I get my info from independent measurements on a Zeiss K8. The same equipment Zeiss uses (albeit a newer, more updated version), and also the same equipment Leica uses for real measurements. Exactly where in that quote do you find a reference to "measured value"
-"MTF is indicated..." - blah, blah, blah.
Nothing about how they arrive at the MTF.
The average when looking at Leica lenses is that they're about 10% (0.1 units of TF 1.00) worse than the "official" graphs at full aperture, and about 5% (0.05 TF units) worse at the "other" aperture (either F/5.6 or F/8 depending on the maximum aperture of the lens).
Here's two samples, the 180 APO elmarit and the 100 APO macro, both are at 20lp/mm and WO aperture - one is the official MTF, the lower is what the Zeiss K8 says about a good sample: http://ezpicshare.com/images/leicavsmea.jpg
Most of their lenses look like this when you compare a real measurement against their reported MTF curves.
By "good sample" I mean a lens that is above average in performance - you get quite a lot of lemons with Leica too - just as you do with other brand's expensive lenses. Nikon, Canon, all have 5% from each batch that really shouldn't have left the factory.
I've found a strange exception to the rule:
The Summi-M 50 actually measures better than what the official graphs would indicate... (!!??)
I guess they had to hold the Summi back in the simulation to give the 'lux an edge (?) in the datasheets.
My conjecture about bandstop filtering in the simulated light is that some lenses with extended bandwidth APO correction should be noticeably better (in the official values) than some other lenses more prone to purple fringes and deep red defocus. They aren't. In a true broadband D-weighted light APO shows up pretty easily.
From that quote I didn't get that the MTF is measured. I showed that quote only to point that their MTF charts don't have only 10, 20 and 40 measurements (ok, I think it is measured), but also a 5 one.
I checked around the internet a bit, you know uncle Google...
LLoyd Chambers says that Leica and Zeiss both publish measured MTF charts while Canon and Nikon don't. No offense, theSuede, but I will believe him.
"Zeiss and Leica publish MTF charts that are measured from a real lens. With the Zeiss ZF, ZF.2 and ZE lines, all lenses must meet that required performance level or they rejected and not sold."
But as I said before, I originally assumed they were measured, because they look measured and are much worse than what I see from calculated Canon and Nikon charts.
The only difference is, that Leica publishes them for the last 7 years while Zeiss was doing so even before.
trdonja wrote:
Regarding measured vs. theoretical, if the MTF charts I am looking at are indeed theoretical maximum, then Leica has some of the worst lenses in the business... this HAS to be measured.
Yes, I didn't understand the "meh" feeling that some photographers felt. That MTF chart looks great to me. The field curvature is there, but definitely it looks well controlled. The bokeh should be good as well and corners are very good.
The first thing that I think about, when I see an above average sized lens is, that it is probably well controlled in the corners as well. I read somewhere, that lens designer can make a lens MUCH smaller by compromising corners slightly. Making no compromise across the whole frame means that lens gains considerably more weight. That is the reason, why some parts of the frame are usually a bit washed up, because the lens designer didn't feel that quality gain would justify the size increase.
Zeiss probably wanted to deliver big time with this lens.
trdonja, in the new lenses (S- and Summarits) a lot of the reported MTF graphs approach AND EVEN SURPASS theoretical maximum (diffraction limited) values by a bit sometimes. They are without a doubt simulated values, unless they're measured in another reality, with different laws of nature than the ones we have to suffer..
The measured line (black) for the 100APO above is as close to theoretical maximum as I would say that any credible measurement will ever be able to get in a production lens. It is slightly lower at the optical center than the APO-elmarit 180 - which in turn is as good as you are going to get in the center. If you have a flat field 0.85 @ 20lp/mm you're so close to perfect that it's almost scary. I wouldn't shake the lens too hard if I ever got a sample that good, it would probably cause a misalignment that ruined the perfection.
Short-wavelength bandpassed light (with very little red, no IR and bandstopped long UV) might get close to or even surpass the MTF-reality as we see it, but those measurements are quite useless for photographic purposes (Canon.... cough...).
The absolute theoretical MTF limit at F/5.6 and 20lp/mm is ~0.90. This means having ZERO light-scatter internally in the lens elements (impossible unless you make all elements out of pure crystalline materials with perfect transmission), having ZERO surface reflection and random dispersion (equally impossible unless you have perfect nanocoatings on all surfaces and surface deltas approaching single molecules of crystal), and ZERO internal light scatter in the lens tubing, mirror housing, film or sensor, and front rim of the tubing.
