This database was created in 2012 and has been developed and curated by Barbara Flueckiger, professor at the Department of Film Studies, University of Zurich to provide comprehensive information about historical film color processes invented since the end 19th century including specific still photography color technologies that were their conceptual predecessors.
Timeline of Historical Film Colors is started with Barbara Flueckiger’s research at Harvard University in the framework of her project Film History Re-mastered, funded by Swiss National Science Foundation, 2011-2013.
In 2013 the University of Zurich and Swiss National Science Foundation awarded additional funding for the elaboration of this web resource. 80 financial contributors sponsored the crowdfunding campaign Database of Historical Film Colors with more than USD 11.100 in 2012. In addition, the Institute for the Performing Arts and Film, Zurich University of the Arts provided a major contribution to the development of the database. Many further persons and institutions have supported the project, see acknowledgements.
Since February 2016 the database has been redeveloped in the framework of the research project Film Colors. Technologies, Cultures, Institutions funded by a grant from Swiss National Science Foundation, see project details on SNSF grant database.
Follow the links “Access detailed information ›” to access the currently available detail pages for individual processes. These pages contain an image gallery, a short description, a bibliography of original papers and secondary sources connected to extended quotes from these sources, downloads of seminal papers and links. We are updating these detail pages on a regular basis.
In June 2015, the European Research Council awarded the prestigious Advanced Grant to Barbara Flueckiger for her new research project FilmColors. Bridging the Gap Between Technology and Aesthetics, see press release of the University of Zurich and short abstract on the university’s research database.
Subscribe to the blog to receive all the news: http://filmcolors.org/ (check out sidebar on individual entries for the “follow” button).
The development of the project started in fall 2011 with stage 1. Each stage necessitated a different financing scheme. We are now in stage 3 and are looking for additional funding by private sponsors. Please use the Stripe interface to pay conveniently online or transfer your financial contribution directly to
Account IBAN CH2509000000604877146
Account holder: Barbara Flueckiger, CH-8005 Zurich, Switzerland
SWIFT Code / BIC: P O F I C H B E X X X
Bank: PostFinance AG, Mingerstrasse 20, CH-3030 Bern, Switzerland
Clearing Nummer: 09000
Read more about the financial background of the project on filmcolors.org.
The author has exercised the greatest care in seeking all necessary permissions to publish the material on this website. Please contact the author immediately and directly should anything infringe a copyright nonetheless.
We have seen that the departures from exact colour reproduction inherent in colour photography are of a fundamental and more or less incurable nature. It is necessary, therefore, to consider how important such departures are in practice, and in particular whether errors in some directions are more important than those in others. Before classifying errors under various headings, however, and weighing their relative importance, the way in which colour rendering is judged must be considered for a moment.
Colour photographs are practically never compared side by side with the original scene. Moreover, very few colour photographs are taken with a view to their being seen only by persons who were present at the time of exposure. So the majority of the criticisms of a colour photograph come from persons who never saw the original scene, and whose judgment must be based on some mental comparison between what the picture looks like and what they think it ought to have looked like. The precision of such mental comparisons will depend entirely on the precision of the mental standard used. For instance, if in a colour photograph there was depicted, amongst other things, a pillar-box, then its colour would be mentally compared with one’s impression of the usual colour of pillar-boxes. The impression will be the average of those given by a large number of different pillar-boxes seen on different occasions. And, of course, the colour sensations received will have been subject to the considerable variations caused by differences in the colour, intensity, and the direction of the lighting, whether the surface was wet or dry, dusty or clean, whether the pillar-box had been recently painted or not, etc. The impression, therefore, cannot be precise; and hence, provided that the reproduction of the pillar-box in the colour photograph compares favourably with what a pillar-box could look like, it will generally be acceptable.3
Similarly, all objects of well-known colour give rise to colour sensations which are not always the same and which exhibit quite wide variations. It is these variations, therefore, which govern the tolerances available in colour photography. Let us now consider some of these variations under the headings, lightness, saturation, and hue.
