texture

As far as we can see, there are at least half a dozen tricks that can be useful when it comes to creating the illusion of texture in a photograph. These are not totally separable -- often, more than one is employed, and it can be hard to analyze precisely which contributes most -- but they are:

1     Expectation

2     Tonality

3     Contrast

4     Lighting

5     Careful exposure

6     Sharp lenses or large formats or both

Nothing ever looks really critically sharp on a monitor, especially as compared with an original print or transparency, which makes this module rather hard to illustrate. The advice is good, though, even if you do have to trust us to a certain extent when it comes to the sharpness of the pictures.

Attic

This uses most of the tricks described in this module. Most of us know what the carpentry of an attic looks and feels like, to say nothing of the texture of old suede boots, old painted wood, etc. The doll's face is rendered smooth by tonality -- a roundness that conveys smoothness -- and there are countless contrasts of texture. The light is mostly glancing, which always helps bring out texture; the exposure had to be careful in order to avoid 'blowing' the brightest areas to a featureless white, and to hold as much in the shadows as possible; and the 6x7cm format (Linhof Technika special) and sharp Schneider 100/5.6 Apo-Symmar helped too.

Whichever techniques you use, you are doing one of two things. Either you are making the photograph as naturalistic as possible, or you are matching artifice to reality. In the former case, the medium (the photograph) is made as 'transparent' as possible, so that those who are looking at it are encouraged to forget that they are looking at a representation instead of at the real thing. In the latter, the medium is clearly artificial, but the conventions of the medium are used to reinforce the expectations of the picture. For example, a 'nostalgic' picture may with advantage be grainy and soft, which is often the way we remember things, rather than the way we see them. Of course we may remember things this way because we have been conditioned to do so by the pictures we have seen...

expectation

This is the least technically demanding trick, and arguably the most effective. It relies on a simple truth. We all know what certain textures are like, be they granite, weathered wood or a child's skin. As long as the photographic rendition does not contradict our expectations unduly, we accept it as 'real'.

 

 

Black and white grain is incredibly forgiving when it comes to expectations. If we can see the grain, we can read it equally well as sand, stone, sandstone or skin -- or for that matter as rusty metal or a wide range of other things.

Colour is a bit less adaptable but we can still put up with a surprising amount until we see the colourful pointilliste grain of a film negative or the 'noise' in a digital image. Even then, if the poitillisme suits the subject we can read it as 'realistic'.

About the only things where we cannot easily accept grain as texturally accurate are those things that have no texture

 

Long Range Binoculars, Redondo Beach

Matte chrome; paint; wood; polished aluminum; the grain of the Kodak Tri-X that Frances used in her Voigtländer Bessa-T for this shot stands in for all of them. Of course it helps that the 50/2.5 Color-Skopar is a sharp lens too; a scan of the instruction plate on the binoculars (from the same print) is given below.

Binocular manufacturer's name-plate

You can see from this what we mean about trying to judge sharpness from the computer screen. Look at the size of the name plate in the picture above, to the left of the coin mechanism and reflect that most people would have some difficulty in reading '...and Distributed by...' in the original print, where the lower case letters are approximately 0.25 mm (1/100 inch) high. Given that the print is slightly over a 4x magnification, this corresponds to about 0.06mm or a little over two thousandths of an inch on the film. At 'South Norwalk', where the letters on the film are around 1/1000 inch high, detail runs out.

 

 

 

 

image colour

Closely related to grain and expectation is image tone. A skin tone that is (lightly) sepia or even selenium toned will seem warmer and more skin-like than one that is a neutral grey or (worse still) cool grey. In colour photography, warm skin tones (tending towards yellow, red or even magenta) are likewise more believable, and therefore realistic, than those which tend towards blue, cyan or green. This applies to subjects other than skin tones, too: to quote the late, great Terence Donovan, "Have you ever had a client complain because the picture is too ****ing warm?".

 

Door and foliage, Germany

It may be a purely personal prejudice, but we find that even with green foliage, a warm (brownish) tone works better for the vast majority of monochrome prints. Yes, we have tried green toners for foliage, and blue toners for winter scenes, and our belief is that they work for (at most) one picture in a thousand.