In short it means that if you look at the lens, holding it in your hand, it should be impossible to see that any glass is mounted in the tube. THEN you can get close to the theoretical maximum.
I don't think I have the S-lux M 35 asph measurements, but the older 35 should be here somewhere.
trdonja wrote:
LLoyd Chambers says that Leica and Zeiss both publish measured MTF charts while Canon and Nikon don't. No offense, theSuede, but I will believe him.
Lloyd Chambers makes mistakes, and he clearly doesn't know as much about this stuff as theSuede above. Furthermore, in this particular case his statement goes against the explicit admission of none other than Peter Karbe. See from about 8:41 in this video:
MTF curves published by the manufacturer can be useful things, especially if an aperture series is available (Leica are good about this, often publishing curves for three f-stops), but I wouldn't trust them to be true for a lens I've bought at the store — whether that lens is a Leica, Zeiss, Nikon or Canon.
The old photodo.com site is full of measured MTF curves that anyone can compare to the manufacturer's published curves. Takes about five minutes per lens.
Andi Dietrich wrote:
from experience I would prefer the lens with less lateral color over the lens with the higher MTF curves. Seems to be the Zeiss here.
The compared curves are for f/1.4, at which f-stop lateral chromatic aberration is relatively unimportant (because it's drowned out by more urgent problems, like coma).
Still, the f/4 curve suggests that Zeiss paid much more attention to fixing lateral chromatic aberration in this new lens than the f/2 model. Consequently its performance at f/8 will be very good across the whole frame.
denoir wrote:
Here are two more MTF charts, more comparable probably and more relevant. The first one is old vs new 35/1.4 Distagon. Black new, blue old:
The second chart is of the new 35/1.4 @ f/1.4 vs 35/2 ZE @ f/2. Black is the f/1.4 version and green the f/2 version:
interesting, suggests (to me at least) that the new 35/1.4 might be closer in rendering to the ZE 35/2 then the contax 35/1.4. i'll be interested to see how that plays out in actual pictures.
trdonja wrote:
"Zeiss and Leica publish MTF charts that are measured from a real lens. With the Zeiss ZF, ZF.2 and ZE lines, all lenses must meet that required performance level or they rejected and not sold."
Somehow, I find it difficult to believe that Zeiss actually test every single lens they sell and reject those that do not meet their MTF charts. Has Zeiss ever officially stated this somewhere? We know that there are some Zeiss lenses sold that suffer from de-centering and other issues so this must not be the case - part of that 5% the Suede was mentioning. If anyone tested every lens, I would have thought it would be Leica given the cost of each lens. Guess that's not the case either.
Tariq Gibran wrote:
If anyone tested every lens, I would have thought it would be Leica given the cost of each lens. Guess that's not the case either.
No, no. The cost of each lens has nothing to do with the quality of it. It's simply that there is a much larger demand for the lenses than Leica can output. As for testing.. well let's put it this way: you should consider yourself lucky if they have looked at the lens to see if the aperture ring hasn't been mounted upside down during assembly..
If you want a tested Leica lens, buy a used one.
Jan 17, 2011 at 07:48 PM
Lars Johnsson Offline • • • • • • • Upload & Sell: Off
denoir wrote:
Well, we already know they succeeded in that. They really delivered BIG. 120 mm length, 72 mm diameter, to be precise!
I can't understand why so many people here are whining about the size/weight of the new lens. It's an f/1,4 lens. People here like to have the best image quality of any 35mm lens. They like the build to be solid and entirely metal. And at the same time they like it to have the same size/weight as the Zeiss 35/2
If anything you could put the lens in a sock and take out would be robbers. Tryna' run my gear? Taste my ZEISS. And then go back to photographing benches
But seriously, I'm psyched to see what this lens produces. Being larger (longer) than the 135L just blows my mind.
Lars Johnsson wrote:
I can't understand why so many people here are whining about the size/weight of the new lens. It's an f/1,4 lens. People here like to have the best image quality of any 35mm lens. They like the build to be solid and entirely metal. And at the same time they like it to have the same size/weight as the Zeiss 35/2
actually i'd like it to have the same size and weight as the leica summilux pre asph.
in all seriousness, i'm much more sensitive to the size of lenses i want to use for shooting people than i am to the size of lenses i use to shoot wildlife. at a certain point people begin to worry that i'm pointing a canon at them.
