sharpness i -- technical considerations

To the inventors of photography, its ability to capture images sharply was fascinating. Suddenly, a degree of detail was available which hitherto would have made great demands on the most painstaking of miniaturists. For half a century or more, this love affair with sharpness grew and blossomed, thanks to ever-sharper lenses and (to a lesser extent) sharper and sharper emulsions.

There were always a few photographers who were not so besotted with ultimate sharpness -- Julia Margaret Cameron is a good example -- and by the end of the 19th century there was a positive fashion for wooly pictures, some of which were very good, but the majority of which were very bad.

But then, as Theodore Sturgeon famously remarked when someone pointed out that ninety per cent of science fiction is rubbish, ninety per cent of anything is rubbish.

Since then, there have been fashions in sharpness and unsharpness, but the default has, for the most part, been sharpness. The main reasons to depart from sharpness are first, for artistic effect, and second, through incompetence. No-one can gainsay the former, but the latter can be overcome.

 

 

Wall, poster, graffiti, Arles

It's little more than a snapshot -- but it's very, very sharp. That's because it uses a sharp film (ISO 100 slide, Kodak Elite Chrome EBX); a high shutter speed for the focal length (1/250 with a 35/1.4 Summilux); an optimum aperture for the lens (around f/8); careful focusing; and the advantage of a pretty good depth of field, thanks to the modest aperture.

It's also bright, contrasty lighting, which always helps -- look at the shadow of the window-stay on the right -- and a contrasty subject we 'know' to be sharp: we can imagine the texture of the wall, and fine print (provided it is crisply resolved) adds still more to the illusion of sharpness, the subject of another module.

The pre-aspheric Summilux is often given a bad press by those who have never used one, or by those who shoot test charts at full aperture, but at f/5.6 or so it is as sharp as any lens; and when you really need the maximum aperture, you are rarely shooting for maximum sharpness.

grafitti & poster

 

defining sharpness

Strictly, sharpness is the abruptness of transition from a light area to a dark area. If an image is perfectly sharp, then there will be an instantaneous transition from dark to light: in the limiting case, black on one side of the line, white on the other.

This can be measured and defined with a microdensitometer and the acutance equation, a fearsome piece of mathematics for which readers are referred to page 17 of Quality in Photography. As we point out in the book, this equation is of limited usefulness even to those who can understand it.

A rather more useful way of thinking about it is visually, and as soon as you do that, sharpness blurs (literally) into resolution.

The topmost chart is the sharpest. It contains eleven line pairs, that is to say, eleven white lines and eleven black lines.

 

The next chart down contains exactly the same number of lines, i.e. the resolution is the same, but the transition from black to white is smeared; it is a lot less clear and the chart is less sharp.

 

The third chart is less clear still, and the bottom chart is the least clear of all. In fact, with the higher-contrast chart on the left, the precise location of the bars is becoming unclear. Although this is a digital re-creation of an optical effect, the parallels are pretty much exact. You can also see how, intriguingly, the deterioration is more obvious in the higher-contrast target.

 

At this point, the easiest way to discuss sharpness is in terms of the maximum number of line pairs per unit distance that can be easily distinguished. This is normally expressed in line pairs per millimetre (lp/mm).

 

Of course, 'easily distinguished' is open to some interpretation. There comes a point where contrast is so low as to be useless, but exactly where this is will be a matter of opinion as much as a matter of hard science.

bar chart 1

bar chart 2

bar chart grey 2

bar chart 3

bar chart grey 3

bar chart 4

bar chart grey 4

sharpness, contrast and resolution

If two lenses (or lens/film combinations) are compared, the one with with higher contrast but lower resolution may appear sharper than one with higher resolution but lower contrast: it may not be able to resolve as many line pairs, but the ones it can resolve, it resolves more sharply.

Before the advent of lens coating, this was a major consideration. Famously, Leica lenses tended to favour resolution over contrast, while Zeiss lenses tended to favour contrast over resolution. Advocates of both marques therefore claimed superiority. Both were equally right or equally wrong.

the useful limit of resolution

The ultimate useful limit to resolution is what the human eye can see. The trouble is that this is a slippery concept. Viewing distance is critical, obviously, and the usual arbitrary choice here is 10 inches, 25cm. At half this distance (at which point most people except the very young will need glasses), standards are doubled; at twice this distance (a reasonable viewing distance for many prints), they are halved.

misterie

 

Even then, there is no absolute agreement about what the human eye can see at a given distance. Quite apart from variation between different people, estimates of what constitutes 'sharp' range from around 5 lp/mm to around 30 lp/mm. The former is 'pretty sharp for most purposes' and the latter is the so-called 'vernier' resolution, the ability to see a discontinuity in a fine line -- something at which the human eye excels.

