Audiovisual accessibility: Evaluating workflows for closed captioning and transcripts; ALA Annual Poster

The Moving Image and Sound Archive and Digital Library Services Departments at the University of Utah's J. Willard Marriott Library received an internal grant to explore methods for generating and displaying closed captions and transcripts for digital audiovisual resources. Although many other institutions are currently experimenting with various methods for creating captioning and transcripts, there is no perfect standard for effectively and inexpensively generating this important content. The University of Utah has worked with a set of archival audio and audiovisual files, which represent a variety of sound qualities, source formats, and genres of amateur and professional audio content. These recordings have been transcribed using three methods - manual transcription, Google Speech API with manual correction, and outsourced vendor transcription. The time, cost, and quality of the results of each workflow has been compared. Meanwhile, partners in Library IT have explored options for searchability and display within the Marriott Library's digital asset management system. By Summer 2019, the Marriott Library expects to have a functioning workflow for creating and displaying text representations of audio and audiovisual materials of various types and qualities for use by digital scholars and the hearing impaired.