Improving the Accessibility of Screen-Shared Presentations by Enabling Concurrent Exploration
Danyang Fan, Sasa Junuzovic, John Tang, Thomas Jaeger · 2023 · Proceedings of the 25th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3597638.3608411
Summary
This paper addresses a fundamental gap in remote meeting accessibility: when presenters screen-share slides during video calls, blind and low vision (BLV) screen reader users have no adequate way to independently explore slide content in real-time. While sighted audience members can freely scan slides, read ahead, or re-examine details while listening to the presenter, screen reader users are locked into a linear, presenter-dependent experience. Existing tools like PowerPoint Live on Microsoft Teams and JAWS OCR technically provide some slide access, but prior research found them too cumbersome for live use due to competing audio streams (presenter voice overlapping screen reader output) and overly complex navigation commands. The researchers conducted a formative design workshop with four screen reader users to identify five core values for concurrent slide exploration: Prioritize Key Information, Reduce Cognitive Load, Provide Exploration Independence, Minimize Control Effort, and Encode Spatial Information. Based on these values, they built two prototypes. The laptop prototype used simple arrow-key navigation to move between slide elements, with the ability to play/pause descriptions on demand, and introduced spatial audio separation between the presenter voice (centered) and screen reader output (positionable along a circular arc). A physical control board with sliding knobs allowed real-time adjustment of spatial audio position, screen reader volume, and reading speed. The phone prototype added touchscreen-based spatial navigation and audio spatialization mapped to element positions on the slide.
Key findings
In a design probe study with 10 BLV participants (ages 25-65+), the laptop prototype significantly improved perceived accessibility of screen-shared content compared to the reference condition (PowerPoint Live on Teams) with statistical significance (chi-squared = 7.52, p < .01). In the reference condition, most participants chose not to interact with slide elements at all because the cognitive burden of managing competing audio streams was too high. With the laptop prototype, participants navigated to 74% of all elements on average, began listening to 65% of descriptions, and fully listened to over 50%. Participants developed ten distinct exploration strategies, including using slide titles to reinforce the topic of focus, exploring typically inaccessible content like images and videos to complement the presenter, synchronizing exploration with the presenter to gain complementary information, and using spatial audio separation to negotiate attention between audio sources. Nearly all participants moved the screen reader audio away from the presenter (mean 71.4 degrees of separation), confirming that spatial separation is valuable for managing concurrent audio. Screen reader and presenter audio overlapped 64.1% of the time the screen reader was playing, yet participants still preferred exploring with temporal overlap rather than having no slide access. The phone prototype received mixed feedback: touchscreen navigation was praised by participants with some visual perception but found too cognitively demanding by others, while audio spatialization of element positions was appreciated for building mental maps of slide layouts.
Relevance
This research has immediate practical implications for the hundreds of millions of remote meetings held daily. The core finding that screen reader users will actively explore slides during live presentations when given adequate tools — even tolerating significant audio overlap — overturns the assumption that concurrent access is impractical. For videoconferencing platform developers (Microsoft Teams, Zoom, Google Meet), the five identified values provide a concrete design framework: strip slide content to key information (text and alt-text without layout metadata), let users control what, when, and how elements are read, use spatial audio to separate presenter and screen reader channels, and keep navigation commands simple (arrow keys, not complex screen reader commands). The study also highlights a broader equity issue: without concurrent slide access, screen reader users must either interrupt meetings to ask about visual content (disclosing their disability) or simply miss information that sighted peers access freely. The tension between values — spatial information increases exploration independence but at the cost of cognitive load — offers a useful lens for any accessibility tool balancing multiple user needs.
Tags: screen reader · blindness · presentations · screen sharing · spatial audio · cognitive load · videoconferencing · concurrent exploration · Microsoft Teams
Standards referenced: WCAG