One of the more astounding moments in the DIASTOR project (https://diastor.ch/, 2013-2015) happened when we were on a work retreat in a rustic sport hotel in the Swiss mountains, and after two long days of conferring, Barbara Flueckiger announced that in the next project she wanted to, drumroll…, build a scanner. The slightly flabbergasted research team members looked at each other and thought ‘build a scanner…?!!’
Barbara Flueckiger has many talents. One is to think BIG. Another is to put together wonderful and complementary research teams. With the holy trinity consisting of Giorgio Trumpy, David Pfluger and Lutz Garmsen in the ERC Proof of Concept VeCoScan, she has outdone herself. What they presented last week in Zurich, at the Expert Days at DIASTOR’s project partner Cinegrell’s facilities, five years after that initial announcement, is nothing short of incredible.
A little bit of context.
Firstly, as the comparative scanner study that was started in DIASTOR (Flueckiger, et al. 2018, as described in detail here) showed, existing commercial film scanners can be ‘colour-blind’, particularly when it comes to historical applied film colours. The reasons are manifold, but are a combination of the films containing numerous silver particles that scatter light in a particular way, and the more or less vibrant tints and tones that have been added to these surfaces, which often fall outside of the three RGB spectral bands that most existing commercial scanners use to scan the material. In practice, this has led, for instance, to post-production houses developing a process called ‘Digital Desmet’ (as described here); digitally adding a layer of colour to the scanned film image in the desired tint and/or tone.
Secondly, all existing commercial scanners use diffuse light sources, which as a business decision can be seen as logical, as those light sources can obscure such deficiencies as surface scratches, and is therefore a ‘blanket’ solution for most modern film material. However, in the case of some historical film material, previous research (Trumpy and Flueckiger 2019) has shown that a focus on a collimated light source –which replicates the light source in projection –might be a very useful addition.
And lastly, there are many choices when it comes to choosing the most appropriate source for colour reference in the film restoration process. Although not without its own limitations, analogue projection is considered to be one of the best sources for colour reference, but in practice hardly ever used. Where it is used, it is often from human memory, and not in direct comparison with projection of the digitised image.
The last few years must have been an incredible journey. What the aforementioned trinity has come up with is not only a bespoke versatile, multispectral scanner, but also a proposal for rethinking what a scanner—and scanning—is. What they have produced is not a ready-to-purchase ‘box’ as such, but a particular set-up that consists of amongst others multispectral illumination, a capturing device, film transportation controls, and processing software; a set-up that can be modified by the respective institution using it.
A side-by-side hybrid projection of the analogue source material next to the digitised images resulting from this multispectral set-up, created some gasping in last week’s Expert Days attendees. The projection showed that scanning an image 50 times each (25 times at set points through the visible spectrum with a diffuse light source, combined and post-processed with 25 images through that same spectrum with a condensed light source) creates a result that is as close as possible to the colour reference in analogue projection to date, without the intervention of a colourist. Of course, not all film material will require this amount of attention; only the most difficult scenarios. Many of the tinted and toned examples, which have been used in the project, and which follow historical recipes, have been created by Ulrich Ruedel, who really turns the trinity into a quartet.
The ERC Advanced Grant FilmColors 2015-2021 (incorporating the Proof of Concept VeCoScan 2018-2020) has not only irrefutably exposed many shortcomings of existing options in film scanning, but it has mainly dared to think BIG; created a scalable and affordable real-world application; and has shown the sum-is-greater-than-the-parts result of truly interdisciplinary research between vendors, film scholars, and scientists.
None of the workshop attendees have a crystal ball, but if the presented results promise anything, it is that they have built it (!!), and come they will…
References
Flueckiger, Barbara; Kamp, Claudy Op den; Heller, Franziska; Pfluger, David (2016): „Digital Desmet“. Translating Early Applied Colors. In: The Moving Image, 16,1, pp. 106–124, (= https://muse.jhu.edu/article/640570).
Flueckiger, Barbara; Pfluger, David; Trumpy, Giorgio; Croci, Simone; Aydın, Tunç; Smolic, Aljoscha (2018): Investigation of Film Material–Scanner Interaction. 88 pp, (= https://diastor.ch/results/#SCAN).
Trumpy, Giorgio; Flueckiger, Barbara (2019): Chromatic Callier Effect and its Repercussions on the Digitization of Early Film Colors. In: Journal of Imaging Science and Technology, 63,1,pp. 10506-1-10506-11 (= https://doi.org/10.2352/J.ImagingSci.Technol.2019.63.1.010506).
Dr. Claudy Op den Kamp is Programme Leader for the BA (Hons) Film, Senior Lecturer in Film, and faculty member at the Centre for Intellectual Property Policy and Management at Bournemouth University. She holds a PhD from Plymouth University on the relationship between copyright ownership, access to archival film, and film historiography. She is a graduate of the University of Amsterdam (Film & Television Studies) and the University of East Anglia (Film Archiving). She has worked as Haghefilm Conservation’s Account Manager; as a Film Restoration Project Leader at the Nederlands Filmmuseum, and as a senior research assistant in the DIASTOR project at the Department of Film Studies at the University of Zurich.