Scholarly video is on the rise today in no small part due to the rise of virtual events– which are, on the whole, a massive opportunity to make conferences more accessible and inclusive. However, the same accommodations that are needed at in-person conferences, such as sign-language interpreters, are also required in the virtual world. If anything, it is easier to make content accessible online than it is in a physical environment: there is more time to prepare (and pre-record presentations), and there already exists web technology and standards that can be applied to make event platforms accessible.
One of those key accessibility requirements is the addition of closed captions and transcripts to video content. Captions are needed to help people with hearing disabilities, but they also help reach global audiences, expand the reach of non-native speakers with strong accents, and make technical content more accessible to anyone unfamiliar with the terminology. Transcripts can also be used to index media content, make it citable, and enable deep search and linking, and synchronized transcripts can make media content “skimmable,” just like a journal article.
However, conferences with large technical programs often find that providing captioning for thousands of presentations is cost-prohibitive. Speech-to-text technology is rapidly improving, but the focus of the big players (Google, Amazon) seems to be on virtual assistant-type applications that are becoming very good at learning to understand single voices, especially from North America, but not so good at interpreting thousands of different speakers with different accents and dialects giving technical presentations.
Captions need to be reviewed by humans to be of scholarly quality – and that quality will often be higher if the reviewer is the author of the content. That is why Cadmore developed a workflow for event organizers to give speakers the opportunity to review, edit, and submit captions and transcripts for their pre-recorded presentations. When an author submits a video to an abstract submission system, such as ScholarOne Abstracts, Cadmore processes it, generates an automated transcript, and sends the author a personalized link where they can access a WYSIWYG editor to review, edit, and approve it.
At a recent AIAA event, presenters were given the option to edit their automatically-generated transcripts prior to the meeting. We shared some best practices and encouraged authors to think in terms of both enhancing accessibility and extending the discoverability of their work, and were pleased to have an over 90% compliance rate.
If you’d like to learn more about automated transcripts, drop us a note.