Chapter 16. The reproduction of colour

248 The reproduction of colour

Table 16.1 Typical reflectances of natural objects in the

three main regions of the spectrum

Colours of

natural objects



Reflectance (%)

Blue



Red

Orange

Yellow

Brown

Flesh

Blue

Green



Green



Red



5

6

7

4

25

30

6



5

15

50

8

30

15

10



45

50

70

12

40

5

7



The colours described in the example above were a

result of the complete absorption of one or more of

the main colour groupings of white light. Such

colours are called pure, or saturated colours. The

pigments of most commonly occurring objects absorb

and reflect more generally throughout the spectrum.

The colours of natural objects (sometimes referred to

as pigmentary colours) therefore contain all wavelengths to some extent, with certain wavelengths

predominating. They are of much lower saturation

than the colours of the papers supposed in our

example. Thus, a red roof, for example, appears red

not because it reflects red light only, but because it

reflects red light better than it reflects blue and green.

The spectral properties of some common surface

colours are illustrated in Plate 8, in terms of the

percentage reflectance of each of the seven spectral

hues.

Table 16.1 shows the percentage reflectance in the

three principal regions of the spectrum of some of the

main colours in the world around us. It will be seen

that the general rules connecting the kind of absorption with the resulting impression are the same as

with the pure colours of the papers considered in the

example above. The figures quoted in the table,

which are approximate only, are percentages of the

amount of incident light in the spectral region

concerned, not of the total incident light.

We see from Figure 14.1 and Plate 8 that the

spectral absorption bands of commonly encountered

coloured objects are broad.



Effect of the light source on the

appearance of colours

We noted above that the colour assumed by an object

depends both upon the object itself and on the

illuminant. This is strikingly illustrated by a red bus,

which, by the light of sodium street lamps, appears



brown. In this context ‘brown’ means, strictly, ‘dark

yellow’. (This is a combination of words that does not

normally occur in English, for cultural reasons that do

not concern us here.) If the illuminant is pure yellow,

then the hue of the reflected light cannot be other than

yellow, even though the object may (as in this case)

reflect very little. A deep red rose reflects no yellow

light at all, and when seen under a (low-pressure)

sodium street lamp it appears black. We are accustomed to viewing most objects by daylight, and

therefore take this as our reference. Thus, although,

by the light of sodium lamps, the bus appears brown,

we still regard it as a red bus. The change from

daylight to light from a sodium lamp represents an

extreme change in the quality of the illuminant, from

a continuous spectrum to a line spectrum. The change

from daylight to tungsten light – i.e. from one

continuous source to another of different of different

energy distribution – has much less effect on the

visual appearance of colours. In fact, over a wide

range of energy distributions the change in colour is

not perceived by the eye. This is because of its

property of chromatic adaptation, as a result of which

the eye continues to visualize the objects as though

they were in daylight.

If we are to obtain a faithful record of the

appearance of coloured objects, our aim in photography must be to reproduce them as the eye sees them

in daylight. Unlike the eye, however, photographic

materials do not adapt themselves to changes in the

light source, but faithfully record the effects of any

such changes: if a photograph is not being taken by

daylight, the difference in colour quality between

daylight and the source employed must be taken into

account if technically correct colour rendering is to be

obtained, and it will be necessary to use colourcompensating filters.

In practice, a technically correct rendering is rarely

required in black-and-white photography, and changes in the colour quality of the illuminant can usually

be ignored. In colour photography, however, very

small changes in the colour quality of the illuminant

may produce significant changes in the result, and

accurate control of the quality of the lighting is

essential if good results are to be obtained.



Response of the eye to colours

Our sensation of colour is due to a mechanism of

visual perception which is complicated, and which

operates in a different way at high and low levels of

illumination. In considering the reproduction of

colours in a photograph we are concerned with the

photopic mechanism, i.e. the one operating at fairly

high light levels. For the purposes of photographic

theory, it is convenient to consider this as using three

types of receptor. A consideration of the way in which

these receptors work is known as the trichromatic



The reproduction of colour



249



Figure 16.1 The three fundamental sensation curves of the

trichromatic theory of colour vision



Figure 16.2



Visual luminosity curve



theory of colour vision. This theory, associated with

the names of Young and Helmholtz, postulates

receptors in the retina of the eye which differ in their

sensitivity and in the regions of the spectrum to which

they respond. The first set of receptors is considered

to respond to light in the region of 400–500 nm, the

second to light in the region of 450–630 nm and the

third to light in the region of 500–700 nm. Considered on this basis, the behaviour of the eye is

probably best grasped from a study of curves in

which the apparent response of the hypothetical

receptors (in terms of luminosity) is plotted against

wavelength, as in Figure 16.1.

