Since the introduction of mirrorless camera with shorter flange focal distances the usage of vintage lenses grew rapidly. With full frame bodies like the Sony A7R II or newer with 42 or 61 megapixel, it is interesting to know whether those lenses are still usable on those cameras and can be a cheap alternative for the manually focussing photographers out there.
I started using digital cameras with buying a Sony alpha 700 DSL-R with 12.2 megapixels. The successors with APS-C- and full frame sensors (alpha 77 and alpha 99) grew in pixel count reaching an interim peak within the Sony lineup in the Sony A7R II and A7R III having 42 megapixels. In September 2019 Sony presented the A7R IV packed with 61 megapixels providing a resolution only available to medium format cameras up to this point in time.
I really like using my vintage Minolta MC-/MD-lenses on my mirrorless cameras – a bit less since I discovered the Voigtländer E-mount lenses lately. Anyway, the results of the Minolta MC-/MD-lenses always positively surprised me, especially on the 42 megapixel Sony A7R II. Irritated by internet and print publications saying that the new Sony A7R IV definitely requires Sony GM lenses to match the requirements of the 61 megapixel sensor I tried to figure out which lenses are still good to go on the A7R IV and which are not – starting with some prime lenses.
Minolta MC W. Rokkor-NL 21mm F2.8
Voigtländer 21mm/1:3.5 Color Skopar ashpärisch
I already can hear you asking: „Serious? You want to compare a lens from 1973 – 1977 with modern, newly designed Voigtländer?“. Well yes, the Minolta is an good lens and was one of the best 21mm at its time. It already had „floating elements“ and therefore was a modern and state-of-the-art lens when it came to construction and design. Check out allphotolenses.com, if you want to know more about the 21mm Minolta.
The size comparison indicates, that in the past, optical quality was heavily depending on a more complex construction leading to a set of 12 elements in 9 groups for the Minolta vs. 9 elements in 8 groups for the Voigtländer. This, the shorter flange distance and being a half stop slower makes the Voigtländer a winner in reference to size – and weight, since it only weights 230 grams in comparison the 510 grams Minolta.
Contrast, sharpness and flares
But who cares about the technical specs – it’s the images that count, right? The following images all have been exported from the RAW using Capture One 12 without any changes unless mentioned otherwise.
The first image comparison is a good example to show the progress in less coating over the last year to reduce flares. Both lenses were stopped down to F4, the Voigtländer on the left, the Minolta on the right:
The Voigtländer 21mm / 1:3.5 Color Skopar easily handles the direct sunlight whereas the Minolta MC W. Rokkor-NL 21mm F2.8 has a hard time delivering an image with contrast. Let me clarify, that the flares and reflections are not caused by the inside of the mount adapter. The one used here is manufactured by Novoflex and is surely beyond any doubt. Stopping further down to F8, there is not much of a change for both lenses. The Voigtländer remains at a top level. The Minolta gains a bit of contrast and even the flares become clearly defined with edges.
I prepared some direct comparisons based on the 100% view of both images by creating screenshots in Capture One 12 at F4 and F8 from an area near the center and from the upper right corner:
Colors and bokeh
Let’s move over to color rendering and bokeh, of which the later is probably not really important when using a shorter focal length of 21mm, again the Voigtländer on the left and the Minolta on the right:
Well, I didn’t expect that the legendary color rendering of the Minolta lenses would disappoint me that much as the MC W. Rokkor-NL 21mm does here. The yellow of the car and the colors in the Abarth sign makes the Voigtländer a clear winner of the color comparison. When it comes to sharpness and contrast, I am surprised that the Minolta MC W. Rokkor-NL 21mm F2.8 can almost keep up with the Voigtländer on a 61 megapixel camera and when it comes to bokeh and blurry backgrounds I would even prefer the Minolta over the Voigtländer since the Minoltas bokeh is a bit softer.
