Basics: Lenses

You can take a picture with a simple magnifying glass, if you can figure out some way of fitting it into a focusing mount. But it won't be very sharp, and it will have all sorts of 'aberrations'. It will bring light of different colours to different foci, so there will be colour fringing: you may remember seeing this with toy binoculars when you were a child. The plane of focus won't be a plane, but a curved surface, so if the centre is in focus, the edges won't be. There are other aberrations too, but these are the most noticeable.





A magnifying-glass portrait

Roger's brother Jeremy took this picture in about 1972, using a magnifying glass mounted on a bellows; the camera would have been a Pentax SV. It's recognizable, but very impressionistic.


pos & neg lenses

Positive and negative lenses

Group a are magnifying lenses: 1, biconvex; 2, plano-convex; 3, meniscus or concavo-convex (positive). Group b are reducing lenses that cannot form a real image: 4, biconvex; 5, plano-concave; 6, concavo-convex (negative).

The more light-gathering power a lens has (the higher the speed, which is to say, the bigger or wider the aperture), the worse these aberrations will be. If you restrict the amount of light passing through a simple magnifying glass, the quality will increase steadily. If you want more speed, you will need more glasses.

Also, a simple magnifying lens won't cover a very wide angle of view. As you get further from the centre of the image, the quality falls faster and faster. If you only want a very narrow angle of view, you can get away with a surprisingly simple design, just two pieces of glass cemented together, but if you want a wide angle, you are going to need more glasses again.

Of course, if you want a fast wide-angle, you will need even more glasses, so the lens gets bigger and heavier and more expensive. As for zooms -- well, you need an awful lot of glasses.

The two most basic considerations with any lens are focal length and aperture.


Focal length

When you were young, you almost certainly focused the sun with a magnifying glass, to set fire to pieces of paper or dry leaves. When you did, of course, you were focusing an image of the sun, which is (in photographic terms) effectively infinitely far away. The distance between the magnifying glass and the paper corresponded to the focal length (f.l.), which is the focusing distance required to get a sharp image at infinity. With most magnifying glasses this is about 4 to 6 inches, 10 to 15cm, but it can be less or more. This is an easy way to tell the approximate focal length of an old, unmarked camera lens.


burning glass


garden roofs

Back garden roofs

This is one of the walls of our back garden, plus the roof and wall of the house next door. It was shot by the light of the setting sun, very rich and red, and underexposed slightly (on a high-saturation film, Kodak Elite Chrome 100 EBX) in order to saturate the colours still more. What may be less apparent is that it was shot with a soft-focus lens, a 1930s Leitz Thambar (90mm f/2.2) on a Leica MP. Broad areas of colour and tone, and dramatic shapes, can have every bit as much impact as super-sharp detail. It depends on the picture...


f-l magnificstion




Focal length and magnification

The longer the focal length, the bigger the image will be projected onto the focusing screen, film or image sensor. As the illustration on the left shows, it is a simple linear relationship: a lens with twice the focal length produces an image that is twice the size.

You can work it out with trigonometry if you like, but you can also just draw it with similar triangles, as in the illustration.

In a moderately dark room, you should be able to form an image of the scene outside the window (during daylight) by holding a magnifying glass (or a photographic lens) close to a sheet of paper and varying the separation between them until you see an image. You need to have the lights off or the light inside the room will be too bright for you to see the image.

Oh, yes; and the image will be upside down.


brick lane


Pub, Brick Lane, London

Users of digital cameras are often poorly served for extreme wide-angles, especially if they use one of those systems that shares a common mount with film cameras. This is because digital sensors are usually smaller than full-frame 35mm film, and therefore only show the central part of the image. Roger shot this with a Leica M8 and 15/4.5 Voigtländer Super-Wide-Heliar, which has the same coverage on the 18x27mm Leica sensor as a 20mm lens on full-frame 24x36mm 35mm film. ISO equivalent was set to 2,500, the maximum possible.


