Colour Photography

White light comprises a continuous range of wavelengths within a band of the electromagnetic spectrum, sections of which we interpret as colour. Individual colours can, though, be represented by mixtures of red, green and blue light (the additive primary colours). This is not a property of the light but of the way our eyes see colour. With a few exceptions colour photography was based on representing light by its red, green and blue components.1

Additive
In additive three-colour processes the scene is photographed separately through red, green and blue filters, positives are made from the three negatives which can be projected behind filters of the same colour onto a screen. The three overlapping beams of coloured light will mix on the screen to reproduce the colours of the original scene. Since, at the viewing stage, we are adding light to light to form the original array of colours this became known as the 'additive method', or in today's terminology the 'RGB colour model'.

Subtractive
An alternative was to print the positive transparencies in the complimentary colour to the taking filter, thus the red-filter negative would be printed as a cyan positive, the green-filter negative would be printed in magenta and the blue-filter negative would be printed as a yellow positive; the three positives could then be superimposed and viewed by white light (either by projection or as a print on paper).2 The cyan image provided 'drawing' to the image and was ideally printed at the front.

Since we are here subtracting or filtering out, colour from white light this became known as the 'subtractive method', or CMY colour model in today's terminology (an additional grey or key image is often included when printing leading to CMYK).

When working with pigments, such as printer's ink, rather than light it is the subtractive colour model that is used.

The terms cyan and magenta are relatively modern, in early descriptions of colour photography blue-green or 'minus red' is mostly used instead of cyan, sometimes blue and red will be used in place of cyan and magenta.

In both the additive and subtractive processes colour is introduced either by filters, pigments or chemical changes and is to some extent simulated. What was meant by red, green and blue light was not defined in terms of wavelength and bandwidth nor was the exact composition of pigments.3 For this reason there were differences in the way individual processes reproduced colour.

Hand Colouring

Daguerreotype and Ambrotype

Soon after their introduction the harsh tones of the Daguerreotype were made softer and more natural by hand colouring. At its simplest faces were 'rouged'; at its most elaborate most of the image was coloured, some of the most striking examples were soldiers in red uniforms. Buttons, watch chains and jewellery were often picked out in gold paint.

The colour was mostly powder mixed with gum, when dry this was applied with a fine camel's hair brush. The colouring was fixed to the plate by breathing on it. The nature of the powder pigment gave a subtle natural look to the plate unlike the full opaque colouring sometimes applied to prints. Much the same process was used on ambrotypes.4

Prints

Salt and albumen prints were coloured with oil or water colours, salt prints were particularly suitable as the uncoated surface would readily absorb paint. The colouring varies from pale washes to the print being completely re-worked and its photographic nature being hidden. Mostly colours were applied to the front of the image, sometimes the print was made partly transparent and coloured from the back. Tylar sold a special collodion plate on which a positive image was made, paper was then placed over the front and broadly coloured with water colours, when dry the paper was attached to the back of the plate.5 A variation was to paint the front in oil, varnish it and while still wet attach it to a sheet of glass. Where a run of prints was made, such as stereograms, stencils might be used.

Later, gelatine-based, prints could be coloured with water colours, colouring sets remained on sale into the 1960s.

Lantern Slides & Films

Lantern slides were often coloured either by hand, stencil or using a lithographic printing process. Cinema films were also hand-coloured or on the most popular films coloured using stencils.6

Tinting & Toning

Prints and films could be given a uniform colour change by tinting (staining) the gelatine base. In toning, the metallic silver image was modified by changing its composition or replacing it with another compound.7 Typically this would be used on, for example, a seascape that would be toned blue. Coloured printing paper was also available.

An interesting use was made in early cinema films where, for instance, a newsreel of a fire would be toned red.8

Crystoleum

A Crystoleum comprises a print stuck to a curved glass, the back of the print is rubbed away and oil applied making the print transparent. Fine detail is then applied in oil colours to the back of the print. A second similarly shaped glass is attached to the back but with a small separation. On the second glass broad areas of colour are applied. Finally a card is stuck to the back and all the components are bound together.

