What do you look for in a lens?

There is a long list of things that can make a lens desirable or undesirable, and some of them are hard to quantify. The best example of this is 'magic', which can be defined in at least two ways, though the two are far from mutually exclusive. One is that it's a lens which gives you -- you, personally -- an unreasonably high proportion of good pictures. The other is that it's a lens which has a unique 'look', which suits the way you work. So let's put 'magic' near the bottom of a list of more than a dozen questions, and begin with much more obvious characteristics.

1   Can it be used with your camera?

2   What is the focal length?

3   Will it cover the format?

4   How fast is it?

5   Is it sharp and distortion-free?

6   Is it contrasty?

7   Does it flare easily?

8   What is the bokeh like?

9   Does it handle well?

10   What is its filter size?

11 Has it any special features?  

12   What is the build quality like?

13   Is it 'magic'?

14   Can you afford it?

Tibetan monk

The 100/2 Zeiss Makro-Planar which Frances used here is a 'magic' lens, in the sense of giving unreasonably many first-class pictures. This is a hand-coloured Ilford Multigrade Warmtone print from an Ilford XP2 negative.


monk fs

Can it be used with your camera?

This is the most fundamental question of all, but it is not necessarily as simple as it looks. It is not even restricted to whether a given lens will fit on a given camera. For example, there are Russuan SLRs with the same 39mm x 26 tpi (turns per inch) thread mount as old rangefinder Leicas -- but a Leica rangefinder lens won't focus to infinity because the flange-to-film distance is too great. A manual-focus Nikon-F or Nikon-F fit lens can be used on a Nikon D70, but without any metering. A Leitz dual-range Summicron will fit on a Leica M8, but won't focus further than 3 or 4 metres because of the rangefinder cam. In all three cases, the lens can be used, but you may or may not be willing to give up such features as infinity focus or metering.



These are only some examples. Many lens mounts evolve to add new features that are needed for new cameras, especially if they are to do with metering, such as auto-indexing or AI on Nikkors. They may also subtract old ones that are no longer needed: sticking with Nikons, the 'rabbit ear' meter coupling. With any long-running mount you therefore need to know whether the features needed by your camera are there -- and, if need be, to decide whether or not you can live without them or not. Some features can be retrofitted or replaced; others can't. A few lenses won't fit at all (as listed above) or can even damage the camera. The easiest way to check any of this is on-line: via either a specialist site dealing with a particular camera, or by asking a question on something like the rangefinder forum.


Both of us forget, more and more, which lens we used to take a given picture. This is not just advancing years: it is that we use a lot of lenses, some our own, some on loan, and often, it just doesn't matter. If we look at an original transparency, that's one thing, but by the time it's on a monitor, quality is levelled out so much that you can't tell.

This was probably taken with a 90mm lens -- a handy working distance, and reasonable image size even with rangefinder cameras -- but it could be one of four lenses for the Leica Ms or a couple for the Nikon Fs. Almost all our lenses, ancient and modern, are used on both film and digital.


In the opposite direction, many lenses can be adapted to many cameras. This may involve simple adapters, optical adapters or some degree of reversible or non-reversible mechanical engineering on the lens mount. A great deal depends on flange-to-film distance. Thus, for example, a Nikon F has a flange-to-film distance of 46mm, so Nikon F-fit lenses can usually be adapted to other SLRs with shorter flange-to-film distances via a simple 'sandwich' adapter between the lens and the camera body. Adaptation in the opposite direction will require surgery to camera or lens or both if the lens is to retain infinity focus. Remember, it may (or may not) be possible to change the mount on the camera, rather than the lens.

There is another way of doing it, even if the lens mount is made for a camera with a shorter back focus. This is to incorporate an optical element in the converter. This will always have an adverse effect on image quality; the only question is whether the image quality is adequate for your requirements.

Then there are lenses that are made for a range of cameras, fitting none of them directly but requiring an adapter according to the brand. The simplest (because there are no linkages of any kind) are T-mounts; more complex versions include such things as Tamron's Adaptall series. We have both. Nowadays, most people prefer lenses that are specifically made in versions for their cameras, but the old systems can work fine too.


Children, Morro Bay

Roger shot this many years ago with either the 600/8 Vivitar Series 1 Solid Cat (which we still have) or the 800/11 (which we don't -- it was just too slow, and too close to the 600). Because it's a fixed-aperture catadioptric lens (using both mirrors and lenses) there's no aperture linkage, and of course it's manual focus, so a simple T-mount suffices. This was on a Nikon F; on our D70 it is the equivalent of a 900mm lens, but we need a new 35,5mm UV filter to replace the one that Roger dropped and broke.


morro bay

Rangefinder base

A point of interest to rangefinder users is that focusing accuracy is very much dependent on effective [rangefinder] base length, usually abbreviated to EBL. This is obtained by multiplying the actual base length by the magnification of the optical system in the rangefinder. Thus, to take a simple example, if the actual base length is 50mm and the magnification is 0.7x, then the EBL is 35mm; if it is 1.2x (not feasible in a combined range/viewfinder) then the EBL is 60mm.

Magnifications below unity tend to make for more accuracy, because the mechanical demands are less. Thus a 40mm base and a 1.5x magnification (EBL 60mm) is less likely to be accurate than a 70mm base and a 0.85x magnification (EBL 59.5mm).