Our own view is that somewhere under 10 lp/mm the picture starts to look unsharp, while beyond about 20 lp/mm there is nothing to be gained in a real-world picture, even if the difference can be demonstrated with test targets. We used to believe that 8 lp/mm was 'sharp' but now we tend to set the standard a little higher: maybe 10-12 lp/mm.

 

 

The Misterie of Exposure

An 8x10 inch contact print such as this is likely to be as sharp as anyone can see, subject of course to correct focusing, depth of field and correct use of camera movements. Any lens should resolve well in excess of 30 lp/mm (see above), even after allowing for the 'bellying out' of 8x10 inch film in the holder. Roger shot this on Ilford Ortho Plus (note the very dark rendering of the mahogany Gandolfi) with our 8x10 inch De Vere and 14 inch (356mm) Cooke Apotal process lens.

 

sniffing the print

The bigger a print is, the greater the normal viewing distance. Where we might look at a postcard-sized print from as close as we can easily focus, maybe eight inches/20cm or so, we are more likely to look at a 20x30 inch/50x75cm print from a yard/metre or more away. The standard of resolution of the big print can therefore afford to be about one quarter or one fifth as great as that of the smaller print, so if we need (say) 20 lp/mm in the postcard, we can get away with 5 lp/mm in the big print.

This leads many 35mm users, and some 120 users too, to argue that no-one should look at big prints from closer; they say disdainfully that what they call 'sniffing the print' is in some way cheating.

We are of the exact opposite opinion. If the print doesn't look sharp when you start looking more closely at details, then it isn't sharp. This is not necessarily a significant fault, but equally, it is foolish to pretend that a print is sharp when you can see that it isn't.

This is one reason -- and only one -- why we often prefer small prints to bigger ones, especially where sharpness is important.

 

 

Bud vase and charger

This is a 3x enlargement from a Linhof 56x72mm '6x7cm' negative. As noted a few paragraphs below, this is within a few millimetres of the size of a whole-plate contact print, from which we believe it is indistinguishable.

The probable resolution on the film is around 70 to 80 lp/mm, so even after allowing for losses in enlarging, the resolution on the print should be in excess of 20 lp/mm. This is more than sharp enough for all practical applications, even if you do 'sniff the print'.

The camera was a Linhof Technikardan 4x5 inch with roll-film reducing back; the lens, as far as we recall, a 210/5.6 Rodenstock Apo-Sironar-N; the film, Ilford Delta 100. The print us on Ilford Multigrade Warmtone.

 

resolving power and enlargement

Almost any modern lens, and the vast majority of older ones too, for that matter, can resolve more detail than the human eye can see. This is one reason, we are convinced, why ultra-large-format contact prints (bigger than 8x10 inch) are increasingly popular. They create a super-sharp 'magic window' with a quality that just doesn't exist with most enlargements.

Even then, we believe that a modest enlargement (a maximum of about 3x) can in fact be indistnguishable from a contact print of the same size. Our favourite example of this is 3x the Linhof 56x72mm format, which is 168x216mm, effectively identical to a whole-plate contact print: whole plate is 6½ x 8½ inches, 165x216mm.

If you want an admittedly high 20 lp/mm on the print, a 3x enlargement with a perfect lens requires only 60 lp/mm on the negative. Allow that no enlarging lens is perfect, and you might need 70-80 lp/mm on the negative, but even so, this is realistically attainable with many medium format cameras.

 

green shutters

 

Shutters, Sibiu (Hermannstadt), Romania

We find that colour pictures can often be enlarged more than black and white, because colour does not suffer from greater degrees of enlargement in the same way as black and white. This allows the sharpness of the lens to be exploited to the full -- and this is the sharpest lens we own, the 75/2 Summicron, here mounted on Roger's Leica MP. The film is Fuji Sensia ISO 100.

