Sony's MTF figures (even for the ones with the blue Zeiss sticker on them) are pointless and mostly fantasy land. Searching on Google will bring up a thread or two with a response from Dr H Nasse on what he thinks about Sony's published MTFs.
What we need are real tests from pros and trusted members of this forum. Guy, Mark, Tim and others on this forum for example.
-- Eeraj
I found this:
Zeiss FE 55mm lens is on par with the Zeiss Otus at medium aperture! | sonyalpharumors
Date: Mon, 29 Jun 2009 12:56:47 +0200
From: “Hoenlinger, Bertram”
Dear Mr. Lau,
Thanks for your request.
The MTF data for all Sony lenses (including the Carl Zeiss ZA lenses) are created and published exclusively by our partner SONY. We do not have had influence on the published MTF charts.
So please contact Sony directly about the measuring methods and further details of their MTF data.
Best Regards
Bertram Hoenlinger
Reply – 2
(From Dr. Hubert Nasse who had published the guide for reading MTF diagrams in Camera Lens News)
Date: Mon, 29 Jun 2009 14:14:05 +0200
From: “Nasse, Hubert”
Dear Mr. Lau,
the curves which are shown in some Sony brochures have been “processed” with some freedom
by the graphics designers of the brochures. They are scientifically not valid, easily to be seen in
some cases where the figures exceed diffraction limits. So they should not at all be compared to
the data which are published by Carl Zeiss.
With best regards
Assuming these letters are legit, they say only the obvious: every manufacturer uses a different methodology. Do we know how Zeiss tests their lenses? Do they have a double-walled test chamber ammortized from external influences?
Is Sony's marketing to be trusted? Hell, sure, like any other salesman.
Are "Guy, Mark, Tim and others" better than Sony and Zeiss in testing lenses?
I don't know. Are they to be more trusted than corporations? I think so.
The thing I know here is that lens testing is more art than science.
MTF charts give an idea, same as a car's specs.
When I see a graph reaching 100% contrast for 5 or 10 lp/mm I know it's "optimistic". No real lens has 100% contrast. But also 5 and 10 lp/mm are pretty meaningless. "Creative graphic design" is more difficult for 20, 30 or 40 lp/mm, where it counts. In cinematography I am interested in 70lp/mm contrast to be around 70% (Zeiss Master Primes). Thats where the picture looks really "sharp" on the screen.
As for "independent" lens testing... Arguably the most popular photo site in the world for years had lens tests that often showed system resolution higher than Nyquist limit (more lines than sensor pixels). All of their tests were worthless because they misunderstood the definition of Nyquist limit.
DxO today claims that Sony FE 55/1.8 has T=f=1.8.
Possible? A lens with no transmission losses? Only if it has a built in amplifier.
PS. Not long ago one CERN team published a "discovery" of some neutrinos travelliing faster than light. After a year's re-testing the culprit was found--uneven length of cables carrying data.