The perceived luminosity of blues is lower than

that of either greens or reds. As a result, spectral blues

look darker than reds, and reds look darker than

greens. It must be clearly understood, however, that

luminosity (or lightness) has little to do with the

vividness or saturation of colours. Thus a dark blue

may look as fully saturated as a green that appears

inherently much lighter; and a differently balanced

set of curves is obtained if colour-sensitivity is

substituted for luminosity as the vertical scale (as

shown in Figure 16.17).

The validity of the three-colour theory was for a

long time a subject of controversy, but it is now

supported by experimental evidence indicating that

there are indeed three types of colour-sensitive

receptor in the retina of the eye. The justification of

the use of the theory for our purposes is that a person

with normal colour vision can match the hue of

almost any given colour by mixing appropriate

amounts of blue, green and red light. Colour photography itself relies on this phenomenon. By adding

together the ordinates of the three curves shown in

Figure 16.1, we obtain a curve showing the sensitivity

to different wavelengths of the eye as a whole. Such

a curve is shown in Figure 16.2. This visual



luminosity curve shows the relative luminous efficiency of radiant energy. It shows how the human eye

responds to the series of spectral colours obtained by

splitting up a beam of pure white light of uniform

spectral composition throughout the visible spectrum

– referred to as an equal energy spectrum – to which

sunlight roughly approximates. It is useful to notice

by how many times the luminosity of the brightest

colour, yellow–green, exceeds that of colours nearer

the ends of the spectrum.



Primary and complementary

colours

When any one of the three sets of colour-sensitive

receptors of the eye is stimulated on its own, the eye

sees blue, green, or red light respectively. These three

colours are known to the photographer as the primary

colours, already mentioned. The sensations obtained

by mixing the primaries, are called secondary or

complementary colours and are obtained when just

two sets of receptors are stimulated (Table 16.2). If all

three sets of receptors are stimulated in equal

proportions the colour perceived is neutral.



Table 16.2 Primary and complementary colours

Primary colour



Complementary colour



Additive mixture



Red

Green

Blue



Cyan

Magenta

Yellow



= Blue + green

= Blue + red

= Green + red



250 The reproduction of colour



It should be pointed out that in painting the theory

of colour is treated somewhat differently. As a painter

is concerned with mixtures of colorants rather than

lights it is customary in painting to describe as

‘primaries’ the pigments blue, red and yellow. These

can be the basic set of Prussian Blue, Crimson Lake

and Chrome Yellow, equivalent to the photographer’s

cyan, magenta and yellow. By mixing these pigments

in various combinations it is possible to obtain almost

any hue.



Complementary pairs of colours

Any two coloured lights which when added together

produce white, are said to be of complementary

colours. Thus, the secondary colours (cyan, magenta

and yellow) are complementary to the primary

colours (red, green and blue), respectively. Some

colour theorists in the field of aesthetics refer to such

complementary pairs of colour as ‘harmonious

colours’.



conditions, becomes much less colour-sensitive at

lower levels of illumination. Under very dim conditions only monochrome vision is possible. This

property is not shared by colour films, so that under

dim conditions the correct exposure will nevertheless

reveal the colours of the subject and often surprises

the photographer, whose recollection is of a far more

drab original. It is of interest that the brightness of

television screens and computer monitors is considerably less than sunlit scenes, and may be

inadequate for the eye to operate in its fully photopic

mode, but much too bright for scotopic vision to

predominate. In this case the eye is said to be

operating in a mesopic mode, which entails a lowered

sensation of colourfulness compared with that from

photopic vision. This is true despite the apparent

success of television in giving vivid colour

reproduction.