The aperture is a measure of the light-gathering power, and is measured in 'stops'. This term dates back to the very earliest days of photography, when the amount of light passing through the lens was controlled by inserting a metal plate with a hole in it, literally a 'stop'. For well over 100 years, the usual way to control aperture has however been via a multi-bladed 'iris diaphragm', so called because when it opens and closes, it rather resembles the iris on the eye. 'Diaphragm' in this context is simply another word for 'stop'.





Iris Diaphragm

The aperture on the left is twice the diameter of the one on the right, and therefore allows four times as much light through: remember, area varies as the square of the linear measure. The middle aperture allows half as much light through as the one on the left, and twice as much as the one on the right, and is therefore about 0.7x the diameter of the larger one or 1.4x the diameter of the smaller one.

At first sight, stop numbers are really confusing. They seem to go backwards -- a bigger opening is a smaller number -- and the numbers appear somewhat arbitrary. Perhaps surprisingly, they are very easy to understand.

They go backwards because they are a written as a fraction of the focal length of the lens. The effective diaphragm aperture of an 'f/2' lens is one-half the focal length, so a 50mm f/2 has an effective aperture of 25mm. Stop it down to f/4, and the effective diaphragm aperture is 12.5mm. With a 100mm lens, f/2 would be 50mm and f/4 would be 25mm. Because the aperture is calculated relative to the focal length, it is called a relative aperture. We will come back in a moment to why this is important.


linda wedding


Lauren Trezise

Lauren is one of Roger's oldest friends and he loves her like a very dear sister. This is why he allowed himself to be persuaded to photograph her wedding to Greg Trezise.

For a traditional yet romantic effect, he used his old 50mm f/1.2 Canon, shooting black and white and sepia-toning the print. The lens is shown below. It dates from the early 1960s and fits any rangefinder camera with a lens mount compatible with Leica screw thread or (via an adapter) the Leica 4-claw M-bayonet. For the picture of Lauren it was mounted on a Leica M4P, not a Voigtländer Bessa R3a as here. It's a big area of glass!


canon 50



The apparently odd numbers of the f-stop sequence are explained by the fact that the interval between 'whole stops' is the square root of two, 1.41 recurring. Thus an f/1.4 lens passes half as much light as f/1 (there are very few f/1 lenses) and an f/2 (1.4 x 1.4) passes half as much as an f/1.4. Multiply f/2 by 1.4 and you get f/2.8, the next whole stop. A much easier way to look at it is that every alternate whole stop is halved (or doubled), thus 1 - 2 - 4 - 8 - 16 etc. and 1.4 - 2.8 - 5.6 - 11 etc.



Whole Stops


Half Stops


One-third Stops

Why the square root of two? Because, as noted in the caption to the iris diaphragm above, apertures are expressed as part of a linear relationship, based on the diameter of the aperture, and the amount of light that gets through depends on the area of the hole, which is square measure. In other words, if the hole is twice as big (linear measure) it lets through four times as much light (square measure).

What is really clever is that a given relative aperture (calculated as described above) will always pass the same amount of light, regardless of the focal length. Thus, if your meter recommends f/8, you can set f/8 on any lens and it will pass the the same amount of light, with only trivial variations for lens design and manufacturing considerations.

In practice, the actual size of the hole and the effective size of the hole may not be the same, depending on the lens design. What is more, again depending on the lens design, the apparent size of the hole may vary. It can even look bigger from one side of the lens than from the other, which is no great surprise when you think that you are looking at it through a magnifying glass. This is why we have been careful to say 'effective size'. But it's all close enough to understand the theory.

To make life a little more interesting, lens manufacturers sometimes use half-stops and third-stops for maximum apertures, or use a maximum aperture that doesn't fit any sequence at all. And very old lenses may not use the whole-stop sequence anyway: old Leica Elmars go f/3.5, f/4.5, f/6.3, f/9, f/12,3, f/18. But the list on the right is of whole, half and third stops, as far as f/16; after that, we've just given whole stops.

You'll notice that there's a lot of rounding in these figures, but that's just the way that it is. In fact, some manufacturers use other numbers for the same aperture: a lot of f/1.4 and f/1.5 lenses are identical in speed, as are f/2.4 and f/2.5, and f/3.4 and f/3.5. For that matter, f/12.3 and f/12.6 are used pretty much interchangeably if they are used at all.