Distortion, vignetting and other lens errors
Finally, let’s have a look on lens errors like distortion and vignetting, starting with F4 and then with F8. Don’t worry about the dark area in the upper left corner which is not part of the vignetting but a shadow of a lamp attached to the facade of the building. As previously, the Voigtländer is on the left and the Minolta on the right, starting with the images at F4
Apart from the slightly different color rendering and the missing contrast, the upper image part and especially the upper right corner shows a strong drop in brightness with the Minolta – nothing severe, since this is quickly corrected in Lightroom or Capture One. In reference to distortion, there is not much of a difference between the two. The Voigtländer looks a little bit better, which is not based on the lens profile correction for this lens, since the lens profile for this lens is still not available in Capture One 12. But there is one thing which makes the Voigtländer a winner in this category, which is the missing chromatic aberration, indicated by the green coloring of the Minolta image.
How does this change when stopping down to F8?
Interesting to see, that both lenses do not really behave a lot different between using them at F4 or F8, except that the chromatic aberration gets a bit better with the Minolta MC W. Rokkor-NL 21mm F2.8:
All images used for the 21mm comparison are bundled in this Flickr collection: https://www.flickr.com/photos/joerg_haag/shares/A3i612
Conclusion for the Minolta MC 21mm F2.8
This is a clear case for me: I would not recommend using the Minolta MC W. Rokkor-NL 21mm F2.8 on a camera like the Sony A7R IV with 61 megapixel. I used it quite often on a A7 I and II and would use it on the A7 III although the downsides of contrast in direct sunlight, flares, color rendering and brightness loss in the corners will remain with these cameras, although the later two can easily be corrected in post production. But on sensors above 24 megapixel I think it is a good advice to go for a modern lens with state-of-the art glass and design.
Minolta MD W. Rokkor 35mm F2.8
Sony Sonnar T* FE 35mm F2.8 ZA
35mm is my favorite focal length for any kind of photography except portrait, but for city walks, street and reportage this is the ideal focal length because it embeds the subject or action into a relationship with its environment.
If you compare the two 35mm lenses, the pure size comparison is a bit different from the two 21mm:
Ignoring the mount adapter these two lenses have about the same size and their construction is very similar. The Minolta MD 35mm has 5 elements in 5 groups and the Sony has 7 elements in 5 groups. Let’s check how different these two designs are in direct sunlight, first using an open aperture at F2.8 having the Sony on the left and the Minolta on the right side:
As expected, the Minolta MD creates a lot more sun reflections where the Sony Sonnar T* 35mm F2.8 ZA is not really impressed by the direct sun hitting the front lens. Let me repeat, that the Novoflex mount adapter is free from reflections caused by the inside tube of the adapter. So let’s have a look how this might change when stopping down to F8:
It’s interesting to see, that there is not much of a change. The Sony Sonnar T* 35mm starts to render a the first visible star rays. The Minolta MD 35mm gains a little in contrast which also makes the flare in the lower center of the image a bit harder and more recognizable.
Another interesting thing about these two lenses is the sharpness at F2.8 and F8 (below), seen in the 100% view:
Keep in mind, these images are shot on a Sony A7R IV with 61 megapixels. They show details of the following images at 100% view. Have look at the bokeh at F2.8 and F8, Sony FE 35mm at the left and Minolta MD 35mm at the right.
Astonishing, right? The Sony FE 35mm is a little sharper, which can be caused by my wrong manual focussing (I doubt it), but I like the colors and the bokeh of the Minolta MD 35mm a bit more, although the images are very close.
The last thing to look at is a comparison of vignetting and distortion.
F2.8, with the Sony on the left and the Minolta on the right:
The center of the images, both at F2.8 and F8 a clear and sharp. Neither lens shows distortion or vignetting, although the Sony lens takes advantage of an existing lens profile where the Minolta MD lens doesn’t. One interesting detail and difference is the sharpness of the lower left corner, which gets better with the Minolta stopping down from F2.8 to F8 where the Sony is loosing sharpness. Maybe this is caused by a curvature of field or just because I have a slightly bad copy of the Sony FE 35mm. In real life photography, I have not yet experienced any problems in my images.