Variations of Lightness
Lightness is probably the attribute of surface colours which varies most from point to point over their surfaces. Any fabric tends to hang in folds, and the troughs will be much darker than the crests. Foliage, and grass, are subject to wide variations in lightness due to the shading of one leaf or blade by another, and due to the difference in orientation with respect to the direction of the incident light. Simultaneous contrast between a patch of colour and its background will also affect its apparent lightness. Thus a given colour will appear much lighter when seen against a black background than when seen against a white background. Thus one would expect errors in lightness reproduction to be relatively unimportant in colour photography, and this is borne out by the fact that there is a considerable range of contrasts over which a colour photograph may vary without detriment.
Variations of Saturation
Saturation also exhibits variations from point to point over a surface, especially on any surface having some sheen or gloss. Such surfaces can also show large increases in saturation when the type of lighting is changed from diffuse to directional, and it is well known that a scene always looks more colourful when the sun is out than when the weather is overcast. The saturation of all colours of distant objects is likely to be decreased by atmospheric haze, and sometimes the effect is strong enough to remove all sensations of colour completely. Reference has already been made to the presence of dust or dirt on surfaces, and while this may result in some changes of lightness, the change in saturation will be considerable. Wetting a matte surface often results in startling increases in saturation. The saturation of the blueness of the sky varies enormously with the direction of viewing relative to the sun, and similar variations occur, of course, in the case of the blueness of seas, rivers and lakes. The apparent saturation of colours also varies with the intensity of the illumination; for instance, at dusk colours are far less saturated than at noon, and by moonlight colour vision has almost ceased, all colours appearing almost completely desaturated.
The use of illuminants of different colours, such as tungsten light and daylight, also results in variations in apparent colour,4 and in the case of blues and yellows the differences in saturation can be considerable. Similar effects also occur, of course, with different phases of daylight, such as noon sunlight, north sky light and late evening sunlight. Again simultaneous contrast can alter the apparent saturations of colours. A pale colour seen against its complementary colour appears more saturated than when against a saturated colour of the same hue. It would thus be expected that errors in saturation would not be of very great importance. This is borne out by the appearance of many water-colour paintings, in which the colours are usually quite pale, but which as pictures are often highly successful. It seems that rather than requiring exact reproduction of saturation, all that is necessary is a reasonable saturation-maximum for each hue, and a uniform desaturation of colours of all hues and saturations, since this is what normally occurs in the conditions mentioned above. In terms of the purity characteristic curve, suggested by Wright,5 this means that the purity gamma can have a value substantially below unity, but that it should have the same value for all hues, and that the curve should be linear with saturation.
Variations of Hue
Let hue as a variable in surface colours now be considered. Simultaneous contrast can cause apparent changes in the hues of colours, but it is clear that most of the phenomena described above, which give rise to changes in lightness and saturation, do not give rise to any changes in the hues of colours. The hues of some objects, however, are quite variable. For instance, foliage varies in hue with time of year and with type of tree, and nearly all fruits change hue with degree of ripeness, as well as being different for different varieties. The colour of flesh varies with type of skin, and, of course, with amount of sunburn. There are, of course, other objects the hues of which vary, but, generally speaking, variations in hue, while important, would seem to be more restricted in surface colours than variations in lightness and saturation. It is, for instance, easier to think of a pillar-box which is a light or a dark red, or a pale or a deep red, than to think of one which is an orange- or magenta-red.
By this type of argument, and by experience, an approximate order of priority in the requirements, as far as colour is concerned, of a successful process of colour reproduction is arrived at, as follows:6
1. Correctness of hue.
2. Approximately equal desaturation of colours of all hues.
3. Approximately proportional desaturation of colours of all saturations.
By way of illustration of these principles it is a well-known fact that in colour reproduction the variable with the least tolerance is the overall colour balance of the picture. If the picture is slightly off balance, pale colours will undergo violent changes of hue, and it is these which make off-balance pictures so intolerable.
