Thursday, 1 December 2016

Brilliant Bargain Basement Lens for Mirrorless Cameras

I’ve had my little NEX 5n marvel now for 5 years and long ago came to grips with its charms and deficits and reached a point where most of its operation is totally intuitive. It’s been my main snapshooter and travel camera companion since early 2012, yet even today the files compare nicely with those from the A6000 and A6300.  More recently, however, the NEX 5n has been relegated to classroom duties as a shiny new Olympus EM 5 Mk 2 has been pressed into service for travel and casual shooting duties.

The EM5 is not the first M4/3 camera I’ve owned, I also have a GH2 which has principally served for video usage and has now become my wifes’ personal camera.

Regardless of which of these mirrorless cameras I have been using one thing has always remained constant, my interest in using legacy lenses as a creative tool. 

Most Sony mirrorless and M4/3 mirrorless owners I meet seem to have at least a passing interest in legacy glass, perhaps we all like some cheap optical fun, or maybe we are just owners of camera cupboards filled with optical orphans from the pre-digital age.  Thinking along these lines, I thought some of you may find the following examination of the old school Nikon 35-70 f3.3-4.5 I’ve been using for almost 5 years helpful. And maybe even enlightening.

Remember there have been a whole raft of Nikon 35-70s over the years, this particular one is the f3.3-4.5 Zoom Nikkor with Macro Mode, I have no significant experience with the other cheap kit 35-70s so don’t assume my findings with this lens will automatically translate to others in the Nikon breed, from some casual testing I am pretty sure they won’t. 

I have included a photo of the lens here just so you know just which one we are talking about. The basic specs are it weighs 250 grams and is 59mm long (not including the mount) and 62mm wide.  It’s a two ring lens, and the front does rotate as you focus just in case you were wondering. Though the size is small and light, just a tad smaller than the Sony 18-55 OSS, you do have to add in the length and weight of the adapter needed for each camera.

Yep, mine is pretty knocked about, so that just shows you how tough this little Nikkor is!

Testing Rationale

Before proceeding I need to point out that I never ever test lenses using JPEGs, I shoot raw almost exclusively and only use JPEGs for class demos or on my iPhone where until recently I had no Raw option. Secondly, I judge a lens on how the images look post editing.  I realise that might be a little controversial, but to me the “out of camera” (OOC) look is irrelevant, what’s far more significant is how will the images look in a final state that represents what I would do using my normal workflow.
The above means that what I usually get out of lens likely exceeds what average consumers will see from their jpegs and honestly my results may not be representative of what even many RAW shooters liberate from their files if they just use “quick down and dirty conversion processes”.  Nonetheless, the attributes of a good lens always shine through and are quite obvious, it’s pretty easy to pre-sort the wheat from the chaff as far as lenses go just by looking at the cameras zoomed in review image and sometimes even the overall preview image. Ultimately it’s the subtle nuances that are not totally obvious at shooting time.

The two pics above were taken on the NEX 5n at around the 70mm mark at f6.3, these have a very 3D look, well that's due to my TLC (True Light Capture) system, and you'll be hearing more about that soon, but check out the 100% crop below.

In Use

This Nikon 35-70 lens is, of course, manual focus but using manual focus lenses on the NEX 5N and my M4/3 bodies is an absolute hoot. On the NEX 5N the touch to zoom in view, reasonable screen, accessory EVF and focus peaking make it both quick and easy, *  I reckon those blokes raving about Leica M9s and M240s have a few lessons to learn when it comes to really good manual focus options.  If anything the Olympus EM5 MK 2 is even better, mainly because the EVF is an excellent step up the ladder.  

And before we leave the manual focus, here’s the elegant part, focusing is all done at the shooting aperture, which means....wait for can see the exact effect of the DOF for the aperture you have chosen!  If you don’t think that’s important well clearly you have not drunk enough developer and fixer, good lord back in the film days we would have almost given our first born to be able to do that without having to contend with a screen that looked like a dark moonless night in winter.

