System and method of testing voice signals in a telecommunication system

• US 6,504,905 B1
• Filed: 04/09/1999
• Issued: 01/07/2003
• Est. Priority Date: 04/09/1999
• Status: Active Grant

• Current Assignee: Qwest Communications International Incorporated (Lumen Technologies, Inc.)
• Original Assignee: Qwest Communications International Incorporated (Lumen Technologies, Inc.)
• Inventors: Tsai, Louis K., Wu, Ying-li
• Primary Examiner(s): Tsang, Fan
• Assistant Examiner(s): Sing, Simon P.
• Application Number:
  US09/289,495
• Time in Patent Office: 1,369 Days
• Field of Search: 379/26.01, 379/26.02, 379/27.02, 379/27.04, 379/27.08, 379/41, 379/67.1, 379/68, 379/88.08, 379/88.09, 379/88.22, 379/1.02, 379/10.01, 379/10.02, 379/10.03, 379/21, 379/22.02
• US Class Current: 379/26.01
• CPC Class Codes: G10L 25/69 for evaluating synthetic or decoded voice signals
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/22 Arrangements for supervision, monitoring or testing
• CPC Class Codes: H04M 3/533 Voice mail systems

Abstract

A system and method of testing voice signals in a telecommunication system, such as a voice messaging system (VMS), are provided. A test signal is generated to elicit a response from the telecommunication system, and a voice signal produced by the telecommunication system in response to the test signal is detected. To evaluate the detected voice signal, the system accesses stored reference phrases, where one or more designated reference phrases can be combined to form an expected voice signal. The detected voice signal is compared with at least one designated reference phrase to determine if the detected voice signal substantially corresponds to the expected voice signal. In a preferred embodiment, the voice signal and reference phrases are divided into component frames of substantially equal time durations, then compared frame by frame to determine if the detected voice signal and the expected voice signal substantially correspond.

Claims

1. A method for testing voice signals in a telecommunication system, the method comprising:

generating a test signal to elicit a response from the telecommunication system;

detecting a voice signal produced by the telecommunication system in response to the test signal;

accessing stored reference phrases, wherein designated reference phrases being combined to form and expected voice signal; and

comparing the detected voice signal with said expected voice signal determine if the detected voice signal corresponds to the expected voice signal.

2. The method of claim 1, wherein generating a test signal includes generating a DTMF signal.

3. The method of claim 1, wherein generating a test signal includes generating a voice signal.

4. The method of claim 1, wherein detecting a voice signal includes detecting a voice signal produced by a voice messaging system (VMS).

5. The method of claim 1, further comprising dividing the voice signal and the at least one reference phrase into component frames of substantially equal time durations.

6. The method of claim 5, further comprising determining an average acoustic intensity of each voice signal frame and of each reference phrase frame.

7. The method of claim 6, further comprising

determining an average acoustic intensity of the voice signal and normalizing the average acoustic intensity of each voice signal frame by the average acoustic intensity of the entire voice signal; and

determining an average acoustic intensity of the reference phrase and normalizing the average acoustic intensity of each reference phrase frame by the average acoustic intensity of the entire reference phrase.

8. The method of claim 5, further comprising determining the number of zero crossings in each voice signal frame and in each reference phrase frame.

9. The method of claim 5, further comprising comparing each voice signal frame with the reference phrase frame of a corresponding time interval.

10. The method of claim 5, further comprising determining a matching percentage between the voice signal frames and the reference phrase frames.

11. The method of claim 10, further comprising comparing the matching percentage with a threshold value.

12. A system for testing voice signals in a telecommunication system, the system comprising:

a processor for controlling the generation of a test signal to elicit a response from the telecommunication network;

a telecommunication interface in communication with the processor and the telecommunication system to generate the test signal and to detect a voice signal produced by the telecommunication system in response to the test signal;

a storage device in communication with the processor for storing the detected voice signal and a plurality of reference phrases, wherein designated reference phrases being combined to form an expected voice signal; and

voice recognition software operable with the processor for comparing the detected voice signal with said expected voice signal to determine if the detected voice signal substantially corresponds to the expected voice signal.

13. The system of claim 12, wherein the telecommunication system includes a voice messaging system (VMS).

14. The system of claim 12, wherein the test signal comprises a DTMF signal.

15. The system of claim 12, wherein the test signal comprises a voice signal.

16. The system of claim 12, wherein the detected voice recognition software is operable to divide the voice signal and at least one reference phrase into component frames of equal time durations.

17. The system of claim 16, wherein the voice recognition software is operable to determine an average acoustic intensity of each voice signal frame and of each reference phrase frame.

18. The system of claim 17, wherein the voice recognition software is operable to determine an average acoustic intensity of the voice signal and normalize the average acoustic intensity of each voice signal frame by the average acoustic intensity of the entire voice signal, and to determine an average acoustic intensity of the reference phrase and normalize the average acoustic intensity of each reference phrase frame by the average acoustic intensity of the entire reference phrase.

19. The system of claim 16, wherein the voice recognition software is operable to determine the number of zero crossings in each voice signal frame and in each reference phrase frame.

20. The system of claim 16, wherein the voice recognition software is operable to compare each voice signal frame with the reference phrase frame of a corresponding time interval.

21. The system of claim 16, wherein the voice recognition software is operable to determine a matching percentage between the voice signal frames and the reference phrase frames.

22. The system of claim 21, wherein the voice recognition software is operable to compare the matching percentage with a threshold value.