Although talking is an integral part of collaborative activity, there has been little computer support for acquiring and accessing the contents of conversations. This talk describes one approach to the problem: ubiquitous audio, or the unobtrusive capture of voice interactions in everyday work environments. Because the words themselves are not available for organizing the captured interactions, structure is derived from acoustical information inherent in the stored voice and augmented by user interaction during or after capture.

Two ubiquitous audio applications capture and structure audio from office discussions and telephone calls. The first application is a digital tape loop that provides short-term auditory memory in the office, with no inherent structure to the recording. The second application allows users to selectively mark and save portions of telephone conversations, and utilizes the conversants' turntaking for structure. These applications were created in the context of a set of desktop audio tools, and this context allows audio data to be used in its original form.

The visual representation of the stored audio is an ongoing research area. Recent work addressing aspects of this problem will be presented.