There have also been a number of f/0.95 lenses. Most are for 16mm and even 8mm ciné use, or video, but Canon made one for their old rangefinder cameras and there was another on an extremely rare sub-miniature camera. The difference between f/1 and f/0.95 matters more to the marketing department than the photographer. Likewise, there have been quite a lot of f/1.9 lenses, presumably because it looks faster than f/2, though the difference (1/6 stop) is trivial. Then there are old Vivitar Series 1 lenses with weird maximum apertures, such as 135/2.3 (1/6 stop slower than f/2.2) and 200/3 (1/6 stop slower than f/2.8) -- both excellent lenses, incidentally. And a very few German lenses were marked f/5.5 for reasons we have never been able to figure: it's effectively f/5.6.





Toho FC45X with 120/6.8 Schneider Angulon

With a few old wide-angle lenses for view cameras, the maximum aperture was for focusing and composition only; neither sharpness nor coverage was really adequate until the lens was stopped down to f/16, the widest recommended working aperture. This is why the older and seemingly faster Schneider Angulon series was replaced with the seemingly slower Schneider Super Angulon f/8 series. The big difference was that the Super Angulons could be used at full aperture, though coverage still increased on stopping down. On the other hand, coverage of a Super Angulon is greater at any given aperture than the coverage of a plain Angulon at the same aperture.









































Variable apertures

Some zoom lenses have variable apertures. This is easy to understand, because obviously the focal length varies, and if the diaphragm remains the same size, the f/number must vary too. Imagine a 100-200mm zoom with a 25mm diaphragm, and you have f/4 to f/8. In practice, the point we made above about the effective size of the diaphragm comes into play, and there's often mechanical adjustment of the aperture too, so you'd be more likely to have f/4 at the 100mm end and f/5.6 at the 200mm end.

A few zooms have variable apertures throughout the focal range, but the majority (through cunning control of the diaphragm) vary only in maximum aperture, i.e. by the time you have stopped down to the maximum aperture at the slow end of the range, the aperture remains constant at all focal length settings.

cyclist times square

Cyclist, Times Square

Roger's favourite lens on 35mm is his old 35/1.4 Summilux, dating from about 1985; the first Leica lens he ever bought new. Objectively, there are many better, newer lenses, and the contemporary 35/2 Summicron was better too; but if you like shooting in low light, sheer raw speed and a compact, unobtrusive lens have their attractions. This was on his Leica M4-P.

Standard, wide-angle and long lenses

Traditionally, a 'standard' lens was one with a focal length roughly equal to the diagonal of the negative it was designed to cover. This was reckoned to give the most natural perspective. Thus, a 'standard' lens for a whole-plate negative (6½ x 8½ inches, 165 x 216mm) was 10 or 12 inches (250-300mm, actual diagonal 271mm) while a 'standard' lens for a quarter-plate (3¼ x 4¼ inches, 83 x 108mm) was 5 inch/127mm.

This worked fine for contact prints, because most people tend to examine a small print from closer up, and a bigger print from further away, but it gets very complicated when you start to make enlargements. There is a (paid) module about perspective, which goes into this much more deeply.


Anyway, a lens that covered a wider angle soon became known as a wide-angle lens (the English language is wonderfully simple sometimes), while one that covered a smaller angle became known variously as a long-focus (or simply 'long') lens or as a telephoto. Strictly the latter applies only to a particular type of optical design (below), so purists prefer 'long', but 'telephoto' and its abbreviation 'tele' are so deeply entrenched that only purists bother.


The Pacific at Morro Bay

All's fair in love and war -- and photography! This is not very sharp; the contrast is not very high; and the exposure is questionable. On the other hand it seems to us to capture very well two children on the beach in front of the vast Pacific ocean. Roger used either our 800/11 Vivitar Series 1 (which we no longer have) or our 600/8 (which we do) on a Nikon F, shooting on Fuji RFP ISO 50 slide film.


