Supporting navigation in auditory interfaces using personalization and multiple synthetic voices

by Peer Shajahan

Abstract

Auditory interfaces commonly use synthetic speech for conveying information. In many instances the information being conveyed is hierarchically structured, such as menus. However, using the auditory channel the structure of the information is not “visible”. In particular, the hierarchical nature of auditory interfaces is not revealed explicitly to users. As a result, navigation can be complex and it is considered to be one of the critical issues in auditory interfaces. In an effort to address the issue of navigation in auditory interfaces, two solutions were suggested and investigated in this thesis.

The first solution utilizes a hierarchy re-structuring approach. It aims at reducing the navigation time in auditory interfaces by allowing users to customize/personalize their favorite menus in the interface. The system developed allows users to bookmark any given node in the menu-based system. Bookmarks provide a method for skipping the entire hierarchy structure to access only a node of interest. By means of this approach, users can access personalized information more efficiently and without spending time listening to prompts before making decisions. The results of an evaluation of this technique show that users can quickly access items of interest. However, if the hierarchy is significantly re-structured by the user, performance results degrade.

The second solution described in this thesis focuses on reducing the error rates in auditory interfaces by providing navigation cues to the users, using multiple synthetic voices. In this study, different synthetic voices were created, in a structured manner, by manipulating synthetic parameters for representing the nodes in the hierarchy. This study consists of two experiments. The first experiment focuses on representing small hierarchies using multiple synthetic voices, whereas the second experiment focuses on representing complex hierarchies using multiple synthetic voices. The experimental results suggest that multiple synthetic voices can be used to represent and provide navigation cues in hierarchies.

Citation

Peer Shajahan. 2005. Supporting navigation in auditory interfaces using personalization and multiple synthetic voices. Master's thesis, University of Manitoba.

Bibtex entry

@MASTERSTHESIS { PeerShajahanMScThesis,
    AUTHOR = { Peer Shajahan },
    TITLE = { Supporting navigation in auditory interfaces using personalization and multiple synthetic voices },
    SCHOOL = { University of Manitoba },
    ADDRESS = { Winnipeg, Manitoba, Canada },
    YEAR = { 2005 },
}