Flexichrome

An interesting colour process was developed by Jack Crawford around 1940 in the USA. A black and white negative was printed on to gelatine silver bromide paper which was developed to form a relief image. This was dyed grey to reveal the image, it was then locally coloured by applying dye. The dyes if overlaid did not mix, the last applied removed the first dye. The tone of the original photograph was preserved in the gelatine relief.

The patent was assigned to Eastman Kodak who marketed the process.9

Direct Methods

Lippmann Interference Process

This is a natural colour process and does not rely on filters or dyes to reproduce the colours. A glass plate coated with a fine-grain emulsion is placed in contact with a layer of mercury and exposed through the glass base. Light which passes through the emulsion is reflected back producing standing waves in the emulsion where the incident and reflected rays interfere. The interference pattern which is unique for each wavelength is recorded in the emulsion layer. When the plate was illuminated in a certain way the recorded wave pattern reflected back only the wavelength with which it was formed. The process was demonstrated by Lippmann in 1891. The Lippmann plate was viewed from the emulsion side by reflected light, a shallow prism was sometimes mounted in front of the plate to eliminate surface reflections. Apparatus for the process was sold commercially.

The image on the right shows a Lippmann process dark slide.

Dispersion Process

There were various suggestions for recording the dispersion spectra produced by a prism. One of the earliest was by F.W. Lanchester in 1895.10 His system was for a lens to form an image on a grating of clear and opaque lines, behind the grating was a second lens that formed an image of the grating on a screen or photographic plate. Associated with the second lens was a narrow-angle prism. The screen would show a series of spectra, one for each line in the grating. The opaque lines in the grating were twice the width of the clear lines, the apparatus was arranged so that each spectrum just filled the space caused by the opaque lines i.e. the spectra were contiguous. After the spectra were recorded on a photographic plate, a positive transparency was produced, this was placed in the same position as when recording the spectra. When the grating was illuminated by white light the transparency would show the colours of the original object.

One of the simplest proposals was by M. Raymond, where a diaphragm having a slot rather than a circular aperture was placed behind the lens. An image was formed on a cross line screen (as used in half-tone work). This would break-up the image into a number of strips with spaces in between. Between the cross-line screen and the sensitive plate was a prism which produced a spectrum for each line in the screen.

Additive Processes

Three-colour Separation Negatives

Separation negatives, made through red, green and blue filters, form the basis of many additive and subtractive processes. The negatives could be produced in special three-colour cameras that exposed the negatives at the same time or by using a repeating back that made the exposures sequentially.

Early in the development of colour photography yellow was sometimes used in place of green (e.g. by Du Hauron) and orange rather than red.

Three-colour Separation Transparencies

'Black and white' positive transparencies are made from a set of separation negatives. These are projected through red, green and blue filters to form an image on a screen. The transparencies could also be placed in a viewer.

The first demonstration of a photograph taken in this way was by J. Clerk Maxwell in 1861.11 The process had though to await panchromatic or at least orthochromatic sensitised plates before use could be made of it. The Ives Kromscope system of 1895, where mono or stereo images were mounted in a viewer, was an early commercial application.

The image on the right shows a Kromogram for use with a Kromscope viewer.

Successive Frames

Rather than project each of the three-colour transparencies at the same time as described in the previous section, the three transparencies could be projected in quick succession, the resultant colour being fused in the eye.

This method has applications in cinematography where a red/green/blue filter wheel is spun in front of the film and synchronised to the shutter and film movement. The Turner & Lee system developed in 1899 used this arrangement. Kinemacolor, an early and successful two-colour system, used only green and red frames.

Mosaic Screens, Screen-Plate Processes

The principle here is the same as in separation transparencies, that is the subject is photographed through red, green and blue filters. Here though, the three filters are microscopic in size and in contact with the plate. Two forms exist:

  • Separate screen - where the coloured screen is on a separate plate to the photographic plate. A similar screen is used to view the image.
  • Combined screen - where the coloured lines or dots are on the same plate as the photographic emulsion.