The Rockwell Challenge

This lens found its way initially onto my Nex 5N as a happy accident. I’d read on Ken Rockwell's’ site that the 35-70 3.5-4.5 were one of Nikon's ten worst lenses, well perhaps Ken got a bad one, or perhaps he didn’t even try it out, he does claim he makes stuff up after all.  Anyway, I stupidly trusted Kens’ opinion on the lens and expected nothing much, in fact, I only tested it out initially because I was curious to see just how crook it really could be and since I was initially given the lens for free what did I have to lose.

Dear old Ken made one valid point that I agreed with, yes, the lens is plastic barrelled and the attendant external bits are plastic, but the mount (and internals I think) are metal.  It doesn’t have entirely that traditional smooth solid Nikkor feel, it’s not the sort of lens to get excited about by just looking at it or holding it in your hand, a Zeiss it is not.  However, when mounted on the NEX 5N or an M4/3 body the balance is actually very nice and using a heavier lens is not the way I would generally want to go from an ergonomics perspective when utilizing these mirrorless cameras handheld.

Despite being significantly plastic, and let's be honest how many modern lenses aren’t, everything is solid, there’s no slop to be found in focus, apertures rings or the zoom control.

Subtle tonal rendering leads to very flexible monochrome conversions, this was captured at 70mm, f6.3 on the Sony NEX 5n near Banff in Canada in 2012.

Initial Testing

Initially, I made a couple of shots out my front door on one of Goulburn's more miserable days, flat light, flat grey sky and flat out cold, I then loaded the pics onto “Mac the Master of the Pixel” and cracked them opened in Raw Developer. 

(Raw Developer is my long-term “go to” application when I really want to see if a Raw file is going cut like mustard or mush out like melted ice cream, but it’s a Mac only deal.)

At first, the pics looked a bit sad, low in contrast with weak colour, but a little adjustment and sharpening using Raw Developers’ super duper algorithms (which have only got way better with intervening years) and those files were showing incredible potential at the 35mm focal length.

My curiosity suitably piqued I decided to give this “worst Nikon lens ever” the benefit of doubt and shoot some proper stuff across all focal lengths on a better day, i.e. one with at least some sun, colour and warmth, and hopefully not the usual Goulburn gale force winds that would blow your dog off its chain.  

A few grabbed shots at the end of lunch after running a morning class gave me what I needed to make some more astute judgments.  The shots were taken at 35mm, 50mm and 70mm wide open and at f8, all handheld. 

I recall that by 7.30 that night I was sitting at my computer trying to lift my jaw off the keyboard, the processed files were utterly sharp edge to edge, even wide open they were pretty good.  The images still had a low contrast “film like” look and the colour was subtle so they would not win over the MTV photographic crowd who like their colour “turned up to 11”, but that filmic look always floated my digital boat and low contrast files are a sweet pleasure to post edit.

Now a few shots does not make a lens test, I needed something more rigorous than that, ( don’t ever expect me to shoot test charts etc.) and as far as I am concerned the only way to judge a lens is to shoot real world pics, which is exactly what I did with the 35-70.

My next set of images were taken at my standard “camera test range” Marsden Weir, which is just a short walk from home and the office. In this case, shots were taken at 35mm, 50mm and 70mm, at third f-stop increments ranging from wide open through to f9**.  The focus was on the same point in each shot and fine-tuned for each image at shooting apertures to eliminate focus shift when stopping down and the variability of focus with focal length changes.  

Incidentally the 35-70 not par-focal, i.e., it does shift focus significantly as the focal length is changed, just in case you wondered.  

Naturally, a tripod was used, and the self-timer enlisted for triggering purposes, the electronic first curtain shutter option enabled, 100 ISO and RAW used of course.  In other words, everything was optimal, if there was going to be a problem, it would be the lens, not the camera set-up.  At the same time, I tested the Sony kit zoom for comparison purposes.

Ah the kit zoom, look it’s not without its charms and I have dealt with that little beasty elsewhere in my old blog, but let’s just say for those who are asking/wondering….it’s not even in the hunt for sharpness compared to the 35-70 Nikon, hell it hasn’t even mounted the horse, but don’t take that as meaning it is rubbish, it’s okay within its design limits.

