The NFSA’s collection includes early coloured films, many of which are tinted. As these films deteriorate with age, their colours are lost.
Steve Clark, Trevor Carter and Bruce Cowell from the NFSA’s Motion Picture Laboratory share their research about using traditional dyes and techniques to restore tinted films.
Films from the Corrick Collection: experiments with red and blue
Following our successful machine tinting experiments, the lab was asked to try to tint copies of Corrick films being sent to the 2011 Pordenone silent film festival. Of the films being copied, one had blue titles and the rest had red titles. The project was not meant to provide an exact match of the tints in the original films but instead to give the feeling, the experience, of watching films tinted in the manner of the early 1900s. While we were able to approximate the tint used in the blue title, the films with red titles had been produced by a number of different companies, each using their own dyes. They had differing shades of red and some of these titles had faded to an extent that made it impossible to tell with certainty what the original colour had been. Because of this, it was impractical to try and re-create every shade of red. We decided to use original dyes and techniques to create an in-house tint, later known as NFSA Red 1, and use this colour for all of the red-tinted films being sent to the festival.
Choosing and testing the dyes
From our previous experiments it was clear that aniline dyes vary chemically. A Kodak film booklet from 1922 broke the dyes down into two kinds, acid and basic: ‘acid dyes being alkali salts of organic acids, while basic dyes are the chlorides, sulphates, etc, of organic bases’.20 We knew that some original dyes were highly corrosive, which could lead to Brittleness in the film. Even in the 1920s it was difficult to tell a dye’s type and properties (exactly) without conducting tests. Of the aniline dyes still commercially available we found two blues and three reds that appeared promising:
- Basic Blue 5GL
- Sumiacryl Blue N3 GL
- Basic Red GRL
- Panacryl Red GTLN
- Sumiacryl Red NGRL
As with the Championship Boxing Contest project, our plan was to hand test the dyes and then modify our spare film processor to enable us to machine produce tinted tiles and intertitles. Hand testing was initially done in a glass beaker using a magnetic stirring system to keep the solution agitated. However, there were problems with uneven tinting, particularly around the sprocket holes, caused by irregular agitation. We overcame the issue by adopting the techniques used in the development of photographic films. Changing to a small stainless steel film developing tank, we gave the tank three inversions, at thirty second intervals, with no agitation in between. This produced an even result.
When we scaled up to machine processing, the time spent dyeing and washing the film (and the temperature of the dye solution) would be determined by limits of the processor. Because of this, we controlled the degree of uptake (absorption) of the dye into the film emulsion by varying the concentration of the dye. Each colour was tested in turn by dissolving a half a gram of powdered dye in one litre of water heated to 35°C. The solution was cooled to 21°C and a clear piece of processed release print film was immersed for eight minutes with agitation. It was then washed for ten minutes in running water. The test strips were compared and we chose the dyes to copy the Corrick titles based on dye stability and staff knowledge of the types of colours historically used in tinted films.
Two colours blue
Only one of the films being copied had blue tinted titles and the processor has a large tank. To avoid wasting dye, this film was dyed by hand. Initial tests showed that between Sumiacryl Blue N3 GL and Basic Blue 5GL, the Basic Blue dye gave a better colour for our purposes. Our trial concentration (0.5 grams of Basic Blue 5GL per one litre of solution) and temperature (21°C) produced a colour without enough Saturation. We increased the concentration of the dye in steps, until we reached 10 grams per litre, which gave a good depth of colour. Further testing showed that we could increase the film’s uptake of the dye by raising the temperature of the dye to 25°C. However, we were still having issues with uneven tinting and the dye washed out too quickly (known as dye ‘bleed’).
Dye bleed diminishes the overall colour of the film and can produce droplets of water that have more or less dye in them. When dry, these irregular concentrations of dye set, showing up as spots or marks on the film when it is projected. We learnt from a 1922 Kodak manual that changing the pH of the dye solution could increase ‘the rate of dyeing’.21 We tried altering the pH and found that adding an alkali improved the film’s uptake of the dye. Adding sodium hydroxide or sodium carbonate created dye solutions that had extremely rapid uptakes. However, these solutions were difficult to control and the results were not repeatable (no two results were the same). After more testing, the pH was controlled by adding a weaker alkali — borax — added at a rate of 2.5 grams per litre of dye. Our final formula for the blue tint was:
10 gm/l Basic Blue 5GL + 2.5gm/l borax: hand dyed for 10 minutes at 25°C and washed for 10 to 12 minutes.