Otherwise, the shorter the EBL, the harder it is to measure distances accurately. With the very short EBL of all Voigtländer Bessas, longer and faster lenses are harder to focus accurately than they are with the longer base of a Zeiss Ikon or Leica, though the Bessa-T (where the magnification is above unity) is better than most and a lot better than the R4 wide-angle series with their very short EBL.

From personal experience, Voigtländer Bessas (except the T) are marginal at close distances and full aperture with 50/1.5 or faster; 75/2 or faster; 90/2 or faster; and all 135mm lenses. Once again, a lens that fits is not necessarily going to work well.

spectacle 07 finale

Finale, Moncontour Spectacle, 2007

The 50/1 Leica Noctilux that Roger used on his Leica M8 for this shot is equivalent to a 67mm lens on full-frame 35mm. At full aperture, rangefinder coupling is at the limits of mechanical precision, even for a Leica, and unless you buy the equipment new (or even if you do), it may not be a bad idea to have the camera and lens specifically calibrated to one another.

What is the focal length?

This is the first question most people ask. Is it a wide angle, a 'standard' lens, long focus/tele or a zoom? The other modules in this series are divided up by focal length, so the main point worth making here is that there is rarely any great advantage in having focal lengths that are too close together. e.g. 35mm and 40mm or 85mm and 90mm. Remember, though, that the shorter the focal length, the greater the percentage difference, i.e. there is more difference between a 15mm and a 21mm lens (6mm but 1.4x) than between 90mm and 105mm (15mm but less than 1.2x).

Remember too the way in which small sensors apply a multiplication factor to the effective focal length as compared with 35mm full frame, so that a 50mm lens on an 16x24mm sensor covers the same angle of view as 75mm on 35mm. For larger formats than 35mm (typically roll-film, though there are some 'medium format' digital sensors), the focal length factor goes the other way: e.g a 50mm lens on 6x9cm nominal (56x84mm actual) is a wide-angle equating to 18-19mm (actually 18.4mm) on 35mm. Similar arguments would be true for (say) a 30x45mm digital sensor, though this size does not actually exist at the time of writing.


vx dive 2

At least you can make direct comparisons between 35mm and 6x9cm, and with 16x24mm and 18x27mm, because they are the same shape. With formats of different shapes, coverage can only be compared roughly because the relationships are different according to whether you compare the short side, the long side or the diagonal of the negative or sensor format. Thus 50mm on 6x7cm nominal (56x72mm actual for Linhof) might equate to 21mm vertical, 25mm horizontal and 24mm diagonal, all as compared with 35mm full-frame.

Finally, many focal lengths are nominal. For example, '50mm' lenses are often slightly longer than 50mm. In fact, many are 2 inch (50.8mm) or even 51-52mm. For a given lens design you will see no detectable difference in angle of coverage, depth of field, etc., so ignore the internet chatter about this.

River Dive

The Dive (pronounced 'Deeve') was canalized from marshland around 1000 years ago; there are now three parallel branches ('Les Trois Dives') which later reunite. Roger shot this on Ilford HP5 Plus using a 1950s Exakta Varex IIa with a 58/2 Biotar. The diagonal of a 35mm negative is just over 43mm, but 50mm/2 inches was established as 'standard' by the Leica in the 1920s.

In order to clear the mirror with fast lenses (f/2 and faster), 55mm and even 58mm lenses were commonplace for reflexes in the 1950s and 1960s. There are perceptible differences in apparent perspective between 50mm and 58mm, especially in portraiture, but the difference between 55mm and 58mm is pretty tiny and more depends on the overall 'signature' of the lens than on the focal length. Print on Ilford MG WT.

Will it cover the format?

This is rarely a concern except with large format lenses, or lenses adapted from a smaller format to a larger one. Lenses for 35mm must have an image circle at least 43.3mm in diameter (the diagonal of the 35mm frame); the corresponding figure for lenses for 6x6cm nominal (56x56mm actual) is 79.2mm.

On the other hand, quite a few lenses are made for the digital SLR format of 16x24mm (minimum image circle 28.8mm) and these will not necessarily cover full-frame 35mm, even if the lens physically fits on the camera, e.g. Nikon digital lenses on a film camera.

With large formats, it's another matter entirely: a given 110mm lens may be a slightly long standard lens on 6x9cm (about 50mm equivalent for full-frame 35mm), a modest wide-angle on 4x5 inch (rough equivalent 32mm), a distinct wide-angle on 5x7 inch (rough equivalent 23mm) and an extreme wide-angle, if it covers at all, on 8x10 inch (rough equivalent 16mm). Or it may not even cover 4x5 inch. It all depends on the design.

Coverage is increased by stopping down, so circles of coverage for most large format lenses are given at f/16 or even f/22, and there can also be quite a big difference between the circle of illumination and the circle of sharp focus. In other words, at the edge of the image circle, you get an image, but it's not very sharp. Some manufacturers are a lot more conservative than others, so 'not very sharp' may still be adequate. And besides, in the words of the late Colin Glanfield, "How do you tell if a moving cloud is sharp?"

How fast is it?

