Vocal Mouse: Various Phases?

Non-speech vocal sounds can be used along with command words as input to enable people (especially those with motor disabilities) to control computer interfaces effectively. The Vocal Mouse will track both speech and non-speech vocal features including pitch, volume, and vowel quality in real time using audio signal processing. This application will enable the user to control the mouse pointer smoothly and continuously by vocalizing various vowel sounds corresponding to the desired direction of movement. Under conventional speech-driven pointer control, only spoken words as commands are used. The rate at which such parameters can be changed is also limited by the speed at which each command phrase can be uttered and recognized by the speech recognizer. Movement direction is specified by vowel sound vocalizations, in which different vowel sounds are mapped to each of the four cardinal directions. By varying the vowel sound and the volume continuously, the pointer’s movement direction and speed can be smoothly controlled using the Vocal Mouse pointer control. This dissertation has combined the functionality offered by the Vocal Mouse with traditional speech recognition engines to explore ways to extend the capability of voice-based interaction with new user interface technologies.