Minolta AF DT 18-70mm F3.5-5.6 D A-mount lens review by Miranda F

reviewer#34758 date: Jan-21-2017
sharpness: 4
color: 4
build: 5
distortion: 4
flare control: 3
overall: 4
tested on:
  • film camera:Film camera
  • APS-C: 6MP6 MP; 10MP10 MP; 12MP12 MP; 14MP14 MP; 16MP16 MP; 20MP20 MP; 24MP24 MP
  • full frame: 12MP12 MP; 24MP24 MP; 36MP36 MP; 42MP42 MP; 61MP61 MP
ownership:I own this lens
compared to:Minolta 50mm f1.7 (RS & original)
Minolta 28-80
Minolta 28-85
Minolta 28-100
Minolta 35-70 (f4 & f3.5-4.5)
Minolta 75-300
Minolta 100-300 (original & APO)
Sony 18-55mm
Sony 30mm f2.8
Sony 35mm f1.8
Sony 50mm f1.8
Sony 55-200mm
Sony 55-300mm
Sigma 10-20mm f4-5.6
Sigma & Tamron 28-200mm
Tamron 90mm f2.8
Vivitar 70-210
Various MF lenses
price paid:Basically free.
positive:Cheap, compact, lightweight, useful range;
Very little CA.
Essentially free from vignetting & partially useable on FF
Acceptable sharpness everywhere and excellent at 18mm. Centre always good and corners never particularly bad.
negative:A little soft in the 24-28mm region and again at 70mm.
Slow aperture (f5.6 from 50-70mm)
Busy bokeh at times.
Sometimes terrible flare at the long end, not helped by useless hood.
comment:I got this lens with a Dynax 5d very recently, and hoped to keep it with the camera as a normal/wide zoom because it has the right focal length range and because as a collector and enthusiast for older cameras I like to mate my cameras with lenses of their correct make and historical period. It came basically free with a Dynax 5d as the shop refused to separate them, claiming they wouldn;t be able to sell the lens on its own! [LOL]
After reading the reviews here I wasn’t optimistic of its performance so immediately did some stress testing with trees or buildings covering the visual field, to see just how bad the reported softness and CA really was and whether I needed to source a better ‘period’ Minolta lens immediately.
And I was quite favourably surprised - it isn’t bad at all!
Let me say immediately that I tested the lens with the 6Mp camera it came with (and which I intend to use it with exclusively), and at least some of the defects will obviously be worse on a 20-24Mp DSLT, especially if you like to crop heavily or pixel peep. However, on the test camera, with jpegs in camera (large, fine, and viewed on an old 1280 x1024 pixel monitor with a fairly low pixel density, so 100% crops are magnified a lot), full-size images are sharp and show good contrast at all apertures up to f11 and all focal lengths.
At 18mm there is significant bowing of horizontals away from the centre line which (unlike the later Sony cameras and lenses) won’t be corrected internally, so if this is an issue expect to correct it later in PP. At longer focal lengths this geometry isn’t obvious and I didn’t notice any significant pincushion at 50-70mm, but doubtless you will find some if you look hard enough.
At 18mm 100% crops also appear sharp at all apertures, but the lens looks weaker between 24-28mm where it is softer in the corners near full aperture and with slight CA that is just about visible at 100% crop; on a 20Mp camera this is more noticeable at 100% crop and compares badly with the 18-55 which gets corrections but even here, where it is worst, it is still miles better than many other lenses I’ve tested and with normal image framing the CA isn’t noticeable (unlike certain other lenses I’ve reviewed …). The lens improves over 35mm and from there to 70mm I couldn’t see any CA unless I looked at much more than 100%. Corners are slightly softer than the centre at full aperture, as you would expect, but it seems to be mostly the extreme corners with quite a wide sharp central region and isn’t very marked.
I do notice that although the lens improves at f8-11, f11 isn’t as much better than f8 as is often the case on other lenses, and there isn’t really any aperture where the lens ‘pops’ as bitingly sharp (we’re talking micro-contrast here) – but then there aren’t any where it is falls apart, either, and the comparatively modest zoom ratio clearly helps to give it good performance everywhere without anywhere being either superb or awful. Maybe that in itself is enough to dissuade some potential users, but it arguably does do what is expected of it. To put this in context, I find that 100% crops from most of the image at any focal length and aperture from full to f11 look pretty good with just a little sharpening in PP, and if this isn’t good enough for you, what are you after?
Colour is natural and likeable, and vignetting is virtually absent which was another surprise; I tested it with no hood fitted, not even the totally useless item it comes with, and which just gets in the way of the lens cap. It’s the same item on the 18-55, 28-80 and 28-100mm. What were they thinking of?
