RESTORATION OF TWO-COLOUR SUBTRACTIVE PROCESSES
NOTES ON SUBTRACTIVE TWO-COLOUR PROCESSES
Two colour subtractive print systems were in use from about 1915 to about 1949, and companies in many countries produced and patented their own versions.
Two separation negatives are produced in a single camera exposing a "bipack", two films emulsion to emulsion in a single gate, or in a camera with a beam splitter and two film elements. The separations are made with a red-orange and a cyan-green complimentary pair of filters, or relied on the different sensitivity of the films [in the case of the bipack] to produce a similar result. Thus complimentary records were generated covering the entire visible spectrum.
The print was a single film base with red-orange and cyan-green positive images each made from a separation negative. The images were on either side of the film base or on the same side [e.g. Dascolour, early Trucolor]. The dyes were produced from the silver images by dye transfer, by chromagenic development, but most frequently by toning, using metallic and mordant dye techniques developed for toned silent films.
The processes were highly successful until 3-colour Technicolor in 1936, and even then hung on until Gevaert and Eastman colour negative films in 1947-50.
From about 1930 Eastman, Agfa, Gevaert, DuPont [who called their bipack "Rainbow Negative" and later DuPack] all made "bipack" films for camera use. One negative geometry was reversed as the films were usually exposed emulsion to emulsion, but almost any camera could be simply modified to supply and take up to two rolls of film.
These same manufacturers also made double-coated print films, usually using their standard black and white emulsion coated on both sides of the nitrate support. These films were called Duplex or Duplicoat [DuPont] or Duplitized [Eastman] etc.
The sound tracks, where present, are usually variable density and a bright Prussian blue, more rarely silver coated with a shiny cellulose nitrate lacquer.
2-colour print technology failed to make the jump to 3-colour although several complicated three colour versions were tried. The last British 2-colour was Cinecolor [Slough, Bucks] and the last 3-colour of this type was SuperCinecolor [processed as a joint venture by Cinecolor, Gevaert and Studio Film Laboratories in 1950!].
Early Technicolor prints were 2-colour, either by cementing two films together or by two dye transfer images onto "blank" film stock.
Technichrome included a complex method of deriving a 3-colour print from two separation films.
Only Harriscolor [USA 1929+], Kelleycolor [USA], Polychrome and Kodachrome [USA, 1920], Fox Natural Color [1924] used beam splitter cameras and these are very rare indeed. All others used commercial bipack films pairs that were not unique to the print process.
[The terms bi-pack and monopack have been confused for many years, since they are used as loose terms for a number of different still and motion picture systems. Technicolor called Kodachrome "Monopack" when they used it as a camera film in 1945, and Ilford produced a "bi-pack", producing three separation negatives from two films, for still use, which they called Monopack!]
Identifying the colour system is essential to produce an authentic restoration as the dyes varied widely. The two colours were intended to be complementary but this can cover a wide range of colour combinations.
Blues are poor, probably as much due to the dyes darkening with time, reds are strong and saturated, but some results are extremely pleasant. Sharpness was usually poor on double sided print films.
Identifying the system can ONLY be done from a print and is partly by identifying geometry and layer arrangement and by identifying the dyes and sound tracks.
Prussian Blue is a common cyan blue used in Dascolour, Harriscolor, early
Kelleycolor ["Prizma"], Colorfilm, Fullcolor, early Cinecolor, Kodachrome 1920ish, Multicolor, Magnacolor? The orange dye was occasionally Uranium Ferro cyanide but more usually a mordant dye mixture. Some processes used chromagenic development.
Measure the density of any pure dye areas on Status A.
Identifying the film stock manufacturer may help or may not. Most US film stock manufacturers made a double-coated stock for these processes.
Some 2-colour, perhaps all, was printed on a yellow film base or the emulsion was tinted to produce a subjectively better result.
The best general source of references and other data is "Colour Cinematography", A Cornell Clyne, and "A History of Motion Picture Colour Technology", R Ryan and most US systems were well described in SMPTE Journals.
Making A Restored Print From Separation Negatives
If only separations exist it is rarely possible to tell from them, which print system would have been used, and different print systems could have been used with the same pair of negatives at different times.
If both separations and a print exist a close match to the original print can be made from the separations as follows:
1 Identify which separation is which. As with any set of sep negs.
2 Prepare test exposures on Eastman Colour Print Film 5386 to achieve a densitometric [and visual] match with the primary colours using short lengths of the separation negatives. Usually areas of pure dye can be seen in splashes, outside the perforations, or at reel ends, which can be used as colour patches. Either additive or subtractive lamp houses can be used.
3 Make a register print on Eastman Colour Print Film 5386 by two separate passes, one from the Blue/Green separation the second from the Orange/Red, producing the B/G primary from the O/R separation and so on.
4 Grade the final print by varying the total exposure of the two separations [not by varying the colours produced as primaries] Neutral density filters are the best method, but scene by scene grading cannot easily be done this way, and some printer light adjustments may be needed.