Back to the “worst Nikon lens ever”…..

70mm on Sony NEX 5n f6.3

100% Crop, no issues here!


The Nex 5n has a very weak aliasing filter, and current Sony mirrorless models generally don’t have one at all, images can easily be aliased if the RAW extraction program is capable of really extracting pixel level detail out of the file and “IF” the lens used is really really sharp.  Nikon's “worst ever lens” easily gets that old NEX sensor aliasing on patterns and man-made objects with repeating structures.  So much so, that for the first time I found myself cutting the sharpening tools right back and turning on the Raw Developers “early stage noise reduction” to control it.

So here we have a lens that actually resolves down to a pixel level on the NEX 5N in the central portion and was very close to showing the same clarity right out to the far corners, regardless of focal length.  In fact having conducted tests throughout the aperture range at each focal length, I am confident in saying this. On APS-C and M4/3 the 35-70 is actually a diffraction limited lens, in other words, once you exceed the diffraction limits at around 5.6-7.1 depending upon focal length, it gets noticeably softer regarding resolution. Generally, it’s sharpest at f5, and that’s pretty amazing for a cheap zoom!  Think about that for a second, at f5.6 the 18-55mm Sony kit zoom set at 50mm can hardly resolve any edge detail and it’s barely any better at 35mm, yet this ancient little Nikon is peaking at an extremely high level.  As an aside, it out-resolves the Minolta f3.5 55mm macro, and trust me that lens is no slouch and is one of the better 50mm lenses I have tried in the Minolta/Sony system, and I have indeed tried them all, yes, that includes the f2.8 macro, f1.7 and f1.4 primes! 

On M4/3 the story is much the same, the 14-42mm series 2 Panasonic doesn’t hold a candle to the Nikon in the 35-42mm range and bear in mind that you can’t open the 14-42mm wider than f5.6 at those settings.

Ah but I will tell you something else as a tease, (there might just be another article on the way) the Nikon 55mm f2.8 micro is even better than this little baby, for the moment though let’s just say that with proper raw conversion the 55mm micro is a scalpel!

35mm @ f5.6, sharp edge to edge, check out the wire fence on the edge of the frame in the 100% view below.

Sharp but Wait There's More

So the 35-70 Nikon is sharp but what about the other characteristics, after all, there’s far more to lens performance than sharpness alone.  

As said the lens displays low contrast, in fact you can easily see this when looking at the histograms of images even when shot under bright contrasty lighting, in other words, the histograms don’t push into the extremes of the highlights or shadows, which makes for flattish images, but we can always add contrast (especially if the exposure is captured with this in mind and the sensor is of the low noise variety).  Taking contrast away from a punchy image is much less satisfactory and often impossible. 

The highlight rendering has a little creamy glow to it which is hard to describe, but regardless I find it rather nice, you mainly notice it when looking at white or near white objects in sunlight, the glow reduces as you close down the aperture but never really goes away.

The colour rendition is somewhat warm, particularly with cyans and blues, in some ways a bit like old school Minolta glass but without that knockout colour punch those old A series lens exhibit.  Shadows tend towards neutral or warm which is nice, I hate just cold shadows!

Creamy Bokeh?

Bokeh? Honestly not really that great, it looks a bit busy to me, and I wouldn’t be choosing this lens if I were shooting a portrait or an image where smooth defocused areas were desired.  Basically on close examination, you end up with a slightly dark outline around out of focus details, and of course, the max aperture limits the blur possibilities.  

I must point out though that in many RAW conversion applications the type of sharpening and noise reduction you apply can have a significant effect on the look of the bokeh and I can certainly get nice results from this lens if I’m willing to make adjustments to push the rendering in a more bokeh influenced direction.  But you know what, not everything needs to be shot at wide apertures and the frame filled with out of focus stuff, sometimes, and I know some folk find this hard to believe, we actually want most of our photos in focus, especially if we actually need to sell them as prints for a living!

So here we have moved to M4/3, The Bokeh wide open at 35mm is not brilliant, but then it's not too bad either, and there is no issue in getting sharp detail rendering even wide open. Consider also this is shot in the shade and there is a fairly wide dynamic range present, but the rendering looks really very analogue.

