Tested on Micro Four Thirds

by Ralph Eckstein, September 2016



  • 1. Lenses

    Prices asked for new manual lenses hindered me to include any of them in my comparison, thus I stuck to older second hand lenses.

    I collected at least two copies of all lenses, with the following exceptions, of which I only had one of each:
    – both Cosina versions
    – both Leica versions
    – both Mamiya 1.4 versions
    – Asahi SMC Pentax 55
    – Asahi SMC Takumar 55
    – Exaktar
    – Hanimex
    – Miranda Auto TM
    – Nikkor Ai-S
    – Porst Auto F
    – Voigtländer 1.4 55 (though, I had the identical Rolleinar MC)

    I also returned following lenses, thus not included in the test:
    – Minolta MD 1.2 50 (both copies had minor issues)
    – Porst 1.2 50 (all three copies defective and optically mediocre)
    – Carl Zeiss Jena Pancolar 1.4 50 (scratched and optically mediocre)
    – Meyer-Optik Görlitz Domiplan 2.8 50 (both copies defective)

    If the duplicates differed in performance in my pre-test comparison, I bought at least one other copy, or even more, to minimize the issue of testing an "exceptional" lens, but I couldn't find major differences in most of the doubles or triples, therefore concluding serial production in general seems reasonably consistant.

    A rare – but even more surprising – exception in the latter regard was Canon! Each of the four tested versions of their lenses showed significant variations in optical performance – from mediocre to very good. My advice – especially with Canon lenses – is to test several copies, before keeping one in your bag!

    Another problem lens was the Carl Zeiss Jena Pancolar 1.8 50 – I had six copies, of which only one showed decent optical results, three of them were mediocre, the other two bad. For the test I had to use one of the mediocre ones, as all other ones were desribed as "clear and without scratches" by their sellers on ebay, but actually had lots of scratches and/or dust and I returned them immediately.

    Then there were the Mamiyas, of which I had 17 in total – and all of them were excellent performers! The only exception were both 1.4 versions, of which I only could get one of each, and both were in bad condition, one of them was even optically bad. During my test I could not figure out to use any other than open aperture with my Mamiya adapter – so there are only wide open images available. The latter applies also to the Canon S.C. and S.S.C. models.

    Minoltas, of which I had the most (23 in total), due to their overwhelming model variety, showed consistantly good results, but were not among the absolute top 10 performers. That said, you cannot go wrong with a Minolta lens, especially at apertures smaller than 2, where they usually have excellent resolution in the whole frame, and a good contrast.
    I wasn't able to get one of the 1.4 MC Rokkors for a decent price, but their optical performance shouldn't be superior to the MD versions anyway.

    I also wanted to include some of the occationally highly praised Tomioka lenses – or what is said they are, branded in a variety of names, like Alpa, Beroflex, Carenar, Chinon, Edixa, Exaktar, Panorama, Porst, Reflecta, Reflex, Revuenon, Rikenon, Sears, Soligor, Vivitar, Yashinon, and their own brand Tominon. I bought many of them (mostly 1.2 55 and 1.4 55), but returned nearly all of them, as they were usually dirty or scratched from inappropriate inner cleansing. And all my pre-tests of Tomiokas showed mediocre results anyway.

    I always kept the best copy of each lens for my test. The only exception, as mentioned above, was the Pancolar 1.8.

    Fugus was another problem: Though I only got three lenses with severe fungus, I got more than twenty which were slightly infected. I returned almost all of them. On three of them, which the seller refused to take back, I attempted to clean them, but wasn't too satisfied with the result, as it is almost impossible to clean lenses at home without having at least a little dust residue.

    2. Adapters

    Initially I testet two copies of several adapters, as precision issues seemed possible, but couldn't find any differences at all. They all worked properly.
    The only exception concerned both Rollei QBM adapters, which I had to correct by bending the adapter mounts, due to the fact that the lenses were wiggly, and further the Mamiya adapter, at which two screws were loose. By the way, the Mamiya adapter was the most expensive one.

    3. Thoughts on "Typical Users of Brand X"

    Interesting is the fact, that from all the lenses I gathered, I could draw a picture of "typical users" by brand:

    Most Olympus lenses were scratched (I had to return several, as they weren't described properly on ebay), most of Canon and Nikon lenses were bumped, most of Asahi lenses had lots of dust and fog, and most Minolta lenses were quite fine.

    Does that suggest, Olympus users are "easy life types", Canon and Nikon users are "action types", Asahi users are "adventure types", and Minolta users are "occasional shooter types"?

    Or does it suggest, Olympus coating or glass is soft, Canon and Nikon barrels are weak, Asahi sealings are leaky, and Minoltas are perfect?

    Probably neither of that ... but nevertheless an interesting observation.