Going in the opposite direction, the very best 35mm cameras, lenses and film can produce a fairly reliable 100 lp/mm on the negative, at least centrally. This implies a 5x enlargement with a perfect enlarging lens to give to 20 lp/mm on the negative, or maybe 4x with a real-world lens: the higher the resolution on the negative, the bigger the 'fudge factor' you need to build in for the imperfection of the enlarging lens. Drop to 12 lp/mm as sharp enough, and a 7x or even 8x enlargement should still look very sharp indeed. An 8x enlargement off 35mm is 192 x 288mm, or about 7½ x 11 inches -- a good-sized picture on a sheet of 11x14 inch or even 12x16 inch (30x40cm) paper.

tonality

In our view, what does for oversize enlargements in black and white is not sharpness but tonality. With colour, you can go on enlarging until you run out of sharpness, but with black and white, the tonality usually starts to look significantly inferior to that of a contact print at somewhere between 4x and 6x, depending on the exposure, film and developer. This is a result of the half-tone effect, which will in due course be the subject of a free module of its own.

donjon de moncontour

 

This, for us, is another compelling reason for not enlarging too much. On the other hand, if you are going to go up more than about 4x to 6x, you will often get better tonality by going much bigger, 8x or more. Again, we will explain why we think this happens in the module on the half-tone effect, but we are not the only ones who have noticed what we call a 'tonality gap' at intermediate enlargement ratios.

 

Donjon de Moncontour

Frances shot this in the course of a camera test, with the Voigtländer NHS R2S and 50/3.5 Heliar, one of the sharpest lenses currently available for 35mm cameras. It is a 'vintage' sort of subject, and for us, it therefore works best as a small print: in the past, large prints were the exception rather than the rule.

As well as giving a classical tonality, this has the double benefit that it is also extremely sharp. With a mere 5x enlargement, resolution all over the print is likely to be in excess of 15 lp/mm.

Frances had also shot the Donjon on 5x7 inch, a few months or perhaps even just a few weeks previously, and she much prefers the 35mm shot -- clear proof that it is composition, not technique, that usually 'makes' a shot. Film was Ilford XP2 Super; the print is on Ilford Multigrade Warmtone.

lens and film sharpness -- and film flatness

These three cannot really be considered separately. Yes, on any given film, some lenses will be sharper than others, and with any given lens of adequate quality to explore the limits of the film's resolution, some films will be sharper than others. But as we are talking about the absolute limits here, the most important factor is film location. There are surprisingly many lenses than can resolve better than 200 lp/mm, and a fair number of films, but the limiting factor is film location and flatness.

This sets a practical limit for reliable resolution of a little over 100 lp/mm. To be sure, you can see 125 lp/mm or better with the best film/lens combinations in 35mm, but you are unlikely to see it reliably, simply because the film cannot be located precisely enough or held flat enough. In practice, anything over 80 lp/mm can be regarded as very good with 35mm.

With larger formats, the resolution falls further. Even 100 lp/mm would be unusual with roll-film, though the 44x66mm Alpa format, masked down from 6x7cm because the 38/4.5 Zeiss Biogon cannot cover the full 6x7cm, returns 90+ lp/mm fairly reliably with the Biogon. In general, 80-90 lp/mm would however be regarded as excellent with roll-film, and 60-70 lp/mm would be good.

Move to cut film, and without a vacuum back (or a dab of adhesive in the middle) the film can belly out most alarmingly. It also tends to curve slightly into a very flat W-shape from side to side. Thus, although 60 lp/mm should be readily attainable with the best lenses and almost all films, lack of film flatness may well drag on-the-film resolution down as low as 30-40 lp/mm.

 

After the Oscars

 

Frances shot this still-life on 6x7cm inch Fuji Astia slide film (RAP ISO 100) with our 4x5 inch Linhof Technikardan (with reducing back) and 210/5,6 Rodenstock Apo-Sironar-N. Because it is a colour film, the advantage of 4x5 inch would have been modest: grain would not be apparent until around 10x, and with the equipment in question, although the likely sharpness on the film is only around 80 lp/mm, the scanner acts in effect as a near-perfect enlarger lens and a 10x enlargement (520 x 720mm, 20.5 x 28 inches) would still deliver 8 lp/mm. At a more realistic borderless 12x16 inches/30x40cm, 5.5x, the resolution should be close to 15 lp/mm.

oscar programme

 

This is why we believe that the biggest single advantage of 4x5 inch film over roll-film is tonality, not sharpness. A top-flight 6x7cm negative can, as we have said above, be enlarged 3x to whole-plate size to give the same sharpness and tonality as a contact print. For the same sharpness (20 lp/mm), the biggest feasible enlargement of a 4x5 inch negative may be less than 2x. At 20x16 inches (40x50cm), on the other hand, both may deliver roughly similar sharpness of around 10 lp/mm, but the tonality of the 4x5 inch shot at 4x should be greatly superior to that of the roll-film shot at 8x.

As for the oft-quoted advantage that large format camera have movements, and roll-film cameras don't, we simply reply "Buy a medium format camera with movements, or put a roll-fim back on a 4x5 inch camera."

grain and sharpness

bar, sibiu

 

Many photographers, especially novices, confuse fine grain and sharpness: they believe that a finer-grained film is automatically sharper.

