I use MTF charts or any resolution/contrast documentation to only understand why I like a lens and a film and a developer more or less, in a situation or another. Although, I make my decisions on what lens or film-developer I like most, only based on my tests in real situations (not photographing charts). And I also tried each of my lenses at different apertures too. One combo works in landscapes with contrasty light, another in low contrast outdoors and indoor portraiture, another (and here I add the camera to the lens-film-developer combo) in low light situations, one works for a reportage style, another for fine-art, not to remember color vs. B&W, and so one.
The idea of “the best…” is a wrong questioning. The good one is “the best… for…” (or the most appropriate). I hate almost everything that is said to be “the best for all”, this makes my nightmare. I have old, cheap lenses in 135 format I wouldn’t dare to get outdoors because of their huuuuge flare, but indoors these lenses gave me the most sublime portraits (beyond my imagination). Since than, I don’t trust any chart or documentation until I try the material myself in real situations. Charts are only to understand what you already see. But if you start with charts, I doubt you’ll ever see something sublime.
Sublime does not always comes from “the most…”, sometimes it also comes from “less…”, or better said for a specific balance between different characteristics of the object. I remember Leica worsted years ago the design of one of it’s sharpest lens, to make it appropriate for portraiture (found this in a Leica equipment book, but I don’t remember if it belonged to the M or the R lens line). They “worsted” it (larger chromatic aberration), but didn’t sold it cheaper… They cold it “specialized lens” and sold it even higher. And if you have a larger budget than I do, why not experimenting with limited series lenses, like the Pentax* (star) ones? I saw two or three B&W pictures taken with the Pentax* 43mm and was astonished.
Last edited by phenix; 05-16-2010 at 10:56 PM. Click to view previous post history.
B&W is silver.
Originally Posted by olleorama
That may hold true for lenses of more or less "normal" length.
Originally Posted by ic-racer
The trouble starts with wide angle, ultra-wide angle and longer telephoto lenses.
Many wides, even the very best, still have less than ideally sharp corners at f/8.0 (and clearly visible even on high speed film - maximum resolution theory notwithstanding).
Same for many longer teles, which will often show various aberrations, even at f/8.0 and again clearly visible even on less than ideal media...
In any case, my favorite lenses do have a big advantage at f/8.0 compared to ones I like less: they are less suceptible to flare and have less distortion.
So while it is true that most 50mm will look pretty similar at f/8.0 in tests, once conditions are less than ideal, real differences will emerge.
I won't even get started on subjective aesthetic parameters such as bokeh and 3-d look...
Last edited by Rol_Lei Nut; 05-17-2010 at 01:50 AM. Click to view previous post history.
M6, SL, SL2, R5, P6x7, SL3003, SL35-E, F, F2, FM, FE-2, Varex IIa
I think the question goes into the wrong direction. How would it help your photography to know that your lenses perform best with a high-contrast micro film? The answer is clear, they do, but these films are not ideal for practical photography, so why test with them. To evaluate lenses, and more importantly, how any difference impacts our photography, they should be tested with the films we use. Also, resolution measured in lp/mm (line pairs per mm, not ln/mm) is a poor measure of lens performance anyway. Contrast and sharpness being more important than resolution, only an MTF will tell.
Originally Posted by kauffman v36
Fully agree with phenix, by the way.
Arent most good films around the 60 lp/mm at 1.6:1 contrast ratio? That is what I call useful resolution, ie: across the texture of an object, the 1000:1 figure is for shooting sharp-edged silhouette backlit tree branches against a white sky :P
There are of course much higher resolution films... Rollei ATP, Adox CMS 20. etc
Reasonably large prints from my 30D look critically good to me, 35mm film is capable of more than my 30D, so therefore would do better and satisfy critical print demands.
Originally Posted by kauffman v36
CMS 20 + The Polar/Samyang 85/1.4 would be a great test if ultimate sharpness makes you happy.
Sponsored Ad. (Subscribers to APUG have the option to remove this ad.)
Kodak lists Tmax100 at 63 lp/mm and Tmax400 at 50 lp/mm at 1.6:1. That's probably as unrealistic as listing 200 lp/mm at 1,000:1. Nevertheless, it shows that film and paper limit the system performance significantly. I'd like to see the figures for 160:1, which is an average scene contrast of roughly 7 stops. I assume it's around 125 lp/mm.
Originally Posted by Athiril
160:1 contrast resolution would be how sharp one edge is next to edge that is 7 and 1/3 stops difference, not the resolution of a scene that is 7 and 1/3 stops wide. You're extremely unlikely to find any 7 and 1/3 stop edges in any kind of normal exterior scene, the transitions are much lower than that, and most transitions are not hard-edged, you wont get that on the hard edge between direct sun and shadow either, you'd have to set up a controlled situation where there is no ambient bounce back into the shadows.
The texture of bark or someone's skin etc, is not going to be 7 and 1/3 stops of variance over those surfaces as a whole, let along as the texture/transition from each "line" of the surface, surface detail is going to be in the 1.6:1 vicinity.
Last edited by Athiril; 05-17-2010 at 05:34 AM. Click to view previous post history.
Good point. I have some RIT resolution targets with varying contrast (three different setting, I believe). They clearly show the influence of subject contrast on resolution measurements. I also measured my USAF 1951 targets, which I bought from Edmund's Scientific. They have a density range (black printing vs white paper) of over 100:1.
Originally Posted by Athiril
Here are two examples. An Olympus 50mm 1.4 SLR lens vs a Leica 50mm 1.4 rangefinder lens.
Error bars represent "similar lenses in class."
Y-axis is contrast at a given 50 line pairs per millimeter. So these read a little more like a contemporary MTF chart.
I think this info is interesting to read and also may be of some benefit to the next generation of 35mm film users who need to sort through all the left-over cameras from the last generation.
Last edited by ic-racer; 05-17-2010 at 04:49 PM. Click to view previous post history.
I checked my files. The RIT targets come in three different contrasts:
Originally Posted by RalphLambrecht
32:1, 6.3:1, and 1.6:1