Frances shot this with her Nikkormat FTn and 35/2.8 PC-Nikkor -- which, as she wasn't using the PC feature, was a pretty basic 35/2.8 lens. The film was Ilford XP2.

Slightly higher than usual contrast has also helped differentiate each leaf from its dark background: we are not sure whether Roger overdid this when scanning the Ilford Multigrade Warmtone print for the site.

 

 

how familiar are colours and textures?

Some textures are all but universally familiar, such as sand, stone and polished wood. So are some colours: green foliage, pink Caucasian skin. But others may be limited geographically or even by age: before you assume too much familiarity, ask yourself if you are not remembering a particular personal experience such as the coarse, open weave of a cheap Indian shawl, or the virulent green of old Kawasaki motorcycles.

 

 

VE Day celebrations, Prague

The woman on the right is wearing a British army blouse of World War Two vintage, the very thought of which makes Roger's skin crawl -- as it will many others who were forced to wear this incredibly scratchy material, next to which a coarse blanket feels as soft as silk. But how many nowadays have direct experience of the beastly stuff? Roger can still smell it... He used his Leica M4P and 90/2 Summicron for this shot.

 

 

 

 

 

 

 

letting one texture 'carry' another

One strong texture can 'carry' another, in the sense that if there is a single strong, recognizable texture, then other textures are assumed to be well rendered too. Rusty iron is particularly useful in this regard, because it is so familiar that other textures will often be accepted even if they are blatantly inaccurate.

Old Swiss Army knife

The picture is not super-sharp, but it doesn't have to be: everyone knows what ancient rusty iron looks and feels like. Roger shot it with the Nikon D70 and Dreamagon soft-focus lens at f/11.

tonality

Even if we can't see (or are not aware of) the texture, tonality can be important, most especially with smooth subjects. This is to a very large extent a continuation of the idea of expectation, above: unless the photographic rendition breaks the expectations of tonality, usually by falling into 'soot and whitewash', tonality helps re-create the ideas of texture.

Astonishingly, this can work as well in black and white as in colour: it really is a matter of tone, not colour. But equally, if you are shooting in colour, you need to match the colours to the subject. Staying with skin tones, there is a world of difference between 'pale and interesting' (slight overexposure, reduced saturation), 'rosy cheeked' (suitable for most children, 'accurate' exposure) and the kind of rich, warm flesh tones that characterized such painters as Boucher and Fragonard (warming filters and the slightest touch of underexposure, suitable only for flawless complexions).

 

Handstand

You can't really see the fine texture in her shirt, and for that matter, her eyebrows and her hair could be sharper. But she is about 10 years old, and we all have a pretty good idea of the texture of a ten-year-old's skin, not least from our own memories of being that age. To a large extent, therefore, this falls under the heading of 'expectation' above, but the light, delicate tonality of the skin, especially in the upside-down lighting of the face, reinforces our expectations. The tiny spot or scrape on her right arm still further reinforces the illusion -- an example of contrast (below).

As far as he recalls, Roger used his 1936 Leica IIIa with uncoated 5cm f/3.5 Elmar for this shot on Ilford HP5.

 

 

contrast

If texture is clear, it is made clearer by more contrast. We are all aware that it is possible to overdo this, especially with the various sharpening options in Adobe Photoshop: sand soon becomes as rough and abrasive as sandpaper, and skin acquires an exceptionally vicious form of acne. Our rule of thumb for sharpening in Photoshop is that if there is any more than a barely perceptible difference between the unmanipulated image and the manipulated one, it is too much. But there's a lot more to contrast than this.

 

 

smooth versus rough

If a picture contains a variety of textures, the contrast of those textures automatically reinforces the impression of each texture.

Broken Treasures

Probably the most dominant texture is the 1920s sherry glass, but the old books, the jet beads and the broken lorgnette each contribute their own impressions. This was pretty much a joint exercise, based on an idea by Frances but with a fair degree of execution by Roger. It was shot on both 4x5 inch and 6x7cm with a Linhof Technikardan and 210mm lens, probably a 210/5.6 Rodenstock Apo-Sironar-N.