All images used for the 35mm comparison are bundled in this Flickr collection: https://www.flickr.com/photos/joerg_haag/shares/Cy2256
Conclusion for the MD W. Rokkor 35mm F2.8
The Minolta MD 35mm F2.8 gets all thumbs up when it comes to sharpness, bokeh, vignetting and distortion and if you have one: try it and keep it if you rate it good, but try to avoid flares.
Using these older ladies on a modern mirrorless camera with high pixel counts is a matter of luck. It can give you very good, astonishing results like with the Minolta MD W. Rokkor 35mm F2.8, but it can also be a sort of a disappointment, like it is in some concerns referring to the Minolta MC W. Rokkor-NL 21mm F2.8. If you can, test before purchase.
Latest posts by Joerg Haag (see all)
- Testing two 45-year-old wideangles on 61 Megapixels (guest post) - January 14, 2020
52 thoughts on “Testing two 45-year-old wideangles on 61 Megapixels (guest post)”
Very interesting article!
I always wondered what the “older lenses don’t perform on 64MP” was supposed to be about. AFAIK 60+ MP still isn’t outperforming film, or is it? Anyways, I’m glad the 35mm performed great, since I love this little lens, I have one in the MDII version (I think) with 55mm diameter and it’s astonishing on the A7II – nice to know that it won’t let me down in the future either.
i thought 35mm film was about 12MP?
From the diafilms/negatives i’ve scanned i’d say thats about right.
But then, i’ve never seen/compared a big print from 35mm film
That sounds way too low. I just googled a bit and found this:
No idea how much water it holds but the mentioned 87 MP doesn’t sound outrageous to me either. I don’t have a 8000 DPI or even 5700 DPI scanner to do test with though 😉
well, thinking of microfilm used in archives i’d agree it’s amazing what detail can be captured with those chemical/optical processes. But from what i’ve seen being produced by common SLRs 35mm film in the hands of amateurs i don’t think it’s that amazing.
Definitely is scanner dependent. Scanning 35mm color film on an Epson V550 I wasn’t capturing even as much detail as my Sony NEX-3 was. My 16mp m43 cameras certainly surpassed it. Not to mention 24mp 35mm DSLR and mirrorless cameras.
However, scanning b&w medium format film (645) was a bit of a different story. It was very detailed. I’m sure on a nice scanner it would be providing tons of detail that would be beyond most top of the line sensors today but even on the that cheap flatbed scanner it was definitely at least as good as my Sony A7 and probably a bit better and closer to the A7Rii.
^ That’s what I was thinking, thank you for the insight!
The good thing is that unless you plan to crop heavily all the time and/or printing literal wall papers it doesn’t really matter. I just think that there’s no such thing as sensors outresolving older quality lenses.
I used to have a Leaf 45 scanner as a student, which is slow, but super high quality. If I have to estimate I got about 15mp of resolution out of a 35mm dia film. I can imagine very slow film can capture more detail.
Back in my film days I was shooting Kodachrome 64 and Kodak Tech Pan b/w (essentially microfilm processed for continuous tone). I used a Minolta Dimage 5400 scanner (5400DPI) that autofocused each scan. On the Kodachrome I was seeing the film color grain resolved in the scans (silver was removed in development). On the Tech Pan I was seeing the film grain resolved. Printing or scanning the Tech pan, it was noticeably higher resolution than any other film I ever encountered.
The general advice I recall from that time was that, for transparency film, there was nothing gained by scanning at more than 2400 dpi.
When my Minolta died, after Sony bought Minolta, discontinued scanners, and there were no longer repair parts available (sigh!!), I bought a Plustec 8200i Ai, that I end up swearing at every time I use it, for memories of what the Minolta was capable of.
The Voigt just destroys the legacy lens. Was not expecting such an enormous difference in literally every area of performance.
Thanks for testing!
I used to run Minolta SLR’s, had the 21mm lens, also 28, 35, etc.. Your pictures look like my recollection that the 21mm images went blurry in the corners.
I do have the Minolta 35/2.8 MD-III and love it. Excellent lens, as your test demonstrates. Some (mild) criticism though: some of your photos were not taken at exactly the same distance (or are these 35mm not exactly 35mm?) and the last photos were exposed slightly differently (the Minolta are a little bit overexposed). But very interesting tests anyway, thanks.
