Improving Digital Accessibility Through Audio-Game Co-Design

This document examines how audio-only, spatial-sound games can be co-designed with blind and visually impaired players to improve digital accessibility, particularly as emerging virtual environments such as metaverse spaces remain largely visual-first and inaccessible to non-sighted users. Drawing on a co-design study with twelve participants recruited through sight loss charities, the authors built a maze-explorer prototype in Unreal Engine 5 with fifteen levels and stereo spatialised audio, designed to run on low-cost hardware such as headphones, a keyboard, and a joystick. The work positions this approach as a counter to audio games' tendency toward narrative gamebook structures with limited replay value.

In a workshop with participants, twelve of fifteen levels were completed, with most players reaching around level eight, suggesting the game was learnable without sight. Heat-map analysis, however, revealed heavy reliance on trial and error and frequent collisions with walls and enemies, especially at corners, pointing to a need for clearer spatial audio cues or haptic assistance. Participants responded enthusiastically, proposing extensions to genres such as racing and audio maps for rehearsing navigation in unfamiliar public spaces, and they highlighted that accessibility barriers often arise before a game even launches, at the level of device home screens and menus.

The authors conclude that accessible audio gaming is constrained by visually oriented hardware and software, including expensive consoles, platforms such as Steam that are incompatible with screen readers, and a scarcity of non-visual peripherals. They propose a dedicated non-visual console emphasising audio and haptic output with lower-fidelity visuals as a means of reducing cost and broadening access to both games and wider spatial digital environments. The evidence presented is largely qualitative, supported by the project's Specsavers Visionary Award (2022) and a participant's account of screen-reader incompatibility, with no formal quantitative data reported.