Project

Voice as Data

In order to utilize digitized speech (a telephone message, recording of a meeting, lecture, etc.) more effectively as a computer-data-type, it is important to understand speech structure and allow rapid scanning of the audio data. Structure is derived from pauses, speaker changes and turn-taking, emphasis, and intonational cues. Presentation may be serial, including audio time-compression, or parallel, using both ears to listen to multiple audio streams simultaneously.

In order to utilize digitized speech (a telephone message, recording of a meeting, lecture, etc.) more effectively as a computer-data-type, it is important to understand speech structure and allow rapid scanning of the audio data. Structure is derived from pauses, speaker changes and turn-taking, emphasis, and intonational cues. Presentation may be serial, including audio time-compression, or parallel, using both ears to listen to multiple audio streams simultaneously.