 

contrasting colours

In the picture on the left, the colours are all pretty much harmonious, but another way of increasing the impression of texture can be to contrast one colour with another: for example, yellow stone with a blue sky.

 

light/dark

A subject against a dark background tends to 'loom' more than one against a background that is closer to itself in tone, thereby concentrating more attention on the principal texture, and entirely removing any textural considerations from the background. In the studio, black velvet or black flock can be an excellent way to create a background that photographs as a solid black; elsewhere, lighting is the main technique.

 

 

Sculpture, Marnes

This centuries-old stone head at the church of Marnes in the Valley of the Dive is lit by glancing light (see below) which brings out plenty of texture but the way that it emerges from inky shadow lends it still more impact and concentrates more attention on the texture.

 

 

 

 

differential focus

Differential focus -- a wide aperture used to put one area in sharp focus while leaving another out of focus -- can be quite a powerful tool for conveying texture via the technique already described of 'carrying' one texture with another. If, for example, the highlights in the eyes in a portrait are sharp, and some of the hair is much sharper than other parts, a considerable illusion of sharpness and therefore texture can be created.

 

Jack

Roger used a 135mm f/1.8 Porst lens for this portrait: not an outstandingly sharp lens, but one which, when used wide open, gives a razor-thin depth of field. Both Jack's shirt and his further curls are convincingly out of focus, while the highlights in the eyes really 'sparkle'. The lens was mounted on a Pentax SV (because that's the only camera we have that takes the Porst) and film was Ilford XP2 Super, printed on Ilford Multigrade Warmtone.

 

lighting

Oblique or 'glancing' lighting, by its very nature, throws shadows caused by variations in surface texture. Indeed, we would commend to anyone a trick for finding things that have been dropped on the floor. Use a powerful torch (flashlight) held as close to parallel to the floor as possible. The long shadows make it much easier to see dropped earrings, small camera parts, etc.

What is less obvious is that other forms of lighting can also help create the illusion of texture, by revealing how shiny or matt a surface is. Even soft, overcast lighting can do this by revealing 'roundness' via what is known as the limb effect.

Armoured vehicle, VE Day celebrations, Prague

The way in which the diffuse skylight is reflected here reveals the paint as satiny or semi-matte, rather than truly matte: the effect is most marked on the head of the shovel. Another point that this picture makes is that imperfections, such as the patch to the right of the star and the minute indentations at the edges of the hinged door above, often do more to convey the illusion of texture than can be achieved via perfection -- and digital images (see below) are all too often too perfect, too flawless: fine detail is masked. A similar contrast-by-flaw effect is seen in the blemish on the arm of the young girl, above. Roger shot this on Fuji RF/RFP ISO 50 using his Leica M4-P and 35/1.4 Summilux.

polarizers

Even where there are no obvious reflections, a polarizer will often 'lose' diffuse white-light reflections which mask detail. An unexpectedly efficient demonstration of this can be with grass, where a polarizer makes a more intense green and often seems to differentiate each blade more convincingly.

Museum, Ramsgate

A polarizer would have rendered this picture even more saturated and thus (apparently) sharper, but one has to be careful with high-contrast lenses (this was a 35/1.7 Ultron on a Voigtländer Bessa-R) and high-saturation films (here, Kodak Elite Chrome 100 EBX) as there is a risk of overdoing things: as long as 20 years ago, one reviewer criticized Roger's 'over-polarized' images in a book review, thereby revealing that he hadn't read the book in which Roger specifically made the point that he doesn't use polarizers with rangefinder cameras because they are too much like hard work...

 

careful exposure

This is particularly important with unusually light or unusually dark subjects. The former may 'blow' to a featureless white, while the latter may 'block up' to an equally featureless black -- or, if you plan to scan a transparency (as we often do), dark tones might as well be solid black, even if they are adequately differentiated on the film, because they are impenetrable to all but the best scanners. You can get around the latter by buying a more expensive scanner but this can be an expensive solution.