Again, a dead easy one: just look at the lens specifications. The smaller the number, the faster the lens (and, usually, the more expensive). For more information on f/stop speeds, take a look at the free Basics module on lens design.

On the other hand, a lot of people draw false conclusions about fast lenses. Some assume that the faster lens is better in every way than the slower one, while others assume that the extra speed is always worth the money, if they can afford it. Neither is true.

service incendie

Old Moncontour Fire Station

Leica makes three 75mm lenses for M-series Leicas: the f/1.4 Summilux, the f/2 Summicron and the f/2.5 Summarit used for this picture. The Summicron is popularly acclaimed as the best, with the Summarit close behind and the Summilux a poor third -- until you want the extra speed. Leica M8, ISO set to 160. (Roger)

Do faster lenses deliver better quality?

It costs more to make a fast lens than a slow one, but this is independent of the quality. Thus the 50/2 Nikkor was a lot cheaper than the 58/1.4 when it came out (it was Nikon's first standard f/1.4 for the F-series). But the 50/2 was also a better lens in just about any way you cared to define 'better' -- until you wanted the extra speed, or unless you liked the 'magic' of the 58mm. Also, 58mm is a surprisingly different focal length from 50mm, despite being only 16 per cent longer. Likewise, among the pre-aspheric Leica lenses, the 35/2 Summicron was a much better lens than the 35/1.4 Summilux at f/2, f/2.8 and f/4. At f/5.6 and below they were hard to tell apart in many situations. At f/1.4 there was no contest.

Generally, the very best lenses are neither the fastest nor the slowest in the manufacturer's line-up, though the slower lenses are often a better bet than the fastest: the 50/2.5 Leica Summarit is better in almost every way (resolution, contrast, distortion, focus shift) than the 50/1 Noctilux at over three times the price, but the Noctilux also 22/3 stop faster.

sonnar digital camera

Digital photography

Roger used the Zeiss 1,5/50 C-Sonnar for this, on (as far as he recalls) the Leica M8. You can see that the depth of field is tiny: her hair and cheek are in focus, but not her bracelet, let alone hands or camera. It is always a trade-off whether to use faster films/ISO settings or faster films. There is plenty to be said for both. Note also the quality of the out of focus image ('bokeh') which is discussed more below.

Is the extra speed always useful?

If you constantly find yourself working at the maximum aperture your lens affords, then maybe you would benefit from a faster lens -- but even then, only maybe. Another possibility is to use faster film or to up the ISO setting on your digital camera, and in our view, it is possible for a lens to be too fast. We have used a number of f/1.2 lenses, and one f/1, but we generally stick with f/1.4 or f/1.5 because the lenses are smaller, lighter, cheaper and deliver better quality. Depth of field with ultra-fast lenses is also so tiny at close focusing distances that it can limit their usefulness. If we could afford it, sure, we'd buy a 50/1 Leica Noctilux and a 35/1.2 Voigtländer Nokton -- but they wouldn't be our only 50mm and 35mm lenses.

If you rarely or never use the maximum aperture of your lens, then you may well do better with something slower, and you almost certainly do not need anything faster. The sole exception to this is manual focus SLRs, where a brighter screen image is always easier to focus.

Is it sharp and distortion-free?

This is one of the great bugbears of choosing a lens, especially with the arrival of the internet. Is lens A sharper than lens B? Well, it's not hard to test sharpness, either with a test chart for yourself or by looking at the MTF figures provided by the manufacturer. To test distortion, just photograph a brick wall -- or again, look at the manufacturers' figures.

bastille day

The question is, how much does it matter? Unless you always use a tripod, you'll rarely see the maximum sharpness that a lens can deliver; if you do use a tripod, the great majority of good prime lenses are sharp enough for all practical purposes, and if you want more sharpness, go to a bigger format: most decent second-hand medium-format cameras, even our old KowaSix, will outperform a new Leica at several times the price.

Bastille day

All right, it's an extreme example, but with a shot like this, hand-held with a 15/4.5 Voigtländer lens on a Bessa-L (both newly introduced at the time), how much do you care about sharpness and distortion? If this is the kind of thing you shoot, why agonize over such matters? As far as we recall, the film was Fuji RFP ISO 50: it's certainly saturated...

Still more to the point, how often do you photograph flat subjects where corner-to-corner sharpness and square edges are of paramount importance? Sure, they exist -- instrument panels in ships' engine rooms, for example -- but how often do you personally photograph such things? If you make a habit of it, you'd better not use (say) a 50/1.5 Zeiss C-Sonnar at full aperture -- but in the real world, Roger is more than happy with his Sonnar. Likewise, the Tri-Elmar 4/16-18-21 has more distortion than a 21mm Biogon, but distortion is still negligible and Frances has no doubt which she would rather have.

This is not to deny that some lenses are better than others, sometimes vastly better, but if you are comparing broadly similar high-end lenses, ultimate resolution and distortion are likely to be secondary considerations as compared with most of the other considerations in this module.


Mearle's Drive-In

The story is told of a professional sports photographer who was asked by an amateur why he used a 280mm instead of a 300mm. "Ah," he replied, "the 280 has an enormous advantage over the 300." The amateur nodded, agog: what secret was about to be revealed? The sports photographer went on: "I own a 280mm, and I don't own a 300mm."