Wide Open at 70mm on the EM5 mk2, yes not the most excellent bokeh but most folk would be okay with it and you can easily fine tune it in the raw converter or in post editing, Oh and that's my wife, Wendy.

So here is Holly the Collie, actually, she's human, it's just someone stuck her in a dogs body!  Anyhow 70mm wide open on the EM5 mk2, I can live with that bokeh, and Holly will happily live with the treat on offer just out of the frame.

Does It Distort?

Geometric distortion? You want figures, sorry I don’t have them, I’ve never deliberately gone out and shot stuff with this lens to show up any distortion issues and rarely does it ever bother me unless I’m shooting architectural images for money.  I’ll say this, I’ve not seen any issue at all in the regular shots I’ve taken other than some extremely mild barrel distortion with “front on” shots of buildings at 35mm, but it should be noted this distortion is very easily fixed.  Realistically being a moderate range, full frame lens, it would be most unusual if any severe distortion issues showed up in shots taken on APSC or M4/3 crop sensors. 

Vignetting, there could be some but again generally with moderate zooms designed for full frame and working within moderate maximum apertures ranges, it’s quite rare to encounter any significant issues.  Compared to the Sony kit lens it’s obvious the 35-70 has no vignetting issues worth considering unless perhaps you like shooting white walls at close range for fun, on M4/3 I simply can’t see any vignetting at all regardless of the focal length or subject type.

NEX 5n, 60mm at f7.1, what's not to like, see those signs under the awning, well in the original you can read those easily!

Does It Play a Level Field?

What about field curvature?  At moderate distances this seems to be a reasonably flat field lens, no significant nasties at all, the focus holding equally across the whole image at the intended plane of focus.

Field curvature is virtually flat between 60 and 70 mm regardless of the focused distance, but the 35-50mm range shows significant field curvature as you focus closer towards infinity. Basically, the curvature comes back towards you at the edges and corners which is tricky to deal with if you are unaware of what is going on.  When you focus the lens at infinity in the centre of the frame, the edges of the horizon will be just slightly blurry (though I have seen far worse) while the corners closest to you will likely be quite a bit sharper than you may expect for any given aperture.

The type of field curvature I’ve described can be a useful creative tool, and it’s definitely easier to work with that the opposite type, where you focus on near object in the centre of the frame and then find the rearward corners of the frame are sharp, but the middle sides and near corners are blurry.

Ultimately to get landscape shots that are sharp throughout the entire frame you need to focus at infinity distance objects about a 1/3 to 1/2 way out from the horizon centre and use an aperture of f6.3-7.1.  Do this at 35-42 and 50mm, and suddenly everything looks incredibly sharp.  I expect it’s this behaviour that has led some people to conclude this is a soft lens, it isn’t, just a bit misunderstood.

As said the 60 and 70mm focal lengths are virtually flat field, so you just focus on the plane you want to use for peak sharpness, and that’s it.

The 50mm field curvature can be very useful as when you focus on the centre of the frame for say a portrait, so long as you have no near objects showing in the corner of your frame it will give a more shallow DOF effect than you would otherwise expect based on the f-stop chosen.  It also means that the 35mm setting can give a somewhat shallow DOF effect that seems to be rather more akin to what you would expect from a lens shooting at say f2.4 or so.

So this is telling, the dynamic range in this pic is huge, it's a single frame on M4/3 off the EM5 Mk 2, but the low contrast rendering of the 35-70 means it works perfectly.  Oh and the pic is sharp, check the 100% crop below.  Stats?...35mm f6.3 @ 200ISO 

Another M4/3 sample, this time taken at 42mm, f6.3 and just a look at that 100% crop below, we have detail!