Because the same focal length will cover different angles on different formats, it is easy to see see how the same lens might function as 'standard' on one format, 'long' on another (where you use only the central part of the image) and 'wide angle' on yet a third (if it will actually cover the third format -- see below). For example, our 360/9 Apotal is a 'long standard' on 8x10 inch -- but we bought it from someone who had been using it as a long 'tele' lens on 35mm, in a custom-made mount. The second table in this module lists a few typical focal lengths and their applications:


morro bay



Standard on small digital imaging sensors; ultra-wide on 35mm


'Wide standard' on 35mm; 'short tele' on small digital imaging chip; ultra-wide on roll-film


Standard on 35mm; wide on roll-film; ultra-wide on 4x5 inch.


Standard on 6x9cm roll-film; 'short tele' on 35mm; wide angle on 4x5 inch; ultra-wide on 8x10 inch


Standard on 8x10 inch; long tele on 35mm; long on 4x5 inch; wide-angle on 11x14 inch


Equivalent focal lengths

You will often see digital camera zooms quoted as 'equivalent' focal lengths for 35mm photography. The only excuse for this is that most photographers have a pretty good idea of the coverage of various focal lengths on 35mm: 21mm as very wide, 35mm as a 'normal' wide angle, 45-55mm as 'standard', 85-105mm as a 'portrait' lens, etc.

The actual dimensions of the so-called "1/2.5 inch" sensor (the smallest standard digital imaging sensor for most still cameras) are 5.7 x 4.3mm: the 1/2.5 inch figure is all but meaningless and refers to the TV camera imaging tubes of the 1950s. Anyway, Pythagoras tells us that the diagonal on this size sensor is 7.1mm, so a 'standard' lens would be 7.1mm, 'equivalent' in 35mm terms to 43.3mm (the diagonal of a 24x36mm frame); a multiplication factor of a bit under 6.1x. A 5-15mm zoom might therefore ore be marketed, after some rounding to tidy the numbers (30.49 to 91.48), as a '30-90mm equivalent'.

It is not apparent until you give it a moment's thought, but strict equivalents are impossible unless the formats are the same shape, because vertical and horizontal coverage will vary. You can therefore compare 35mm and 6x9cm (true 56x84mm), both 2:3, and say that 20mm on 35mm is very close to 47mm on 6x9cm. Likewise if you enlarge 56x72mm (Linhof '6x7') three times you get 168x216mm, almost identical to the 165x216mm of whole plate (6½ x 8½ inches). But you can't really compare 38mm on 44x66mm and 56x56mm (both near enough 80mm diagonal) because the results do create a different impression.


sturdivant house


Sturdivant House, Selma, AL

A 14mm ultra-wide on 35mm (this is a Sigma 14/3.5 on a Nikon F, shooting on Ilford XP2) is pretty much exactly equivalent to an unobtainable 33mm on 6x9cm (56x84mm actual), but it doesn't necessarily mean a lot to equate it to 97mm on 8x10 inch because 8x10 inch is a different shape: not as long and thin.

What lenses do you need?

No-one can tell you -- and you should never trust anyone to tries to do so.

There is a romantic appeal in the idea of 'one camera, one lens', and a very few photographers use exactly that. They really are very few, however, and there are a lot of untrue legends about. In particular, Henri Cartier-Bresson did not restrict himself to a 50mm lens, though it seems to have been his favourite.


tree roots



A lot depends on what sort of photography you do. As already mentioned, Roger uses a 35mm lens (on 35mm) for the great majority of his pictures, followed by 75mm, followed (in no particular order) by 21mm, 50mm, 135mm and 200mm. He rarely uses a 15mm and almost never uses 28mm. Frances's favourite lens is a 50mm, followed by 21mm or 28mm and 90mm; she almost never uses 35mm, or anything longer than 90mm. It's such a personal matter that even husband and wife, with access to the same lenses, make different choices.

Tree roots, Montreuil s/Mer

Quite apart from preference, there's a question of what you have and can afford. Frances shot this many years ago on 6x7cm Kodak colour negative film, using with one of our 'baby' (6x9cm) Linhofs and, as far as we recall, a 105/3.5 Linhof-selected Schneider Xenar. The Xenar is not a particularly well regarded lens but we got a number of good pictures with it, and to be honest, the extra quality obtainable from our current 100/5.6 Schneider Apo-Symmar is rarely as important as the compositional skill with which we use the camera and lens.

