Method and system for the automatic segmentation of an audio stream into semantic or syntactic units

An article of manufacture comprising a computer usable medium having computer readable program code means embodied therein for causing segmentation of an audio steam into semantic or syntactic units, wherein the audio stream is provided in a digitized format, the computer readable program code means in the article of manufacture comprising computer readable program code means for causing a computer to effect:

• determining a fundamental frequency for the digitized audio stream;
• detecting changes of the fundamental frequency in the audio stream, wherein detecting the changes of the fundamental frequency includes providing a threshold value for estimates of the fundamental frequency's voicedness and determining whether the voicedness of the fundamental frequency estimates are higher or lower than the threshold value, and wherein the voicedness of the fundamental frequency estimates lower than the threshold value equals no voice, and wherein the voicedness of the fundamnental frequency estimates higher than the threshold value equals voice;
• determining candidate boundaries for the semantic or syntactic units depending on the detected changes of the fundamental frequency;
• extracting a plurality of prosodic features in an environment of the audio stream where the voiceness of the fundamental frequency estimates are lower than the threshold value, wherein the environment is a period of time between 500 and 4000 milliseconds preceding and following the candidate boundaries;
• combining the plurality of prosodic features;
• and
• determining boundaries for the semantic or syntactic units depending only on the combined plurality of prosodic features.

A digital audio processing system for segmentation of a digitized audio stream into semantic or syntactic units comprising:

• means for determining a fundamental frequency for the digitized audio stream;
• means for detecting changes of the fundamental frequency in the audio stream, wherein detecting the changes of the fundamental frequency includes providing a threshold value for estimates of the fundamental frequency's voicedness and determining whether the voicedness of the fundamental frequency estimates are higher or lower than the threshold value, and wherein the voicedness of the fundamental frequency estimates lower than the threshold value equals no voice, and wherein the voicedness of the fundamental frequency estimates higher than the threshold value equals voice;
• means for determining candidate boundaries for the semantic or syntactic units depending on the detected changes of the fundamental frequency;
• means for extracting a plurality of prosodic features in an environment of the audio stream where the voicedness of the fundamental frequency estimates are lower than the threshold value, wherein the emvironment is a period of time between 500 and 4000 milliseconds preceding and following the candidate boundaries;
• means for combining the plurality of prosodic features;
• and
• means for determining boundaries for the semantic or syntactic units depending only on the combined plurality of prosodic features.