When Roger took this in the 1980s, he used the lens he owned: a 21/4.5 Zeiss Biogon in (original, 1950s) Contax fit, custom adapted for his Leica M4-P.


Comparing like with like

It is however true that some people don't compare like with like. A favourite comparison among rangefinder camera users, for example, is a second-hand 50/1.2 Canon from the 1960s versus a new 50/1 Noctilux at maybe ten times the price. Even assuming that the Canon is in perfect condition, which is far from certain in a lens that is several decades old, it will be nothing like as sharp and contrasty as the Noctilux. This is hardly astonishing.

bookshop, poitiers

But people say, "Yeah, but it ain't ten times better." This is a meaningless statement. It's better, and it's half a stop faster, and ten times the price is what it costs if you want the extra quality and speed. Only you can decide whether you want it or can afford it.

Bookshop, twilight

Actually, the Canon 50/1.2 isn't a bad lens under the right conditions, but at full aperture (as here, on an M8) it is never very sharp. By f/5.6 is it surprisingly good, though still flary. Its advocates, of course, tend to show pictures which flatter it.

A further consideration is that it is very hard to find the proper 'ventilated' (cut-out) lens hood for it, and non-ventilated hoods block a good deal of the viewfinder. It also has an infinity lock: see below, 'Does it handle well?'.

There are also some truly awful lenses out there, especially old ones. Many Communist Bloc lenses from the 1950s to the 1980s demonstrate very poor resolution, even if they are in perfect order and were assembled right to begin with -- and many weren't. Roger once had a 50/2 Jupiter where one of the lenses inside was so poorly mounted that it flopped to and fro. What was amazing was that it produced an image at all, let alone one that was not too bad.

Old, fast standard lenses (faster than f/1.7) are often pretty bad, too. So are old wide-angles: cheap lenses from the 1960s often had horrific barrel distortion. Roger remembers looking through a cheap 28mm at a building in the early 1970s, and involuntarily taking a step backwards because the building looked as if it was bellying out and about to collapse. Then there are old zooms, most of which are awful. Actually, many low-cost zooms are fairly unimpressive to this day: the performance of our Nikon D70 is transformed by replacing the 'kit' 18-70mm zoom with a Zeiss prime, but the kit lens is still adequate for many purposes.

Adequate for many purposes

This is an important phrase. If you only ever make postcard size prints, or (worse still) only ever view your images in electronic form, almost any lens is adequate. Move up to 8x10 inch or A4 prints and you need better lenses. Make big enlargements and you need the best lenses you can get -- or you need to move up to bigger formats.

A funny story here is one that we have heard several times in different forms. Someone 'upgrades' his lens and sees no difference in his 4x6 inch/10x15cm machine prints. He consults a 'knowledgeable' friend who assures him that the difference is there, but that it is only visible to experienced and knowledgeable photographers. In a sense, this is true: the knowledge in question is knowing which lens was used, at which point the 'knowledgeable' friend says, "Oh, yes, it's much better."


Back yard

Even a 'big name' is no guarantee of high (conventional) image quality, if the lens is old enough. The first series of 90/2 Summicron from Leica is frankly soft at full aperture, as compared with the later and much more compact version. But there are two important qualifiers to add to this condemnation.

One is that many people actually like this slight softness, especially for portraiture though it can also be handy for 'nostalgia' shots like this one. If you can use the 'shortcomings' of a lens to your advantage, they cease to be shortcomings. Many internet users, obsessed with test targets, tend to forget this.

The other is that it can be difficult or impossible to see the difference on a monitor like this, even though the difference might be abundantly clear in an original 35mm transparency or a well-made print.

pump, woodpile

Is it contrasty?

This is another favourite subject for internet chatter. Minute (and often imaginary) differences in contrast are discussed and dissected as if they were the be-all and end-all of choosing lenses. Well, it's true that some lenses give a flat, dull image: again, old zooms and Communist Bloc lenses are favourite, though our 135/1.8 Porst (Japanese-made for a German photo chain) is well up there. Old, uncoated lenses exhibit even less contrast.

boss 135-1.8

HH Dalai Lama

Before we had the Porst 135/1.8 for the Pentax, we had a Soligor 135/1.8 for the Nikons, which we are reasonably confident was essentially the same lens from the same manufacturer. As is so often the case, we wish we had never sold it, which is why, when the Portst came up cheaply, we bought it. There was very little light, so even with the Ferrania 1000D film which Roger used for this shot, as much sharpness has been lost to camera shake as to the inherent softness of the lens.

Nowadays, we'd love to try that old Soligor on digital, shooting the Nikon D70 at 1600 or even (if we could afford it) buying a newer DSLR with more speed. Actually, for the same shot today, we'd probably use the Leica M8 with either the 90mm f/2 (120mm equivalent, and only 1/3 stop slower) or the 135/2.8; a stop and a third slower, but we can set the sensor to ISO 2500 equivalent.

But when you look at the brilliant photography in the magazines of the 40s and 50s, you have to ask yourself how much lens contrast really matters, at least in black and white. Yes, the tonality is different, because they had to develop the film for longer in order to regain the contrast and then print on harder paper to compensate for uncoated enlarging lenses. But does it ruin the pictures? Hardly.

