The EVIA Project collects two primary sets of metadata. Depositors create the first set by segmenting approximately ten hours of video with the Annotator's Workbench and providing extensive annotations of the segments. For each video segment, this descriptive metadata includes start and stop points; short, long, and detailed descriptions; participant names and roles; and controlled vocabulary.

We originally had planned to use standard METS/MODS for descriptive metadata, but the need for specialized data structures for annotated video made this impossible to do. Our solution has been to closely align our descriptive metadata for video segments with METS/MODS and to enable METS files to be derived from our scheme. Additional bibliographic data will be contained in the Dublin Core Schema—supported by the Dublin Core Metadata Initiative—for interoperability with the Open Archives Initiative (OAI) and for the ability of FEDORA to support OAI harvesting through the OAI-PMH protocol. All the metadata collected and transformed will be stored in the Digital Libraries FEDORA Repository.

In addition to descriptive metadata, we also collect technical metadata about the original tapes or files, the preservation masters, and the derived digital files. The audio portion of the technical metadata is based on the draft proposal from the American Engineering Society's Schema for Audio Metadata. For the video portion, we use the Library of Congress Video Technical Metadata Extension Schema. We have developed the Technical Metadata Collector, an online tool used to collect this technical metadata as each video tape or file is processed by technicians.