Measuring Lens "Sharpness" across the image field: MTF

[Bernard Delley]

Yes, the color aberration on the135dc is strong. Can you post the wide open chart for it?

I am just about to take a plane. But the linked supplementary material has a some on DC 135 wide open , also figs S15 and S16.

There are some talk about decentered lens elements and that it is quite visible shooting wide open in MTF charts.
Are most Nikkors decentered?
Or how can you be sure to get a "perfect" lens?
Probably a dealer will not give you 10 samples to select from?
Maybe some "famous" persons can make special agreements with Nikon to get "hand picked" lenses?

Here I can buy a lens and I can return it if I don't like it within.....I think 14 days if I deliver it back in original box etc.
But I would not be able to detect if a lens is decentered unless it is quite bad wide open. Else I would not know if the performance wide open is "normal" or it cold be better. Then I would need some samples to test and select from.

There is no perfect realization of a designed lens. It is a gradual matter. The map assembly for the 105mm f/1.4 shows an example of a very good lens sample. The full field MTF analysis is awfully sensitive to the slightest departures (say on the order of 100 nm = 0.1 micro meter) from the ideal. There may also be minor striations in the glass, same in principle as with bad sighting conditions in warm, damp air. Usually there are some small departures affecting the axial symmetry, and these are the obviously showing in the maps as not due to limitations of the lens design.

The maps in my article underline this story. In the case of an Irix 15mm f/2.5 (shown in supplementary) after my routine MTF test  for a incoming lens, I returned it at the predefined financial loss of 20%. I found it so bad across the board, that I did not trust that another example might have all the tolerances right. In the case of the Sigma 40mm f/1.4, I had the special opportunity to take 2 extra samples from the distributor for the purpose of the publication. And I was allowed to return whichever two I chose. So I kept the one with the least perturbation of high micro-contrast showing in the image center. After calibration, it became the best reference lens with the smallest deviation from the diffraction limit, see article for detail.

The trouble is, once you have seen a problem in the maps, you look out for it in the images. In the end I have tossed several lenses that were not satisfactory.

If you have the nerve to look at  the maps for your lenses, you might set up such measurements with some perseverance. It is not that difficult to do as you have mtfmapper to start with, and my article going into most of the fine points.  If you have a particularly interesting lens, we might discuss the options, and we could plan a measurement session on your next stop in Zurich.

A more accessible method for stronger decenterings is to take an  images of a printed Siemens star  at shooting distance say 30-50 x FL, in the center and near the 4 corners. If decentering strikes obviously in eyeball inspection you can take consequences. You may judge the lenses in your possession first, to gain experience.

[John Geerts]

In the case of an Irix 15mm f/2.5 (shown in supplementary) after my routine MTF test  for a incoming lens, I returned it at the predefined financial loss of 20%. I found it so bad across the board, that I did not trust that another example might have all the tolerances right.

It is a very strange lens-design.  You had to enter manually the infinity point of the specific lens.  I had the feeling this was not a very solid construction.  After a while I also sold the lens.

[MEPER]

There is no perfect realization of a designed lens. It is a gradual matter. The map assembly for the 105mm f/1.4 shows an example of a very good lens sample. The full field MTF analysis is awfully sensitive to the slightest departures (say on the order of 100 nm = 0.1 micro meter) from the ideal. There may also be minor striations in the glass, same in principle as with bad sighting conditions in warm, damp air. Usually there are some small departures affecting the axial symmetry, and these are the obviously showing in the maps as not due to limitations of the lens design.

The maps in my article underline this story. In the case of an Irix 15mm f/2.5 (shown in supplementary) after my routine MTF test  for a incoming lens, I returned it at the predefined financial loss of 20%. I found it so bad across the board, that I did not trust that another example might have all the tolerances right. In the case of the Sigma 40mm f/1.4, I had the special opportunity to take 2 extra samples from the distributor for the purpose of the publication. And I was allowed to return whichever two I chose. So I kept the one with the least perturbation of high micro-contrast showing in the image center. After calibration, it became the best reference lens with the smallest deviation from the diffraction limit, see article for detail.

The trouble is, once you have seen a problem in the maps, you look out for it in the images. In the end I have tossed several lenses that were not satisfactory.

