The Media Lab's ALIVE project, Artificial Life Interactive Video Environment, is a testbed for research into remote-sensing, full-body, interactive interfaces for virtual environments, [4]. A plan view of the ALIVE space is shown in Figure 1. A camera on top of the large video projection screen captures images of the user in the active zone, these images are fed to a visual recognition system where various features are tracked. A mirror image of the user is projected onto the video screen at the front of the space and computer-rendered scenes are overlayed onto the users image to produce a combined real/artificial composite image. One of the applications in the ALIVE space is an autonomous agent model of a dog, called ``Silas'', which posseses its own system of behaviors and motor-control schemes. The user is able to give the dog visual commands by gesturing. The visual recognition system makes decisions on which command has been issued out of the twelve gestures that it can recognize, see Figure 2. This section describes the design and implementation of an audio component for the ALIVE system that allows the user to give the artificial dog spoken commands from the free field without the use of body-mounted microphones.