I took pictures of a clear blue sky at 18mm and various apertures and at f3.5-4 I could just about see a [I]very[/I] slight darkening of the extreme corners which you would hardly notice on a real picture, which is far better than I usually find at the wide end of a zoom. I guess this is because the 18-70mm is based on the 28-100mm FF lens of the later Dynax film cameras – indeed, put a late Konica Minolta 28-100 next to the KM18-70 and you can’t tell the two apart except for the writing – they even extend the same amount at the same positions of the zoom ring. I assume they simply repeated the same optical and mechanical design with all the glass elements just a bit stronger, though it was clever of them to maintain the same registration distance. Stick the 18-70 on a film camera and the image ends vignette at 18mm and corners up to 24mm but the rest of the range is clear, so you can use it on film, perhaps even at 18mm if you want a square crop and don’t mind the corners being a bit soft. On FF digital you’d probably see a bit more vignetting, but I don’t suppose many FF aficionados are going to be tempted. :-)
Field curvature seems to be fairly low – not enough to be obvious, anyway – just like its FF brother, and a lot less than some of the older Minolta kit zooms. People seem to hate field curvature and soft corners today, though they were once thought quite desirable for portrait lenses …
Flare and veiling are good to bad depending on focal length set. I took a picture with a strong winter sun in one corner of the image at 18mm focal length as a fairly extreme test, and the result was large but weak semicircular ghosts in the upper quarter of the frame rather than the little spots you often get, but little veiling and a quite useable picture if you’re tolerant of those things. Zooming progressively in to 70mm without changing the camera axis resulted in increasing flare and veiling (despite the sun now being out of shot) until by 50-70mm the veiling was so bad the image looked like more an 18% grey test card than a scenic view :-). Shielding the sun with a hand, however, immediately brought the picture back to a normal contrast, and the lens doesn’t seem to show much veiling due to sky alone, which is nice. If you actually [I]want[/I] some veiling to create an atmospheric summer’s day effect, just try this lens and vary the focal length until you get the amount you want!
I expect to do some better tests of bokeh later, but first impressions are a rather busier background at the long end than I’d like, which probably won’t bother me too much because if I want thin DOF I’d almost certainly pick another lens – I just want things sharp on a basic zoom.
BTW I do this kind of testing not merely to see whether a lens is any good, but in order to check what apertures and focal length combinations to avoid when I need a very sharp picture or when (as is often the case) a background of distant bare tree branches makes any CA more noticeable. Many lenses (and not just the cheap ones …) are basically unusable anywhere near full aperture unless you actually [I]want[/I] a soft-focus effect, and quite a few of the older ones have terrible CA at one end, so the key is not get taken by surprise.
Maybe I’m lucky with my copy of this lens (I expect there is considerable variation) but it isn’t actually bad enough anywhere in the aperture/focal-length map to generate any must-avoid regions, though as I generally use aperture-priority or manual I would probably stick to f8 from 28mm upwards to keep it sharp and use full aperture sparingly at the long end. Full aperture performance at 18mm is however very good which is the best news, because 18-23mm is the region I can’t easily cover with the Minolta FF lenses I’m likely to have around*, and if 18mm is good you can always crop a little to suit. The slight softness at 70mm won’t bother me much as I won’t be cropping heavily here (I will usually have longer tele-zooms available). The slow aperture with this and other similar lenses is always a potential issue, but with ISO and IBIS to help it’s not a major problem, and when light levels are low or I want thin DOF I have a selection of fast primes to choose from.
So though I wouldn’t call it a great lens (it’s not as good as the Sony 18-55 SAM II which is better to start with [I]and[/I] benefits from in-camera corrections) but I don’t think it deserves the calumny often heaped upon it. And, to be realistic, if I want maximum resolution and heavy cropping I will use other lenses and another camera anyway. I haven’t tested the Sony version, and it is arguable whether subsequent issues of a lens years apart have greater tolerances and more variability as the manufacturing tools wear out, or benefit from improvements in manufacturing as the tools are replaced and updated (if they are …) . I also can’t compare it to the 18-200mm Minolta because I don’t buy super-zooms :-)
*The 5d is not just a backup camera but intended to be paired with a Minolta AF film camera in my gadget bag, and the low pixel count and lack of live-view suit my way of doing film photography.

rating summary

lens image
  • total reviews: 125
  • sharpness: 3.69
  • color: 4.01
  • build: 3.07
  • distortion: 3.51
  • flare control: 3.73
  • overall: 3.60
Dyxum.com - Home of the alpha system photographer
In memory of Cameron Hill - brettania