They are almost completely wrong. Yes, slower films with finer grain often are sharper than faster, coarser-grained films, but if you take the same film and process it in two developers, one for the finest possible grain and the other for the maximum possible sharpness, the finer-grained negative will not be as sharp. Yes, the grain is finer, but instead of a crisp, clear resolution of detail, you will get the same effect as in the resolution charts at the beginning of this module. In colour, too, it is rare that the finest-grained films are the sharpest: the emulsion designer must needs choose between the two, and different designers have different opinions.

Also, if you can see clear, crisp grain in a black and white picture, it may appear sharper than one with significantly higher resolution but either mushy grain or no visible grain -- though this belongs in the other sharpness module, 'The Illusion of Sharpness'.

 

 

Bar, Sibiu (Hermannstadt), Romania

 

The lens -- our 75/2 Summicron -- is super-sharp, but this wonderful painted inn sign is painted in a slightly soft way, without hard edges. As this is the principal subject matter, neither the lens nor the film (Fuji Sensia) is shown to its best advantage.

exposure and sharpness (black and white and colour)

More exposure means less sharpness. It's as simple as that. This is true even when you reduce exposure to the point where tonality starts to suffer. Exposure is always a compromise, therefore, between the best possible tonality and the best possible sharpness. With larger formats, you can better afford to sacrifice sharpness in return for tonality, whereas with 35mm the losses rapidly become more important. This is yet another reason why we do not care for big enlargement ratios.

As a rough guide, an extra stop can reduce maximum resolution by around 20 per cent, dropping 100 lp/mm to 80 lp/mm or 80 lp/mm to 64 lp/mm. These figures are based on tests we have carried out with a number of films and are a pretty fair indication.

exposure and grain

With conventional black and white films, more exposure means bigger grain. With colour negative films and chromogenic black and white films, more exposure means finer grain (but still reduced sharpness, another illustration that fine grain and meximum sharpness are far from synonymous).

In fact, chromogenic films well illustrate the choice between fine grain and maximum sharpness. Kodak's chromogenics are significantly finer grained than Ilford's -- but Ilford's are significantly sharper and (in our opinion) tonally much superior.

 

ft. delaware

 

Fort Delaware, Pea Patch Island

All chromogenic films have a different tonality from conventional films, not least because grain goes 'backwards': the most exposed areas (such as the sky) have the finest grain, unlike a conventional film where the grain is at its most coarse in the sky. This leads some 35mm users, in particular, to dismiss chromogenics as 'unsharp' because they cannot see the grain that they expect in the light mid-tones. We have been using Ilford's XP-series since the mid-to-late 1980s; this is actually the original XP1. It was probably shot with a Nikon F and 17/3.5 Tamron but quite honestly we can't remember.

diffraction limits to resolution

Absolute limits to resolution are set by the laws of optics. There is room for dispute about what consitutes 'useful contrast' and about the precise wavelength of light to use, and pedants point out that you should use the numerical aperture rather than the relative aperture, but a very useful limit to the maximum aerial resolving power of a lens can be obtained by dividing the marked aperture into 1500.

Thus, at f/2 the diffraction-limited resolution is 750 lp/mm (few if any lenses will achieve this); at f/4 it is 375 lp/mm (just about feasible); and at f/8 it is about 190 lp/mm. Life starts to get interesting at around f/16, where it obviously drops under 100 lp/mm, and by f/22 it is a mere 68 lp/mm.

On-the-film resolution will always be lower, so a top-flight lens is likely to be diffraction limited by about f/5.6 to f/8 and to deteriorate after f/11.

This also explains why relatively few 35mm lenses stop down to f/22 and even fewer to f/32, at which point the diffraction limit is under 50 lp/mm. With larger formats it matters less, because degrees of enlargement are smaller, but again you wouldn't normally want to stop an MF lens down to f/45 (33 lp/mm) and even with ultra-large-format contact prints, f/64 (23 lp/mm) or maybe f/90 (16 lp/mm) is as far as you want to go if sharpness is important.

By the same logic, don't go below f/11 (and stick to f/8 or so if possible) with your enlarger lens. Because of the reproduction ratios, effective apertures are smaller than marked apertures so the effect is worse.

sharpness, image size and movement

From all of the above, it is fairly clear that you need to choose your lens and film carefully; work at the optimum aperture; and then focus carefully, if you want maximum sharpness in your negative. Once you have your sharp negative, you do not want to enlarge it too much. But so far we have assumed that you are photographing a stationary subject with the camera mounted on a tripod. We have not yet considered subject movement or camera movement.

