AVD-LV: An Accessible Player for Captioned STEM Videos
Raja S. Kushalnagar, John J. Rivera, Warrance Yu, Daniel S. Steed · 2014 · ASSETS '14: Proceedings of the 16th International ACM SIGACCESS Conference on Computers & Accessibility · doi:10.1145/2661334.2661353
Summary
This demo paper presents AVD-LV (Accessible View Device interface for Low Vision), a Chrome browser extension that enhances YouTube lecture video accessibility for deaf and low vision (DLV) viewers. The core problem is that captioned videos require DLV viewers to process two simultaneous visual streams—the lecture video content and the caption text—but their restricted field of view prevents them from seeing both at once. When important visual and textual information occur simultaneously, DLV viewers inevitably miss one source. Unlike approaches that use magnification or minification (which trade off field of view against resolution), AVD-LV uses a visual gaze management approach: it allows viewers to pause either the video or the captions independently with a single key-press, so they can focus on one stream at a time without losing information from the other.
Key findings
The interface provides several key features for DLV viewers: independent pause controls for video and captions, display of both synchronized transcripts (multiple lines representing several seconds of audio) and standard captions (1-2 lines representing 0.5-1 second of audio), and compatibility with any captioned YouTube video. The Chrome extension fetches caption files and displays them alongside the video, adapting to multiple caption and transcript standards. The approach avoids the fundamental tradeoff of magnification-based solutions where improving resolution reduces field of view and vice versa. By allowing DLV viewers to control the temporal flow of each information stream independently, the interface lets them process dense STEM content at their own pace.
Relevance
This work addresses an underserved population at the intersection of two disabilities—deafness and low vision. While approximately 50 million Americans use closed captions, standard captioned video interfaces assume viewers can see both the video and captions simultaneously, which excludes DLV viewers. The problem is especially acute for STEM lectures, which tend to be visually dense with slides, diagrams, and equations that are essential to understanding. For accessibility practitioners, the gaze management approach—letting users independently control the timing of different information streams—offers a design pattern applicable beyond video. The work also demonstrates that browser extensions can provide meaningful accessibility improvements to existing platforms like YouTube without requiring platform changes, making it a practical and immediately deployable solution.
Tags: deaf · low vision · captions · video accessibility · STEM education · deafblind · online learning
Standards referenced: ADA · Communications and Video Accessibility Act 2010