare minute once the film is developed, check it out under the enlarger using a 25x scope and good enlarging lens.
I think the first thing I need to do is retest the lens, with a F Zuiko 50/1.8 and a Zuiko MC Japan 50/1.8 for comparison. Will post back in case anyone ends up reading this thread and finding it vaguely interesting to know the difference between the various incarnations of the slower Zuiko 50's.
I'm sure Olympus maximized every detail in producing this print for an advertisement poster. It was B&W, no doubt a slow, very fine grained film, and impeccable darkroom work. You could see quite clearly every tiny bump and wrinkle in the makeup applied to the subject's face. If, however you walked right up to it and looked it over with a 5X loop.....then no, the detail would probably not have matched the same shot taken with a then current Hasselblad 500C and 80mm Zeiss lens.
I shoot lots of half frame, and seldom go larger than 8X, or 6X8 inch on 8X10 paper. I have to set up the enlarger in a small, very cramped bathroom, and my 35mm only Durst enlarger was free and is very modest. Still, if you are willing to put in the effort, 35mm can be pushed beyond 8X10 for the occasional outstanding shot.
I personally need 8lpmm in a print, which necessitates EFx8 in the negative, EF being the enlarging factor. Now if I am looking at a 12x16, which I suspect may be a sensible upper limit, I would need 96lp/mm on the negative. There are documents floating the web about using high res Adox CMS film which means this is entirely possible with 135, just google gigabit film. However, I'd rather just use my 120 negatives for that size of print, given arguments about tonal information being lacking prior to resolution limits being hit. Back to the point in hand though, the 50mm lens I have been testing scored 35lp/mm as its resolution limit. This means the upper limit of a print would be 5x7, but at that point, the lens is already being stretched, meaning it would lack contrast and tonal information in a 5x7. There is a benefit in having a lens which could reach higher resolution levels used in a smaller print, as the tonal information will be more detailed for any given level of resolution.
Originally Posted by pen s
The Mij is better than thought, here are some updated results. Make of them what you will. I am reading a book which I could recommend in terms of spirit of testing, written by Richard Henry "Controls in black and white photography". Notably, he makes the point that no one can be more accurate than 1 group in rating the resolving power of a lens using a tricolumnar test pattern. In this spirit, a key lens in my testing is the Tamron 90mm SP 50B, which I have as a reference and can be used as a control as it can be mounted on different camera bodies through the various available adapters. Further, its a notoriously sharp lens in an absolute sense, so it gives an indication of the upper limits available due to other factors in the testing such as the resolving power of the film used. So in this instance, I used the Tamron in identical testing of a MX about half an hour later, with a fleet of Pentax M lenses. I dont include the results as they are not really relevant to an Olympus group. The conclusions I would draw are: a) the MIJ is capable of hitting 60lp/mm at f8 b) at f5.6 its softer than the Tamron SP c) The Tokina made Vivitar S1 lens is pretty respectable compared to the Oly 135 prime. Further testing could include impact of tripod, cable release, filters, hoods, developer, film, exposure.
25/05/2013, 15h10, sunny
190 tripod fully extended, Foba ball head, manual shutter (not cable release), no MLU
target 36 inch wide Edmund test patterns, contrast reported to be 100:1
metered using Lunasix F incident reading
XTOL 1:2 10.5mins 20.0deg, 30s + 3 inversions every 30s
Nikon 50/2.8, Microsight 25x
f4 shots were overexposed by a stop, as the meter would have recommended shooting at 1/2000.
name lens, focal length, distance to target, f4, f5.6, f8, f11
Olympus 28mm 2.8, 28, 3000, 27, 67, 67, 53
Olympus 50mm MIJ, 50, 3000, 42, 47, 59, 53
Tamron SP 90 2.5, 88.5, 6000, 60, 60, 53, 60
Olympus 135mm 3.5, 135, 6000, 43, 43, 55, 55
Olympus 135mm 3.5, 135, 6000, 39, 43, 55, 55
Vivitar S1, 70, 6000, 48, 48, 60, 67
Vivitar S1, 210, 6000, 35, 44, 49, 49
NB step in line spacing between patterns is 12%, so that 6 increments results in a doubling of the line frequency
If anyone is still following this thread after the detour to print size ... I have tested the 50f1.8 MIJ against an earlier version 50f1.8 (though not a silver nose). There was a noticeable difference, and the slides from the MIJ were sharp enough to cut yourself on. I have not tested the f1.4 against the MIJ, but really fail to see the advantage in the slight difference in aperture.
Don't forget, a proper lens hood in daylight can make a tremendous difference.
Just as a follow up, I have continued looking for a sharp 50mm...the Pentax M 50/1.4, contrary to popular opinion, is sharper by my testing than the 50/1.7, which in turn was sharper than the Oly MIJ 50/1.8. The M 50/1.4 hit its peak at f4, ie, it was diffraction limited by f5.6. In contrast, pun intended, the MIJ didn't hit its peak contrast/resolution until f8.
Doesn't do a bit of good if it doesn't mount on an OM-1
There were five different mechanical versions of the /1.8 OM 5cm.