High Performance Sound Technologies for Access and Scholarship Research and Development with Repositories

FAIN: PR-50200-14

University of Texas, Austin (Austin, TX 78712-0100)
Tanya E. Clement (Project Director: May 2013 to May 2018)

The development of software that uses machine learning to help users automate descriptive metadata for spoken-word sound collections.

In order to increase the preservation of significant spoken word (such as poetry, storytelling, speeches, and oral histories) sound recordings, the UT Austin iSchool and the Illinois Informatics Institute (I3) are requesting two years of funding for HiPSTAS Research and Development with Repositories (HRDR) to develop software (ARLO) that uses machine learning and visualization to help users automate metadata description for undescribed sound collections. Products will include: (1) open source software (ARLO) that can be used with a variety of repositories; (2) a Drupal module for Mukurtu, an open source content management system designed for indigenous communities worldwide; (3) workshops and documentation for wider dissemination and training; and (4) a white paper detailing best practices in generating descriptive metadata for audio collections in the humanities.