Fast lenses are obviously more versatile than slow ones, but equally, they are bigger, heavier and more expensive, and often slower-handling too, so unless you regularly need the speed, you may be better off with a slower lens. Even though we have 50/1.5 and 50/1.2 lenses (and access to a 50/1), Frances is especially fond of her 50/2.5 Color Skopar, which has the further advantage that it takes the same 39mm filter size as her 21/4 and 90/3.5.

Old and New Lenses


Lenses are subject to what we call the 'quality threshold'. Up to a certain level, a better lens means better pictures, and it is worth spending the money. Beyond that level, a lot more depends on the photographer's skill than on the lens. Of course the level varies from photographer to photographer, but broadly, we'd suggest the following:

With standard lenses, anything from a major manufacturer should be good or very good since the 1960s (or even 1950s for the very best such as Zeiss or Leica). Be wary of older lenses from lesser manufacturers, especially fast lenses beyond about f/2, and of most East European lenses, though you may find that they give you a soft, romantic look that you rather like. The very fastest lenses (f/1.5 and faster) should always be treated with suspicion before about the 1970s.

When it comes to wide-angles, there were still plenty of rather nasty lenses about even in the 1970s, especially at 28mm or wider, though (for example) a 21/4.5 Zeiss Biogon from the 1950s is still a superb lens even today.



Mearle's Drive-In

Fortune, it is said, favours the prepared mind. It also favours the prepared (or just plain lucky) photographer. Roger had his late-1950s Zeiss Biogon 21/4 on his Leica M4-P (via an adapter) when he found this Corvette parked outside Mearle's in Visalia and shot it on Fuji RFP ISO 50.




Long and tele lenses are pretty easy to design and build, so with the exception of very fast lenses (f/2 and faster) before about 1970, the quality should be subject to the same considerations as standard lenses.


ber zoom girl



Zooms are another matter. Even the finest zooms of the 1970s tend to lack contrast by modern standards, and anything earlier or worse should be treated with great suspicion unless you want a soft-focus portrait lens. The very best zooms from about 1980 onwards are pretty good, and most have improved steadily since that date, but they are never going to be as good as the best prime lenses from the same manufacturers.



Girl, Bermuda

This was taken with a truly rotten lens, a 90-190mm f/5.8 Yashinon which was Roger's first zoom, bought second-hand in Bermuda in about 1967. Then again, he was a truly rotten photographer in those days...

The film was ex-government Ilford FP3 that had gone out of date in 1963 (and had been stored in sub-tropical heat before and since); the camera was a Pentax SV. Actually, he wishes he still had this lens, because it would be great for portraiture and other soft, impressionistic effects.

Finally, remember that even the finest lenses can succumb to old age. Scratches and other mechanical damage are one thing, but focusing mounts can stiffen up and glass can cloud up (especially with internal deposits of distilled-off lubricants). Only the best (or best-loved) lenses warrant the expense of professional cleaning, but they can be transformed if you are prepared to spend the money. For example, our 50/1.2 Canon rangefinder lens went from effectively unusable to being pretty good under contrasty lighting with a deep lens shade, after the ministrations of Optical Instruments (Balham) Ltd., who rather confusingly are in Croydon.

Red trough, Spain

Quite a lot of the pictures so far in this module were shot with old or very old lenses, so here is something that is at least 21st century, taken (by Frances) with a brand new 28/2.8 Zeiss lens on the then-new Zeiss Ikon camera. If you can afford it, new lenses are almost always better than older lenses of the same overall specification from the same manufacturer: better resolution, more contrast, less distortion... This does not of course mean that they will automatically give you better pictures. That's down to you...

At this point, we've said most of what needs to be said about the practical side of lenses, so you may care to leave this module here: we've put in the options on the right. Alternatively, if you are interested in a bit more background, read on.

