ExplATeD: Exploring ASR-enriched Television Debates

The coverage of the drug ecstasy in recent current affairs programs on Dutch television is explored to find out how to analyze a visual data archive using text-based search and visualization strategies based on Automatic Speech Recognition metatada.

Principal investigator
  • Berrie Jens van der Molen
Resource types
  • Tools
  • Data
  • Workflows

About the project

Recent Media Suite research about ecstasy in historical media debates has shown that ASR enrichment can be used to explore radio debates in a way that is comparable to how newspaper debates can be researched using Optical Character Recognition metadata, offering great potential for cross-media public debate research: topics can be researched in their socio-historical context across print news and radio news. This means that (historical) public debate analysis has moved beyond the stage of searchable print media, which was heralded as a ‘digital turn’ for digital history. Now, digitized television data can also be explored, if enriched with ASR. This opportunity will be used to explore the recent reputation of ecstasy in Dutch radio and television coverage. This is an urgent topic, as recently the question whether the illegal yet widely used party drug should be regulated instead of banned is recurring in political, public and scientific contexts.

The recently ASR-enriched current affairs items of the Sound & Vision Television Collection will be explored with the leveled approach in ExplATeD to address the issues expected to arise in this context: ‘ASR terms have no necessary relation to what is shown in the media item, meaning that they are even further removed from television items than ARS terms are from radio items’ (Van der Molen, 2022: 199). This is particularly salient on the distant reading level: here, all visualizations of the data, and by consequence the research results, are based on the ASR terms. If distant reading will be similarly effective for visual data such as television data, digital history might be closer yet to a “cross-media turn” (ibid: 202). To get there, the question of how visual media can be effectively analyzed with search and analysis technologies based on the spoken word needs to be addressed. This known issue has received attention, for instance with ‘temporal visualizations of the content metadata’ (Huurdeman et al., 2019). What other Media Suite functionalities on the distant and/or close reading level could even the analysis level between ASR-enriched radio and television news?

Relevance of this project extends beyond the ecstasy case study: cross-media public debate analysis of ASR-enriched data on any topic will benefit. For future development, user requirements for improved accommodation of leveled approach analysis of television data will be collected. The results will be disseminated to the Digital Humanities community with a conference presentation at EADH 2023 on how ASR can be used to search and analyze digitized television archives and a journal article on the reputation of ecstasy in Dutch media and the crossmedia turn for digital history for TMG: Journal for Media History. Lastly, two Media Suite tutorials will be created. The first of these will be a hands-on tutorial for cross-media public debate research in the Media Suite. This tutorial will be aimed at university level students with an interest in exploring the public debate surrounding their research topic. Secondly, a tool criticism tutorial will be created in which media scholars are invited to reflect on the limitations of exploring a visual data archive using search and visualization techniques of ARS metadata.

Project info


Berrie Jens van der Molen
Berrie Jens van der Molen

Postdoctoral Researcher, Utrecht University


Van der Molen, Berrie. Talking XTC: Drug discourse in post-war Dutch newspaper and radio debates. Doctoral thesis. Universiteit Utrecht, 2022. https://dspace.library.uu.nl/handle/1874/415753

Van der Molen, Berrie, van Gorp, Jasmijn, Pieters, Toine. Operationalizing “Public Debates” across Digitized Heterogeneous Mass Media Datasets in the Development and Use of the Media Suite. Selected Papers from the CLARIN Annual Conference 2018, Pisa, 8-10 October 2018, vol. 159, Linköping University Electronic Press, 2019, pp. 200–08, https://ep.liu.se/ecp/article.asp?issue=159&article=021&volume=0.

Huurdeman, H. C., Liliana-María Melgar-Estrada, Ordelman, Roeland, Noordegraaf, Julia. Supporting the Interpretation of Enriched Audiovisual Sources through Temporal Content Exploration. Conference paper. DHBenelux. Liège. 2019.