In combined screen processes the filter elements could be either random, as in the Autochrome process, or regular consisting of lines and squares. In separate screen processes the screens had to be regular.

The images on the right show the random pattern of an Autochrome plate and the regular arrangement of Dufaycolor.

Screen plates were first proposed by Du Hauron, the first application was by Joly who introduced a plate in 1894. The Autochrome process of 1907 was the most successful.12

The image below right is from an Autochrome plate, the image below left is from a Joly plate.

Lenticular

This is an additive process first proposed by Berthon in 1908 and further developed by Keller-Dorian and Berthon. Immediately in front of the camera lens is a filter consisting of three bands - red, green and blue - the film base was moulded into cylindrical lenses which formed an image of the filter on the emulsion. The effect was for the emulsion to contain narrow bands corresponding to the three filter bands. To project the image the filter was placed over the projector lens.

Kodak introduced the process as Kodacolor in 1928 in 16 mm cine format. AGFA sold lenticular film for still and cine use as Agfacolor in 1933.

Wood's Diffraction Process

Three ordinary separation negatives were produced and from these three positive transparencies. Each transparency had associated with it a diffraction grating (2000 lines per inch for the red image, 2400 for the green and 2750 for the blue). The transparencies together with their diffraction grating were printed in turn onto a single gelatine bichromate plate, after development in warm water this contained an image made up of records of the diffraction gratings impressed into the gelatine, the images of the diffraction grating may overlap to record a colour mixture, e.g. where the green and red gratings overlap the eye will see yellow. The image is viewed by placing it behind a simple condenser and illuminating it with white light. At the focus of the lens is a screen with two small holes for viewing (an image is formed each side of the centre line). The process did not lead to any practical applications.13

Subtractive Processes

Subtractive Assembled Prints and Transparencies

Three-colour separation negatives, red, green and blue, are produced. The positive image is made by overlaying positives made from each of the negatives printed in their complimentary colours. The final image can be viewed as a transparency or a print.

The individual coloured positives were produced by a number of methods:

  • Using pigmented relief images e.g. carbon.
  • Dye transfer from gelatine matrices.
  • Chemically toning the silver image.
  • Using the cyanotype process to produce the cyan image.
In practice these methods were often combined into a single process.

Du Hauron first showed prints made by the subtractive process in 1869. Commercial applications appeared in the late 1890s.

Carbon, Carbro
When exposed to light bichromated gelatine becomes insoluble, the unexposed areas remain soluble in warm water. In the carbon process a layer of pigmented gelatine bichromate is exposed under a negative, after exposure the unexposed areas are removed in warm water. The gelatine image comprises a pigmented gelatine sheet that varies in depth in proportion to the tones of the negative. In three-colour work three such sheets are produced in the subtractive primaries from separation negatives. These are superimposed on to a paper or glass support.14

Gelatine in the presence of a bichromate becomes insoluble when in contact with metallic silver. In the carbro process a bromide print is placed in contact with a gelatine bichromate sheet, after a few minutes the gelatine will have become insoluble where it was in contact with the metallic silver of the bromide print. The insoluble parts were then washed away. The gelatines could then be treated as ordinary carbon sheets.15

Imbibition, Dye transfer
The basis of this process was that bichromated gelatine will selectively harden when exposed under a transparency. The harder areas will be less receptive to dye whilst the less hardened areas will take up dye in proportion to the exposure. Bichromated gelatine sheets were exposed under separation positives, after exposure the bichromate was removed by washing and the gelatine dyed in the appropriate subtractive primary. The individual gelatines were then pressed into contact with a gelatine coated surface (e.g. paper or glass) where the dye was transferred. Examples of this process were Pinatype (1903) and Eastman Wash-off Relief (1935).16

Separable Tri-packs & Bi-packs

The major problem in producing separation negatives was the requirement to use special cameras, repeating backs or suffer a delay while three exposures were made in succession. Tri-packs allowed a single exposure in an ordinary camera. The pack consisted of, for example, two sensitive plates placed either side of a sheet of cut-film with interleaved filters to control penetration of the primary colours.