Go back to Basics or to the Home Page

Go back to the lists of Photo School modules, illustrated or unillustrated


If you take a lens that is designed for a smaller format, and try to use it to take a picture on a larger format, you will often find that all you get is a circular image that deteriorates fast towards the edges. All lenses have a 'circle of illumination' and a (somewhat smaller) 'circle of sharp coverage'. Both terms are pretty self-explanatory. What is slightly less obvious is that a wide angle has a bigger-than-usual circle of sharp coverage for its focal length. Most non-wide-angle lenses cover about 45-50 degrees, but ultra-wide-angles may cover anything up to about 120 degrees.

Lens Design

From all of the above, you can see why some lenses are more complex than others. The bare minimum of glasses you can use for a critically sharp lens is two, and then, only if you are using a very long-focus lens of very modest aperture. There have been a couple of Leica lenses of this design, such as the 400/6.8 and 560/6.8 Telyts. In both, the glasses are cemented together in a single 'group': a 'cemented doublet'.


If you want more angular coverage, still with a pretty slow lens, you can take two cemented doublets (four glasses in two pairs), or just four separate glasses (positive, negative, negative, positive) and place them symmetrically on either side of the diaphragm. The cemented version of such a 4-glass lens is a 'rapid rectilinear' or RR design, though 'rapid' is a relative term: you're looking at a maximum speed of f/8 or so before quality starts to suffer. You can however increase both quality and coverage to a considerable extent, and speed to a more modest extent, by using still more glasses. The Goerz Dagor uses two groups again, but they are cemented triplets, and there have been 2-group, 8-glass (cemented quadruplet) and even 2-group 10-glass (cemented quintuplet) designs.

In the days before lens coating (see below) it was was much more important to keep the number of glass/air surfaces to a minimum, because losses by reflection at cemented surfaces are far lower than at glass/air surfaces. Modern lens designers are much more willing to use multiple groups as well as multiple glasses.


Sunshade, South of France

The bigger the format, the harder it is to build a fast lens that will cover -- and the bigger, heavier and more expensive the lens becomes. The 35/5.6 Rodenstock Apo Grandagon that Frances used on her Alpa 12 S/WA to shoot this on 6x9cm Ilford HP5 Plus has almost identical coverage to a 15mm lens on 35mm -- but for 35mm you can buy a 15/2.8 (from Zeiss), two stops (four times) faster.




4 glass 4 group

Symmetrical 4-glass 4-group lens

With a fully symmetrical lens, many of the aberrations introduced by the left-hand group are cancelled out by the right-hand group, which has the same aberrations, but in the opposite direction. Perfect cancellation occurs only at same-size reproduction. If these glasses were cemented in pairs, instead of separated, this would be a Rapid Rectilinear.

6 glass 4 group

Symmetrical derivative 6-glass 4-group

For extra speed, the two middle groups of a symmetrical design have been split into cemented doublets, and strict symmetry has been abandoned in the interests of better performance with distant subjects. This is the 85/2 Jupiter, based on the pre-war Zeiss Sonnar of the same focal length and speed.

At f/6.8 (half a stop faster than an RR), the six-glass, two-group Dagor is still a cult lens today, but there are countless newer designs. The Rudolph Planar was again a symmetrical design, 4-group, 6-glass, singlet-doublet-doublet-singlet, allowing f/5.6 or better, and later derivatives of the Planar (with new glasses) went to f/2.8 and faster. The most complex modern symmetrical derivatives may use eight or nine or more glasses for maximum speed and coverage.

But symmetrical derivatives are not the only option. The classic f/6.3 'Cooke Triplet' uses three glasses, all separate, though you get much better quality and more speed if you split one of those singlets into a cemented doublet. The famous Zeiss Tessar uses this sort of design. Once again, there are endless Cooke Triplet derivatives, with more and more glasses. Even in the 1930s, while the Zeiss Sonnar 50/1.5 had only three groups, it had seven glasses: the first 'group' (the front glass) was a singlet and the other two were cemented triplets.


cooke triplet

The Original Cooke Triplet Type


section tessar

Zeiss Tessar Type


50-1.5 sonnar

Zeiss 50/1.5 Sonnar Type

Obviously more glasses cost more money, but there is even more to it than this. Some of the more exotic glasses are very expensive indeed, and the fun really starts when you use aspheric surfaces. These allow better correction of aberrations with fewer glasses, but grinding common (spherical) surfaces is very, very much cheaper than grinding aspheric surfaces. Several tricks have however been evolved to reduce the cost, including moulded blanks that require less finishing; turning blanks in a lathe; and moulded plastic aspheric elements attached to conventional glasses, the so called 'hybrid aspherics'.