Thanks for your comment. All photos where taken at exactly the same distance and with the same metering. 35mm is not always 35mm. Differences in exposure may be due to either clouds changing the light situation or slight differences in real F- or T-values (T for transmission).
An (obvious) big draw for me and using the vintage glass is the pricepoint. I just bought the minolta 28mm, 50mm, and 58mm for a fraction of the price of one of the new Voightlanders. I’d love to use some of the new manual glass, but didn’t Voightlander just release a new 50mm lens that was like a grand?! I’m sorry but that’s just not reasonably in my opinion. I’m sure others will disagree, but there’s no way I’m dropping that kind of $ on a manual 50mm lens….
A warm welcome with wishes that this comparison will drive you into the team! Bravo and Success!
Thank you Bozee. It may be not my last post here on phillipreeve.net….
Thanks for the interesting article. My 21 3.5 Voigtlander is one of my favorite lenses. I love the size, build, and handling and the images I get out of it are constantly excellent. I have fancier, faster, and more expensive lenses, but none provide me the pleasure of that lens.
The old knowledge is just reiterated in this article: that on modern mirrorless, you can use vintage glass really well from 35mm focal length onwards. For 28mm, you have to pick carefully the ones that work (e.g. Pentax-K 28mm f3.5), and even these have to be stopped down very much to be sharp across the frame. For FL <28mm, vintage glass is not the greatest idea, use modern glass instead for consistent results.
The Canon FD 2.8/20mm is actually not that bad if you know how to handle the field curvature you can get decent results @f11. Flare resistance is quite bad though
Yeah, but isn’t flare resistance equally important to sharpness on ~20mm FL? It can be sharp and tiny, but if the whole image is ruined by flares and lack of contrast I don’t really see the point. Not for the current price point of vintage UWA lenses, no.
This “old knowledge” mostly refers to rangefinder lenses. SLR wide angles with retro-focus construction always caused less problems. It’s just what this comparison revealed: The manufacturers had learned to make really great lenses in the 35mm to 200mm range, but wider angles were still difficult. Even in the 80s, some 24mm lenses were called “ultra wides”
beyond 100mm CA became a real issue though and only a few lenses used ED-glass to correct for it so I would say that we have pretty good lenses from 28mm to 100mm.
While the fixed focal lenses up from 100 to 300 mm show a significant amount of CA, the RF lenses (250 and 500 mm) do not show any (as well as the zoom 100-500 mm f/8 APO, the only Minolta manual lens that used a ED glass element).
This should a mandatory read for those looking for UWA lens and considering to buy vintage one. Just DON’T.
I mean, it’s ok to spend under £$€100 but current prices are ridiculous, for example – highly regarded MC 21mm Zuiko from Olympus goes for about £350 shipped. A little more than that and you can get equally tiny modern Voigtlander 21/3.5 for e-mount which is Loxia territory if it comes to performance. Why bother then?
28mm to 50mm range is a different story and, as you can see, for peanuts you can get a decent performer.
This is my route: spend more on modern UWA and save on 28mm to 50mm lenses.
Modern lens coatings are really impressive!
Thank you for contributing to this blog!
I’d be interested in sharpness at infinity of those older lenses on high-MP Sensors.
I’ve heard people claim older lenses are not able to resolve more than 24MP but i doubt that. Especially in case of the 35mm lens i think it might yield a more than adequate corner to corner performance, even on the a7Riv for landscape.
look at the performance of the FD 2.8/35mm @f8 on 24MP
it’s so good, theres definetly some margin left, if you ask me
I’ve tested the nFD 3.5/135 on my a7riii now and stopped down a bit it still looks fine on 42 MP 🙂
Why is it so bad if your sensor outperforms your lenses? Suppose I upgrade from my current a7 to the a73 or the a74. Are the pictures using, say, the canon nfd 20 f2.8, going to be worse because the sensor is better? Not really! Whilst I won’t be using the sensor to its full potential, the image quality, all other things being equal, is going to be better, and while it’s true that the lens’s shortcomings will be more easily detectable if I pixel peep, that is because I will be looking at a larger image. So, if you like your lens before upgrading, you should definitely keep it and just be aware of its limitations, but in terms of the pictures you’re going to take, the lens should perform actually better in the new sensor.