 

Both subject movement and camera movement have the same effect, viz., that the image moves relative to the film during the exposure. Usually, though, only parts of the subject will move, so the rest of the subject -- the part that isn't moving -- will be sharp. If the camera moves, then usually everything will be more or less blurred. There is one major exception to the latter, though, which is panning, a subject to which we shall return later in this module and which will be the subject of a short free module of its own.

 

Dancer, Bristol Arts Centre

 

Photgraphing dance in poor light requires a lot of compromises, especially if you are shooting at close quarters. There's a good chance of subject movement; a fair likelihood of camera movement; and (if you are shooting at wide apertures) there is not much depth of field to play with. The answer, invariably, is to shoot and hope. Sometimes it works. If you don't shoot, it certainly won't work.

 

This was in the days before modern fast films such as Kodak TMZ and Ilford Delta 3200 and was in fact shot on Ilford HP5 (pre-Plus in those days) pushed to 1600 by being all but boiled in Ilford Microphen: you can see the grain even on screen. Roger used the old 58/1.4 Nikkor on his Nikon F.

arts dancer

 

subject movement

In older photographic books, there are wonderful lists of the shutter speeds necessary to arrest subject movement. Often these are tabulated by distance (ten feet/3 metres, 20 feet/6 metres, 50 feet/17 metres, 200 feet/60 metres) and by direction of movement (at right angles to the camera/subject axis, at 45 degrees to the camera-subject axis, straight towards the camera). This could conjure up remarkable images of sang-froid, for example in the case of an express locomotive travelling straight towards the photographer at 100 mph (160 km/h).

In reality, the only way to choose the best speed to freeze movement is to choose whatever aperture and shutter speed you think feasible, on the basis of experience. Such experience is quite quickly gained, and we find it a great deal easier than memorizing huge and forbidding tables.

compound movement

An important thing about subject movement, though, is that often, different parts of the subject move at different speeds: a runner's feet move faster than the runner's body, for example, while the wheels of a vehicle move faster than the vehicle. This is where a degree of blur in some areas can create a far greater impression of movement than 'freezing' the movement completely.

 

ice skater

 

Ice Skater

Her right leg is pretty much at rest: you can just about count the laces in her boot. Everything else is more blurred, depending on how fast it was moving. Her right hand is recognizable only from context; her left hand is slightly stretched from moving in the same direction as the focal plane shutter. It would be possible, with enough light and a fast lens and fast film, to take a completely different picture in which she was completely 'frozen' and every detail would be clear. Different, yes; but would it necessarily, with the same composition, be any better? Roger used his Nikon F and (probably) 58/1.4 Nikkor with Ilford HP5 pushed to 1600 in Microphen.

camera movement

There is a popular rule of thumb which says that the longest 'safe' speed for hand-holding any given lens is one-over-focal-length. In other words, if you have a 50mm lens, you can safely hold it at 1/50 second; with a 100mm lens, it's 1/100 second.

As a rule of thumb, it's better than nothing, but it is subject to so many qualifications and provisions that it is not a lot better than nothing.

bouncy castle

 

First, it applies only to full-frame 35mm. If you use a smaller format, such as half-frame or digital, the original image will have to be enlarged more for the final print, so you'll need faster speeds. If you use a larger format, the standards can be relaxed: this is how the news-men of yore, shooting on 4x5 inch, could shoot 1/10 or even 1/5 second with a 127mm (5 inch) or 135mm lens and still get acceptably sharp pictures.

Second, there can be a big difference between 'acceptably sharp' and 'as sharp as can be'. Try shooting a test target (or a sheet of newspaper) with a 50mm lens at 1/30, 1/60 and 1/125. The 1/125 will almost certainly be sharpest.

 

Bouncy Castle

There is a lively competition here between subject movement, camera shake, lens resolution and poor focus. The camera was a Nikon F with 70-210mm f/2.8 Sigma Apo zoom, at 210mm. The shutter speed must have been 1/500 or even 1/1000, in order to arrest the movement, but the girl in mid-air still isn't super-sharp.

There's also the point that we have never seen a table that gives the appropriate shutter speed to arrest a girl in mid-air on a bouncy castle. On the other hand, there is always a 'dead point' in any reciprocating motion, where the subject reverses position (up-down or left-right).

 

Bouncy castle detail

The point of sharpest focus appears to be slightly behind the girl in mid-air, perhaps on the girl behind her, but microscopic examination reveals that she isn't that sharp either. In fact, nothing is really, critically sharp, which leads to the suspicion at even at 1/500 or 1/1000 second, there is still a tiny bit of camera shake. Also, even though it is a very good zoom, it is still a zoom, and not as sharp as a prime lens.

