System and method for combined non-linear and late echo suppression

• US 10,481,831 B2
• Filed: 10/02/2017
• Issued: 11/19/2019
• Est. Priority Date: 10/02/2017
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Schalk-Schupp, Ingo, Faubel, Friedrich, Buck, Markus, Desiraju, Naveen Kumar, Wolff, Tobias
• Primary Examiner(s): Gauthier, Gerald
• Application Number:
  US15/722,367
• Publication Number:
  US 20190102108A1
• Time in Patent Office: 778 Days
• Field of Search: 324 7633, 345440, 375224, 37940608, 37940612, 381 231, 381 58, 381 66, 381 711, 381 716, 381 86, 381 942, 381317, 381303, 435 612, 455 6711, 455434, 455440, 600 13, 600500, 704219, 704226, 607 9
• CPC Class Codes: G06F 13/10 Program control for peripheral devices G06F13/14 - G06F13/42 take precedence
• CPC Class Codes: G06F 3/0659 Command handling arrangements, e.g. command buffers, queues, command scheduling
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A method, computer program product, and computer system for receiving, by a computing device, an input signal. A first power spectral density estimate may be generated for a linear reverberant component associated with the input signal. A second power spectral density estimate may be generated for a non-linear reverberant component associated with the input signal. A power spectral density estimate may be generated by combining the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component. One or more parameters for at least one of the linear reverberant component and the non-linear reverberant component may be updated. One or more undesired signal components in an output signal resulting from the input signal may be reduced via residual echo suppression based upon, at least in part, updating the one or more parameters.

Claims

1. A computer-implemented method comprising:

receiving, by a computing device, an input signal;

generating a first power spectral density estimate for a linear reverberant component associated with the input signal, wherein the linear reverberant component comprises a first late reverberant component;

generating a second power spectral density estimate for a non-linear reverberant component associated with the input signal, wherein the non-linear reverberant component comprises a second late reverberant component;

generating a power spectral density estimate by combining the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component;

updating one or more parameters for at least one of the linear reverberant component and the non-linear reverberant component; and

reducing, via residual echo suppression, one or more undesired signal components in an output signal resulting from the input signal based upon, at least in part, updating the one or more parameters.

2. The computer-implemented method of claim 1 wherein updating the one or more parameters includes alternating updates of the one or more parameters between two frames.

3. The computer-implemented method of claim 1 wherein updating the one or more parameters includes alternating updates of the one or more parameters in one frame.

4. The computer-implemented method of claim 1 wherein updating the one or more parameters includes:

comparing two or more power spectral density errors; and

utilizing a power spectral density error of the two or more power spectral density errors resulting in a larger improvement of the output signal.

5. The computer-implemented method of claim 1 wherein at least one parameter of the one or more parameters is used to generate both the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component.

6. The computer-implemented method of claim 1 further comprising ceasing adaptation of an acoustic echo cancellation filter when the one or more undesired signal components are present.

7. The computer-implemented method of claim 1 further comprising utilizing a power ratio between the linear reverberant component and the estimate for the non-linear reverberant component as control information.

8. A computer program product residing on a computer readable storage medium having a plurality of instructions stored thereon which, when executed across one or more processors, causes at least a portion of the one or more processors to perform operations comprising:

receiving an input signal;

generating a first power spectral density estimate for a linear reverberant component associated with the input signal, wherein the linear reverberant component comprises a first late reverberant component;

generating a second power spectral density estimate for a non-linear reverberant component associated with the input signal, wherein the non-linear reverberant component comprises a second late reverberant component;

generating a power spectral density estimate by combining the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component;

updating one or more parameters for at least one of the linear reverberant component and the non-linear reverberant component; and

reducing, via residual echo suppression, one or more undesired signal components in an output signal resulting from the input signal based upon, at least in part, updating the one or more parameters.

9. The computer program product of claim 8 wherein updating the one or more parameters includes alternating updates of the one or more parameters between two frames.

10. The computer program product of claim 8 wherein updating the one or more parameters includes alternating updates of the one or more parameters in one frame.

11. The computer program product of claim 8 wherein updating the one or more parameters includes:

comparing two or more power spectral density errors; and

utilizing a power spectral density error of the two or more power spectral density errors resulting in a larger improvement of the output signal.

12. The computer program product of claim 8 wherein at least one parameter of the one or more parameters is used to generate both the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component.

13. The computer program product of claim 8 wherein the operations further comprise ceasing adaptation of an acoustic echo cancellation filter when the one or more undesired signal components are present.

14. The computer program product of claim 8 wherein the operations further comprise utilizing a power ratio between the linear reverberant component and the estimate for the non-linear reverberant component as control information.

15. A computing system including one or more processors and one or more memories configured to perform operations comprising:

receiving an input signal;

generating a first power spectral density estimate for a linear reverberant component associated with the input signal, wherein the linear reverberant component comprises a first late reverberant component;

generating a second power spectral density estimate for a non-linear reverberant component associated with the input signal, wherein the non-linear reverberant component comprises a second late reverberant component;

generating a power spectral density estimate by combining the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component;

updating one or more parameters for at least one of the linear reverberant component and the non-linear reverberant component; and

reducing, via residual echo suppression, one or more undesired signal components in an output signal resulting from the input signal based upon, at least in part, updating the one or more parameters.

16. The computing system of claim 15 wherein updating the one or more parameters includes alternating updates of the one or more parameters between two frames.

17. The computing system of claim 15 wherein updating the one or more parameters includes alternating updates of the one or more parameters in one frame.

18. The computing system of claim 15 wherein updating the one or more parameters includes:

comparing two or more power spectral density errors; and

utilizing a power spectral density error of the two or more power spectral density errors resulting in a larger improvement of the output signal.

19. The computing system of claim 15 wherein at least one parameter of the one or more parameters is used to generate both the first power spectral density estimate for the linear reverberant component and the second power spectral density estimate for the non-linear reverberant component.

20. The computing system of claim 15 wherein the operations further comprise at least one of ceasing adaptation of an acoustic echo cancellation filter when the one or more undesired signal components are present and utilizing a power ratio between the linear reverberant component and the estimate for the non-linear reverberant component as control information.