 

Sacre Coeur, Paris

This is perhaps an exaggerated example of underexposing a white surface in order to hold texture -- a stop lighter would probably still have been adequate, and half a stop lighter almost certainly would -- but it's dramatic and makes the point rather well.

Of course, a lot depends on how you meter. Many through-lens meters would automatically suggest this degree of under-exposure, but as far as Frances recalls she shot this with a Nikon F and 35/2.8 PC-Nikkor, and metered with an incident light meter, giving one exposure at a stop down (now lost) and another, this one, at two stops down.

 

 

sharp lenses/large formats

Rather more than in 'expectation', above, this leads us into the realms of psychology and perception. There are at least three commonly accepted standards for 'sharp', all for an 'average' viewing distance of 25-30cm or 10-12 inches. The lowest standard equates to five line pairs per millimetre (5 lp/mm) and is 'pretty sharp for a picture examined at a reasonable distance.' For many pictures, this is entirely adequate. Next comes 8 lp/mm, which allegedly corresponds to the average resolving power of the human eye, one second of arc, at 10 inches/25 cm. Then there is a big jump to 30 lp/mm, the so-called 'vernier' resolution, the (average) ability to see a discontinuity in a straight line.

 

 

None of these standards is totally reliable, because they depend on both 'average' resolving power (variations of 2x either way, even with glasses, would not be unexpected) and 'average' viewing distances. The latter is often the subject of howls of rage from traditionalists, especially those who use 35mm, because they accuse some people of looking at the print from 0too close: 'sniffing the print', as it used to be termed. To us, this is a non-excuse. If your photograph isn't sharp when examined from closer than usual, well, it still isn't sharp.

Our own view, based on a mixture of our own experience and some years of study, is that there is a detectable difference between 10 lp/mm and 20 lp/mm. As we shall see, this is of considerable relevance.

Forcalquier

Original Polaroid prints -- this is Polaroid Sepia -- are by their nature contact prints. The resolution of the print is not necessarily all that high, but it is normally well above the 30 lp/mm that represents the maximum resolution of the human eye. We shot this as a joint enterprise with our Toho FC45 with 120/6.8 Schneider Angulon.

image size

No matter how sharp an image is to begin with, it will sooner or later start to look soft if you blow it up too much.

First, let's assume a whole-plate (6½ x 8½ inch, 165 x 216mm) contact print with negligible loss of sharpness between the negative and the print. You'll see why we chose the old whole-plate size in a minute, but suffice it to say for now that this is a fair size for the actual image area on a sheet of 8x10 inch (20x25cm) paper. If you have 35 lp/mm on the negative, you will have 35 lp/mm (or close enough to it that no-one can tell the difference) on the print. This is well above even the vernier resolution, but remains well within the capabilities of the vast majority of lenses ever made.

Ruined mill

This is where monitors really fall down. This shot, which Frances took on 5x7 inch Ilford FP4 Plus using our Linhof Technika V 13x18cm and a 165/6.8 Goertz Dagor that probably dates from the 1920s, is very, very sharp indeed, but on the screen it's still just the same old 550 pixels high as all the other shots.

Of course you often have to stop down well with large formats, in the interests of depth of field, but this does not matter much with a contact print: even at f/32 you still have well over 35 lp/mm, and at f/45 you are doing better than 20 lp/mm even at the more stringent rule of thumb of 1000/f (see 'Camera Shake and other Thieves of Sharpness', below).

 

Next, let's assume a 6x7cm image with 80 lp/mm on the negative: many top-flight roll-film cameras can achieve this, or better. Blow it up and you will inevitably lose some resolution. At an enlargement of 3x (linear), a 56x72mm negative enlarges to almost exactly whole plate (actually 168 x 216mm).

But the maximum possible resolution from 80 lp/mm at 3x enlargement is 80/3 lp/mm, or just under 27 lp/mm. Allowing for losses due to the enlarging lens, you might realistically expect anything from 20 to 25 lp/mm.