Lakeside, Slovenia

Voigtländer makes considerable use of hybrid aspherics; Frances shot this with the 35/1.7 Ultron on her Bessa-T, with an orange filter for contrast on a slightly hazy day and a tripod for maximum resolution at f/8. Film was Ilford HP5 Plus -- you can tell from the tonality, even on the screen -- and the print is on Ilford Multigrade Warmtone.

Telephoto and reverse-telephoto lenses

We have already mentioned two major groups of lens design, the symmetrical and triplet derivatives, but two other important groups are telephoto and reverse-telephoto. The former are lenses of long focal length, physically shortened by the incorporation of at least one negative (concave) element behind the image-forming group(s): most have two, though a few have three glasses in the telephoto group. This brings the image-forming group closer to the film or image sensor, shortening the lens overall. The sole advantage of a true tele lens is that it is more compact than a long-focus of the same focal length; the tele component will invariably reduce image quality, though with the best designs, this reduction is negligible.

Reverse-tele lenses are, as their name suggests, the opposite: they are physically lengthened via a negative group in front of the image-forming group(s). This is necessary with single-lens reflex (SLR) cameras where otherwise there would not be enough room for the mirror between the back of the lens and the film or image sensor. Again, the reverse-tele component necessarily reduces image quality as compared with a plain no-reverse-tele wide-angle lens, but if you have to fit in the mirror, there may be no choice. They also make it easier to ensure that the light leaving the lens arrives at the film or image sensor in a more parallel state than from a non-reverse-tele lens. This reduces vignetting (darkening of the corners) with film cameras and is essential with most digital sensors which rapidly lose efficiency as the angle of incidence of the light falling upon them increases.

Reverse-tele lenses are often known as Retrofocus, which was the name of the first successful reverse-tele lens (designed by Angénieux) for still photography.


section tele

Telephoto lens (Tair-3)

The left hand group, at the front of the lens, is the image forming group; the right hand group is the negative (tele) group.


reverse tele

Reverse-telephoto lens (Mir 1)

The singlet on the left (the front of the lens) is the negative Retrofocus lens; the symmetrical derivative group on the right is the image forming group.

Other designs

There are may other kinds of lens design, including the Petzval design (the very first purpose-designed photographic lens was designed by Joseph Petzval in 1840) which is ideal for fast, long-focus lenses, though they have very small coverage for their focal length; mirror lenses that use reflecting elements instead of refracting elements, like an astronomical telescope; catadioptric lenses that incorporate both reflecting and refracting elements (most modern 'mirror' lenses are of this type); and of course zooms, which can be thought of as combination telephoto/reverse teleophoto lenses, where the variations in focal length are achieved by varying the position of the zoom groups. This is not perfectly accurate, but it gives a pretty good idea of how they work. Once again, the best zooms can never be as good as the best 'prime' (fixed focal length) lenses, but the best can be amazingly good. The worst, on the other hand, can be amazingly bad. Then there are fish-eyes, varifocals (like zooms except that they need to be refocused whenever you change the focal length) and more; but the above brief summary is enough to be going on with.



door loches



The design process

Originally, lenses were designed via manual ray-tracing. The extent to which the ray was bent as it hit each surface was calculated, and the curves and thicknesses of the glasses determined accordingly. This was very labour intensive. Reputedly, artillery students were used in large numbers, as being a readily available source of those who could perform the necessary calculations. The process was greatly speeded in the 1920s and 30s by the adoption of mechanical calculators (like comptometers) but it was still a painfully slow process.

When computers came in, a third era arrived, and calculations were speeded again. There has since been yet a fourth era. Lens designs are now done using wave-front models instead of ray-tracing models, which is one reason why modern lenses are so amazingly good, even at the bottom end of the market (cheap zooms). No matter how good the cheap lenses become, though, the best will always remain better.