If you have the nerve to look at  the maps for your lenses, you might set up such measurements with some perseverance. It is not that difficult to do as you have mtfmapper to start with, and my article going into most of the fine points.  If you have a particularly interesting lens, we might discuss the options, and we could plan a measurement session on your next stop in Zurich.

A more accessible method for stronger decenterings is to take an  images of a printed Siemens star  at shooting distance say 30-50 x FL, in the center and near the 4 corners. If decentering strikes obviously in eyeball inspection you can take consequences. You may judge the lenses in your possession first, to gain experience.

I have a number of Nikkor-H 50/2 lenses. Could be fun to check them out and see if there are significant differences between them. I was told if I got a good one it could resolve 400 lp/mm. Maybe a bit optimistic?
I used them on a film called "Gigabitfilm" many years ago and was amazed how much detail this film could resolve (and the lens also). In their marketing material they showed an analog x1000 enlargement from a 24x36 negative. I was told that the 24x36 was shot using a hand picked Vivitar Series 1 90/2.5. So a Tokina lens.   
Now I also learned something new.......a Siemens star......I had to look it up :-)
https://en.wikipedia.org/wiki/Siemens_star

[Bernard Delley]

A fun application of MTF measurements is in conjunction with AF fine tuning, it allows not only to set the fine tune value, but also to quantify AF consistency.

You set AF-fine-tune values from -20 to +20 in increments of 4 for example and perform auto-focus via optical viewfinder once coming from the a little defocus on the far and once coming from a little defocus from the close side. And you take record to where the focus fell on the ruler. This setup is sensitive enough that you can really know. From the results you learn the optimal AF fine tune setting. For a blur circle less than 20 micrometers at the sensor, you need to hit focus to better than 28 microns at f/1.4 with f/2.8 it would be to better than 2.8*20 = 56 microns.
The defocus can also be inferred from the MTF data from the planar target without eyeballing optimum sharpness on a ruler.
Such measurements can be summarized in a figure like this typical example for an AF-S 105mm lens at f/1.4 on a D850 :
[AF-S 105mm f/1.4 E ED at f/1.4 on D850, blue dot measurements are coming in from a little back focus, red dot measurements are coming in from a little front defocus. The scatter away from the regression lines is due minor focus inconsistencies because of residual mechanical friction and photon+electronic noise. The combo profits from an AF fine tune value of +10. This value is confirmed within +-1 tune units even years later, provided no damage occurred to camera or lens.]

[Bernard Delley]

I have a number of Nikkor-H 50/2 lenses. Could be fun to check them out and see if there are significant differences between them. I was told if I got a good one it could resolve 400 lp/mm. Maybe a bit optimistic?
https://en.wikipedia.org/wiki/Siemens_star

Bjørn Rørslett certainly rated this lens very high and mentioned excellent center sharpness. How the result of 400c/mm was obtained and details of procedure and results would be more interesting to me than just impressively high numbers. I think this 50mm f/2 lens was also shown in use on a bellows. perhaps it remains quite sharp for closeup use.  If so you could set up a test to compare your 50mm f/2 lenses: You image a razor blade fixed at an angle in a slide holder and analyze the raw image using the free open source program mtf mapper, which can give you an accurate result for MTF performance in a slanted edge test. I attach such results, posted at dpreview 2years ago for two closeup lenses at imaging ratio 1:1   : AI Micro Nikkor 55mm f/3.5    and AF-S Micro Nikkor60mm f/2.8

[Bernard Delley]

MEPER's mention of old lenses reminded me of my old pre AI Nikkor 85mm f/1.8 (K) which I bought 1976 or 77 and which was along on my kayak trip on the Euphrates river in 1977.  As a manual lens it hangs on as retired lens in the cabinet. It came up again in digital age for my first attempts at 35mm slide digitalization with a camera, where I needed a slightly longer focal length for use on the bellows. It turned out as a not so good choice for the purpose and went back to the cabinet.  Bjørn Rørslett rated it a 4 in his famous lens ratings. I realized that this non AI-modified lens can now be used on the FTZ adapter without problem and can be put through my MTF test.  It turns out to be a well centered lens with moderate focus shift and moderate field curvature at the imaging ratio of 32:1.  It has quite a bit of astigmatism, and a relatively fast falloff of meridional sharpness, typical for older lenses. The set of through focus (left front - right back) MTF maps also stopped down shown some further interesting features.