Even so, 20-25 lp/mm is probably still more than sharp enough to pass for a contact print; the difference between this and 30 lp/mm is not usually worth worrying about. Even if you have only 55 lp/mm on the image, which is as low as you might decently expect at a middling aperture, you have 55/3 lp/mm or 18 lp/mm with a perfect lens, maybe 14 to 16 lp/mm in the real world: the lower the degree of enlargement, the less serious the losses. This is still extremely sharp.

 

 

Next, let's imagine a 35mm negative with a uniform 125 lp/mm across the negative. This is unrealistically high, not because of lens resolution, but because of the difficulty of holding the film flat in exactly the right place. To be more realistic, 100 lp/mm is the maximum you will see from most of the finest lenses, with 125 lp/mm in the sharpest plane of focus, which certainly won't be over the whole of the negative.

Even so, blow 35mm up to whole plate, and you need just under a 9x enlargement factor (linear). This implies a maximum resolution of 125/9 lp/mm with a perfect lens, or a bit under 14 lp/mm: call it 10 to 12 lp/mm in the real world. Drop the on-the-film resolution to 100 lp/mm and you are down to 11 lp/mm with a perfect enlarger lens, or around the 8 lp/mm resolution in the real world. We hasten to add that even 100 lp/mm, all across the film, is all but unheard of: edge resolution of 80 lp/mm would be excellent. This equates to 9 lp/mm with a perfect lens and 7 or 8 in the real world.

Graffiti and poster, Arles

The 35/1.4 Summilux is often denigrated by those who have never used it, or by those who have compared it with the 35/2 Summicron at f/2. But the truth is that by f/5.6 to f/8, which are the normal shooting apertures for most pictures by most people, it is very sharp indeed. It is also a lot sharper at f/1.4 than the Summicron. Roger used his M4-P loaded with Kodak Elite Chrome EBX ISO 100 for this picture. Colour contrasts (see above) help here too.

 

Drop down in quality from the finest lenses available to the majority of zooms, and 80 lp/mm centrally is pretty good, with 60 lp/mm or less at the edges. Drop to a cheap zoom, or old and inferior prime lenses, or shoot at full aperture with any but the very best fast lenses, and you may be lucky to see 60 lp/mm centre, 40 lp/mm edge.

Suddenly it becomes clear why, if texture is important, a 12x16 inch/30 x 40cm enlargement from 35mm (over 12x linear) often looks very sick indeed next to the same size enlargement from 6x7cm (around 5x linear), and why even the enlargement from 6x7cm may not look too good alongside one from 4x5 inch (just over 3x) or for that matter 5x7 inch (a little over 2x) or 8x10 inch (1.6x). Yes, big enlargements from small negatives can look great -- until they are compared with the same size enlargements from bigger negatives, even though the bigger negative may not boast quite as many line pairs per millimetre: the degree of enlargement is just so much less.

camera shake and other thieves of sharpness

All of the above resolution figures -- except 125 lp/mm all across the image area on 35mm -- should be reasonably easily attainable with first-class lenses, a fine-grain film, a tripod and the lens at its optimum aperture, usually f/5.6 or f/8.

Open up or stop down too much, and resolution will fall. The better the lens, the faster it will reach its peak performance. As you stop down, the very best lenses may not improve significantly beyond f/4 or f/4.5; the worst may still be getting better at f/11. A good rule of thumb, to get the maximum possible theoretically useful resolution of the lens in lp/mm (the diffraction limited resolution) is to divide the aperture in use into 1000 or 1500 (it depends on the assumptions you make). Thus, f/22 gives 45-66 lp/mm, explaining why so few lenses for 35mm cameras stop down this far.

As for camera shake, put not thy faith in 'one over focal length' for hand holding. A 50mm lens should indeed give better sharpness at 1/50 or 1/60 second than at 1/30, and this sharpness may well be acceptable for most purposes; but there are wide individual variations, and for ultimate sharpness, even the steady-handed may see a significant improvement at 1/125 second and even a further improvement at 1/250 second. Or, of course, you can use a tripod.