Black-and-white processes



At low light levels a different and much more

sensitive mechanism of vision operates. The darkadapted, or scotopic, eye has a spectral sensitivity

differing from the normal, or photopic, eye. The

maximum sensitivity, which, in the normal eye, is in

the yellow–green at about 555 nm, moves to about

510 nm, near the blue–green region. This change in

sensitivity is referred to as the Purkinje shift (Figure

16.3). It is because of this shift that the most efficient

dark-green safelight for panchromatic materials has a

peak transmittance at about 515 nm rather than at

555 nm.

It is also found that the eye, which can distinguish

a large number of colours under bright lighting



A monochrome photographic emulsion is considered

as reproducing colours faithfully when the relative

luminosities of the greys produced are in agreement

with those of the colours as seen by the eye. So far as

a response to light and shade is concerned – and this,

of course, has a very large part to play in the

photographic rendering of the form and structure of a

subject – it is clear that a light-sensitive material of

almost any spectral sensitivity will do. If, however,

our photograph is to reproduce at the same time the

colours of the subject in a scale of tones corresponding with their true luminosities, then it is essential that

the film employed shall have a spectral response

corresponding reasonably closely to that of the

human eye.

We have seen that, although the ordinary silver

halide emulsion is sensitive only to ultraviolet and

blue radiation, it is possible by dye-sensitization to

render an emulsion sensitive to all the visible

spectrum. To assist us in evaluating the performance

of the various materials in common use, Figure 16.4

shows the sensitivity of the eye and the sensitivity to



Figure 16.3



Figure 16.4



Low light levels



The Purkinje shift



Response of photographic materials to daylight



The reproduction of colour



(2)

(3)



Figure 16.5

light



Response of photographic materials to tungsten



(4)



daylight of typical examples of the three main classes

of photographic materials. Figure 16.5 contains

curves for the same materials when exposed to

tungsten light, the visual response curve being

repeated.

Table 16.3, which lists the primary and secondary

hues and indicates how each is recorded in monochrome by the three main classes of emulsion,

illustrates the practical effects of the differences

shown between the visual luminosity curve and the

emulsion spectral sensitivity curves in Figures 16.4

and 16.5. The descriptions given in the table are

relative to the visual appearance of the colours.

Less saturated colours than the primaries and

secondaries follow the general pattern shown in Table

16.3 but to a lesser extent. Thus, brown can be

regarded as a desaturated and dark yellow, and pink

as a desaturated and light magenta.

It is clear from the curves in Figures 16.2 and 16.3

and from Table 16.3 that:

(1)



No class of monochrome material has exactly

the same sensitivity distribution as the human

eye, either to daylight or to tungsten light. In the



251



first place, the characteristic peak of visibility in

the yellow–green is not matched by the photographic response, even with panchromatic materials. On the other hand, the relative sensitivity

of panchromatic materials to violet, blue and red

greatly exceeds that of the human eye.

The closest approximation to the sensitivity of

the eye is given by panchromatic materials.

Of the three groups of light-sensitive materials

listed, only panchromatic materials can be

employed with complete success for the photography of multicoloured objects as only these

respond to nearly the whole of the visible

spectrum.

Even with panchromatic materials, control of

the reproduction of colour may be needed when

a faithful rendering is required, as no type of

panchromatic material responds to the different

wavelengths in the visible spectrum in exactly

the same manner as the human eye. Control of

colour rendering is also sometimes needed to

achieve special effects.



The control referred to under (4) above is achieved

by means of colour filters. These are sheets of

coloured material which absorb certain wavelengths

either partially or completely, while transmitting

others. By correct choice of filter it is possible to

reduce the intensity of light at wavelengths to which

the emulsion is too sensitive. The use of colour filters

is considered in detail elsewhere in this book.



Colour processes

The principles of the two major types of photographic

colour reproduction have been described in Chapter

14. The processes considered there illustrate the ways

in which additive and subtractive colour syntheses

can be carried out following an initial analysis by

means of blue, green and red separation filters. It was



Table 16.3 Recording of primary and complementary colours by the main types of

photographic emulsion

Blue-sensitive

(daylight)



Orthochromatic

(daylight)



Panchromatic

(daylight)



Panchromatic

(tungsten)