For me the point is rather if the old lenses can use a 61MP Sensor to its full potential or not.
What newest sensors offer is more resolution and even a cheap Minolta MD 50mmf2 will probably give you incredible detail corner to corner (stopped down). In my opinion there is no threshhold at which new digital tech outperforms good old optics.
But of course if you invest in an A7riv you should probably rather invest in a CV 2/50 APO instead of a MD 2/50, i just think depending on your usecase (e.g. hi-res landscape) you probably don’t have to…
actually the comparison CV 2/50 vs. MD 2/50 on 61MP i’d really like to see. No question the CV will win in every regard but in terms of sharpness i wonder by what margin.
Look at Loxia 2/50 vs MD 2/50 for example:
Thanks for your work. I really enjoyed reading it. What surprise me is that lower contrast indeed make the bokeh smooth(F4 bokeh comparison).
I also use A7r4 and some old lenses.I hope to read your artcles more often.
Very interesting article. How about Minolta MD 35mm 1.8? 🙂
Well, I would be interested in that comparison as well, but I don’t have that lens and I don’t have a FE 35/1.8 either.
Did you check the length of your Novoflex adapter?
All my Novoflex adapters have been 0.2-0.3mm too short.
Since the MD 21 has floating elements, there is a very good chance that the presumably too short adapter is significantly degrading the lens’ optical quality, especially in the corners.
In the case of one of my UWA lenses, the aperture at which good corner sharpness was reached went from f/11 to f/5.6 after shimming a too short Novoflex adapter.
I’m not sure if this is the cause. Even, if the lens adapter would be short by 0.2 – 0.3mm, that would only affect the far distance objects. Additionally, I’d expect Novoflex having the highest accuracy available in the market and since I use the same Novoflex adapter for all my MC/MD lenses, I would have seen this degraded lens quality also with other lenses, don’t you think.
Many/most adapters are too short by design, so that even lenses which are slightly out of calibration can focus to infinity.
Of my 10 or so Novoflex adapters, ALL have been too short by that amount.
What happens with a lens having floating elements, the more so the shorter the focal length, is that the floating elements align (optimize) themselves for the wrong distance, since the focusing scale is also set for the wrong distance. That means that the entire optical system isn’t working as designed. Basically, when the lens is focused to infinity, the focusing scale AND the optical system will be set for about one meter…
OK, but this should not matter if using focus support with 11-times magnification on the Sony A7 RIV.
Of course it does!
Maybe have a look at some of our former articles covering the influence of adapter thickness on the image quality of lenses with floating elements.
The point is not that you get a focussing error, but as Rollei Nut correctly says, the elements in a floating system will be non optimally arranged if the adapter is too short, so that even when correctly focussed, the quality will suffer.
Thank you all for the clarification. I should write more articles here, since this will add to my knowledge as well.👍
You won’t notice it with lenses without floating elements, which will be most of them.
And yes, all the Novoflex adapters I tried are also too short (still better than some K&F adapters).
Many thanks for the review and the other materials on this site – I enjoy reading here a lot.
Re. adapter length:
For some Leitz R lenses, I’m using a K&F adapter which as per my measurement is between 0,15mm and 0,20mm (measured on oppsite sides) too short. All my lenses focus beyond infinity with the adapter but are spot on with the Leicaflex camera.
I reached out to Novoflex to ask if their adapters would be more accurate. Their reply was that while their tolerances are *significantly* smaller (-0,03mm) they still purposely make the adapters shorter by -0,02mm two safely allow infinity focus. However taken together with 0,05mm (five hundreths) this is significantly better than what K&F adapters deliver with about one to two tenths mm of play.
