“What did I just miss?!” Presenting ClassTranscribe, an automated live-captioning and text-searchable lecture video system, and related pedagogical best-practices

Captioning of recorded video content has been available for over four decades, but its importance and pedagogical use has seen a recent revival due to the legal accessibility compliance requirements of MOOC and online content, and secondly, due to efforts by universities to bring digital accessibility to their on-campus learning environments. We present an open-source web-platform, ClassTranscribe, and related pedagogical practices, to enhance student learning both during lectures, and during later searching and viewing of recorded lectures. Using recent advances in automated Speech-To-Text (STT) technologies, it is possible to provide low-cost, accurate, and timely, text-searchable recorded lecture videos, and in-class live-captioning that is informed by domain-specific words. Using a crowd-sourced approach, captions of lecture recordings can be incrementally improved above the limitations of automated STT approaches. Secondly, we report on suggested technology and practices to support live class captioning for instructors who use PowerPoint, a document camera, live coding, and other presentation styles. Using Universal Design for Learning, we discuss lecture captions and recorded lectures as an assistive technology that can also assist all students. As an example of this, we describe how distractions in a lecture and a student's limited attention span may cause students to lose focus and context ("What did I just miss?!"), and fail to grasp the topic. By providing a live caption history on student devices, students can refocus, reconnect, and thus have an opportunity to learn the current lecture topic being presented. The design of the ClassTranscribe platform is extensible and scalable. We demonstrate captioning of content by integrating with two websites used to host lecture videos, youtube.com and echo360.com.