In fact, in the 1950s many people preferred uncoated lenses for colour, too, as it helped to control what was then seen as the excessive contrast of Kodachrome.

As a general rule, slow lenses have more contrast than fast ones, though a lot depends on the lens design and the coating. Fewer air-glass surfaces mean more contrast -- this is why the Sonnar in the 1930s had only three lens groups, implying six air-glass surfaces -- and multi-coating suppresses flare better than single coating, which in turn suppresses flare better than no coating. Reflections at a single uncoated glass-air surface can top 20 per cent (the exact figure depends on the refractive index of the glass); reflections at the same surface, multi-coated, can be under 1 per cent.

Clean lenses are contrastier than dirty ones, and lens shades (lens hoods) allow increased contrast in the vast majority of conditions, by cutting out non-image-forming light that shines into the lens from outside the image area. For maximum contrast, use the deepest hood you can, preferably rectangular (or square, with square formats) to cut out as must stray light as possible. If you have to shoot straight into the light, remove any protective filters you may have on the front of the lens.

To sum up, of course contrast matters, and it is something you need to be aware of. But it is rarely if ever the most important characteristic of a lens.


The 21 inch (533mm) f/7.7 Ross that Roger used for this 8x10 inch shot probably dates from the 1920s at the newest, and it may well be over 100 years old. In the great days of Hollywood portraiture, all the studios shot on 8x10 inch with uncoated lenses: the easiest way to recreate the tonality of the period is to do the same thing. The film, from memory, was Ilford FP4 Plus; the camera, a De Vere monorail with custom-made rail (a metre of 25mm box-section light alloy) and a custom, extra-long bellows from Camera Bellows.

holly 3

Can you have too much contrast?

Our own view here is a resounding 'no'. True, the combination of a very contrasty lens and a very contrasty or hypersaturated colour slide film may look garish, but there are several easy ways to fix this. The most obvious is to use a less contrasty or saturated film. Two others are to desaturate in Adobe Photoshop or similar or (the most expensive option) to use a Tiffen Ultra Contrast filter on the lens. Contrary to what you might expect, these reduce contrast, rather than increasing it. With lenses that lack contrast, you can adopt the opposite strategies: contrastier or more saturated films, increasing contrast or saturation in Photoshop, using polarizing filters in colour or yellow/orange/red filters in mono, increasing development times...

zeus' birthplace

Does it flare easily?

Low contrast is the result of one form of flare (veiling flare, which 'fills' the shadows), and arguably, it is the most important form. But if you say the word 'flare' to most people, they think of the (usually) colourful shapes you get when you shoot into the sun. These are reflections of the diaphragm or lens elements, and depend upon lens design and coating. The only easy way to see if a given lens flares badly is to try it.

Dikteion Cave

The Dikteion Cave in Crete was the birthplace of Zeus himself -- and it certainly feels like it, as you clamber down to the bottom of the cave via the slippery, clammy staircase with its wooden hand-rails.

When we arrived, it was a few minutes before they closed for lunch, so we did not really have enough time to take the pictures we wanted. Also, Roger's 21/2.8 Elmarit-M had been stolen the previous year and he was using a 1950s Contax-fit 21/4.5 Zeiss Biogon: not the same as the one used for Mearle's Diner, above, as he had sold that when he bought the Elmarit-M.

The flare in this shot, hand-held at 1/15 or maybe even 1/8 second, is impressive. The unexpected blue is of course the deep blue of the Cretan sky outside. The lens was wide open, so there is no hard-edge shape such as you get from a diaphragm -- but the multi-blade diaphragm on an old Biogon is prety close to round anyway.

What is the bokeh like?

This is another internet special. 'Bokeh' (pronounced bow-kay, also spelled boke) is a Japanese word meaning, in this context, 'the quality of the out-of-focus image'. Various terms have been used over the years: in the 1930s, 'plastic rendition' (meaning a three-dimensional look in the picture) was a variant on the same thing. 'Bokeh' appeared in English in the 1990s.

Bokeh is definitely something that can contribute to the overall 'look' of a lens, for good or ill, and it is something you can train yourself to look for. On the other hand, some people seem to be much more sensitive to bokeh than others. It has to be really nasty and 'wiry' before we notice it at all, but the internet is full of discussions about, and alleged examples of, bokeh. 'Alleged' because there is a limit to what you can see on a monitor, and even where people point out the differences, we can't always see it.

Daisy, Thambar

The grass in the background is rather dark and murky, but below the flower you can just about see the exceptionally nasty 'wiry' bokeh that the 90/2.2 Leitz Thambar delivers with the centre stop in place. The remedy: don't use the centre stop, and don't have lines in the background...

thambar bokeh

2 chairs

If you really like (or dislike) the effect that a lens gives, but cannot say why, it may be that you are noticing the bokeh. Or it may be something else entirely: 'magic' (see below), or 'sparkle'


Believe it or not, this is quantifiable. Research conducted independently by Ilford and Zeiss reveals that 'sparkle' corresponds to very high MTF values (in effect, very high contrast) at quite low frequencies.

Garden chairs

You can't say much about the bokeh in the garden shed behind the chairs, but we do believe that the chairs have 'sparkle', at least in the sense in which it is used of photographic prints.