It is concluded, therefore, that, owing to the way in which the colours in a photograph are judged, and owing to the large changes in colour which well-known objects so often undergo, the discrepancies inherent in present-day methods of colour photography can be tolerated. That is not to say, of course, that improvement is not desirable, and in certain types of process special devices have to be resorted to in order to overcome some of the discrepancies because they have exceeded the admittedly very wide tolerances.
Most of the effects described in the paper were demonstrated during the lecture, either by actual experiments or by means of colour transparencies.
3Phot. J., 91B, p. 2, 1951.
4G.E.C. Journal, 18, Apr., 1951.
5J. Brit. Kine. Soc., 13, p. 1, 1948.
6Phot. J., 91B, p. 107, 1951.”
(Hunt, R.W.G. (1951): Colour Cinematography and the Human Eye. In: British Kinematography, 19,6, pp. 173–180, on pp. 176–178.)
“Tati, ad esempio, nei suoi film a colori ripensò cromaticamente l’universo diegetico dei suoi personaggi, trasformandolo in un formidabile campionario di oggetti colorati20. Mio zio (Mon oncle, 1958) fa dell’automobile uno dei tanti mezzi tecnologici presenti nel quotidiano del ricco industriale Arpel, ansioso di sperimentare tutto quanto si configuri come nuovo e funzionale, dall’automazione dei dispositivi alla conversione ai nuovi oggetti di design e ai più aggiornati criteri costruttivi. In una delle sequenze iniziali, il personaggio esce di casa con una elegante berlina grigia e bianca, mescolandosi visivamente con le altre automobili sulla strada, accomunate dagli stessi colori. Il montaggio compone una successione di immagini impostate sulla dominanza di tinte metalliche, metallizzate o cementificate, cui si conformano, oltre alle auto, i muri della casa, gli abiti, l’asfalto, l’esterno della scuola e dell’azienda. Già attraverso una simile ricorrenza del grigio il film riesce a marcare una distanza incolmabile rispetto alla visione in bianco e nero. Tutti questi grigi piatti e a campiture nette, attraversati da vari riflessi cromatici, con la loro brillantezza od opacità, levigatezza o granulosità, hanno ormai poco a che fare con i grigi delle immagini in bianco e nero. La relazione significante tra grigio e colore viene spostata su un piano più raffinato, meno scontato e banale rispetto a Racconti romani.
Il film apre al grigio una nuova visibilità, sottraendolo al dominio dell’incolore per riqualificarlo a dimensione cromatica costitutiva del mondo tecnologico. Il grigio non è solo diventato un colore tra i colori, ma quello che, più di ogni altro, sembra capace di riconfigurare la fisionomia degli spazi e degli ambienti moderni, industriali e urbani. Il film esprime la particolare consapevolezza che soltanto la pellicola a colori possa restituire visivamente le qualità specifiche di questa nuova serie di grigi cromatici portata da nuovi materiali, plastiche, smalti, lacche, vernici, tinture, cementi, bitumi e prodotti sintetici21. In bianco e nero, tutte queste varianti pigmentarie del grigio avrebbero perso la loro forza. Tanto i grigi quanto i moderni colori consumistici sembrano così accomunati dai medesimi valori di superficie e di contorno, dalla disposizione ordinata, quasi chirurgica, dei rispettivi materiali e vernici. In tutte le sequenze dedicate al côté ultramoderno il film fa coabitare sistematicamente questi grigi con i colori più accesi. La serie degli oggetti domestici dispiega un ricco campionario di tinte industriali, frutto di un’ironica rilettura delle più recenti proposte del design: i divani verdi e le sedie con imbottiture rosa, gialle, verdi, l’ombrellino blu, gli abiti sgargianti della vicina di casa, la sciarpa a quadri rossi e neri di Arpel e del cane, il tosaerba, l’idrante e la palla rossi, la sedia a dondolo gialla. Sul finire del film, la gamma dei moderni colori è completata da una sgargiante berlina verde, lilla e rosa, che aggiorna ai dettami della moda del tempo il precedente modello grigio e bianco.
