[Different film stocks] can present different colour degradations depending on the chemical nature and individual properties of the dyes, the photochemical processing and the storage conditions. This is why the development of a general technique for the restoration of the image is not possible. Each film constitutes a particular case and requires the choice of the most adequate technical process of restoration. ((Preservation and Restoration of Moving Images and Sound, FIAF, 1986, 81.))
Having suggested methods of testing and offered an analysis of all the hand-painted films, I shall briefly outline three possible routes of duplication.
Duplication by Eastman Colour Internegative Film
This is the most common method of preserving a film. In many cases it is perfectly adequate and most economical. The availability of technical information on this method is excellent and film archive staff are familiar with the labour involved. The illustration overleaf shows how this work might be undertaken. ((Preservation and Restoration of Moving Images and Sound, FIAF, 1986, 84.))
For the production of a colour master positive from a colour negative, I would recommend using Kodak 2242/3242 ((The first 4 digit number refers to the 35mm stock and the latter to the 16mm version.)) polyester intermediate film. The alternative would be the acetate 5242/7242. Polyester stock is much more stable than acetate stock, is believed to last longer, does not shrink to the degree that acetate does, and is very resistant to micro-organisms. ((“Gelatine emulsion layers have the potential to last for centuries; together, polyester base, gelatine emulsion layers, and the correct choice of image substance will yield a pictorial recording medium that should survice five centuries or more at room temperature and moderate RH.” Reilly, 1993, 15.)) For the production of a colour master positive from a colour reversal film, the polyester intermediate stock would be 2272/3272. The resulting colour negatives are made using process ECN-2 and printed onto Kodak Vision colour print stock 2383/3383 resulting in a reasonable copy of the film in its current state.
Duplication by Black and White Separation Positives
By contrast, for all the films under discussion the production of black and white separation positives is a much better method of ensuring accurate duplication and excellent preservation elements. The illustration overleaf is helpful to understand this process. ((Preservation and Restoration of Moving Images and Sound, 1986, 72.))
Separations from the original hand-painted films will last far longer than the first method of duplication by colour internegative. With separations, when a new print is needed the three black and white films are printed onto a single Eastman (or Fuji) colour Intermediate film to make a new colour negative from which a new print can be made. ((Paraphrasing an email from Read, 6th July 2002.)) Currently, the preferred choice of film stock would be Kodak Panchromatic Separation Film 2238, a polyester black and white stock for making separation masters from both colour negatives and positives. The film can be developed as either positive or negative, depending on the original. At this point, the films would be very well preserved but a restoration would require another set of separations from the new negative or even better, straight from the originals again; this time, varying the levels of exposure and contrast to restore the original dye colours. With regards to the sound track, a new optical track would be made on polyester panchromatic Kodak 2374 stock (the modern version of stock which Tait used to paint on!).
Digital Restoration for Duplication to Film
A route for ensuring the most accurate restoration would be to scan the films to data. The basic route of digital restoration is as follows:
The film image → film scanner → digital image store → workstation → digital image store → film recorder → film ((Read & Meyer, 2000, 219.))
In this method, the film is scanned to data files when most of the colours are restored by the telecine operator. The restoration is then completed at a workstation with the operator precisely matching the data to colour samples obtained from the dye tests previously mentioned. The data may be digitally ‘retouched’ and any deterioration in the form of scratches and marks can be removed. The files are then re-recorded back to film on a colour intermediate stock from which prints are made. It may be noted that this method is much closer to the practice of restoring other forms of art such as a paintings. Typically, these are physically worked on by the restorer unlike traditional film restorations which are basically variations on duplicating.
There are several choices to be made once the digital route has been decided upon, as seen from the following diagram: ((Diagram from a presentation by Paul Read at The East Anglian Film Archive, July 2002.))
Restoration via this method allows the Archive the greatest range of flexibility for manipulating the films from their current condition in order to produce new prints which are judged to be as close as possible to the films Tait originally showed. It necessitates close communication with the workstation operator whose job it is to ‘restore’ the image as there is a greater amount of flexibility over photochemcial restoration and duplication. The digital route is recommended should the SFTVA carry out the dye test I discussed since this will provide exactly the kind of technical data which digital restoration favours. Whereas photochemical restoration by duplication is greatly governed by the limits of the narrow range of film stock available to carry the job out, practically anything is possible with digital restoration (the only ‘limitation’ in the process occurs when recording back to film stock which has given physical attributes) and thus in practice it requires confident technical decisions based on thorough research of the films. This might seem like an obvious requirement for any kind of duplication but because digital manipulation of the image introduces an additional ethical dimension which is still very much under close peer scrutiny, any archive undertaking this kind of work at the moment should expect to be fully accountable and transparent in their decision making. Not a bad thing by any means!
The main obstacle to digital restoration is funding. The cost of scanning and workstation time remains well out of the range of most small archives although it should be noted that the cost of a full restoration via black and white separations is often comparable if not more costly. Within the next year or so, it may be feasible to operate an affordable workstation within the archive although the all-important scanning and recording is likely to remain with the commercial post-production companies. The SFTVA were quoted the following prices in July 2002:
- €1.57/frame scanned
- €170/hr of workstation time
- €1.36/frame recorded to 35mm film
Special rates may be available for lengths longer than 10 minutes. At these prices, Calypso (404ft) would cost €10,148 to scan and €8791 to record back to film with additional costs for the workstation time. Clearly a very expensive process although it has been suggested to me that because of the uniqueness of this project, partial funding would be readily available.