    4. Method

    Using image correction software on digital photos enables to correct lens issues like contrast, distortion, vignetting and chromatic abberation to a very high degree. Hence, my attention is focused on resolution, bokeh, and flare.

    Therefore I only tested the most important point for me – resolution. Because I did all the work in my spare time – collecting, testing and comparing way over 200 lenses is a big deal, which, by the way, cannot be found anywhere else, neither on the web, nor in any magazine.

    I want to remind that on Micro Four Thirds there is the "crop factor", meaning these full frame lenses are all used on their "sweet spot", and, naturally, the effective specs are doubled, meaning a 1.4 50 mm lens acts like a 2.8 100 mm lens on Micro Four Thirds sensors.

    I used a lens test chart of 516 mm in diagonal. I kept the shooting distance for all lenses the same – even for the 52 mm and 55 mm lenses, thus the samples differ slightly in size.

    I shot the lenses at up to three aperture settings: wide open (from 1.4 to 2.8, depending of the lens' specs), at 2 (if printed out on the aperture ring), and at 2.8. Smaller apertures are not relevant (at least not for me), as any lenses perform almost equal at apertures 4 and above.

    This leads me to a short distraction: I recently read an article by Ken Rockwell, called "Lens Sharpness":

    "No one shoots at f/2.8 in decent light. Stop down and the lens is marvelous."

    "Only idiots shoot wide-open in daylight."

    To what kind of photography is Rockwell referring?? ...
    For sure, not portrait photography, nor sports photography, nor art photography ... even nature photography often benefits from a shallow depth of field. (No, I do not dislike Ansel Adams' work – on the contrary!)

    Is Rockwell referring to casual point-and-click photography with zoom lenses, shot at f:8 and ISO 800, everything almost sharp, no need for a good 1.4-lens ... ? But okay, I am an idiot, searching for lenses with high resolution to shoot wide open in daylight. Back to work ...

    I fixed the test chart to a board on the wall, mounted the camera on a tripod, determined the exact middle point with a mirror, put all the lighting on (though, on purpose, not too evenly spread on the test chart), set the camera to manual exposure (using 1/160th of a second for all shots to compare real brightness – resulting in darker pictures with closer apertures, naturally), ISO 200, electronic shutter, 2 seconds release delay, and manual focus with magnifier and indicator, focusing on the middle of the frame. By the way, my entire photographs are taken in 3:2 RAW AdobeRGB format principally, but the samples for the Web are converted to JPEG "100 % Quality" full size sRGB in Adobe Lightroom.

    I made one shot at each aperture with every lens. This was accurate enough, as I was focusing really pedantic, and I never touched anything whithin the release countdown. I also could not find any questionable images within the entire shootout, which wouldn't have matched my pre-tests and after-test.

    I loaded all the images into Lightroom, using no image correction added. Yes, I know, all RAW coverters have to interpret the files somehow, and each of them doing that differently, hence, whatever is seen are never real raw images. Anyhow, this is completely fine, as these basic corrections are applied equally on all samples, due to the use of the same camera and image software. There is additional further correction applied on electronically coupled Micro Four Thirds lenses, but I have used my Panasonic 42.5 mm only for personal comparison – I did not include it into this legacy lens test.

    5. Results

    ... are matching common expectations: the clear winner is the Leica Summilux 1.4 50.

    But: considering purely resolution, it isn't at all that clear. Regarding just the 1.4 lenses open wide, Konica and Olympus have the best center and overall resolution, Agfa and Fuji following behind, and only then comes the Leica. And: the Summilux isn't flawless at all – the image corners and borders are mediocre, even on this crop factor!
    But another But: Leica has by far the best contrast – and this weights heavily regarding image impression. The very low contrast of the Konica and Olympus images cannot ever match the Leica's very high contrast, not even when applying extrem image correction, not to mention completely uncorrected images, which, obviously, are still widely used – on one hand by film users, naturally, and on the other hand even by digital users. This leads to the conclusion:

    The Leica Summilux is the best lens in this comparison. But also the most expensive by far, ranging € 500 – 1500. All the other 1.4 lenses are ranging € 50 – 150.
    That is a price difference – a tenfold difference, to be clear. Is it worth?

    If you're into wide open shooting – the Summilux is it (especially on film).
    If you usually shoot digital at 2.8 or smaller – nearly all other lenses are good, especially all Minoltas and Mamiyas, which are good even wide open (at 1.7 to 2.0), and very good at 2.8 or smaller.
    The Carena and the Pancolar are not usable up to 2.8.
    The Tessar is not usable wide open (2.8).
    The Exaktar is not at all usable – in fact it is that bad, that there must be a defect, like alignment issues, or else.
    The Canons produce quite high contrast (almost like the Summilux), but in terms of resolution they stand clearly behind.
    The Nikons were disappointing – none of all my copies impressed me.

    All lenses auctioned off on ebay Dezember 13th, 2016.