Primaries

Blue

Green

Red



Very light

Dark

Very dark



Light

Rather dark

Very dark



Rather light

Slightly dark

Slightly light



Correct

Slightly dark

Light



Secondaries

Yellow

Cyan

Magenta



Very dark

Light

Slightly light



Slightly dark

Very light

Correct



Correct

Slightly light

Rather light



Rather light

Slightly dark

Rather light



252 The reproduction of colour



Figure 16.6 Simplified cross-section of an elementary

integral tripack film



stated that most colour photographs are made using

the emulsion assembly known as the integral tripack

and the procedure of colour development in which

yellow, magenta and cyan image dyes are formed to

control the transmission of blue, green and red light



Figure 16.8

film



Figure 16.7

film



Layer sensitivities of an elementary tripack



Effective layer sensitivities of a typical tripack



respectively by the final colour reproduction. As

shown in Figure 16.6, three emulsions are coated as

separate layers on a suitable film or paper base and

these emulsions are used independently to record the

blue, green and red components of light from the

subject. The analysis is primarily carried out by

limiting the emulsion layer sensitivities to the spectral

bands required.

The sensitivities of typical camera-speed emulsions are illustrated in Figure 16.7. All three layers

possess blue sensitivity. Blue light must be therefore

be prevented from reaching the green- and redrecording layers.

In practice the film is coated as shown in Plate

9, with the blue-recording layer on top of the other

two layers, and a yellow filter layer between the

blue-recording and green-recording layers. The

supercoat is added to protect the emulsions from



The reproduction of colour



253



Figure 16.9 Cross-section of a typical tripack

colour-printing paper. Note the different order of emulsion

layers when compared with Plate 9



damage. The filter layer absorbs blue light sufficiently to suppress the blue sensitivities of the

underlying emulsions, and an interlayer is usually

positioned between the lower emulsion layers. The

resulting (effective) emulsion layer sensitivities are

shown in Figure 16.8.

A more elegant solution to the problem of

inherent blue-sensitivity of the green- and redrecording layers is to suppress it within the emulsion layers themselves. This may be achieved in

materials for camera use by, for example, the use

in those layers of tabular emulsion crystals of very

high surface area:volume ratio. The blue response

depends on the amount of silver halide in each

crystal, a volume effect; but the dye-sensitized

response depends on the quantity of adsorbed dye,

a surface-area effect. A high area:volume ratio

therefore increases the sensitivity in the sensitized

region compared with the inherent blue sensitivity,

and with modern sensitization techniques this may

be so low as to require no yellow-filter layer in the

coated film.

In print materials it is often possible to reduce

and confine the inherent sensitivity of red- and

green-sensitized emulsions to the far blue and

ultraviolet spectral regions, and to exclude these by

a suitable filter during printing; no yellow filter

layer is then required in the print material and the

order of layers may be changed to suit other needs.

A commonly employed order of layers in colourprinting materials is illustrated in Figure 16.9. The

most important images for the visual impression of

sharpness are magenta and cyan, the latter being

particularly important. The red- and green-sensitive

layers are therefore coated on top of the bluesensitive emulsion so that the red and green optical

images reach the appropriate layers without suffering any prior blurring due to scatter in the bluesensitive layer.

The sensitivity distributions of two colournegative tripack films are shown in Figure 13.5;

they illustrate how the sensitivities of the individual

layers contribute to the overall sensitivity balance

of the film. There is little need for any particular

speed balance in a colour negative, so the actual

sensitivity balance is arranged to suit printing



Figure 16.10 Spectral sensitivities of reversal colour films

exposed to daylight. (a) A subtractive tripack film. (b) An

additive film with integral r´ seau

e



rather than visual criteria. This is not the case for

positive images, which are generally designed to be

viewed directly or by projection, and must therefore possess at least a satisfactory neutral reproduction. Wedge spectrograms of reversal colour

films may thus be considerably different from those

of negative films. Two such reversal films are

illustrated in Figure 16.10, which compares typical

subtractive and additive film sensitivity distributions. The differences in speed separations of the

three records in the two films may be of little

practical significance as the useful exposure range

of the additive film in particular is rather short,

i.e. it has a restricted latitude, and all practical

images may therefore possess adequate colour

separation.