Sweet News for CA Haters

And now the big issue, the one that really gets on my digital goat, “lateral chromatic aberration”.  Few prime lenses and even fewer zooms are what I would call CA benign, in other words, have so little CA that it is not worth fussing over.  Normally you can correct it in the editing application, so it doesn't show up in the final print. But, and dear reader this is the bit most folk don’t really account for, usually, a lens with significant CA is far less sharp on the edges and corners because the offending colour channels are not focused properly at the sensor plane.  Fixing the CA will never regain the clarity in the offending channels, though it will cure the obvious colour fringing issue, so…..if you want peak sharpness all “Lateral CA is a big NO NO”!

This “worst of all NIKONs” has absolutely no sign of lateral CA at anything approaching sensible levels of magnification, in fact, to find any and thus correct it I have to view the image at 300%!  Seriously, I am going on record as saying this lens has the least CA of any lens I have tested on the NEX 5N and is line ball with the best I have ever tested on the M43 format. 

On the other hand at the mid to tele end, the lens does exhibit some longitudinal CA, which is much less of an issue as it only appears on out of focus areas and in any case it’s quite mild.  Longitudinal CA is much harder to fix, and most Raw converters don’t have a tool to do so, but I rarely find the need to anyway.

 35mm at 5.6 on the Sony NEX 5n, it can pick focus distances quite nicely, check the 100% crop below, local separation is not bad at all.

Flared Out

The one significant failing of this lens is flare, actually, there are more flares than a mid 70s disco if you point it in the wrong direction and it’s not a nice type of flare either, it sort of looks bluish and mushy to me.  Yes this baby needs a really good lens hood because bright lights even at angles right out to the side of the lens can cause it to glow like a neon sign

Oddly for a zoom, its worst performance seems to be bang smack in the middle of the focal length range with peak performance at the very extremes, a characteristic I have not previously encountered, and no I cannot explain why.  However, it’s worst performance is still pretty okay all things considered.

Why Use a 35-70mm Zoom?

Many argue that 35-70 is a stupid focal length for APS-C formats but I disagree, or perhaps my needs are different to others.  35-70mm covers the old 50-105 focal length, and I find that very useful indeed, in fact, if you gave me a 24mm lens and a 35-70mm zoom I reckon I could cover about 90% of my personal APSC shooting needs. On the M4/3 format, you end up with a 70-140mm lens, which is actually quite interesting for portraiture and other subjects where you want to shoot a bit longer, think of it this way the two big sellers for M4/3 in the tele range are 45 and 75mm!

Ultimately this is not an easy lens to use, images will need some levels and curves tweaking, the colour saturation will not knock your socks off out of the box, and the field curvature needs to be accommodated if you want to get all over sharpness for landscapes etc.  But compared to the 18-55 or 16-50 kit lenses Sony kit lenses in the 35 to 55mm range this lens “eats em up and spits em out” and of course you still have 70mm as a very nice performing option. 

For M4/3 users its a bit different, though it’s a small lens, the M4/3 versions are much smaller (Ok the 75mm f1.8 is not so small) and you can’t compare it to the kit lenses at all because the shortest focal length is equivalent to 70mm, but still there are some compelling reasons to use it, for me at least.

For Sony APSC, unlike the kit lenses this one can actually be shot wide open and still perform quite well, and of course, even at 50mm, it's still an f4 rather than an f5.6 lens, which is also true compared to the standard M4/3 kit lenses.  

So we have no OIS on the Sony (unless you are lucky enough to have a shiny new A6300 ) but then in the real world, this lens needs 2 stops less on the aperture to give sharp images so much of the OIS advantage may be negated. For M4/3 users, should you use an Olympus you have the OIS in the camera and combined with the short physical length and lightweight it’s easy to get sharp results at shutter speeds in the 1/4 to 1/15 range!  For the record, it balances perfectly on the EM5 mk2 (including the mount adapter of course).

Yep, the files convert nicely to monochrome, it's also sharp in the middle, as you might expect, 35mm f5.6 on the Sony NEX 5n @100 ISO.

Does It Go Close?

One aspect I have not mentioned is that this wee little zoom has a so-called macro capability, compared to modern lenses it’s not that close focusing, like about 30 cm from the subject, but the focus is quite even across the frame. The 35-70 does, however, respond very well to the use of close up filters, 2 or 3 dioptres being good enough to get you nicely close.  It is probably worth noting also that the extra crop factor of the M4/3 format make the macro function a bit more useful than on the APS-C models.