I have only carried this procedure out once, as separations are rare [or overlooked, the occasion on which we used this method was when a client said he had lost a separation negative! However the can was labelled with a Cinecolor label, film was Eastman bipack and one negative geometry was reversed].
Making A Restored Print From An Original Print
Most 2-colour materials only exist, at least in Europe, as a release print.
1 Carry out the identification process as above, and identify the dyes if possible. Measure the Status A densities of the dyes.
2 Make two separation negatives from the print. Filters we have used for this are Wratten 16 [Orange] and Wratten 44 [Cyan], and also the narrower cut pair Wratten 22 and Wratten 45. There are increases in saturation in the final print with the latter, but they reduce exposure considerably. Time-gamma curves are needed to produce matching gammas of 0.6 for both separations.
[We have also tried using an additive lamp house to achieve the separation exposure wavelengths. This was successful but only used for a test.
3 The print is made as for the restoration above from original separation negatives, which as far as is known, were all made to an approximate [natural] gamma of about 0.6. We did find some earlier films with much higher contrasts and have needed separation gammas as low as 0.4, especially where optical printing was used from black and white elements.
Choosing The Colours Of The Two Primaries For Prints
The following dyes are known to have been used for 2-colour prints
Cyan/Blue primaries - Prussian blue/Iron-tone blue/Iron blue tone/Ferric Ferro cyanide. This is the classic Prussian blue formed by the action of ferric chloride and oxalic acid on metallic silver. All the processes we have seen described suggest the silver was "fixed out", and sometimes "bleached out" with the far more effective Ferri cyanide bleach rather than fixer. Prussian blue is more stable than many chromagenic cyan dyes, but darkens and desaturates in time. All 2-colour films using this primary should be restored with a colour that matches "fixed out Prussian blue".
In Dascolour potassium ferricyanide bleach removed all silver. Probably a very saturated blue, but have not seen this. A single sided emulsion process, probably one separation colour originally printed through support.
In Harriscolor, "Prizma" Kelleycolor, Colorfilm, Fullcolor, Cinecolor [early version], Multicolor various formulations resulted in Prussian blue, with various levels of "fixing out", probably with various levels of effectiveness. Kelleycolor went through many generations using a wide range of dyes including a vanadium salt, although this may have been part of an elaborate mordant dye tone process.
Cyan/Blue primaries - Mordant cyan/blue dyes
Few patents list the precise dyes used. The only cyan/blue mordant dyes certain are for Polychromide, a mixture of Malachite Green and Helio Saffranine. The later Cinecolor used a mordanted cyan/blue [unknown?].
[Cinecolor, the longest running process, 1931? to 1950, and the most widely used, especially for B Westerns, went through many technical changes].
Orange/Red primaries
Harriscolor and Multicolor used Uranium toner solution and the resulting metallic dye included or consisted of Uranium ferricyanide.
With the exception of the dye transfer dyes and chromagenic dyes listed below all others were mordanted dyes. The mordant was usually silver iodide.
The only documented dyes we have found so far were for Cinecolor - a mixture of Auramine O and Fuchsin.
Chromagenic dyes
Trucolor used chromagenic dyes based on "CD1", di-ethyl p-phenylene diamine [for both primaries].
Dye transfer dyes
These were used for Kodachrome, 1918, and Technichrome, but we have no data on either dye or colour.
Most 2-colour prints had Prussian blue sound tracks exposed from a separate sound negative but onto the cyan/blue image emulsion and developed with it. In some cases [e.g. some Cinecolor] the silver seems to be retained by lacquering before the "fix out /bleach out" stage.
Some [e.g. Trucolor] have pure silver sound tracks but no data is known as to how this was achieved, probably by some form of redevelopment.
Restoration of the sound is straightforward. It can be rerecorded and a new track produced which is printed at the same time as one of the printing passes.
The yellow or pale orange tint present in some or most 2-colour prints can be simulated by a flash print pass [with no image in the gate] solely to "tint" the image area. This is identical to the tint pass in the Desmetcolor method.
The origins of this tint can be related to the Ive’s Polychrome process of about 1918. Ive’s 2-colour still process used cyan/blue and orange/red and then, since the result produced a poor, rather cold, neutral scale, the whole print was tinted yellow. The grey scale became neutral and the skin tones warmer, but still acceptable. Ive’s called it a "two-and-a-half colour process". We have not found any other references to it than this although nearly all 2-colour prints are tinted amber or yellow. Also many early-stencilled prints are also tinted overall yellow, which increased the subjective effect.
Preservation Of Two Colour Systems
Preservation of the separation images can be undertaken by making positive protection masters in exactly the same manner as that used for preparing Technicolor protection masters. Whenever a negative is required for a restored print a new silver separation negative can be made. Contrasts can be treated in the same way as Technicolor as most original separations were about gamma 0.6.
Where only the print exists the new separation negatives made from the print will become the primary preservation material.