System and method for segmenting audio files for transcription

1. A method for segmenting an audio file at optimal points that allow efficient transcribing, comprising:

analyzing an audio file, wherein the analyzing includes (i) identifying number of speakers in the audio file based on generating a signature for each voice determined to be unique within the audio file, and (ii) identifying accent of each speaker based on at least one of a Gaussian mixture model (GMM), a GMM Support Vector Machine (GMM-SVM) and GMM Universal Background Model (GMM-UBM);

generating metadata based on the audio file, wherein the metadata includes language spoken by each speaker; and

segmenting the audio file at optimal points without contextual or linguistic interruptions between segments, based on the number of speakers, the accent of each speaker, and the metadata.

2. The method of claim 1 , further comprising identifying that the audio file contains confidential information, limiting the pool of eligible candidates for providing the transcription services to include only those who have been identified as having passed a confidentiality clearance sufficient for the relevant audio file, and selecting at least one candidate from the pool of eligible candidates.

3. The method of claim 1 , wherein a first speaker within the audio file is speaking English, a second speaker within the audio file is speaking French, and the audio file is segmented into a first segment and a second segment, such that the first segment includes the speech by the first speaker that is assigned to a first transcription provider capable of English transcription, and the second segment includes the speech by the second speaker that is assigned to a second transcription provider capable of French transcription.

4. The method of claim 1 , wherein the analyzing of the audio file further utilizes at least one of the following techniques: tonal context, linguistic context, voice activity, Linear Predictive Coding (LPC), Perceptional Linear Predictive Coefficients (PLP), Mel-Frequency Cepstral Coefficients (MFCC), Linear Prediction Cepstral Coefficients (LPCC), Wavelet Based Features and Non-Negative Matrix Factorization features.

5. The method of claim 1 , wherein at least one of the optimal points comprises a transition between a first speaker and a second speaker.

6. The method of claim 5 , wherein at least one of the optimal points comprises a shift in recording location.

7. The method of claim 1 , wherein at least one of the optimal points comprises a change in spoken language.

8. The method of claim 1 , wherein the analyzing the at least one audio file further comprises: employing a deep learning technique, and further comprising forwarding a first segment of the segments to a first transcription provider and forwarding a second segment of the segments to a second transcription provider.

9. The method of claim 8 , wherein the deep learning technique includes as least one of: a neural network algorithm, decision tree learning, clustering, homomorphic filtering, wideband reducing filtering, and sound wave anti-aliasing algorithms.

10. A system for segmenting an audio file at optimal points that allow efficient transcribing, comprising:

a processing circuitry; and

a memory, the memory containing instructions that, when executed by the processing circuitry, configure the system to:

analyze an audio file, wherein the analyzing includes (i) identifying number of speakers in the audio file based on generating a signature for each voice determined to be unique within the audio file, and (ii) identifying accent of each speaker based on at least one of a Gaussian mixture model (GMM), a GMM Support Vector Machine (GMM-SVM) and GMM Universal Background Model (GMM-UBM);

generate metadata based on the audio file, wherein the metadata includes language spoken by each speaker; and

segmenting the audio file at optimal points without contextual or linguistic interruptions between segments, based on the number of speakers, the accent of each speaker, and the metadata.

11. The system of claim 10 , wherein the system is further configured to identify that the audio file contains confidential information, limit the pool of eligible candidates for providing the transcription services to include only those who have been identified as having passed a confidentiality clearance sufficient for the relevant audio file, and select at least one candidate from the pool of eligible candidates.

12. The system of claim 10 , wherein at least one of the optimal points comprises a transition between a first speaker and a second speaker.

13. The system of claim 12 , wherein at least one of the optimal points comprises a shift in recording location.

14. The system of claim 10 , wherein at least one of the optimal points comprises a change in spoken language.

15. The system of claim 10 , wherein the analyzing the at least one audio file further comprises: employ a deep learning technique, and further comprising forwarding a first segment of the segments to a first transcription provider and forwarding a second segment of the segments to a second transcription provider.

16. The system of claim 12 , wherein the deep learning technique includes as least one of: a neural network algorithm, decision tree learning, clustering, homomorphic filtering, wideband reducing filtering, and sound wave anti-aliasing algorithms.