Annotation - The EVIA Digital Archive Project


home research & development cataloging and documentation

For centuries, scholars have used annotation to describe and interact with documents of all kinds. Writing in the margins of ancient manuscripts or creating new explanatory editions of older works has long been part of the academic enterprise. Annotation of time-based media is a more recent phenomenon, but it draws on these older practices. Only in the last few decades have scholars had digital tools that allow for precise alignment of text with time in a multimedia environment. The EVIA Digital Archive Project has been among the first projects to create tools and procedures for the annotation of video for scholarly purposes.

It became clear early in the development of the EVIA Project that the annotation of ethnographic video required more than a single layer. The Project thus created a multilayered hierarchy of segmentation for the organization and description of contents. Within this hierarchy, the "Collection" represents a group of video recordings. On average, one collection represents ten hours of video from a given research project, with the depositing scholar annotating entire recordings, not only selected clips. Within a collection, recordings typically are sequenced in chronological order. For cases in which a scholar attempted to maximize tape use by recording new material on remaining blank space of previously used tapes, our tools allow us to place portions of the recordings in correct chronological order.

Within a collection, the secondary organizing segment is called an "Event." While the first level of segmentation is a way of grouping recordings into a coherent organization of time, events often are specially marked and bounded human interactions like ceremonies, rituals, or concerts. They may span several days or last less than an hour. Within ethnographic disciplines, events often constitute objects of study because they are performances or a complex of performances in which culture is displayed, negotiated, and contested. For the purposes of annotation, we define "Event" a little more broadly so that it can accommodate interview sessions or simply segments of time that qualify as the broadest possible divisions within a given recording. All of the time within a collection is segmented into contiguous events.

Within an event, we further subdivide time into "Scenes," which offer a means by which to create intellectually relevant segments for description and for searching. Scenes are rather dependent on the nature of the scholar's research. They may represent different performances at a concert, liturgical sections of a ritual, or thematic sections within an interview. While events describe activity in a general way, scenes allow the scholar to provide focused description of what is happening in the video. Scenes also are the basic units returned in our search results. As with the relationship of collections to events, all of the time within an event is segmented into contiguous scenes.

Finally, an "Action" is the smallest unit of division. Its purpose is for pointing out elements or activity in the video worthy of special description. Actions are used to draw attention to a particular dance movement, costume element, gesture, or manner of speaking, among other possibilities. Unlike events and scenes, actions are not a contiguous subdivision of the level above; actions are used only as needed and appear distributed throughout a collection. Depending upon the research interests of the depositing scholar, a collection may have many or no actions.

Within each of these four segmentation levels—Collection, Event, Scene, and Action—several layers of description are possible:

  • Title—a short reference title of the segment

  • Brief Description—a short description of the segment content

  • Detailed Description—an extended description or analysis of the segment content

  • Participant Listing—references to individuals and their roles

We offer a transcription function that allows annotators to define segments of time within or across other segment boundaries in order to provide text transcriptions of what is being spoken or sung. Multiple language translations of transcription texts are allowed.

Within any annotation text field, annotators can create three different kinds of hyperlinks:

  • Glossary—the glossary creates a reference source inside each collection that defines terms and concepts or provides biographical information about persons. The glossary offers a way to address the non-linear manner of access to segments. It allows the annotator to describe important terms and concepts only once and then refer to them by hyperlink within the annotation texts.

  • Citations—full bibliographic citations can be created and referenced from within annotation texts and linked to online versions if they exist.

  • Inter-segment linking—annotators can create links to other segments within a collection for the purpose of making comparisons or contrasts. This feature is particularly useful if multiple performances of a particular song or ritual exist in a collection.

Copyright © 2001-2022 The Trustees of Indiana University | Copyright Complaints. Address comments to PARTICIPANT LOGIN