254 The reproduction of colour



Formation of subtractive image

dyes

Having analysed the camera image into blue, green

and red components by means of the tripack film

construction it is then necessary to form the appropriate image dye in each layer. Conventionally this is

achieved by the reaction of the by-products of silver

development with special chemicals called colour

couplers or colour formers. Developing agents of the

p-phenylene diamine type yield oxidation products,

which will achieve this:

(1)

(2)



Exposed silver bromide + developing agent →

metallic silver + oxidized developer + bromide

ions

Oxidized developer + coupler → dye



Thus to form a dye image alongside the developed

silver image a special type of developer is required

and a suitable colour coupler must be provided either

by coating it in the emulsion or by including it in the

developer solution.

The dyes formed in subtractive processes would

ideally possess spectral absorptions similar to those

illustrated in Chapter 14. However, they are not in

fact ideal, and possess spectral deficiencies similar to

those also shown in Chapter 14. The consequences of

these dye imperfections can be seen when the

sensitometric performance of colour films is

studied.



Colour sensitometry

Using the sensitometric methods examined in Chapter 15, it is possible to illustrate important features of

colour films in terms of both tone and colour

reproduction. Tone-reproduction properties are

shown by neutral exposure, that is exposure to light of

the colour quality for which the film is designed,

typically 5500 or 3200 K. This exposure also gives

information about the overall colour appearance of

the photographic image. Colour exposures are used to

show further colour reproduction properties not

revealed by neutral exposure.



Negative–positive colour

Colour negative films and printing papers have tone

reproduction properties not unlike those of black-andwhite materials (see Figure 16.24), except that

colour-negative emulsions are designed for exposure

on the linear portion of the characteristic curve. The

materials are integral tripacks and generate a dye

image of the complementary hue in each of the three

emulsion layers: yellow in the blue recording,



Figure 16.11 Neutral exposure of a simple colour-negative

film



magenta in the green recording and cyan in the red

recording layer.

The densities of colour images are measured using

a densitometer equipped with blue, green and red

filters, and all three colour characteristic curves are

drawn on one set of axes. Typical simple negative

characteristics obtained from neutral exposure are

shown in Figure 16.11, which shows that the three

curves are of similar contrast although of slightly

different speeds and density levels. The unequal

negative densities have no adverse effect on the

colour balance of the final print because colour

correction is carried out at the printing stage.

Neutral-exposure characteristic curves are not

very informative about colour reproduction, but

characteristic curves obtained from exposures to the

primary colours are better. Saturated primary-colour

exposures could be expected to reveal the characteristics of individual emulsion layers. The results of

such exposures of a simple colour-negative film are

shown in Figure 16.12. The characteristic curves

show a number of departures from the ideal in which

only one colour density would change with colour

exposure. Two main effects are seen to accompany

the expected increase of one density with log

exposure: the first is the low-contrast increase of an

unwanted density with similar threshold to the

expected curve; the second is a high-contrast

increase (similar to that of the neutral curves) of an

unwanted density with a threshold at a considerably

higher log exposure than the expected curve. The

former effect is due to secondary, unwanted, density

of the image dye: thus the cyan image formed on red

exposure shows low-contrast blue and green secondary absorptions. The second effect is due to the

exposing primary colour filter passing a significant

amount of radiation to which one or both of the

other two layers are sensitive. The blue exposure has



The reproduction of colour



255



Figure 16.13 Neutral and red exposures of a

masked-negative film



Figure 16.12

film



Colour exposures of a simple colour-negative



clearly affected both green- and red-sensitive layers.

In this case of blue exposure the separation of layer

responses is largely determined by the blue density

of the yellow filter layer in the tripack at exposure

as all emulsion layers of the negative are bluesensitive. The results of green exposure show a

combination of the two effects, revealing secondary

blue and red absorptions of the magenta dye and the

overlap of the green radiation band passed by the

filter with the spectral sensitivity bands of the redand blue-sensitive emulsions.



Modern colour negative films yield characteristic

curves different from those shown in Figures 16.11

and 16.12 owing to various methods adopted to

improve colour reproduction. Sensitometrically the

most obvious of these is colour masking, which is

designed to eliminate the printing effect of unwanted

dye absorptions by making them constant throughout

the exposure range of the negative. The important

characteristics of a masked colour-negative film are

shown in Figure 16.13. The most obvious visual

difference between masked and unmasked colour

negatives is the overall orange appearance of the

former, and this appears as high blue and green

densities, even at Dmin . The mechanism of colour

masking is described later, but its results can clearly

be seen in the case of the red exposure illustrated in

Figure 16.13. The unwanted blue and green absorptions of the cyan image are corrected by this masking.

