Variables of MTF lens testing
Variables of MTF Test
I bought a 1954 FED1 with an Lens 4871 Serial number Coated 50mm f:3.5 FED , an Elmar 5cms f:3.5 Copy.
I want to test the lens MTF.
I found from internet 48x31 cms Resolution Chart which prepared to ISO Standarts.
I will order an printout next week from a printer which has a large selection of big industrial printers . I will take pictures of chart for each aperture and order a HR 17MB Tiff Scans and test the big H printouts at the chart with MTF Extraction Software.
Scanner resolution are the two things to itch me.
Do they effect the MTF test and what is the way to overcome this problem without sending the lens to Zeiss option
How about using the chart mentioned in this website and a high power loupe and estimate the MTF yourself as demonstrated.
If you want to scan, here are some guidelines from that site:
Scan the film image. Observe it on your computer and analyze it using ImageJ or PixelProfile. Works well with 4000 dpi scanners and digital cameras but doesn't quite cut it for measuring lens resolution with 2400 dpi scanners. If you print digitally this technique has the advantage that it accurately represents your imaging system. Using the |sinc|3 approximation for a properly functioning film scanner, the 50% MTF frequency (f50) is about 0.366*dpi (or dpmm), and the 10% MTF frequency (f10) is 0.629*dpi. For a 2400 dpi (94.5 dpmm) scanner, f50 = 34.6 lp/mm and f10 = 59.4 lp/mm. Not impressive. For a 4000 dpi (157.5 dpmm) scanner, f50 = 57.6 lp/mm and f10 = 99 lp/mm. This is good enough to distinguish between good and excellent lenses, but I'd want more if I could get it. Sharpening boosts a scanner's f50 and f10, but it can confuse the results if you don't apply it with consistency.
I posted this today at another site. This is the only meaningful test I have devised for my lenses:
"My criterion for judging the sharpness of a lens is to make one negative, focussed at infinity on something with a good deal of fine detail. It could be tree limbs, bicycle wheels, signs, almost anything. Then I look at the negative with a dissecting microscope. If it renders the fine detail, it's sharp for my purposes, since few hang photos of test charts on their walls."
This isn't the best way to arrive at a quantitative comparison, though.
Last edited by E. von Hoegh; 07-19-2012 at 06:18 PM. Click to view previous post history.
My criterion for a lens to get anatomical detail from the portraits or few meters faraway human subjects. Skin details , muscles like an Greek sculpture drapery and below the skin the blood flow. For that color saturation , I test the subjects under oldest technology light bulbs at night. I must see the clean red blood flow and green blood flow. I only get this result from Leica lenses , especially Summitar. And my Rollei 35S with Sonnar 40 have a similar nature.
When it comes to scanning and dpi thing , I hate that stuff. I dont think their printers are not well calibrated and free charts would print bad. Buying a chart is more expensive than original Leica .
Sponsored Ad. (Subscribers to APUG have the option to remove this ad.)
Ok, since I've already responded nastily to one !@$ today I'll do it again.
Mustafa, what does MTF have to do with what you want the lens to do for you? Why screw around with trying to measure MTF when you can simply take a few shots and know whether the lens will do what you want it to? And why do you think a Soviet-made imitation Elmar will do what a Summitar does? Why not just use your Summitar and get on with your work?
I've said it once today, I'll say it again. There are too many damfools here.
I wouldn't try to measure MTF with some kind of chart you found on the internet and printed outtoo many variables in the process to allow for a measurement of any accuracy. I would suggest you purchase an ISO-12233 chart or similar and use slant edge MTF (although this will require that you digitize the image).
That being said, I completely agree with Dan. Why the do you need to know the lens MTF anyway?
MTF curves are irrelevant in practical photography. Sure, you learn about these in art school with photography major, but other than that, pffft. The real thing to concentrate on is quality of your work, knowledge of your subject, an holistic and consistent approach to presentation and disregard for distractions that serve no useful purpose other than to get you hot and bothered when that is not what photography is about.
The photographer must determine how he wants the finished print to look before he exposes the negative.
Before releasing the shutter, he must seek 'the flame of recognition,' a sense that the picture would reveal
the greater mystery of things...more clearly than the eyes see." ~Edward Weston, 1922.
Have you tried http://www.quickmtf.com/ ?
The idea of measuring MTF by shooting charts is somehow flawed at its root. A proper MTF test should be performed on an optical bench.
I've read somewhere that MTF tests are heavily influenced by the quality of lights, so that proper results are obtained by a properly equipped laboratory, and results are not even constant between laboratories for light quality problems.
You don't need to send the lenses to Zeiss, any laboratory equipped for this kind of tests will do, they should be a bit everywhere, this is 1940 technology after all. I have no idea of how much this would cost.
You're correct. I think the OP wants a quantifiable way of comparing two lenses, which is fair enough.
Originally Posted by Poisson Du Jour
But, comparing MTF reminds me of the audiophools, who sit around comparing unquantifiable qualities to the extent of forgetting that the gear was meant as a way to hear music.
Last edited by E. von Hoegh; 07-20-2012 at 03:29 PM. Click to view previous post history.