The secret is simple: a bigger format. Roger shot this on Maco Cube 400 film using Frances's late father's KowaSix (and 85/2.8 standard lens, the only one we have for it). Sometimes we think about shooting more with this camera, but the truth is that like most people we find 35mm (and sometimes digital) to be more convenient -- and the easiest way to get more, good pictures is simply to shoot more pictures.

Does it handle well?

This is an absolutely fundamental question, as far as we are concerned. A lens that is too big or heavy or ill-balanced, or that has the controls in the wrong places, is less desirable than a small, light lens where the controls fall naturally at your fingertips.

Admittedly, 'fall naturally at your fingertips' will be a different matter for (say) a habitual Nikon F user (aperture control ring at the back, anti-clockwise twist to lock the lens in the mount) and Leica M-user (aperture control ring at the front, clockwise twist to lock the lens in its mount), but the important thing is not to have lenses that are wildly out of step with all your others, or have inconveniently narrow focusing rings, or anything else you find hard to use.

Clear alignment marks for mounting the lens (such as Leica's red dot) are much preferable to hard-to-see marks or no marks at all. Further desirable features for good handling, as far as we are concerned, are depth of field markings and infra-red focus indices. One feature we particularly like is focusing spurs, nubs or finger-grips, as they allow you to focus by touch, but some people prefer focusing collars. A feature we particularly dislike is infinity locks, found on many old rangefinder lenses. These do exactly what their name suggests, locking the lens at infinity. You have to depress a button or tab in order to release the lock. We normally disable these (reversibly) on any lens we own.



The painterly quality of this picture is the result of a number of things, chief among which are the lighting, the white balance, the exposure, the 50/1 Noctilux (at full aperture, on an M8, ISO equivalent set to 1250)...

The thing is, the Noctilux is a big brute, but if you're used to a film-based M-series Leica, an M8 is pretty chubby too, and the Noctilux feels much more at home. There's also the simple fulfillment of the old saying. "Take what you want, and pay for it, sayeth the Lord." If you want f/1, well, the bulk and weight of the Noctilux (or its Canon equivalent on an SLR) are the price you pay. And if you're not willing to pay for it, well, don't take it.

young danseuse

What is its filter size?

This may sound like a trivial concern, but if you have three lenses with three filter sizes, buying filters is not only expensive: they are also a bloody nuisance to carry. Nikon's standardization for many years on 52mm was very welcome, as is Leica's weakness for 39mm. If you have three lenses and are considering a fourth, the filter size may be quite an important factor if it represents the choice between buying a new set, and using the same size as one of your existing lenses.

ir st. martin

'System' filters -- square ones that fit in holders with interchangeable filter rings -- are OK for large format, some medium format and some 35mm and digital SLRs, but they are very bulky and slow-handling compared with screw-in filters.

A few lenses, mostly extreme wide-angles, require special filter adapters: the 12/5.6 Voigtländer and 16-18-21/4 Tri-Elmar spring to mind. Others take filters on the back, whether screw-on (for example our Series 1 Solid Cat 600/8, which has to have a filter in the light path in order to maintain correction, or our 110/5.6 Super-Symmar) or cut-down gel (some Sigmas). A very few take drop-in filters. And a few take no filters at all, such as the 15/4.5 Voigtländer.


Church of St. Martin, Noizé

We have only two screw-in infra-red (IR) filters, both B+W 092, in 39mm and 52mm fitting. Most of our Nikon lenses take 52mm fitting, and about half our Leica or Leica-fit lenses are 39mm. Obviously this governs which lenses we use for IR; we are not about to go out and buy 46mm filters, for which we have both Leica/Leitz and Zeiss lenses, when we can use 39mm and 52mm filters instead.

Stepping rings are always useful, and we can use 52mm filters on 46mm lenses via a stepping ring, but fitting a lens shade (hood) can be a problem.

Has the lens any special features?

What we're talking about here is macro lenses, shift or shift/tilt lenses, soft focus lenses... These are all special-application lenses, but often, they are a lot more useful than a general-purpose lens with some form of adapter or accessory. There are penalties too, though. For example, macro lenses are normally bigger and heavier (because of the long-throw focusing mount) and slower (it's harder to correct fast lenses for close-ups, though the Zeiss 100/2 is amazingly good). Nikon's manual shift lenses don't have auto diaphragms and are slow: a 35/2.8 is hardly a speed king. All tend to be expensive: a reflection of much reduced demand as compared with general-purpose lenses.

Even so, if you use macro or shift lenses a lot, you may well find that they are at least as good as general-purpose lenses for non-specialist photographs. In the days when she used Nikkormats, Frances's 90/2.5 Vivitar Series 1 Macro was her only 90mm, and her 35/2.8 PC-Nikkor was her only 35mm, though we have to admit that when we used Contax (we had a loaner outfit for a year), we loved having both the 35/1.4 and the 35/2.8 shift.

Soft focus lenses are another matter. Either you like 'em or you don't, and besides, subtle soft focus effects are very hard to show on the internet. Few if any soft focus lenses can do double duty as sharp, conventional lenses, so they have to be in addition to other kit. We have a Dreamagon and two Lensbabies, and had the the long loan of a Thambar. The last is the subtlest, and the one we would most like to own, but they have not been made since about 1940 and are incredibly expensive. Ah, well, one day...