Ideal masking would result in blue and green

characteristic curves of zero gradient, indicating

completely constant blue and green densities at all



256 The reproduction of colour



exposure levels. In practice it can be seen that the

blue and green absorptions are slightly undercorrected so that both blue and green densities do rise

a little with increased exposure.

A second method of colour correction is to make

image development in one layer inhibit development

in the other emulsion layers. If an emulsion has an

appreciable developed density this may be reduced by

development of another layer, giving a corrective

effect similar to that of colour masking. Such interimage effects are not always simply detected by

measuring integral colour densities but they are

sometimes easy to see. Inter-image effects, for

example, would appear to be operating in the negative

film illustrated in Figure 16.12. Blue exposure results

in development of the yellow image in the top layer

of the tripack and this, in turn, inhibits development

of the green- and red-sensitive layers so that the green

and red fog levels fall while the blue density

increases. Such unexpectedly low image densities are

usually the only clue to the existence of inter-image

effects to be found using integral densitometry.

Most recent colour-negative films (Figure 16.13)

use a combination of colour masking with interimage effects, and this allows excellent colour

correction with much lower mask densities than were

previously used. The mask density and its colour,

however, remain far too high for such a colour

correction method to be used in materials such as

reversal films and colour print materials, designed for

viewing rather than printing. The consequences of

dye deficiencies in such systems are, however, less

important than in negative–positive systems where

two reproduction stages are involved with a consequent reinforcement of colour degradation.



Reversal colour

Colour-reversal films are constructed similarly to

integral tripack negative films. The positive nature of

the image springs from a major difference in the

processing carried out, not from the emulsions used in

the film. As in the negative, the dye image generated

in each layer is complementary to the sensitivity of

the emulsion: a yellow dye is generated in the bluesensitive layer etc.

Sensitometry is carried out as for the negative film

and characteristic curves plotted. The results of neutral

exposure are illustrated in Figure 16.14 and show

typical positive characteristic curves. Unlike the

colour-negative material it is important that the results

of a neutral exposure on reversal film appear neutral;

no simple correction can be carried out once the image

has been developed. In the case illustrated the curves

show sufficiently similar densities for the result to be

visually satisfactory. The major difference is at so high

a density as to be imperceptible – there is rarely any

perceivable colour in regions of deep shadow.



Figure 16.14



Neutral exposure of a colour-reversal film



Colour exposures (Figure 16.15) show various

departures from the ideal and are sometimes quite

difficult to interpret. The two main effects are,

however, as easily characterized as they were for

colour negatives; the unexpected decrease of density

with log exposure at low-contrast and the decrease in

a colour density other than that of the exposure at

high contrast. The former effect is generally caused

by the secondary absorption of an image dye, and the

latter shows the overlap between the spectral region

passed by the exposing filter and the sensitivity band

of the emulsion concerned. Red exposure yields

characteristic curves showing the effects of secondary

absorptions and inter-image effects. The expected

result of red exposure is a decrease in red density with

increasing exposure. This is seen in Figure 16.15; but

other results are also observed. The initial large

decrease in the density of the cyan dye image with

increased exposure is accompanied by a small

decrease in the green density, caused by the correspondingly reduced unwanted secondary green

absorption of the cyan-image dye. A considerable

increase in blue density takes place over the same

exposure range, and this indicates the existence of an

inter-image effect in which the development of an

image in the red-sensitive layer inhibits the development of an image in the blue-sensitive layer. This

effect may take place at the first (non-colour)

development stage, but is more likely during colour

development, and represents a colour-correction

effect akin to masking. The apparently high minimum

density to red in the case of red exposure is due to

unwanted red absorptions of the yellow and magenta

image dyes present in maximum concentration. At

high exposure levels, when the blue- and green-



The reproduction of colour



Figure 16.15



Colour exposures of a colour-reversal film



257



sensitive layers respond and the yellow and magenta

dye concentrations fall, the red density is also

reduced.

The response of the blue- and green-sensitive

layers to exposure through a red filter is due to the

small but appreciable transmission of the red filter in

the green and blue regions of the spectrum.