Tati rivelava dunque come la costruzione del nuovo visibile potesse sostanziarsi tanto nelle più moderne cromie, spinte al massimo grado di saturazione, quanto nelle nuove zone di grigio che, paradossalmente, proprio l’immagine a colori sembrava poter aprire allo sguardo.
20 Tati si era interessato al colore fin da Giorno di festa (Jour de fête, 1949) che aveva girato ricorrendo al sistema additivo lenticolare francese Thomsoncolor. Era però dovuto ricorrere alle riprese di sicurezza in bianco e nero per far approdare in sala il film: dai negativi del Thomsoncolor non si era stati capaci di tirare fuori alcunché di proiettabile e sarebbero occorsi oltre quarant’anni prima che l’operazione risultasse possibile. Giorno di festa sarebbe rimasto per decenni un film in bianco e nero (per approfondimenti, cfr. Ede 1995).
21 Il medesimo discorso può essere esteso al bianco: si pensi alle sequenze ambientate nella cucina.
Ede, Francois (1995), “Jour de fête” ou la couleur retrouvée, Cahiers du cinéma, Paris.”
(Pierotti, Federico (2012): La seduzione dello spettro. Storia e cultura del colore nel cinema. Genova: Le Mani-Microart, on pp. 223–224.) (in Italian)
Our knowledge of the way in which the retina responds to light of different colours is in many ways incomplete. But it is generally agreed that there must be three different types of receptor: one type mainly sensitive to orange and red light, another mainly sensitive to green light, and a third mainly sensitive to blue light. Moreover, from various experiments on colour matching,1 the spectral sensitivity curves of these three types of receptors are known approximately and are as shown in Fig. 1, the three sensitivities, β, γ, and ρ being plotted against wavelength. When light of any colour falls on the retina, the different sensitivity curves of the three receptors will result in three signals, β, γ, and ρ being sent to the brain, and if, for instance, the ρ signal is larger than the other two signals, then a reddish sensation will be experienced; if the γ signal is the largest, then the sensation will be greenish, and so on.
If, now, the three layers of the colour film had sensitivity curves closely matching those of the eye, then it might be thought that correct colour reproduction must necessarily result. But three layers of photographic emulsion are only capable in themselves of producing black-and-white images; the colour has to be put into the system elsewhere. The simplest way, from the theoretical standpoint, of doing this is to strip the three black-and-white negative images apart,* print (with perfect tone reproduction) black-and-white positive transparencies from them, and insert them in three separate projectors, one of which is fitted with a blue filter, another with a green filter, and the third with a red filter. If the red filter were chosen so that the light transmitted by it affected only the ρ-receptors of the eye, the green filter only the γ-receptors, and the blue filter only the β-receptors, then on superimposing the three images on a screen exact colour reproduction would result. For instead of the colours of the original scene producing their appropriate ρ, γ, and β signals, they have produced areas of photographic density on the positives in the projectors such that at each point the transmission is proportional to the ρ, γ, or β signal, as the case may be. Hence light from the red lantern produces the correct ρ signal at each point in the picture, that from the green lantern the correct γ signal, and that from the blue lantern the correct β signal, and hence the eye, receives the same signals ρ, γ, and β from the reproduction as it would have done from the original.
In order for the above state of affairs to obtain, it is essential that:
(a) the three layers of our film should have the sensitivity curves, β, γ, and; ρ
(b) perfect tone reproduction be achieved in our positives, and
(c) the blue, green, and red filters in our three projectors be such that they stimulate respectively only the β-receptors, only the γ-receptors, and only the ρ-receptors of the eye.
