Using nonspeech sounds to provide navigation cues

Reviewer: Martha Elizabeth Crosby

Advances in the understanding of how people use computers have led to the greatest innovations in software development. Brewster describes three experiments that show how structured nonverbal audio messages called “earcons” can provide navigational cues in a nonverbal user interface. Earcons are abstract musical tones that use repetition, variation, and contrast of qualities such as timbre, register, intensity, pitch, and rhythm in structured combinations to create sound messages. Experiment 1 used a file system hierarchy of 25 nodes on 3 levels. The participants correctly recalled 81.5 percent of the 14 earcons they heard, showing the viability of an aural interface. Experiment 2 was designed to generalize these results by addressing the influence of sound quality, method of training, and time on the percentage of earcons recalled. Brewster reports that participants could still recall the earcons after a week, but that the quality of sound and the type of training influenced how well they recalled them. Experiment 3 used compound rather than hierarchical earcons to represent the structure from experiment 1. Earcons were created that did not require the participants to remember more than seven rules. The new design improved the number of earcons recalled from 81.5 percent to 97 percent. This type of earcon design has the advantage of creating arbitrarily sized hierarchies, and participants do not need to be retrained for each new structure's size and shape. This paper provides interface designers with valuable information. It will be particularly valuable to those who are responsible for designing complex displays to clearly present large amounts of changing data in ways compatible with users' information needs. In additio n to developers of applications that use telephone-based interfaces, mentioned by Brewster, this work should be of particular interest to designers of multimodal computing environments.