Methods
From a film studies perspective, all methodological questions concerning video analysis are preceded by the challenge to engage descriptively with a fleeting subject-matter that solely exists within the time of its perception. Accordingly the claim to precisely describe the filmic image in its temporal unfolding and to make these descriptions explicable and usable for other analysts marks a central methodological and theoretical challenge for film studies that has been met with different forms of annotation and visualization not just since Sergei Eisensteins film-scores in the 1940s.
Visualization of audiovisual dynamics of four shots in accord with С.М. Эйзенштейн vertical montage (CC0).
Our research group wants to contribute to these basic questions and challenges for film analysis – specifically in respect to large comparative corpus studies which not only explore individual films but aim for wider contexts. In close cooperation with the computational sciences and against the background of new potentials for (semi-)automatic video analysis as well as software supported video annotation we work on an analytical view on the project’s large corpus which allows for a fluent alternation of close and distant reading: Following first breakthroughs in the digital humanities on the macro level of a digitally based distant reading of data and visualizations, new software solutions for the (semi-)automatic detection – e.g. of cuts, objects within the image or camera movements – raise hope to explore new ways for (scientifically) describing the aesthetic and poetological micro level of audiovisual images.
By exploring aesthetic patterns on the micro level with digitally based methods and empirical data our aim is to develop a typology of figures of audiovisual rhetorics of affect. Such a typology would be a milestone for the exploration of media communication in audiovisual images. What could be gained by this research on the one hand are comparative perspectives on the nexus of individual topics and topoi with various rhetorical figures of affect, which means emotional attitudes solely constituted in the media aesthetical composition. On the other hand methodological and theoretical expertise from film studies on the analysis of affective qualities of audiovisual images could hereby be contributed to the research on audiovisual images within media and communication studies, social or historical sciences, namely in the form of identifiable, abstract types.
The methodological development of a typology of audiovisual crisis coverage and communication includes several aspects:
The methodological work of the project follows film scientific reflections on the empirical reconstruction of audiovisual images and draws on the eMAEX-method which seeks to systematically describe processes of how viewers are affected by rhythm and movement-profiles. In this perspective basic structures of affecting viewers can be grasped as aesthetic patterns in the composition of audiovisual movements. Based on these patterns our typology of rhetorical figures of affect should be developed.
The detailed analysis of large corpora leads inevitably to the accumulation of large sets of data. Due to their scope their evaluation can hardly be managed without machinable support. Therefore we will develop a machine-readable analysis systematization (ontology) based on our methodological framework as outlined above. This ontology is essential for intertwining manual and automatic analysis as well as for data evaluation, and can facilitate perspectively the exchange of annotations between different research projects.
The manual and semi-automatic annotation is carried out with the open-source video annotation software Advene which will be adjusted and extended within the project to meet the specific requirements of detailed film analysis which will be directly interlocked with the development of our ontology.