Unfortunately, this last condition is impossible to achieve. If reference is again made to the curves of Fig. 1, the reason will be obvious. Even if light of monochromatic purity is used in our three lanterns, it is only possible for the red lantern to meet the requirement. It is seen that light of wavelength R, 6500A (or any longer wavelength), stimulates neither the γ- nor the β-receptors, but only the ρ-receptors. Light of wavelength B, 4400A, is seen to stimulate chiefly the β-receptors, but unfortunately the ρ- and γ-receptors are slightly stimulated in addition. When a search is made for a wavelength of light which stimulates only the γ-receptors, none can be found which does so even approximately. The best that can be done is to choose the wavelength G, 5050A, which stimulates the γ-receptors strongly, and the ρ- and β-receptors to almost the same extent.
If, then, using the sensitivity curves of Fig. 1 for the three layers of the film, the three transparencies are projected in lanterns filtered to give the monochromatic radiations, R, G and B, how would the reproduction differ from the original? Any portion of the picture which was before reproduced by only red light, would still be correct, of course, for the red projector meets the requirement. But wherever the green projector was shining, there would be an unwanted excess of ρ- and β-signals, and wherever the blue projector was shining there would be an unwanted excess of ρ- and γ-signals. Thus, in general, the three signals in any part of the picture will now be more nearly equal to one another than before, and this means that the colours will be less saturated than they should be, and also that changes in hue and brightness will occur.
But triple projection with monochromatic light is not a commercially practicable method of colour reproduction, and, while no additive method of colour kinematography has yet met with much commercial success, there has recently been a renewed interest in the lenticular method of projection.2 If the strips of red, green, and blue filter fitted to the projection lens transmitted only very narrow bands of light, the desaturation of colours would not be much worse than with the monochromatic radiations. But one of the principle difficulties with the lenticular process is that the filters over the projection lens result in a considerable loss of light, and, in order to minimise this, filters transmitting broad bands of light have to be used. This means, of course, that the unwanted ρ and β signals given rise to by the green filter will be larger, as will also the unwanted ρ and γ signals given rise to by the blue filter.
Furthermore, even the red filter will give rise to some γ signal, in addition to the required ρ signal. So the result will be that the desaturation will be worse than in the case of monochromatic radiations.
At the present time the majority of modern processes of colour photography do not use the additive method at all, but the more convenient subtractive method. Fundamentally, however, both methods depend on the same principle.
* It is interesting to note that a practicable method of doing this has recently been developed (see Reference 8).
Most of the effects described in the paper were demonstrated during the lecture, either by actual experiments or by means of colour transparencies.
1 See, for example, W. D. Wright, “Researches on Normal and Defective Colour Vision,” Kimpton, London, 1946.
“Animators found it particularly suitable as a film stock capable of rendering deep, saturated colours, as can be seen in films such as Len Lye’s Rainbow Dance which featured a silhouetted figure (Rupert Doone) dancing his way through various scenarios to a piece played by Rico’s Creole Band. The film’s opening very much sets the tone whereby the credits are accompanied by a rainbow shape which cascades with colours. A storm (with rain represented by pulsating, coloured vertical streaks) is followed by a rainbow which sets the figure on his travels, as he dances his way around the world. The film is an advertisement for the Post Office Savings Bank, but along the way it shows a number of vibrant scenes in which the music synchronises with colour and movement. It is endlessly inventive and dazzling. The dancer’s silhouette progresses across frames and through colours, mostly moving but with the occasional moment of stasis. It is a wonderful example of how colour was used by such artists for abstraction in a playful spirit of extending the colour box beyond notions of restraint and narrative cinema.”
(Street, Sarah (2012): Colour Films in Britain. The Negotiation of Innovation 1900-55. Basingstoke, Hampshire: Palgrave Macmillan, on p. 46.)