Du Hauron was the first to propose tri-packs consisting of three plates in 1895. Ives introduced a tri-pack system commercially in 1916, another commercially available process was Colorsnap of 1928.17

The separation of the emulsion could cause a lack of sharpness, light scatter and halation. None of the commercially available processes were very successful.

A bi-pack consists of only two emulsion layers and records only two colours. Because the layers could be closer together it did not suffer problems to the same extent as tri-packs.

Integral Tri-packs

To overcome the problems with ordinary tri-packs caused by the distance between emulsion layers ways were sought to coat the three colour-sensitive layers on one base.

The first practical realisation of this were Eastman's Kodachrome of 1935 and Agfa's Agfacolor Neue of 1936. Both used dye-couplers, the final image being made of dyes - cyan, magenta and yellow, the original silver images being bleached out. The process could be adapted to produce prints on paper as well as transparencies. Tri-pack coatings were used for transparencies (reversal), negatives and prints. Negative/positive cine film was introduced by Agfa in 1939.18

Bleaching-Out Print Processes

In this process dyes in the three subtractive colours are coated on a base. This is exposed under a colour transparency (such as an Autochrome). The dyes faded on exposure to light of their complimentary colour i.e. red light passing through the transparency faded (bleached) the cyan dye but left the yellow and magenta dyes unaffected, thereby producing a red image.

The first application was a printing paper introduced in 1904 as Utocolor which was not successful.19 There were few other applications.

Silver-Dye-Bleaching Print Processes

Dyes can be destroyed when in the presence of the developed silver image and a catalysing compound. The process has been successfully used to produce prints where three emulsion layers are coated on a single base and exposed through a transparency. The emulsion layers are sensitive to one of the three primary colours and contain dyes of the complementary colour. For example, on exposure through, say, a red area of the transparency the cyan dye in the red-sensitive layer will be destroyed leaving yellow and magenta dyes present in the other two layers.

A variation of the process, Gasparcolor, was developed for printing cine film for animations using separation prints in 1933. Gasparcolor Opaque for producing prints from transparencies was introduced around 1949. Later, print processes were introduced by Ilford and CIBA (Cibachrome).

Photo-mechanical Printing

When printing using a photographic process the three images can superimpose to produce the basic hues. The three separate images will vary in density which will alter the tint of the resultant colour in the superimposed image. In photo-mechanical printing, with the exception of photo-gravure, there is no provision to modify the tint, there is either a presence or absence of ink. The most common solution was to use a half-tone screen to break up the image into discrete dots that varied in size according to the density of the original photographic image. At the normal viewing distance the dots would merge into a single colour.

Trichromatic Colour Vision and Reproduction

The theory of trichromatic colour vision has its basis in Thomas Young's conjecture (1802) that the eye contained three types of cells that were each responsive to different bands of the spectrum.20 The theory was further developed by Hermann von Helmholtz in the 1850s and 60s.21 It was only in recent times that the existence of cone cells that respond differently to light of short, medium and long wavelengths was proven.

The diagram to the right shows how the three receptors - labelled rho, gamma and beta - respond to different wavelengths of light.

These functions may be normalised to enclose the same area or so that the peaks are the same height or drawn to an actual rather than relative scale.

A colour sensation can, then, be described by three parameters. This provides the opportunity of replicating a colour sensation by mixing three lights red, green and blue in the proportions shown by the ordinates at a particular wavelength in the diagram.

Maxwell's Colour Top

James Clerk Maxwell devised a Colour Top where overlapping paper discs in the three primary colours (red, green and blue) could be placed on a disc in such a way that their proportions (the amount of each that was visible) could be altered, in the centre of the disc was the reference colour that was to be matched. By altering the proportions of the paper discs and spinning the top it was possible to match the reference colour and quantify the colour into its red, green and blue components. The edges of the discs carried scales, the three values could be used to find the reference colour on a Colour Triangle. The triangle showed the three primaries at each vertex with graduations of colour produced by mixing the primaries in the interior.

