Automatic speech analysis

• US 20070250318A1
• Filed: 04/25/2006
• Published: 10/25/2007
• Est. Priority Date: 04/25/2006
• Status: Active Grant

• Current Assignee: Nice Ltd
• Original Assignee: Nice Systems Limited (Nice Ltd)
• Inventors: Waserblat, Moshe, Eilam, Barak
• Granted Patent: US 8,725,518 B2
• Time in Patent Office: Days
• US Class Current: 704/236
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/14 using statistical models, e.g. Hidden Markov Models [HMMs] G10L15/18 takes precedence

Abstract

A system for providing automatic quality management regarding a level of conformity to a specific accent, including, a recording system, a statistical model database with statistical models representing speech data of different levels of conformity to a specific accent, a speech analysis system, a quality management system. Wherein the recording system is adapted to record one or more samples of a speakers speech and provide it to the speech analysis system for analysis, and wherein the speech analysis system is adapted to provide a score of the speakers speech samples to the quality management system by analyzing the recorded speech samples relative to the statistical models in the statistical model database.

Claims

1. A system for providing automatic quality management regarding a level of conformity to a specific accent, comprising:

a recording system;

a statistical model database with statistical models representing speech data of different levels of conformity to a specific accent;

a speech analysis system;

a quality management system;

wherein said recording system is adapted to record one or more samples of a speakers speech and provide it to said speech analysis system for analysis;

wherein said speech analysis system is adapted to provide a score of the speakers speech samples to said quality management system by analyzing said recorded speech samples relative to the statistical models in said statistical model database.

2. A system according to claim 1, wherein analysis of said recorded speech samples by said speech analysis system comprises:

preparing a statistical model from said speech samples; and

comparing it to other statistical models in said statistical model database to determine the closest model and a level of conformity to the closest model.

3. A system according to claim 1, wherein said quality management system is adapted to provide feedback to said speaker.

4. A system according to claim 1, wherein said quality management system is adapted to provide feedback to a supervisor.

5. A system according to claim 1, wherein said quality management system is adapted to provide feedback regarding the performance of the speaker based on the provided score and previous provided scores of the speaker.

6. A system according to claim 1, wherein said feedback further comprises specific speech segments with a maximum level of deviation from the model which most conforms to a specific accent.

7. A method of providing automated feedback to a speaker regarding conformity of their accent to a specific accent, comprising:

recording one or more samples of the speakers speech;

producing a statistical model from said speech samples;

comparing said produced statistical model to previously prepared statistical models of speech samples with different levels of conformity to a specific accent;

determining the model that conforms the best to said produced statistical model and a level of conformity to the best conforming model; and

providing as feedback a score for said speech samples regarding its conformity to the specific accent based on said determining.

8. A method according to claim 7, wherein said feedback further comprises specific speech segments with a maximum level of deviation from the best conforming model.

9. A method according to claim 7, wherein said feedback takes into account scores of the speaker from previous recordings.

10. A method of automatic accent identification for quality management comprising:

creating one or more statistical accent models representing accents from one or more collections of training speech data;

inputting a speech signal for analysis;

preparing a statistical speech model representing the input speech signal;

comparing the statistical speech model with said one or more statistical accent models;

calculating a score resulting from the comparison of said statistical speech model with each statistical accent model;

determining a closest statistical accent model to said statistical speech model; and

providing the scores to a quality management system to provide feedback.

11. A method according to claim 10, wherein said determining is performed substantially in real-time.

12. A method according to claim 10, further comprising giving indication regarding segments of said speech signal which do not conform to the determined closest statistical accent module.

13. A method according to claim 12, wherein said indication is given substantially in real-time.

14. A method according to claim 10, wherein said quality management system provides indication regarding the quality of the accent of a user.

15. A method according to claim 10, further comprising notifying a user if the determined closest statistical accent changes during a conversation.

16. A method according to claim 10, wherein said determining further determines other characteristics of a speaker.

17. A method according to claim 10, wherein said one or more statistical models are updated based on speech signals from groups of substantially equal scoring users.