Speaker separation in diarization

1. A method of separating speakers in an audio file, the method comprising:

obtaining the audio file:

transcribing the audio file into at least one text file:

identifying homogeneous speech segments within the at least one text file;

segmenting the audio file into homogeneous audio segments that correspond to the identified homogeneous speech segments;

separating the homogeneous audio segments of the audio file into a first speaker audio file and a second speaker audio file, wherein the first speaker audio file only contains audio attributed to the first speaker, and wherein the second speaker audio file only contains audio attributed to the second speaker;

determining which combination of the homogenous speaker segments has a greater likelihood of matching an agent speech model by at least comparing the agent speech model to audio found in first speaker audio file and audio found in the second speaker audio file;

based on the comparing, identifying one of the first speaker audio file and the second speaker audio file as being an agent audio file and identifying the other of the first speaker audio file and the second speaker audio file as being a customer audio file; and

transcribing the agent audio file and the customer audio file to produce a diarized transcript.

2. The method of claim 1 , further comprising filtering the audio file to remove non-speech segments.

3. The method of claim 2 , further comprising;

identifying long speech segments in the homogenous audio segments: and

splitting the long speech segments based upon contextual information from the identified homogeneous speech segments.

4. The method of claim 2 , further wherein filtering of the audio file comprises energy envelope filtering to remove segments with energy determined to be below a lower energy threshold or above an upper energy threshold.

5. The method of claim 1 wherein separating the audio file into a first speaker audio file and a second audio file further comprises:

clustering the identified homogenous segments:

creating a first speaker model and a second speaker model from the clustered identified segments; and

identifying unclustered segments by comparing an unclustered segment to the first speaker model and the second speaker model.

6. The method of claim 1 further comprising:

splitting the audio file into a plurality of frames;

segmenting the audio file into entropy segments based upon an entropy of each frame; and

clustering the entropy segments and the homogeneous audio segments.

7. A method of claim 1 , further comprising:

segmenting the audio file into a plurality of frames;

calculating an overall energy speech probability for each frame;

calculating a band energy speech probability for each frame;

calculating a spectral peakiness speech probability for each frame

calculating a residual energy speech probability for each frame;

computing an activity probability for each frame from the overall energy speech probability, band energy speech probability, spectral peakiness speech probability, and residual energy speech probability;

comparing a moving average of activity probabilities to at least one threshold; and

identifying a speech and non-speech segments in the audio data based upon the comparison.

8. A method of audio diarization, the method comprising:

obtaining a digital audio file;

transcribing the digital audio file with an automated transcription to produce a text file;

identifying homogeneous speech segments in the text file;

segmenting the digital audio file into digital audio file segments based upon the identified homogenous speech segments in the text file;

performing a blind diarization that separates the homogeneous audio segments of the digital audio file into a first speaker audio file and a second speaker audio file, wherein the first speak audio file only contains audio attributed to the first speaker, wherein the second speaker audio file only contains audio attributed to the second speaker, and wherein the blind diarization comprises clustering the digital audio file segments into groups that discriminate between speakers;

identifying one of the first speaker audio file and second speaker audio file as being an agent audio file and identifying the other of the first speaker audio file and the second speaker audio file as being a customer audio file; and

transcribing the agent audio file and the customer audio file to produce a diarized transcript.

9. A method of claim 8 , further comprising: segmenting the audio file into a plurality of frames;

calculating an overall energy speech probability for each frame;

calculating a band energy speech probability for each frame;

calculating a spectral peakiness speech probability for each frame; calculating a residual energy speech probability for each frame

computing an activity probability for each frame from the overall energy speech probability, band energy speech probability, spectral peakiness speech probability, and residual energy speech probability;

comparing a moving average of activity probabilities to at least one threshold; and

identifying a speech and non-speech segments in the audio data based upon the comparison.

10. The method of claim 8 , further comprising:

filtering the audio file to remove non-speech segments;

identifying long speech segments in the homogenous speech segments; and

splitting the long speech segments based upon contextual information from the identified homogeneous speech segments.

11. The method of claim 8 , wherein transcribing the digital audio file comprises applying an agent model to the digital audio file;

wherein identifying one of the first speaker audio file and the second speaker audio file as the agent audio file, comprises comparing the first speaker audio file and the second speaker audio file to the agent model.

12. The method of claim 8 , further comprising:

splitting the audio file into a plurality of frames;

segmenting the audio file into entropy segments based upon an entropy of each frame; and

clustering the entropy segments and the homogeneous audio segments.

13. A system for audio diarization, the system comprising:

a transcription server that receives audio data, transcribes the audio data, and creates an information file that identifies at least one homogeneous speech segment from the transcribed audio data:

a blind diarization module operating on a computer processor, the blind diarization module receives the audio data and the information file, the blind diarization module separates the homogeneous audio segments of the digital audio file into a first speaker audio file and a second speaker audio file, wherein the first speak audio file only contains audio attributed to the first speaker, and wherein the second speaker audio file only contains audio attributed to the second speaker;

an agent diarization module operating on the computer processor, the agent diarization module receives an agent speech model, the agent diarization module determines which combination of the homogenous speaker segments has a greater likelihood of matching the agent speech model by at least comparing the agent speech model to audio found in first speaker audio file and audio found in the second speaker audio file, and based on the comparing, the diarization module identifies one of the first and second speaker audio files as being an agent audio file and identifies the other of the first and second speaker audio files as being a customer audio file; and

a transcription server that receives the agent audio file and the customer audio file and transcribes the agent audio file and the customer audio file to produce a diarized transcript.

14. The system of claim 13 , wherein the blind diarization module further filters the audio file to remove non-speech segments, identifies long speech segments in the homogeneous speech segments and splits the long speech segments based upon contextual information from the identified homogenous speech segments.

15. The system of claim 13 , wherein the blind diarization module further splits the audio file into a plurality of frames, segments the audio file into entropy segments based upon an entropy of each frame, and clusters the entropy segments and the homogeneous audio segments.