For today’s audiovisual archival custodians, access is a core mission. Despite the major opportunities presented by digital techniques and extensive investment globally, the majority of such collections remain largely inaccessible to society. Reasons for this include sheer volume and lack of extant metadata, and extensive copyright restrictions on open access necessitating in situ viewing within archives or museums. These issues are compounded by the lack of intuitive interfaces to accommodate meaningful public engagement. Narratives From the Long Tail: Transforming Access to Audiovisual Archives intends to address this major gap in audiovisual research to resolve access for diverse audiences. Through its methods, tools and systems, Narratives investigates diverse ways to open up the ‘long tail’ of cultural archives as well as how to activate these as communal resources available to the broadest spectrum of society.
Narratives intends to address this major gap in audiovisual research to resolve access for diverse audiences.
The resolution of three interrelated theoretical and experimental problems will be critical to remove existing barriers to access and advance public engagement: Firstly, archival metadata focuses on curatorial information in a unimodal classification system, which fails to encompass the social, aesthetic, ephemeral and emotional nature of the audiovisual content. Secondly, search functions in these archives are developed for specialists, as query-response or text-driven ‘single-shot’ paradigms, requiring prior knowledge and accurate expression in order to retrieve relevant items. Thirdly, contemporary museology has increasingly shifted to performative, participatory and immersive learning as powerful models for audience engagement in which polyvocal narratives are central. Existing interfaces, being analytic and emblematic of digital humanities and specialist archival retrieval, fail to satisfy the interpretive needs of museum audiences or the indeed wider cultural sector.
Three interlocking goals
- Archival semantics: to transform conventional archival data curation and associated knowledge systems with computational machine learning and visual analytics processes that can operate at corpus, film, segment and inter-image and intra-image level leading to an ‘operationalization’ of the archive and creating a far wider range of semantic descriptions as the basis for interactive search.
- Narrative visualization: to develop a visualization framework that supports user interactions to produce ‘narrative coherence’, promoting insight through discovery within the three broadly defined experimental schemas: spatio-temporal, social, and affective and aesthetic.
- Experimental museology: to design scalable public-facing interfaces for audiovisual archives, advancing frameworks for participatory interaction through which users can navigate, explore and creatively reorganize the archive through immersive visualization paradigms.
Three main goals: archival semantics, narrative visualization and experimental museology.
Narrative Visualization Engine
These goals will be collaboratively investigated through the development of experimental Narrative Visualization Engine, a novel system that exists at the intersection of artificial intelligence, data curation, experimental and speculative visualization and performative archives. This ‘engine’ is defined by a combination of technologies: machine readable metadata, machine learning, the extraction of relationships among the elements (visual analytics) and a visualization framework, based on a set of preconditions that give rise to emergent narrative scenarios as users interact with its system. Narratives is founded on the complexities of contemporary participatory culture and, through our novel interdisciplinary framework of computational museology, we will address the three interrelated aspects of the engine’s organism: archive, museum, public. Our ultimate objective is that this research will generate profound applied and theoretical results toward the conceptualization of knowledge creation in symbiosis with intelligence systems. Advancing museological and archival theory in tandem with practice towards a systems approach, Narratives has the potential to set important transdisciplinary precedents across science, culture and industry.
Narratives is founded on the complexities of contemporary participatory culture.
It is only through a systemic shift and interdisciplinary effort spanning computer vision, visual analytics, interactive visualization and archival science, alongside a specialized knowledge of museology and audiovisual collections, that the issues of public access to the cultural memory of audiovisual collections can be analyzed, and sustainable solutions found at all levels of this complex problem. In order to address the challenge, Narratives From the Long Tail: Transforming Access to Audiovisual Archives unites four key research disciplines (and their subdomains), placing computational methods at the nexus of interchange between them. We propose to drive systemic change in existing methods of digital museology and audiovisual archive curation, through advances in machine learning and visual analytics, and interactive visualization, with the aim of creating a platform for computational museology grounded in systems thinking. This computational approach to museology, together with its theoretical framing, is essential to overcome the trifold challenge of access of audiovisual archives, which hinges on:
- the need for semantic richness across immense and constantly growing archives,
- the need for narrative engagement to make sense of these archives
- the need for participatory platforms of engagement.
Our interwoven experimental project methods will allow us to integrate, operationalize and visualize audiovisual material, as demonstrated in the proposed experimental Narrative Visualization Engine, which intends to transform existing visitor discovery experiences in museums and archives into open ended and serendipitous realms, replete with polychronic routes of discovery. By addressing imperatives of scalability and interoperability, this engine forms the basis for next-generation visualization systems for use across the cultural sector and related industries. Our methods and results also transcend the state-of-the-art across its plural domains to define computational museology.