Image editing for enamelists and silversmiths
Introduction
It has been a long time since the first photographs, and hence portraits, appeared on copper, silver, gold and other metal, employing a technique known as photogravure or heliogravure. In this section we shall concentrate on applications for etching and for casting objects; however, it is also possible to produce enamel photographs. This process has been employed extensively for over a century by printers, originally also by photographers, in reproducing photographs, but its use has been superseded by the introduction of rotogravure. As the technique provides ample possibilities to influence the end result, photogravure has been used by artists. The process is different from the techniques presently used in reproducing images, which entail rasterisation (applying screens) as a means of reproducing all shades of black and white (or colour). Rasterisation may in principle also be useful in the silversmith's trade; however, we shall leave that alone and review how images are formed on computer screens, as this has bearing on the techniques that we apply.
We stand on the shoulders of our forebears, who
made the most of the techniques they had developed by applying it with
great skill, ingenuity and dedication. This therefore presents a good
opportunity to go into the difference between the process of rasterisation
and that of photogravure, as it helps in understanding why the two processes
do not lead to the same image quality. A thorough review of photogravure
is presented by J. de Zoete 
.
About photogravure and rasterisation
When reproducing a black and white photograph, with all the shades from black to white, one can make use of rasterisation and of photogravure. The original technique of photogravure for reproducing pictures entailed applying gelatin and dichromate salts to pigment paper. Gelatin becomes photosensitive by the addition of dichromates, a discovery made by Mungo Ponton (1839), Becquerel (1840) and Fox-Talbot (1850).
The technique developed by Fox-Talbot was perfected by Klic (1870). After exposure (through a photographic image) of the gelatin 'sensitised' by the addition of dichromate salt it turns hard and insoluble (this process is called 'tanning'). The depth to which the gelatin hardens is proportional to the intensity of the light transmitted. The gelatin (face down) is then transferred to a polished and clean (degreased) copper plate onto which very fine asphalt powder has been adhered by heat. The pigment paper and untanned gelatin are subsequently washed out with water. What remains is hardened gelatin on copper; the thickness of the gelatin layer varies in proportion to the exposure, so that the surface is made up of innumerable tiny pits of variable depth. This surface is subsequently exposed to ferric chloride, a fluid with etching properties. The thinner the gelatin, the sooner will the copper surface be reached and the deeper the copper will be etched.
So in the end one has obtained a copper plate (an intaglio plate) with innumerable tiny pits of variable depth: the finer the asphaltum powder, the more detailed the image in copper. The small pits represent a negative of the original image. When these tiny holes are filled with ink, which is then transferred to high quality paper by a press, the paper surface will have finely dispersed ink dots of varying heights; this causes the eye to perceive different shades of white, grey and black. As the number of dots per unit of surface area is so large, we in fact have an unlimited (continuous) scale from black to white, with the greatest possible detail in black and white reproductions.
Briefly, therefore, photogravure is a process whereby a photographic image is etched with varying depth into a copper plate, using a photographic procedure. A somewhat similar process is used widely for the production of electronic boards. A thin layer of copper is fixed to a plate made of synthetic material and covered with a UV-sensitive layer ('photoresist'); this layer is subsequently exposed to UV light through an appropriate negative and then developed. The developing fluid dissolves the unexposed photoresist, exposing the copper. The exposed copper is then removed with an etching fluid. The UV-sensitive film can subsequently be removed either by prolonged exposure to developing fluid or by some other means, usually acetone. What remains is the electronic circuitry in the form of copper strips. The process can also be used to transfer images to silver or copper. The procedure is explained elsewhere.
 |
| Figure 1 - Magnified view of a photographic raster. |
As the image is depressed in the plate, printers call this the intaglio technique (rotogravure). It is also possible to mimic various shades of grey by varying the diameter of equally spaced black dots on a white background; areas with the largest dots are not perceived as individual points but as a black area, and areas with smaller dots are perceived as various shades of grey. Print makers achieve this effect by rasterising (applying a screen to) an image. The raster or screen resembles a slightly out of focus representation of a regular pattern, for example of black and white squares, or black and white lines. By projecting a B/W negative or positive onto a light sensitive surface via for example a screen with a blocked pattern (figure 1), an image is created made up of black dots surrounded by white space. This is how it works.
 |
| Figure 2 - Magnified lateral view of a photographic screen. The intensity of transmitted light (bottom red line) increases from the centre of the dark square to its edge and is maximal at the centre of the transparent area. |
Let us start with the lateral view of the screen (figure 2). We measure the light intensity transmitted through the screen at various intensities. The centre of the black squares blocks all light, but the edges are not entirely opaque. At great light intensity the light transmitted increases from the centre of the square towards its edge, and all light is transmitted in the transparent part of the screen. If the intensity with which the screen is illuminated diminishes the black square will block proportionately more light from reaching the surface below, and at low intensities even block all the light in its path.
