Speech Synthesis Integration in Open Journal Systems

Abstract

PDFs of research papers are often characterized by poor accessibility and lack of speech synthesis compatibility. In this poster, we present our project that addressed these challenges by introducing a tailored EPUB format optimized for speech synthesis. The project was carried out at UiT's publishing service for open access journals (Septentrio Academic Publishing) and was partially funded by the National Library of Norway.

During the course of the project, we explored two approaches to article formats compatible with speech synthesis.

The first approach involved the implementation of specialized markup intended for speech synthesis software. However, this method proved to be excessively labor-intensive for our university's publishing service, prompting us to seek a more efficient alternative.

Our second approach entailed converting documents from Word to EPUB via JATS XML, leveraging the versatile Pandoc tool. This streamlined approach not only simplified ongoing management but also ensured the creation of accessible files enriched with robust metadata. A noteworthy aspect of this project is its reliance on open source tools, making the approach readily transferable to other publishing services, particularly those employing Open Journal Systems.