ARIA: Demonstrating Transfer from AR-Based Spatial Audio Training to Speech-in-Noise Performance

Sungyoung Kim

Abstract: We demonstrate ARIA, an iOS augmented reality application that addresses engagement and accessibility limitations of traditional auditory training systems. The app integrates 6-degree-of-freedom spatial audio with gamified auditory localization and speech-in-noise training in users’ real environments. Our feasibility study with 11 normal-hearing adults (age: 53±3 years) showed significant improvements in Korean Matrix Sentence Test (KMST) performance across challenging listening conditions (-6dB: 7.3±4.5%, p=0.001; -9dB: 10.5±8.0%, p=0.004). Training duration demonstrated dose-response effects, with correlations between engagement time, in-game localization improvements, and transfer to standardized assessments at -6/-9dB where real-world benefits are most critical