 

Third, a lot depends on the photographer. If you are relaxed, well fed and well rested, and have been in one place for a few minutes, you can hold the camera a lot steadier than if you are tired, hungry, exhausted or frightened, or if you have just run 100 yards. That's quite apart from individual variation which can imply doubling or halving 'safe' speeds, or more. Because of her 'benign essential tremor', Frances prefers to use 1/250 second with a 50mm lens and 1/500 with a 90mm.

bouncy castle detail

 

Fourth, a lot depends on the camera. Some cameras are just easier to hold steady than others. Zeiss research showed that the same lens, the 38/4.5 Biogon, was sharper on an Alpa than on a Hasselblad because the Alpa is easier to hold steady. Longer, heavier lenses, with their greater polar moments of inertia, are harder to hold steady than shorter, lighter ones -- at least until the lens becomes so short and light that it hasn't enough inertia to stabilize it. And most photographers agree that it's easier to hold a rangefinder camera steady, rather than a reflex. Why? No-one is sure, but enough people agree that there seems to be some truth in it.

Fifth, there's an inevitable temptation to push your luck. If you're using a 50mm lens, well, 1/30 is near enough to 1/50, isn't it? But 1/60 is closer to 1/50 and (as noted above) likely to be sharper.

Sixth, the rule is safer to relax with wide-angles and more dangerous to relax with longer lenses. With a 21mm lens on 35mm, many people can get away with 1/15 and even 1/8 second, or sometimes as much as 1/4. With a 200mm lens, even 1/250 is unlikely to be as sharp as 1/500.

Ultimately, the only guide is experience. Use the rule of thumb, sure, but don't imagine it is anything more than a very rough guide.

 

noodle eater

 

Noodle eater, Beijing

Frances has a 'benign essential tremor' and therefore needs to shoot at faster shutter speeds than most people. She therefore needs to focus more carefully, because she must perforce use wider apertures. If we could afford one, we would buy her a Zeiss Ikon, which she finds much easier to hold still than her Voigtländers, which in turn she finds easier than reflexes. This was shot with an R2 on Kodak Tri-X with her 50/2.5 Color-Skopar at or near full aperture.

focal plane shutters

A focal plane shutter is, at its faster speeds, a travelling slit. This means that while each part of the film may only be exposed for 1/250 second at a marked speed of 1/250 second, the actual exposure from beginning to end can take as much as 1/50 second (on a Leica) or even 1/10 second (on a big, old reflex). It is therefore possible, though not common, for camera shake to be apparent in some areas of the image and not others.

The direction of travel of the shutter can also have an effect. A horizontal-travel shutter will elongate a horizontally-moving subject that is travelling in the same direction as the shutter, and shorten one that is travelling in the opposite direction. A vertical-travel shutter can 'stagger' a horizontally moving subject, exposing the top before the bottom. With modern fast-moving shutters these effects are all very unusual but there are several famous pictures of old racing-cars where the wheels are strange ovals with the top of the oval offset from the bottom of the oval.

panning

Blur is caused by relative movement of the image and the film. If you move the film so it keeps in step with the movement of the image, there will be no blur. Sure, the background is likely to be blurred, because the film is now moving relative to it; but the subject can be sharp. This is the basis of panning, which will be the subject of a separate module. Essentially, panning consists of smoothly following the subject in the viewfinder, then pressing the shutter release without interrupting the panning movement (novices often fail to do this, and stop the camera with a jerk when they fire the shutter).

 

pedal car

 

 

Pedal car, Taize

 

 

Panning with the direction of the pedal car has resulted in a blurred background but absolute sharpness in the middle of the pedal car, extending backwards to the driver. The longer the shutter speeds, the greater the background blur; this also allows smaller apertures for greater depth of field. Roger shot this with our Nikon D70 and standard 18-70mm zoom lens, with the ISO set to the minimum possible of 200.

Note also how the spokes that are at right angles to the direction of panning are sharply rendered, while those that are in the direction of panning (right to left) are blurred to nothing.

image stabilizing lenses

These again move the image relative to the film, compensating for movement of the camera in the photographer's hands. We have no experience of them, so can say very little, except that the extent to which they work will obviously vary. From what we understand, they normally allow photographers to get comparable sharpness at double or even four times the sort of hand-held speed they would normally use without the image stabilization. What they do with panning, we hate to think.

sharpness all over?

We have already remarked that the best 35mm lenses should be able to deliver 100 lp/mm centrally, but that edge definition is unlikely to be as high: 80 lp/mm would be excellent in the corners.