Green exposure causes the expected reduction in

green density giving a green characteristic curve with

a long low-contrast foot. This is not due to a lowcontrast reduction of magenta image at the foot, as

there is probably no magenta dye present. It is due to

a fall in the total secondary green absorptions of the

yellow and cyan dyes as they decrease in concentration at high exposure levels. This is caused by a

significant overlap between the pass-band of the

green exposing filter and the spectral responses of the

red- and blue-sensitive layers. It will also be noticed

that the initial fall in magenta dye concentration at

low exposure levels is accompanied by reduced red

and blue densities. This is due to the unwanted

secondary red and blue absorptions of the magenta

image dye; as the concentration of magenta dye

decreases so do its absorptions, both wanted and

unwanted. In the case of blue exposure all three layers

show the effects of actinic exposure, the speed

separation of the emulsions being mostly due to the

yellow-filter layer present at exposure. As with

colour-negative material, all the emulsion layers are

sensitive to blue. The low-contrast shoulder of the

green characteristic curve shows the reduced green

absorption associated with decreasing yellow dye

concentration. The yellow image clearly has a green

secondary absorption. The long low-contrast foot of

the blue characteristic curve is due, not to the yellow

image dye which is probably present at minimum

concentration, but to the blue secondary densities of

the magenta and cyan dyes which are decreasing in

concentration with increased exposure.

In essence, each of the colour-exposed results may

be analysed in terms of three main regions. The first,

at low exposure levels, shows one dye decreasing in

concentration; its secondary densities are revealed by

low-contrast decreases in the other two densities. The

second region, at intermediate exposure, shows no

change in any of the three curves, but the level of the

lowest is well above the minimum found on neutral

exposure, owing to the unwanted absorptions of the

two image dyes present. This is exemplified by the

red exposure (Figure 16.13) at a log relative exposure

of 2.0. Lastly we find, at high exposures, the region

where sufficient actinic exposure, that is exposure to

which the emulsion is sensitive, of the remaining two

emulsions leads to a decrease in the concentration of

the corresponding image dyes. The secondary absorptions of these two dyes are shown by a low-contrast

decrease in the third curve – for example the blue

density at log relative exposures of 1.5 or more. For

reversal film, departures from this scheme are caused



258 The reproduction of colour



by inter-image effects which appear as unexpected

increases in density with increased exposure, or by

overlap of filter pass bands with sensitivity bands of

the emulsions. This last departure from the ideal

results in a compression of the three regions with the

possible loss of the middle one and overlap of the

other two. In the examples shown, the middle region

is not evident on green or blue exposure owing to the

sensitivity separations of the three emulsions being

only as much as is needed for adequate colour

reproduction. Only the red exposure has sensitivity

separation to spare. Indeed the extensive range of

exposures over which no change is seen in the image

may be the cause of poor reproduction of form and

texture of vivid red subjects. Red tulips or roses, for

example, may simply appear as red blobs with no

visible tonal quality or texture.

Inter-image effects are extremely common in

colour processes, either by design or accident. They

are not usually detectable by integral densitometry,

although there are a number illustrated in Figures

16.12 and 16.15.



imately 3.0, so an upper limit is therefore set on the

density scale of a reversal film. The necessarily high

minimum density caused by the integral r´ seau of

e

additive colour filters imposes a downward limit on

the density of highlights that is considerably higher

than that of subtractive reversal films. Transparencies

made using additive films tend therefore to appear

rather dark and to have a shorter tonal scale than

subtractive materials. Not only does this short tonescale limit the latitude of the film, but it may also

limit the range of colours available.



Imperfections of colour processes

Additive system

Typical spectral sensitivity curves of the eye are

shown in Figure 16.17(a); the sensitivities of the three

colour receptors overlap considerably. While the red

receptor alone is stimulated by light of a wavelength

of 650 nm or greater, there are no wavelengths at



Additive systems

The neutral characteristics of a typical additive

colour-reversal film are shown in Figure 16.16. A

number of the limitations of such processes are

immediately obvious. The useful upper limit of

density in images designed for projection is approx-



Figure 16.16 Neutral characteristics of a typical additive

colour-reversal film



Figure 16.17 Spectral sensitivities of the human eye and a

colour tripack film