Trichromatic Matching Functions

From the diagram showing the spectral sensitivity curves it can be seen that there is overlap of the responses especially in the rho (red) response. Whilst it is possible to accurately record the spectral sensitivity responses using a particular set of colour filters and film sensitivity/characteristics there is a problem in reconstructing the original response from the three records.

For example using trichromatic lights a colour sensation resulting from monochromatic light in the region of 480 nm will be recorded on the three images as strong blue, strong green and also a red value. In playing these records back to the eye the green record will trigger a red response as it should but there is a further red response triggered by the red record, the result is that the reproduced colour is less saturated than the original.

The best at the reproduction stage would be three lights that only stimulate their respective cones, unfortunately this is not possible.

The second diagram shows the colour matching functions which with three light sources of 650, 530 and 460 nm wavelength will reproduce the colour sensitivity of the human eye, it will be noted that in some areas the stimulus is negative - which is awkward. The upshot of this is that certain colours cannot be reproduced by three lights i.e. where one or more of the trichromatic values is negative.

In a demonstration of trichromatic colour matching the negative amount of colour can be added to the sample that is to be matched.

In non-photographic printing further colours could be used such as pure yellow and reds that would increase the range of reproducible colours.

The perceived colour provided by the trichromatic values need not be unique, the same colour may be perceived as a result of a different set of stimuli, a phenomenon known as metamerism.

References & Notes

General references
BJA 1899, p. 690. BJA 1902, p. 867. BJA 1908, p. 557. Eder, History, p. 639. Hasluck, Book of Photography, p. 420. Kodak Mus. Cat. p. 47. Friedman, History of Color Photography. Wall, Three-Color Photography. Coote, History of Colour Photography. Coe, Colour Photography.

A nice demonstration of complimentary colours is given in Spencer, Colour Photography in Practice:

Look at the cross on the left for around 30 seconds then look at the cross on the right, the cross on the right should be surrounded by the complementary colours of those on the left.

Hand Colouring

Crystoleum: BJA 1907, p. 923. YBP 1888, p. 183. Cyclopedia of Photography, p. 154.

Direct

Interference: Bolas, Phot. in Colours, p. 317. BJA 1901, p. 819. Eder, History, p. 668. Hasluck, Book of Photography, p. 421. Photographic News 1896, p. 258.

A 'biographical sketch' of Lippmann is contained in The Photogram 1894, p. 186.

Dispersion Process: Friedman, History of Color Photography, p. 24. Wall, Three-Color Photography, p. 659. BJA 1908, p. 714.

Additive

Mosaic Screen: Coe, Colour Photography, p. 46. Clerc, Photography. Theory and Practice, p. 544. Eder, History, p. 660.

Lenticular: Clerc, Photography. Theory and Practice, p. 541. Eder, History, p. 672. BJA 1911, p. 627. PJ 9/1929, p. 402.

Three-colour Separation Transparencies: Coe, Colour Photography, p. 28. Clerc, Photography. Theory and Practice, p. 541, Eder, History, p. 656.

Subtractive

Separable Tri-packs & Bi-packs: Coe, Colour Photography, p. 110. Eder, History, p. 647. Coote, History of Colour Photography, p. 111.

Integral Tri-packs: Coe, Colour Photography. Clerc, Photography. Theory and Practice, p. 543. Coote, History of Colour Photography, pp. 154, 158. BJA 1952, p. 136.

Subtractive Assembled Prints and Transparencies: Coe, Colour Photography, p. 84. Eder, History, pp. 654, 642, 539. Hasluck, Book of Photography, p. 429. BJA 1907, p. 628. Coote, History of Colour Photography, p. 84.

Bleaching-Out Process: Coe, Colour Photography, p. 130. Clerc, Photography. Theory and Practice, p. 553. Eder, History, p. 673. Hasluck, Book of Photography, p. 435. Neblette, Principles and Practice, p. 537. BJA 1907, p. 648.