Lets Sum This Show Up

Really this lens is about rendering and creating editable files under high contrast light, the resulting images to my eyes anyway look nicely analogue, think of it as another handy tool to have in the shed, one that costs so little it’s easy to justify. Of course, it’s not a replacement for the kit lenses, it is a 35-70 after all, but in so many photographic instances it is a vastly better choice.

So to sum up, the 35-70 Nikon is very far from being a bad lens, it’s rather good in fact, lightly but nicely built, and very sharp across the full range.  Compared to the kit lenses at the same focal lengths, it is vastly better, and that includes the rather nice Panasonic 14-42 series 2 kit lens.  When it comes to the Sony Kit options in the post 35mm range, well let’s just say there is no comparison.

It’s probably better suited to Olympus M4/3 owners because the focal length is handy for portrait and short tele duties and the sensor based IS radically expands the low light shooting envelope.  I have no trouble getting very sharp results on the Panasonic M4/3 or Sony APS-C mirrorless bodies with this lens at normal shutter speeds despite the lack of image stabilisation.  If like me you have cameras in both formats then it’s an even bigger bargain, just buy the extra adapters.

One type of shooting I would not use this lens for is night photography, it just hasn’t got enough contrast to do that well, and frankly, there is a reason why lens makers produce lenses with the “Noct” moniker to their name….nocturnal..get it.  Anyway, this is not a nocturnal lens, it loves the daylight.

TLC image converted to mono using mainly the blue and some of the green channel, (a Combo that rarely works well at all) taken at 35mm f7.1 on the NEX 5n somewhere along the Tioga Pass in California a few years back.

Want One?

A quick check on eBay today revealed there are good to excellent copies going for just $50 to 80.00!  You may uncover some AF versions as well but honestly, I have no idea what the quality and performance of that version is like.  Note:  Don’t get the 3.5-4.8 version it is a much lower grade lens and does not perform near as well in my brief experience.

There are also constant f2.8 and f3.5 versions of the 35-70, these are vastly bigger and heavier and not really a good match for the small mirrorless bodies, though I can confirm they both perform excellently.  If you decide to go down this heavyweight route be very careful with the f2.8 version, it’s a legend that commands a premium price but alas it’s also very prone to internal dust, haze, mould attack and glass separation. Most of the f2.8 lenses have led a hard life being the pro versions and in addition to the mentioned issues, good ones start at about $400.00 and with use will just as likely end up with the exact same problems the cheap one have right now.

The fixed f3.5 version is a bit lighter and vastly more reliable, it’s just as sharp and the icing on that cake being you can snag one for less than $150.00 in near mint condition!

Now promise me this, say after me now  “I will not go out and bid the price of the 35-70 f3.3-4.5 lens up on eBay, and I promise to keep that I now know a big secret”. 

(That way I can go out and corner the market on cheap “worlds worst Nikon lenses” so when the world wakes up to this little jewel I can get rich quick!)

Whats that? You’re going to break your promise, ah well happy hunting.

Full Frame Usage?  Honestly, I’m not too sure, I tried it on a friend's Sony A7r Mk 2, and the results looked really promising, but I didn’t have the camera long enough to do a proper test.  If you try it out on one drop me a line and let me know how you fared, I’d be happy to share the results on the blog.

The Small Print

*Can someone please explain to me why Sony, having had an adequate touch screen operation on the NEX 5 series all those years ago, decided to remove the feature from the later 6000 series and then introduce a hobbled version on the A6500…anyone, surely someone knows? This quirk may explain why I can’t seem to buy a used NEX 5n/r or t for a sensible price as a backup for the day the end of my NEX 5n finally comes!

** I use f9 as the smallest setting for APSC and M4/3 lens tests as invariably the quality always falls off a cliff due to diffraction after that and generally if I focus in the correct spot f9 is a small enough aperture to get good overall focus in almost all situations. I will use smaller settings if absolutely necessary, but I very rarely need to.

No comments:

Post a Comment