WP5: Media studies
Media Scientists generally work with existing datasets: audiovisual sources (film, radio and television programs, online audio and video), published sources (newspapers, broadcasting directories, social media content) and structured data (such as amount of TV-watchers, or a database like Cinema Context with data derived from newspapers and archival materials, about movies, cinemas, screenings and active individuals and companies from the film industry).
Due to the specific nature of audiovisual resources (combination of picture and sound, time-bound), additional measures are needed to make them searchable and accessible on a larger scale. This is done by linking them to descriptions containing the content and some contextual data (metadata, subtitle files) and with the help of new technology such as automatic speech recognition to generate transcripts and image recognition to search for visual patterns.
Recent user studies (Bron et al 2013; Bron et al. 2015) have shown that media scholars want to examine the various sources together. An good understanding of audiovisual resources requires detailed knowledge of the context in which they are produced, distributed and observed. Moreover, researchers who work with audiovisual resources are often interested in the role of media in the construction distributions reception of a particular social or cultural phenomenon, which requires different perspectives on the same phenomenon can be tracked over time and analyzed in conjunction. In addition, the research process of media scholars is characterized by an exploratory phase of the contents of the material (which is developed in the question and/or tightened) followed by a more focused collection and analysis of particular contextual data (contextualiseringsfase) - a cycle that often or more times repeated and concludes with the presentation of the results found (presentation phase) (Bron et al.2015).
The biggest challenge for scientists, media access to the source material, which is often copyrighted and is scattered across different repositories. Also, there is a need for tools to investigate the various sources together.
|Werkpakket 5 (NL)|