Brandon Woodward Defends Computer Science Thesis
Brandon Woodard is defending his Computer Science dissertation, “Characterizing Constraints for Audio Feedback Dimensions in Extended Reality,” this coming Tuesday, April 21st, at 4pm in the Center for Information Technology.
Below is a description of the thesis:
Interactive extended reality (XR) systems let users manipulate virtual content by moving, rotating, and reaching with their bodies, extending interaction beyond traditional 2D interfaces. XR systems increasingly rely on coordinated audio and visual cues to help users locate and interact with virtual content. For this, XR systems frequently use models of human hearing or vision that simulate audio-visual signals in the real world. However, realistic simulation is only one possible presentation, and constraining such presentations is also possible, such as limiting audio variation to a single axis, collapsing depth cues, or using metaphors instead of accurate spatial mappings. While many studies have examined constraints on visual content in XR, few studies have examined constraints on audio and how it relates to its corresponding visual elements or interaction control space. In particular, it remains underexplored when it is beneficial to deliberately reduce or constrain the dimensions encoded. To address this gap, this dissertation explores the effects of selectively constraining the dimensions in existing auditory content rendered in response to users’ actions and visual content.
After exploring the effects of constraints across three studies, constraints on audio feedback mechanisms were found to be non-inferior to spatial audio or superior in instances where XR tasks did not rely on precise audio localization. The first study derives design lessons for audio-visual feedback that support immersion and task performance when expanding the control space of smartphone-based augmented reality using an interface called Cam-2-Cam. Users highlighted how the audio-visual feedback made the two camera views used in AR feel like a single unified experience, thus enhancing immersion. Next, the second study presents an empirical evaluation of AudioMiXR, a 2D desktop interface with an AR component, for mixing audio that encodes three axes of information (i.e., spatial audio). AudioMiXR is an AR interaction concept in which audio is rendered in response to a user’s movements, closely reflecting how people listen in the real world. Compared to the 2D interface, users generally felt the audio-visual feedback afforded by AR provided a more immersive experience, improved usability, and supported their creativity. SoundFleXR, the final study, compared audio presentations with reduced or fuller spatial cues across XR tasks with different interaction demands. The results showed that simpler presentations with dimensions aligned to the task could sometimes work just as well as, or even better than, more realistic spatial audio.
Together, all three studies characterize constraints on audio feedback dimensions and the XR scenarios in which they may be non-inferior to more realistic rendering methods. Future studies should evaluate these design lessons in more general XR systems (e.g., operating-system-level interfaces) to assess their ecological validity.
Congratulations, Brandon!
