, , , , , ,

Information – and especially news – are increasingly broadcasted by modern media such as television and internet videos. However, audiovisual data has not yet received as much attention as written sources in the field of Digital Humanities, although they may be the most profuse ones in the coming decades. In the current situation, it is clearly necessary to draw attention to the opportunities offered by audiovisual content archives.

For someone who would like to work with video or sound data, the main issue is how to obtain valid search results. Some authors point out 3 main reasons there is so little interest in video sources for the moment [1]: the young age of audiovisual media, the issues with copyrighted sources and mainly, the relatively poor existence of audiovisual material analysis methods. There is indeed no automatic accurate annotation of metadata and, knowing the amount of audiovisual content produced every day, this becomes a very limiting factor. To better employ their content, we need to improve the way we mine, catalogue and archive videos. Several progress can also be made on speech recognition techniques and computer vision for example and in developing new methods of interpretation and analysis.

To fill this technology gap, some teams have developed their own idea on how to improve data cataloguing in video content.

A group from UCLA (University of California, Los Angeles) created a software application, ClipNotes [2], that allows tagging video files with metadata and uses XML files to store it. The advantage of XML format files is that they permit a precise analysis with easily readable results and are convenient to share, so that this resource can be accessible to everyone. ClipNotes can be used for example to catalogue all the camerawork or to follow the apparitions of a selected object or pattern throughout the film. The possibility of tagging every scene, every detail of a video is interesting for comparisons between movies and/or scenes, and it also could be useful to analyse and study shooting or editing techniques. This kind of tool offers a large variety of possibilities to catalogue and archive audiovisual content.

Another team, from UCSD (University of California, San Diego), focused on video news [3]. In fact, there is all kind of information that can be collected from these videos. Aspects such as the speaker’s gestural (its facial expression or its bodily gesture for instance), the settings (lighting, background objects, etc.) or even the speech itself can be analysed. This project, called “Visualizing Global News”, concentrates on political speech videos and studies how they are reused and remixed by other channels on their own news broadcast and how these archives could be used for the purpose of historical study. They used the collection “Weekly Addresses” from the Obama Administration which talked over current events, and worked also on videos from the US public diplomacy campaign. An example the team pointed out is the evolution of US-Iranian relations through 3 annual addresses performed by Barack Obama: “Each address is fundamentally different in tone and diction, as the U.S.-Iranian diplomatic relationship seems to deteriorate, as well as in shot composition and White House location.” They also tried to spot the usage of Obama’s official publications on foreign broadcasts and compare the original with the remixed one. This work may be of interest for historians and political scientists since a good cataloguing and retrieving process could facilitate the comparison of “historical world events from different national or political perspectives” [3].

These are only two of the countless examples in which techniques of video mining and archiving could be leveraged, as audiovisual data material possesses many advantages. They have the interesting property of being “multi-layered” since you can have information “regarding language, emotions, speech acts, narrative plots and references to people, places and events. This richness provides interesting data for various disciplines and holds the promise of multidisciplinary collaboration.” [1] And they also could give the opportunity to have archives accessible from almost everywhere in the world. These are some of the reasons we should improve our techniques and knowledge in video archiving process.



[1] Roeland Ordelman, Max Kemman, Martijn Kleppe, Franciska de Jong ; Sound and (moving) images in focus – How to integrate audiovisual material in Digital Humanities research

[2] Andrew de Waard ; ClipNotes: Digital Annotation and Data­Mining for Film & Television Analysis

[3] Elizabeth Losh, Lev Manovich ; Visualizing Global News