This is quite apart from depth of field, the variation between the focused point and other points that are closer or further away. In other words, some parts of the picture are likely to be sharper than others -- sometimes, very much sharper. This can be used creatively, or it can be a disadvantage.

 

Tibetan Musician

 

 

Depth of field does not get much shallower than this: a 50mm f/1 Noctilux used at full aperture. Although a slower lens -- even f/2 -- would have been feasible, the full aperture was used to throw the rest of the picture, everything except the musician's face, out of focus. If you cannot hold everything in focus, it sometimes makes more sense to avoid 'almost-in-focus' images which can be distracting.

You can also see that the quality of the out-of-focus image, the 'bokeh' (see below) is very smooth; another advantage of this wonderful lens, which we would buy if we could afford one. This was a concert by the musicians of the Tibetan Institute for Performing Arts (TIPA) at Niort in France; Roger used the Noctilux on his Leica MP, shooting on ISO 100 daylight-balanced slide film (Kodak EBX).

 

tibetan musician

 

edge-to-edge sharpness

This is really only critical when you are photographing either a flat subject at any aperture, especially the test target so beloved of some kinds of people, or when your entire composition is within the depth of field of the lens (assisted by camera movements if necessary).

Some lenses offer very much better edge-to-edge sharpness than others, and a few (such as the squeeze-focus Novoflex lenses that were for many years beloved of sports and wildlife photographers) fall off quite spectacularly. The narrower the angle of view of the lens, and the smaller the maximum aperture, the easier it is to ensure edge-to-edge sharpness, even with very simple designs. The leading exponents of this were probably Leica, with their two-glass cemented long-focus lenses, the 400/5.6 and 560/5.6 Telyts. A great advantage of these lenses is that they are inherently very contrasty.

Fast lenses, and wide-angles, and above all, fast wide-angles, are hardest to design with good edge-to-edge sharpness.

 

 

Landscape, Aragon

Some landscapes can probably stand being sharper in the middle than at the edges, but for the most part, they look a bit odd if you can't see the corners as clearly as the middle. Very few modern lenses are designed to give anything other than good sharpness all across the field, so this is not really a major concern nowadays. Zeiss Ikon, 25/2.8, Kodak Tri-X, print on Ilford Multigrade Warmtone. (Frances)

depth of field

There are two ways to use depth of field creatively: with selective focus, where the principal subject is sharp and the rest is left to go hang, and 'deep field' composition where everything appearrs sharp. Both have their difficulties.

With selective focus, you generally need either a fast lens or a long lens, or both, or the shallow depth of field may not look deliberate enough or effective enough.

With 'deep field' you have the twin potential problems of reduced definition as a result of diffraction limitations (if you stop down too far) and long exposure times leading to the risk of camera shake, though a tripod will obviously help with the latter.

Objectively, depth of field is dependent on the magnification of the subject on the film -- which tends to be related to both format and focal length, but is not necessarily so -- and on the aperture. Thus, the smaller the format and the wider the angle of coverage of the lens, the greater the depth of field tends to be, even though strictly, a 105mm lens at 1 metre gives exactly the same depth of field on 35mm, 6x9cm or 4x5 inch. The big difference is that it will normally be used in very different ways on the three formats, being a long-focus on one, a standard lens on the second, and a wide-angle on the third.

Subjectively, depth of field also seems to vary with the lens design, as does the 'roundness' or three-dimensionality of the subject. It is not entirely clear why this should be, but it is probably related to two things: the maximum sharpness of the focused part of the image, and the quality of the out-of-focus image or 'bokeh'.

 

 

Sturdivant House, Selma, Alabama

As noted above, depth of field depends principally -- perhaps solely -- on the magnification on the film; but as one of the main reasons for using a wide-angle is to get more in, the magnification is lower and the depth of field is greater. We shot this (on Ilford XP2) with a 14/3.5 Sigma on a Nikon F, probably at f/11 or even f/16 in order to ensure absolute back-to-front sharpness, even if this meant sacrificing the last morsel of sharpness to diffraction limitations. Objects at the edge of the field of view are unavoidably stretched, though only the door looks unnaturally wide.

lower maximum sharpness

If the maximum sharpness is not all that high, it can create an impression of greater depth of field and 'roundness', because the difference between 'sharp' and 'not sharp' is less and the transition is smoother. This is not something we have ever seen discussed; it's just the result of some decades of experience and thinking about it.

bokeh

This is a word that suddenly came into vogue around 2000. Pronounced 'boke-ay' or 'boke-eh' (a short 'e' as in 'ex'), it is apparently Japanese.