Into the unknown

Actually it's not unknown; it's ten or twenty minutes' walk from our house. But the Lensbaby, used here on a Nikon D70, makes it look pretty mysterious.

lensbaby frances

What is the build quality like?

The only real way to find this out is to buy the lens and use it for years. On the other hand, you do tend to get what you pay for, so that among our lenses, for example, Leica lenses (the most expensive) tend to be more durable than Cosina-built Zeiss (medium price) which are in turn more durable than Voigtländer (least expensive). At the time of writing, we were on the point of sending four lenses to Optical Instruments, Balham, for repair: a 35/2.8 PC-Nikkor and 90/2.5 Vivitar Series 1 (loose focusing mounts), a 50/2.5 Voigtländer Color-Skopar (loose finger-grip) and a 28/1.9 Voigtländer Ultron (stiff, juddery focusing mount). We should add that we also have 15/4.5, 21/4, 35/1.7, 35/2.5, 50/1.5 and 90/3.5 Voigtländers, none of which has given a moment's trouble, though some of them have been used hard. Nor have we had any trouble with other Nikkors or Series 1 Vivitars.

stairway & door

Chateau de Berrie

Again and again in this module, we realized that the secret of getting the best out of a lens is to use it for what it is best for, not what it is second-best for. This was taken with a Zeiss 18/4 on a Leica M8 -- but despite the excellence of the lens, we suspect we could have taken substantially the same shot with most good 15mm, 17mm, 18mm, 19mm (Canon!), or even 20mm or 21mm lenses -- though the 20/5.6 Russar would not have been in the same class.

The high point of lens construction, from a mechanical point of view, was probably the 1950s and 1960s, but even the finest lenses from this period are now several decades old. Diaphragms and focusing mounts may be sticky; lubricants may have distilled onto the lens, reducing contrast and resolution. A good repairer (such as Balham, who are actually in Croydon nowadays) can strip, clean and overhaul a lens, returning it as good as new; a bad amateur repair can effectively write it off, and such repairs are quite commonly encountered, especially among cheaper, older lenses being sold on the internet.

Is it 'magic'?

We have already touched upon this several times, hinting that it is very, very difficult to analyze 'magic'. As a general rule, 'magic' lenses tend to have a reputation for extraordinariness and acquire 'cult' status. As a result they are often correspondingly expensive. On the subject of magic, though, there four things worth pondering.

The first is that some people never see the magic. Neither of us has ever been able to understand the adulation of 50/2 Summicrons, of any generation, for example, nor any Tessars except Roger's long-gone 150/6.3. Switching to lenses where we differ, Roger finds the Zeiss C-Sonnar 1,5/50 magic, while Frances doesn't, but Frances finds the 4/16-18-21 Tri-Elmar magic, whereas Roger merely likes it very much. And while Roger thinks his 38/4.5 Biogon (on the Alpa) is magic, a friend of ours is even more impressed: he very nearly worships it and can instantly spot any picture taken with it.

viaduct, manchester

Railway viaduct, Manchester

One of the great disappointments of the internet is the difficulty of making black and whites look good on screen. Even on the same monitor, this looks a lot better in Adobe Photoshop than in our internet assembly program. It's all very well to talk about monitor calibration and workflow but the truth is that the vast majority of monitors that people use for internet browsing are not calibrated anyway, and besides, all calibration is to some extent local. But the original print, shot on 44x66mm Ilford HP5 Plus with an Alpa and 38/4.5 Biogon, then printed on Ilford Multigrade Warmtone, is indeed magic.

The second thing to beware of is that those who have tried lenses that are held up as magic, but have not seen the magic, sometimes bad-mouth either the lenses or those who praise them. We do not do this. All we say is that we've never seen the magic (or sometimes, that one of us has, and the other hasn't).

The third is that surprisingly many people who have never tried the lenses in question are downright offensive about either the lenses or the people who use them. This is (as ever) especially true of the internet. Accusations of snob appeal, stupidity, and seeing what you want to see are hurled around freely. Because of what we do for a living, often reviewing top-flight equipment, we believe that these people are mistaken as well as rude and ignorant. We often handle more top-flight lenses in a year than most people handle in a decade, and some are magic, and some aren't.

slovenia waterfall

The fourth thing to remember is that for the right application, for the right person, almost any lens can be magic. We are not talking about cult lenses here, such as the 38/4.5 Biogon or 58/1.4 Nikkor, or even the very rare (but usually inexpensive) 150/6.3 Tessar, but some right junk.

One London photographer in the 1960s, for example, reputedly made his name with a misty, romantic style that owed a lot to his having bought, very cheaply, a 150/5.6 Sonnar for his Hasselblad where the front element looked as if it had been cleaned with a steel scouring pad. Our 21 inch (533mm) f/7.7 Ross is our favourite lens for 8x10 inch portraits, despite (or perhaps because of) the fact that it is uncoated and around 100 years old. And so forth. If you think a particular lens will do something you want, buy it. But don't buy it just because it's cheap, or because it has a good reputation. Be especially wary if others call it magic.

