The Cloud Innovation Lab is developing a new media archivist tool in collaboration with Special Collections & University Archives (SCUA) to improve how archival audiovideo collections are processed, described, and prepared for access.
Video archives present a distinct challenge for archivists. Unlike text-based materials, identifying the people, topics, events, and significant moments contained within a video often requires viewing the recording in its entirety. Creating detailed metadata, locating meaningful segments, and preparing clips can therefore require substantial staff time, particularly when working with large historical collections.
The Cloud Innovation Lab recently created a solution (VAST) for processing television news broadcast footage for the Vanderbilt Television News Archive. The media archivist project has built on that solution to expand its capabilities. Already, the staff at SCUA can process a wide variety of videos using the tool. The team is continuing to work on processing purely audio files.
The SCUA platform provides an integrated workspace for working with digitized video. Archivists can review recordings, create and refine descriptive metadata, identify and segment meaningful portions of a video, and perform editing tasks without moving between multiple disconnected tools. AI-assisted capabilities can help surface information from video content and accelerate initial processing, while archivists review, correct, and enrich the resulting metadata and segments.
Key areas of the project include:
- Metadata creation — Assist archivists in generating and refining descriptive metadata for audiovisual materials.
- Video segmentation — Identify and define meaningful sections within longer recordings so individual topics, speakers, events, or moments can be described and accessed more precisely.
- Video editing — Provide tools for preparing clips and edited derivatives as part of the archival workflow.
- AI-assisted analysis — Explore the use of cloud-based AI services to extract useful information from audiovisual content and reduce repetitive processing work.
- Archivist review and control — Keep professional archival judgment at the center of the workflow, allowing generated information and suggested segments to be reviewed and modified before they become part of the archival record.
By bringing these capabilities together, the project aims to reduce the effort required to process audiovisual collections while enabling richer, more granular description. More detailed metadata and time-based segmentation can also make archival video easier to discover, navigate, and use by researchers.
The AI-assisted archival tool is being developed as an applied Cloud Innovation Lab project, combining experimentation with emerging cloud and AI technologies with the practical requirements of Special Collections & University Archives. The project will help evaluate where these technologies can provide meaningful assistance to archival workflows and where human expertise remains essential.
Ultimately, the project is not simply about automating video processing. It is about developing a workflow in which technology handles more of the time-intensive mechanics of audiovisual processing and routine tasks so archivists can concentrate their expertise on more complex and nuanced work.