It is not a new concept, merely a new word. The only real difference is that people pay a lot more attention to it nowadays. To listen to some, you would think it was the most important characteristic of a lens. Some people, it seems, are very much more sensitive to it than others. We are not particularly sensitive: we cannot imagine saying that a lens has 'good bokeh', though we can spot really bad bokeh.

An important factor in bokeh appears to be the circularity of the diaphragm: the closer it is to a perfect circle, the better the quality of the out-of-focus image. This may be one reason why rangefinder lenses are often praised for their bokeh. A lens for an SLR, with its auto stop-down, is obviously easier to build with as few diaphragm leaves as possible: the more leaves there are, the likelier they are to bind.

 

Garlic stall

 

There are many reasons to use selective focus, and this is one that is seldom addressed. French law on the subject of photographing people is downright perverse, and using a wide aperture and a high shutter speed allowed Roger to render the man in the background unrecognizable. This was almost certainly shot with a 35/1.7 Voigtländer Ultron at maybe f/2; the camera was probably a Bessa-R2 and the film was definitely Kodak Elite Chrome EBX ISO 100.

garlic stall

 

what are you going to do with the picture?

There is little point in striving for maximum sharpness if it will not be visible in the final picture. On the web, in particular, sharpness is severely compromised by the relatively low resolution of even the very best monitors, and the other sharpness module, dealing with the illusion of sharpness, is rather more relevant.

spain mono

 

Likewise, there is little point in creating a super-sharp negative or slide, and then either scanning it with a low-resolution scanner or enlarging it with a poor-quality enlarging lens.

Finally, the picture needs to be big enough for the sharpness to be admired, but not so big that it ceases to exist. Almost any half-decent lens on a 35mm camera will suffice for a 10x15cm/4x6 inch picture, but going up even to 16x24cm or 6x9 inches will make what appear to be disproportionately greater demands.

 

 

Daroca, Spain

 

Roger shot this with our 75/2 Summicron on his Leica MP; the film was Ilford HP5 Plus. The sharpness of the original is impressive but this cannot really be conveyed on screen except via a sectional blow-up -- which doesn't really tell you anything about the picture as a picture.

The original print is approximately a 5x blow-up (5 x 7½ inches, or 12x18cm) in the middle of a sheet of 8x10 inch paper, and the tonality works very well indeed.

Frances made the print on Ilford Multigrade Warmtone using our 50/4.5 Leitz Focotar enlarging lens at its optimum aperture of between f/5.6 and f/8. Even the best enlarging lenses have a 'sweet spot' at which their performance is fractionally better than at other apertures.

the bottom line

There are at least a dozen things you can do to ensure maximum sharpness:

1   Choose a film known for its sharpness

2   Use a tripod...

3   ...or a shutter speed high enough to arrest motion (subject or camera)

4   Choose the sharpest lens in your armoury

5   Use the lens at its optimum aperture...

6   ...or at whatever aperture will give you the depth of field you want...

7   ...but do not stop down too far because of diffraction limitations

8   Focus carefully

9   Do not over-expose

10  With black and white films, use an acutance developer

11  Use the best enlarging system (optical or scanner) that you can reasonably afford

12   Do not over-enlarge

 

Guadalupe Passion Play

Be prepared to break whatever rules are necessary in order to get the picture. Yes, this is one of the sharpest films of its day (Fuji RFP ISO 50, #1) but a tripod would have been too restrictive (# 2) so Roger used the fastest shutter speed he could (1/250, #3) with a very sharp lens (90/2 Summicron, # 4) but at full aperture to allow the fast shutter speed (breaking # 5).

Depth of field was adequate at this distance (#6) and diffraction limits were not a problem (# 7) but very careful focus was essential (# 8) and under-exposing by 1/3 stop (# 9) not only gave maximum sharpness but also rich colours that reflected the evening light. It was colour, so # 10 didn't apply, and it was scanned with a Konica Minolta Dimage 5400-II (# 11) and because it's colour it can stand about a 10x enlargement before sharpness becomes a problem (# 12), and then it's mainly a question of depth of field with the girl on the right.

 

passion play

 

As the above caption demonstrates, all twelve tricks are ultimately secondary to a successful picture. Religious determination to achieve maximum sharpness at all times, at the expense of all else, is foolish in the extreme. Do what you can, when you can, but if in doubt, shoot, and ask questions afterwards.

Also, remember that there are a number of ways to create the illusion of sharpness, and that these are not necessarily the same as technical sharpness. But that is another module...

 

go back to the 'quality' introduction page

or go to the list of modules

or go to the home page

or support the site with a small donation.

© 2007 Roger W. Hicks