Waterfall, Julian Alps

Sometimes, 'magic' is not just a lens but a veritable concatenation. A few years ago Roger put a Soviet-era orange filter on his old 200/3 Vivitar Series 1 and took a series of pictures on Ilford HP5 Plus. That set -- lens, filter, film -- pleased him so much that he has used the same combination repeatedly since. The magic is far less visible on the monitor than on the original prints, but he really feels that it has given him a disproportionate number of good pictures. And, of course, it's cheap. He'd be lucky to see £50/$100/70 euros for the lens; the filter is worth maybe a tenth of that; and even the Nikon F with plain prism that he uses with the lens is hardly expensive nowadays.


Can you afford it?

It may seem odd to leave this to last. After all, for most people (including us), it is one of the most basic concerns. If we could write a cheque without worrying about it, we'd have a 16-18-21 Tri-Elmar, a 50/1 Noctilux and a 90/2.2 Thambar in the blink of an eye. But we'd be looking at a good wedge out of of £10,000, maybe $15,000 or 12,000 euros, and that's just too much for us. There is, however, some logic to leaving price until last

First, it is never a bad idea to work out what you would buy if cost were no object. You can then consider the gap between what you want and what you can afford, and consider whether it is actually worth buying second-best. Increasingly, our own view is that it isn't. This is in very large measure because we already have a lot of good stuff, and there is no sense in buying a slight improvement at a price we can't afford. Until we can afford a Tri-Elmar, we'll live with our exiting 15mm and 21mm lenses. Until we can afford a Noctilux, we'll live with our other 50mm lenses. And the Thambar... Well, there is no substitute, though we do have a Dreamagon and a couple of Lensbabies. If you don't already have the good stuff we do, taking the approach we suggest will enable you to decide whether to go for second best in the medium to long term, or for third or fourth best as a stopgap until you can get what you really want. Or of course just to wait.



When/if we can afford one, we intend to buy a Leica Tri-Elmar 4/16-18-21mm. If you look at the upright of the book-case on the right you can see that there is a tiny bit of distortion. Now put your hand on your heart and tell us it was the first thing you noticed about the picture... The great thing about the Tri-Elmar, as far as we are concerned, is that it gives the same angle of view on the M8 at 16mm (as used here) as it does on full-frame at 21mm. It is also uncommonly sharp -- pixelation limits resolution, not the lens -- and reasonably compact. Alas, it is also damnably expensive.

Second, keeping your eye on what you really want can discourage you from buying cheaper lenses merely because you 'can afford' them. We put 'can afford' in quotation marks because it's all too easy to spend a bit here, a bit there, a bit more somewhere else, on 'bargains' that aren't really bargains at all: if you hadn't spent all that money on 'bargains', you'd be half way to the lens you really wanted, and the extra financial stretch to find the other half would be a lot less painful than finding the whole amount.

Third, as our friend Red Denner once pointed out, the pain of buying something will eventually pass, unlike the thing itself. Thus, although our 75/2 Summicron was horribly expensive in 2006, by 2008 the pain had passed and we still had the lens. Of course, it's not worth bankrupting yourself to buy something, but if you can just about afford it, it may be worth looking at whether you can stretch 'just about' into 'oh, all right'.

Fourth, having a goal can be a powerful incentive either to save up, or to sell equipment you have acquired over the years but never use any more. Ebay and PayPal can be quite good for this: just let the PayPal stack up until you can afford the lens you really want, or at least, until you are close enough to it to top it up with some money from elsewhere and buy the lens.

Exakta Varex IIa

All right, it's not worth much. But if you don't use it at all, fifty quid ($100, 70 euros) in the piggy-bank is a bit under 10 per cent of the price of a 1,5/50 C-Sonnar. Which is more use to you? Your ambition may not be the Sonnar but something else: this can help you get there.

exakta varex 11a

The bottom line

As so often, writing and illustrating this module clarified our own thoughts. Some opinions remained unchanged: for example, that you should postpone the purchase of a new lens (or indeed any new equipment) until you are reasonably sure of two things: first, that its lack is hindering your photography, and second, of how you will use it to overcome those hindrances.

The main things we realized, though, were also twofold. One is that it is surprising often possible to turn the disadvantages of a lens (especially a lack of ultimate sharpness) into advantages, by trying to take pictures that work with the shortcomings rather than pretending they don't exist. The other has already been mentioned, too: the fact that most of the time, it doesn't really matter very much which lens we used, at least from the point of view of resolution, distortion, etc., because the picture is the thing. We are much more concerned to have a lens we can use fast and comfortably, and which has the features we want, than with whether that lens or another of similar specification offers the ultimate performance when shooting test targets.

la purisma kitchen

Mision de la Purisima Concepcion, Lompoc, California

It's quite hard to take a really bad picture of this scene, which is why it was, when we lived in California, one of our standard test shots for wide-angle and ultra-wide-angle lenses. We shot it with 21mm (f/4.5 Biogon and f/2.8 Elmarit-M); with 17mm (Tamron SP); and with 14mm (Sigma f/3.5). We don't actually remember which lens we used here but it looks like 21mm. The film was Roger's all-time favourite in colour slide, Fuji RF/RFP ISO 50.

Go to the list of modules

or go to the home page

or support the site with a small donation.


© 2008 Roger W. Hicks