Silver-Dye-Bleaching: Friedman, History of Color Photography, p. 405. Coote, History of Colour Photography, pp. 131, 184.

[1] White light is roughly the wavelengths of electromagnetic radiation between 390 - 700 nanometre (millimicrons).
Objects can selectively absorb, reflect and transmit light frequencies. In white light an object appears coloured due to the electro-magnetic wavelengths that it reflects or transmits. Some objects may absorb certain wavelengths of radiation and emit the energy at a different wavelength, a property termed fluorescence. This is particularly noticeable if the absorbed wavelength is in, say, the ultraviolet range and re-emitted as visible light. Iridescence is a phenomena seen, for instance, in soap bubbles where the colour is seen to change with the viewing angle or as the object changes dimension, this is caused by interference from multiple reflections from surfaces of semi-transparent objects. Iridescence can also be formed by diffraction such as is seen in some cloud formations close to the position of the sun.

[2] The transparencies act as filters, as an example, where the yellow and magenta images overlap, the yellow transparency stops blue light, the magenta transparency stops green light thus a red image results.

[3] More accurately there were many different opinions as to the theoretically correct bandwidths to use.

[4] Beard's patent 9292/1842.

[5] Called Chameleotype. The Photogram 1894 p. 168.

[6] A stencilling machine is shown in The First Colour Motion Pictures by D.B. Thomas, p. 3.

[7] Clerc, Photography. Theory and Practice, gives details of ferrocyanide and uranium toning.

[8] A description is included in Came the Dawn by Cecil M. Hepworth.

[9] US Pat. 2244905/1941, US Pat 2324069/1943. Coe, Colour Photography, p. 17.

[10] Frederick William Lanchester BP 16548/1895.

[11] The demonstration was to the Royal Institution on 17/5/1861. Although it produced the correct result it did so for the wrong reasons, the photographs were made using wet collodion plates that were insensitive to the red end of the spectrum, by chance the liquid filter used to record the red image also transmitted ultraviolet light, the red dye used in the cloth being photographed emitted ultraviolet, thus a 'red' image was secured.
Thomas Sutton, who made the exposures, also made a photograph through a yellow filter, what part this played in the demonstration is not clear.

[12] For details of the Autochrome and Dufaycolor processes see Autochrome and Dufaycolor. For details of the Joly process see Joly.

[13] BP 4601/1899. BJA 1900, p. 830. BJA 1901, p. 829. Wall, Three-Color Photography, p. 670.
The principle behind the process is that gratings with different spacing (lines per inch) produce spectra at different angular distances from their centre. By choosing three gratings with different spacings it is possible for a red band in one spectrum to have the same angle of diffraction as a green band in the second grating and a blue band in the third grating.

[14] See Carbon for details of the carbon process.

[15] For details of the Trichro Carbro process see Trichrome Carbro.

[16] Coe, Colour Photography, pp. 100, 109. Friedman, History of Color Photography, p. 462.

[17] Tarbin, William Thomas, BP 283765/1928.

[18] For details of the Kodachrome process see Kodachrome.

[19] Developed by J. H. Smith.

[20] Bakerian Lecture: On the Theory of Light and Colours. Phil. Trans. R. Soc. Lond. 1802 92, doi: 10.1098/rstl. 1802.0004. See hypothesis III.

[21] Helmholtz's opinions on three-colour theory were changing over this period.

Colour Photography

Hand Colouring

Daguerreotype

Prints

Slides

Tinting

Crystoleum

Flexichrome

Direct Methods

Lippmann

Dispersion

Additive Processes

3-colour Separation Negs

3-colour positives

Successive Frames

Screen-Plate

Lenticular

Diffraction

Subtractive Processes

Assembled Prints

Tri-packs

Integral Tri-packs

Bleaching Out

Silver-Dye-Bleaching

Photo-mechanical

Colour Vision

Maxwell's Colour Top

Matching Functions

References & Notes


Return to Glossary

  Top