It follows, therefore, that if the screen is exposed to decreasing light intensities, the intensity of transmitted light diminishes on two accounts: first simply because of diminishing exposure, but secondly because the screen blocks light more effectively. If the light intensity during exposure of a screen decreases, therefore, smaller dots will be formed in the underlying photosensitive material.
 |
| Figure 3 - Image formed when photosensitive material is exposed to light transmitted through a screen at different light intensities. |
Let us look at a magnification of the image formed (figure 3). When exposing photosensitive material via a photographic image (negative or positive) the intensity of the light reaching the screen depends on the transparency of the overlying film: the more light is transmitted, the greater the resulting black dot, and conversely smaller dots are due to less light being transmitted through dark portions of the film. If the photographic image is made up of 4 shades of grey only, the result would be as in figure 4.
 |
| Figure 4 - Representation of 4 grey areas when a negative is filtered by a fine screen, and blow-up of a small section. |
When we apply a screen to a photographic image with all shades of black and white, we do not obtain a continuous tonal scale but rather an image made up of tiny black dots of varying diameter. A screen made up of lines gives rise to an image with varying line widths. This process cannot give the same rich black and white tonal scale as obtained with photogravure. With the latter the image is reproduced by transferring a tanned and washed out gelatin layer to metal onto which we have previously stuck a thin layer of very fine asphaltum powder ('screen'); upon etching this, minute pits of varying diameter and depth are created (figure 5). With a more recent technique the same is achieved by the use of lasers. Rotogravure or intaglio printing is therefore quite different from printing processes in which the ink is transferred from elevated parts of a printing plate.
Figure 5 - Rotogravure is an intaglio process, whereby pits of varying diameter and depth are formed. |
One can also transfer the image to a UV-sensitive resin that becomes very hard upon exposure. Non-exposed portions are washed out, leaving a thin but very hard resin. This 'mould' is put on silver or other metal (or marble) and then sandblasted, imitating the process of etching. A limitation of this technique is that an image containing crossing lines leads to the resin mould falling apart into pieces.
 Figure 6 - Example of a photograph rasterised by computer. The screens used were 80, 40 and 20 lines per inch, respectively; this demonstrates the progressive loss of detail. |
Halftone screens can also be applied by computer (figure 6). Bear in mind that printer images are also made up of dots. If the printer resolution is 600 dots per inch (236 dots per cm) then all shades of grey arise from varying the density and the diameter of the dots. In fact, each computer image when printed is transformed into dots. The greater the number of dots per unit of length (resolution), the more detailed the printed image and the more photorealistic it will be.
Bitmaps on the computer screen
The screen of a computer can be compared with a mosaic frame that can hold pieces of stone of the same dimensions. Each piece of stone is called a pixel. Screen resolutions range from 640x480 (width x height) pixels, 800x600, 1024x768, 1152x870 and up. Generally, the higher the screen resolution, the smaller the pixel size. Each pixel, just like mosaic stones, has a colour. The information about colours is stored in bits. One bit has either the value zero or 1, “true” or “false”, "on" or "off". In black and white colour is represented by either black or white: one bit can assume two values. If colour information is stored in 2 bits, then 2x2 = 4 shades of black and white can be stored, so that we have white, black, and two shades of grey. In practice in black and white illustrations it suffices to use 8 bits (2 to the power 8 = 256 tones: black, white and 254 shades of grey). On the screen this is achieved by varying the strength of the electronic beam. The screen can also display colours. This is because it is made up of 3 layers where pixels, when hit by electrons, light up in either red, green or blue. All shades of colours can be reproduced by varying the intensity with which each pixel emits light. If there are 8 shades of red, green and blue, respectively, (3 bits per colour), this leads to 8x8x8 = 512 different shades of colour. This represents only a limited colour scale. Modern computers use screens that can display 256x256x256 = 16.777.216 colours (8 bits colour resolution for red, green and blue each) or more.
A screen resolution of 640x480 pixels implies that there are 307.200 pixels; this suffices to reproduce a figure recognisably. At a screen resolution of 1024x768 pixels (hence 786.432 pixels) image quality is appreciably better. Present digital photo cameras use 1800x1200 = 2,3 million or more pixels per image. The same image in a photogravure, black and white or colour negative, or a colour slide, is made up of a far greater number of "pixels". Standard screens mayuse 72 or 96 pixels per inch, but in fact these figures vary. Suppose one scans a 35 mm negative (24x36 mm = 0.94x1.42 inch) at a resolution of 300 points per inch, then this will only partially fill a 640x480 pixel screen, since the digital photo represents only 426x283 pixels; the same digital picture displayed at a higher screen resolution will fill an even smaller portion of the screen. In order to fill a 1024x768 pixel screen horizontally the 35 mm negative or slide should be scanned at a resolution of 722 dpi, vertically at 813 dpi.
Printers similarly reproduce a picture by printing it at a resolution of 300 or more dots per inch. The reasoning applied to screens also applies to printers.
Further considerations
If only the highest quality and photorealistic representation of an image is acceptable, then one should obviously resort to photogravure. The original technique entails the use of sensitised gelatine and fine asphaltum powder, ingredients that are not easily available; the technique is also demanding and laborious. Other techniques require that the image is reduced to pure black and white: colours or shades of grey cannot be reproduced on silver. High resolution screens, suitable for photorealistic images, produce such small dots that they cannot possibly be reproduced upon printing, let alone in the casting. Usually low resolution screens give rise to unsatisfactory images. Therefore the images need to be more or less transformed into line drawings. A gifted artist will not need a computer to that end. However, even gifted artists may find that computers can be helpful in carrying out the various procedures, such as mirroring and inverting the image, or quickly changing line width.