Film Analysis as Annotation: Exploring Current Tools

Film Analysis as Annotation:Exploring Current Tools Liliana Melgar Estrada (bio), Eva Hielscher (bio), Marijn Koolen (bio), Christian Gosvig Olesen (bio), Julia Noordegraaf (bio), and Jaap Blom (bio) [End Page 40] With the VCR's increasingly widespread use in the 1980s and 1990s, film scholars and historians acquired new techniques for doing film analysis. Whereas previously, analysis performed on celluloid films had been restricted to on-site research in film archives,1 the VCR technology's playback mode allowed scholars to fast-forward, rewind, pause, and segment films in settings outside of the archive. Also, since the 1990s, the rise of the internet and the appearance of a great variety of computational tools have spawned scholarly practices for segmenting and annotating films. Digitization also has allowed linking digitized films to contextual film-related materials in CD-ROM and DVD editions as well as presenting them on curated websites. In the past twenty years, the possibilities of hypermedia and enhanced CD-ROM [End Page 41] editions of films have resulted in groundbreaking projects such as Lauren Rabinovitz and Greg Easley's Rebecca Project (1995) and Yuri Tsivian's CD-ROM Immaterial Bodies (2000)—also called a Cine Disc—released through Marsha Kinder's Labyrinth project,2 and, more recently, DVD editions like Digital Formalism (2007–10), based on the digitized collection of Soviet filmmaker Dziga Vertov's films at the Austrian Filmmuseum, which originated from experiments with automated semantic content analysis of big data. The tool kits were developed by media theorist Lev Manovich, with software for multimedia annotation such as ANVIL.3 Although film analysis and audiovisual publication formats that relied on video annotation and hyperlinking in past decades may have seemed somewhat exotic and offered a limited range of choices,4 now, though, annotation and editing are becoming increasingly accepted in scholarly research projects allowing scholars to annotate larger corpora of films and to produce increasingly personal interpretations. There is a wide range of software to choose from, not only for presentations or publication but also as research tools in their own right. From being what film scholar Raymond Bellour in the 1970s famously bemoaned an "unattainable text," film—in its electronic avatar—became attainable.5 This article describes a study of the information systems (so-called tools) that have made this possible. We begin with the assumption that it is important that scholars discover and use these tools to understand and critically discuss how the systems influence the process of analysis and its results. This closer approach will, as Olesen claims, "elucidate the epistemological underpinnings of digital film history's methods."6 Because not much is presently known about the implications of using these computational tools for film scholarship, this article aims to contribute to the initial effort to understand these implications. We focus on one side of the problem, which is the lack of an overview of existing tools and a description of their conceptual models and affordances7 (i.e., easy replaying, segmenting, annotating, or collaborating) based on empirical experiments. Reporting about the connection between the characteristics and functionalities of the tools with research goals will help in comprehending the implications of their use.8 This work is the result of a collaborative effort between media scholars and information scientists. We are motivated by the need to codevelop systems and information services for audiovisual media-centered research within the Dutch National Digital Humanities infrastructure.9 This article evaluates two types of video annotation tools that can be used during the analysis phase of the research process, to demonstrate their affordances and pitfalls. The following section presents an inventory of the different types of existing [End Page 42] tools. The next section describes a case study wherein we tested these tools in a small film analysis project. Specifically, we evaluated the tools ELAN and NVivo for analyzing an excerpt of a 1930 film. The third section synthesizes the main methodological implications of the use of these systems in media scholarship. In the conclusion of the article, we evaluate the implications of using these tools for film analysis. DIGITAL TOOLS FOR MOVING IMAGE ANNOTATION The digital humanities community uses the word tool to refer to any software component...