Providing a Structured Method for Integrating Non-Speech Audio into Human-Computer Interfaces

This thesis provides a framework for integrating non-speech sound into humancomputer interfaces. Previously there was no structured way of doing this, it was done in an ad hoc manner by individual designers. This led to ineffective uses of sound. In order to add sounds to improve usability two questions must be answered: What sounds should be used and where is it best to use them? With these answers a structured method for adding sound can be created An investigation of earcons as a means of presenting information in sound was undertaken. A series of detailed experiments showed that earcons were effective, especially if musical timbres were used. Parallel earcons were also investigated (where two earcons are played simultaneously) and an experiment showed that they could increase sound presentation rates. From these results guidelines were drawn up for designers to use when creating usable earcons. These formed the first half of the structured method for integrating sound into interfaces. An informal analysis technique was designed to investigate interactions to identify situations where hidden information existed and where non-speech sound could be used to overcome the associated problems. Interactions were considered in terms of events, status and modes to find hidden information. This information was then categorised in terms of the feedback needed to present it. Several examples of the use of the technique were presented. This technique formed the second half of the structured method. The structured method was evaluated by testing sonically-enhanced scrollbars, buttons and windows. Experimental results showed that sound could improve usability by increasing performance, reducing time to recover from errors and reducing workload. There was also no increased annoyance due to the sound. Thus the structured method for integrating sound into interfaces was shown to be effective when applied to existing interface widgets.