 

Smith, Czech Republic

Photography is a boundless set of compromises. Here, Roger was using a Leica M4-P because it is smaller and lighter than roll-film, faster handling, and cheaper to run. The lens was the 35/1.4 Summilux, chosen in case of low light, and the film was Kodak Elite Chrome 100 EBX, chosen for fine grain and an amazingly good price/performance ratio.

Could it have been better? Possibly. With 35mm, even an ISO 400 film might have allowed adequately small grain, together with more depth of field and a shorter shutter speed to reduce the risk of camera shake.

Better still might have been a 6x9cm back on our Polaroid 600SE with the 75/5.6 lens; then, an ISO 400 film would not have been a problem, and with the camera on a tripod, a speed as long as 1/15 second would have been fine as long as everyone stayed still. The trouble is that with a tripod and asking people to hold still for longer, the mood of the picture might well have been lost.

 

 

 

the digital disadvantage

To begin with, most digital cameras have a far lower overall resolution than film cameras. The actual sensor resolution may be higher, but the sensor is usually so much smaller that the total amount of information captured is significantly less. Direct comparisons are actually theoretically impossible, because the sensor array in a digital camera is arranged in a regular raster, like a TV or monitor screen, instead of a random array like film grain. Most people who work in the field are convinced that a 35mm slide equates to around 18 to 20 megapixels, though valid arguments can be advanced for equivalences of 12 megapixels or below and 30 megapixels and above. Obviously a low-resolution system has more difficulty in capturing fine detail than a high-resolution system.

In practice, the great majority of digital cameras suffer from a further handicap. The regular sensor array introduces the risk of interference or Moire patterns, those familiar and usually unpleasant patterns you get when two identical or near-identical patterns are superimposed. To get rid of these, many digital cameras incorporate a so-called 'low pass' filter to scramble fine image detail and stop it interfering with the sensor array.

 

 

Exakta and decanter

This picture relies very much on expectations -- we all know what old chrome looks and feels like, and glass decanters, and so forth -- but it also relies on contrasts and on differential reflectivity: the way that the Exakta name-plate catches the light, in particular, shows off the 'chromium-ness' of it. The sharp lettering on the Biotar lens is an example of 'texture by association' -- if we can read text, we assume everything else is sharp -- and the picture is run fairly small, which always helps to make things look sharper, especially with older people who are never sure how much is down to the picture and how much is down to their eyesight. Roger shot this with the Nikon D70 and Dreamagon 90/4 soft focus lens at f/11.

If you insist on using a digital camera to capture texture, therefore, you must work under several constraints unless you use a camera with a very high pixel count (30+ megapixels) and, preferably, no low-pass filter. Expectation, our first heading above, is your greatest friend. Coarse texture (peeling paint) will be easier to capture than fine texture (peach fuzz). And over-enlarging your pictures is to be avoided at all costs. At trade shows we are regularly astonished by the abysmal quality of 12x16 inch/30x40 cm enlargements from digital originals, even on manufacturers' stands, when the same manufacturer is often showing enlargements from film as well. The digital pictures from anything less than 30 megapixels always have an 'airbrushed' quality. This may work fine for wedding pictures and portraits of little girls, where flawlessness is expected or at least politely assumed, but they are not much use when fine detail is required.

the bottom line

Whether you are shooting silver halide or digital, the viewer's expectation that the picture is faithful to the subject is always your strongest weapon. After this, you have to use as many of the other tricks as you can, if you are aiming for maximum sharpness.

Remember, though, that sharpness and texture are not everything. Many pictures can be perfectly successful if they are not as sharp as theoretically possible, and a few may even be improved by a modest (or even considerable) loss of sharpness: art, not science, rules at the limits of photography.

 

Marc Porter

Marc was a talented writer and poet who died young, at around 40; here he is probably in his early 20s. Roger took this picture with a Nikon F and 58/1.4 Nikkor on Ilford HP5 at the old Bristol Arts Centre. The exposure must have been somewhere between 1/8 second and 1 second. This picture is very much how we both like to remember him. Cheers, Marc.

 

 

Go to the list of modules

or go to the home page

or support the site with a small donation.

© 2006 Roger W. Hicks