Loco - Radio – Designing High - Density Augmented Reality Audio

In this dissertation research, we explore ways of using audio on AR applications, as it is especially suitable for mobile users when their eyes and hands are not necessarily available and they have limited attention capacity. While most previous mobile AR audio systems were mostly tested in sparse audio maps, we want to create a system that can be challenged by a city load of information. We design and implement Loco-Radio, a mobile augmented reality audio browsing system. It uses GPS and a geomagnetic-based sensing module to provide outdoor and indoor location sensing. To enhance the audio browsing experience in high-density spatialized audio environments, we introduce auditory spatial scaling, which enables users or the system to adjust the spatial density of perceived sounds based on context. The audio comes from a custom geo-tagged audio database, which contains a set of channels designed for different use cases. In the first scenario, iconic music is assigned to represent restaurants. As users move in the city, they encounter a series of songs and the perception enhances their awareness of the numbers, styles, and locations of restaurants. It is tested by car drivers, bikers, and pedestrians. In the second scenario, audio clips of media lab research demos are tagged around the building. As a result, users can participate in an augmented reality audio lab tours. We argue that AR audio systems should consider not only where users are but also how they move. Discussion will be focus on strategies of using spatial audio in high-density audio environments and how they should change in different moving modes. Thesis Supervisor: Christopher Schmandt Title: Principal Research Scientist, MIT Media Lab