Door, Loches

Voigtländer's lenses for their rangefinder cameras includes a large number of extremely fine objectives, but the 90/3.5 Apo Lanthar that Roger used to take this shot (on Ilford XP2 Super, printed on Ilford Multigrade Warmtone) is one of the finest. They have been able to introduce so many excellent designs, made in relatively small numbers, because of modern computerised lens design/optimization programs.

Flare and coating

In any lens, a small percentage of the light is reflected back from each air/glass surface. Except at the front of the lens, where it is simply reflected back out, this non-image-forming light then bounces around inside the lens. Some of it is absorbed by baffles and blacking, but some ends up on the film (or image sensor) in the form of flare. This lightens the darkest areas of the image, thereby compressing the brightness range as compared with the original subject brightness range (there is a free module about this). This is known as veiling flare, and is at least as important in many shots as the effect that most people think of as flare, which is images of the diaphragm reflected from the various elements of the lens.

The more air-glass surfaces, the more flare, and until lens coating came into widespread use in the 1950s, this was a strong incentive to lens designers to keep the number of groups to a minimum: flare is much less at a cemented glass-glass interface, hence the 3-group, 7-glass design of the 50/1.5 Sonnar mentioned above.


sunset flare




Chapel, Monolithos, Crete

When you are shooting straight into a powerful light source -- and they don't come much more powerful than the sun -- it's a good idea to remove any protective UV filters you may have on the front of the lens, but there is still a strong likehood of internal reflections of the diaphragm being imaged on the film, as seen here. Nor will a lens hood necessarily make much difference.

Veiling flare, on the other hand, can be at its most insidious on a bright, overcast day, when the deepest possible lens hood can make an enormous difference. Roger shot this with his Leica M4-P and 35/1.4 pre-aspheric Summilux.

As coating improved, the penalties for extra air-glass interfaces decreased, and contrast increased to the point where large numbers of groups became commonplace. For a zoom, a 14-glass, 10-group design would not be regarded as unusual, and for a fast wide-angle, a 9-glass, 7-group lens is quite likely.


Even so, fewer glasses and (above all) fewer groups still make for a contrastier lens, so quite apart from cost considerations, it makes sense to use the minimum possible number of glasses and groups for maximum quality -- which leads us back to super-expensive glasses and aspheric surfaces.

It is also worth adding two more things. The first is regardless of the quality of the lens design or the coating, a good lens hood (lens shade) will add contrast as well as protecting the lens from rain, bumps, spraying champagne, etc. The second is reiterating that while a clear filter makes an excellent 'optical lens cap' to protect the lens, it necessarily adds to flare and should be removed whenever flare is likely to be a serious problem, especially when shooting into (against) the light.



Liberté 1789-1989

It doesn't matter how good your lens is: you need to put it on a tripod; focus very carefully; and use a slow, sharp, fine-grained film if you want to get the best resolution. Roger shot this with the 75/2 Summicron we had on loan from Leica, before we bought it.

Although it is sharp and contrasy, it could have been even sharper if Roger had put it on the tripod instead of hand-holding it. Film was Kodak Elite Chrome 100 EBX; the camera, our Leica MP. You may also care to form your own opinion about the 'bokeh', the quality of the out-of-focus image, which some people consider a very important characteristic of any lens.



Manufacturing quality

The very last point to make is that when you buy an expensive lens, one of the things you are paying for is sheer durability. A cheap lens can be knocked out of alignment by a fairly slight jolt; an expensive one can withstand extraordinary abuse. Two stories about the old 35/1.4 Leitz Summilux illustrate this. Roger dropped his some five or six feet onto cobbles in Prague in the early 1990s, while Geoffrey Crawley lost his in the bilges of his boat for six months (he thought it had gone overboard) in the 1980s. Both lenses are still in use...


Shrine, Poland

The Blessed Virgin is especially highly regarded in Poland. This taken with Roger's Summilux (probably on the M4-P) about a decade after he had dropped it on the cobbles in Prague. Film was Kodak Elite Chrome 100 EBX.

Go back to Basics

Go to the unillustrated list of modules (in either alphabetical or date order)

or go to the illustrated list of modules

or go to the home page


© 2007 Roger W. Hicks