Another aspect to look at is how large are the production tolerances at all on the camera bayonets and is it worthwhile to worry about several hundreths of a mm?
I checked with Leica and they replied that the R-mount tolerances were very tight indeed with about +/- 0,02 mm. I’m not sure about the Sony e-mount; it might have more play given that it’s designed for AF lenses, but I think that the Novoflex adapter (for R mount) is about in the same tolerance range as the original camera mount production tolerances were.
Also, I wasn’t too worried about the adapter being short from a optical standpoint since the old lenses I use don’t have floating elements, but a) I want to use them on both cameras and if I adjust infinity for the Sony, I won’t get to it with the Leicaflex and b) the fact that K&F tolerance varies around the circumfence might well affect the focus accuracy across the frame.
Long story short – I’ll receive the Novoflex adapter next week and will see how it performs. If interested I’ll be happy to report back then.
I tested my ancient Canon FL 55mm f/1.2 wide open on the A7Riv a few months back. If anyone’s interested you can see the results here – https://jamesclark1991.wordpress.com/2019/11/16/sony-a7r4-tested-with-canon-fl-55mm-f-1-2/
I think it held up really well considering the age and f stop.
Did the coating improve between the 21 mm NL and the MDIII 35 mm which was issued years later?
The 35 mm seems to be holding better than the 21 mm not only in sharpness but also in colour rendering.
“Of course it does!
Maybe have a look at some of our former articles covering the influence of adapter thickness on the image quality of lenses with floating elements”
I went to the former article on tuning adapters and found a huge impact on flare, etc… this raises a question about the adapter used for this test especially for the 21 mm.
(both about the focusing and the flare). Is this lens so bad or has the adapter had an influence on the comparison?
I wonder too.
I suppose it is the lens, since I used the same adapter with the 21mm and 35mm. If you want to find out yourself: I used a Novoflex adapter.
The whole point is that the 21mm has floating elements (and is a much shorter FL as well), so a too short adapter will make a difference.
Do a simple test: visually focus the 21mm on something far away (operational infinity) and see what the focusing scale shows. If it’s on or close to the infinity mark, your adapter should be close to the right length. If the focusing scale is closer to the 1 (or 2) meter mark, that means that the adapter is too short and is very likely negatively affecting your results.
If you shim your adapter (typically by 0.2-0.3mm) so that focusing to “infinity” corresponds to the infinity mark on the focusing scale, you can do a “before & after” comparison to see what difference a too short adapter makes…
Thank you for your thorough review! You (along with the reviews on Fred Miranda) have inspired me to get a 21mm f/3.5 Color Skopar to test against my 21mm f/3.5 MC Zuiko and 20mm f/4 SMC Pentax K lenses. I have high hopes for the Pentax, because of its relatively slow maximum aperture combined with a complex 12 element, 10 group construction, but I would be delighted if the Voigtlander lens proves to be better.
On another note, I previously tested my Novoflex adapters with and without Protostar black flocking material, and found a slight improvement in contrast and reflection control.
Well, I just received the Voigtlander 21mm f/3.5 Color-Skopar this weekend. I promptly used an A7RIV to run a set of tests against the Pentax 20mm and Olympus 21mm lenses, using some targets on a bookcase illuminated with 3200K lights,.
The results were a little disappointing. Although the Voigtlander was slightly sharper than the other lenses across the frame, and also had less chromatic aberration, it suffered from pronounced vignetting, the worst of the group. The Pentax lens was slightly sharper than the Olympus in the middle of the frame, but even at f/8 the image deteriorated considerably when moving towards the corners. After a white balance, images from all the lenses exhibited similar colors.
So . . . rather than testing for flare, infinity performance, bokeh, etc., I have decided to return the Voigtlander, sell the Pentax, and keep the tiny Zuiko in my camera bag for now, at least until I test a 21mm Nokton or a Loxia .
The NL was released in 1971 and thus 50 years old in two months. Based on sample images, the lens used for testing seems to have issues, possibly haze, separation or manufacturing problems. The contrast is not as depicted.
actually that 21 appear to be a bad copy.
never seen color that bad from a minolta lens