Report

Teaser, summary, work performed and final results

Periodic Reporting for period 2 - Filmcolors (Film Colors. An Interdisciplinary Approach.)

Teaser

Why is film color research so important at this moment? As a result of the digital revolution, the last decade has seen the greatest media transition since the introduction of photography and film in the 19th century. With digital encoding of images, colors have become a...

Summary

Why is film color research so important at this moment? As a result of the digital revolution, the last decade has seen the greatest media transition since the introduction of photography and film in the 19th century. With digital encoding of images, colors have become a completely different topic from a technical perspective. Yet the film archive vaults are full of historical films even as we rapidly lose the knowledge of traditional techniques and aesthetics in film. We need to investigate and develop methods to digitize and restore historical color films to keep film heritage alive and preserve it for future generations. While every serious art restoration connects scientific analyses of pigments, binding material, and a painting’s foundation with art-historical and aesthetic investigations, a similar approach is only rarely applied to film. The groundbreaking nature of this project lies in its truly interdisciplinary approach. Progress in the research on film colors requires that all relevant aspects of film colors be connected in an interdisciplinary fashion, because technology and aesthetics are closely linked.

A novel methodology: With her professional background in engineering and film production Barbara Flueckiger (PI) has developed a technobole approach to investigate the interaction between technology and aesthetics in her previous projects. While the methodology makes use of a wider set of approaches, the focus remains firmly rooted in the humanities. Beyond its interdisciplinary approach, the project aims at the implementation of research tools in the emerging methodology of digital humanities. With her digital humanities project Timeline of Historical Film Colors (http://zauberklang.ch/filmcolors/), the PI has established a comprehensive resource for the investigation of film color technology and aesthetics. This internationally renowned platform contains hundreds of primary and secondary sources, patents, links, and downloads in combination with thousands of images.

The Timeline of Historical Film Colors http://zauberklang.ch/filmcolors/ forms an integral part of the ERC project and an important contribution to a novel citizen science approach.

Work performed

Thanks to a very diverse team consisting of film scholars, historians, a technician and restorer, a physicist with a special degree in material analysis of cultural heritage, the project’s aim to connect the various, hitherto separated domains in film colors research can be fully implemented. The following steps were executed or are in progress:

Seminars with students and phd-candidates from the University of Zurich: In collaboration with PostDoc Researcher Dr. Bregt Lameris, the PI organized three comprehensive and intensive seminars and about 20 workshops with external experts from universities and archives in Europe and the United States.

Analysis of a group of at least 400 films: For the computer assisted analysis the team devised a workflow consisting of a video annotation system with the software ELAN and a custom-built offline database in FileMaker (analysis database). The analysis database is connected to an online database for the complete group of films (corpus database) that allows to define relations and to import data directly from the corpus database. Each film is segmented in sequences with consistent color schemes and then analyzed according to the protocol defined in the analysis database. In addition, there is a keyword database and a glossary database, the latter elaborated by the students to support consistency across the analyses. Six analysis work stations with calibrated 4K monitors are installed at the University of Zurich to allow six persons to work simultaneously on the analyses (fig. 1). To date, 376 films of the planned corpus of 400 films have been analyzed and evaluated.

Development and application of a computer-assisted tool with a web interface for crowdsourcing of film color analyses: Based on the computer-assisted method for the analysis of films, the PI initiated a collaboration with the Visualization and MultiMedia Lab at the University of Zurich to support the development of the tools for the (semi-)automatic analysis of film colors. At the core is a platform called VIAN that enables semi-automatic segmentation, visual and verbal video annotation, a screenshot manager, the implementation of all the databases and data developed in the project, and the automatic tools for the analysis of the film colors’ aesthetics (fig. 2).

Study of film color technology in combination with chemico-physical analyses of historical color films to understand their influence on films’ aesthetic appearance: Dr. Trumpy has established a multispectral imaging workflow that enables the measurement of spectral absorbance distribution of dyes and color compounds in historical color films, regardless of their composition (fig. 3).

Digitization and restoration of historical films: Complete digitization and restoration equipment and a workflow have been installed. These consist of a professional scanner specifically developed for sensitive and brittle archival film, plus a color grading workstation and one for image stabilization, optical restoration and retouching. More than 30 restoration case studies have been undertaken in the course of the project, 14 of which are completed, while the other ones are underway in different stages (see http://www.research-projects.uzh.ch/p21207.htm) (fig. 4 + 5).

Exemplary studies of three periods in the development of film colors: The emergence of film colors from early applied colors to so-called “natural” colors 1896–1930; development of standards in film color technology and aesthetics 1930–1955; the dominance of chromogenic processes 1955–1995. Three PhD students have been working on the three periods and applied the method presented in section Aim 1. In March 2018, all of them presented their results at the third international Color in Film conference in London (see: http://www.research-projects.uzh.ch/p21207.htm).

Final results

Development and application of a computer-assisted tool with a web interface for crowdsourcing of film color analyses through recent advancements in digital humanities (advanced deep-learning and computer vision, image and video processing with concepts from aesthetics, theories of representation and perception, color systems from art and science).

Successful submission for the Proof-of-Concept VeCoScan – Development of a Multispectral, Versatile Film Scanner. The PoC develops a new core film scanner unit capable of changing both the spectral distribution – favoring color separation or color accuracy – and the directional arrangement – diffuse or directed – of its illumination system, adapting these features to the specific type of film to be scanned.

Development of an exhibition at the renowned Fotomuseum Winterthur that will be shown in November 2019 and curatorship for a film program at the Internationale Kurzfilmtage Winterthur in November 2019.

Website & more info

More info: http://www.research-projects.uzh.ch/p21207.htm.