Method and apparatus for reducing echo and noise

• US 7,359,504 B1
• Filed: 12/03/2002
• Issued: 04/15/2008
• Est. Priority Date: 12/03/2002
• Status: Expired due to Fees

• Current Assignee: Plantronics, Inc.
• Original Assignee: Plantronics, Inc.
• Inventors: Weeks, William A., Reuss, Edward L.
• Primary Examiner(s): Kuntz; Curtis A
• Assistant Examiner(s): Jamal; Alexander
• Application Number:
  US10/309,695
• Time in Patent Office: 1,960 Days
• Field of Search: 37940601-40616
• US Class Current: 379/406.02
• CPC Class Codes: H04M 9/082 using echo cancellers echo cancellers per se H04B3/23

Abstract

The present invention provides a solution to the needs described above through a method and apparatus for reducing echo and noise. The apparatus includes a microphone array for receiving and audio signal, the audio signal including a voice signal component and a noise signal component. The apparatus further includes a voice processing path having an input coupled to the microphone array and a noise processing path having an input coupled to the microphone array. The voice processing path is adapted to detect voice signals and the noise processing path is adapted to detect noise signals. A first echo controller is coupled to the voice processing path and a second echo controller is coupled to the noise processing path. A noise reducer is coupled to the output of the first echo controller and second echo controller.

Claims

1. An apparatus for processing a signal comprising:

a microphone array for receiving an audio signal, the audio signal including a voice signal component and a noise signal component;

a voice processing path having an input coupled to the microphone array, wherein the voice processing path comprises a beamform voice processor adapted to detect voice signals;

a noise processing path having an input coupled to the microphone array, wherein the noise processing path comprises a beamform noise processor adapted to detect noise signals;

a first echo controller coupled to the voice processing path, wherein the first echo controller determines a near end speech talk state, a far-end speech talk state, a no speech talk state, or both a near-end and far-end speech talk state, wherein a first echo controller output comprises a voice signal with reduced echo;

a second echo controller coupled to the noise processing path, wherein the second echo controller determines the near end speech talk state, the far-end speech talk state, the no speech talk state, or both the near-end and far-end speech talk state, the second echo controller comprising;

an adaptive filter for generating an approximated echo signal in the audio signal; and

a subtractor for subtracting the approximated echo signal from an output of the noise processing path, wherein the output of the second echo controller is a noise signal with reduced echo;

a noise reducer coupled to the output of the first echo controller and second echo controller, wherein the noise reducer utilizes the noise signal with reduced echo and the voice signal with reduced echo in a noise reduction process to generate an output signal with reduced noise; and

a first voice activity detector, wherein the first activity detector receives a beamform voice processor output, beamform noise processor output, and the output signal with reduced noise from the noise reducer, and the first voice activity detector outputs a first voice/no voice signal to the noise reducer; and

a second voice activity detector, wherein the second activity detector receives a far end audio signal, and the second voice activity detector outputs a second voice/no voice signal to the beamform voice processor and the beamform noise processor.

2. The apparatus of claim 1, wherein the first echo controller comprises:

a first adaptive filter with adaptive weights for generating a first approximated echo signal in the audio signal; and

a first subtractor for subtracting the first approximated echo signal from an output of the voice processing path, wherein the output of the first echo controller is a voice signal with reduced echo.

3. The apparatus of claim 1 further comprising an analog to digital converter for processing the audio signal to generate a digitized signal.

4. The apparatus of claim 1, wherein the audio signal is a signal in a bi-directional communications system, and wherein the noise signal component comprises an acoustic echo signal or background noise.

5. The apparatus of claim 4, wherein the voice signal is generated by a near end user of a near-end communication device speaking into the microphone array, the acoustic echo signal is an echo signal generated by a speaker of the near end communication device broadcasting a voice signal generated by a far end user of a far-end communication device.

6. The apparatus of claim 1, further comprising a double talk detector coupled to the output of far end communication device and the output of the first voice activity detector and the second voice activity detector for determining when the near end user and far end user are simultaneously speaking, wherein the output of the double talk detector is provided to the first and second echo controller and the noise reducer.

7. The apparatus of claim 1, wherein the voice processing path, noise processing path, first echo controller, second echo controller, and noise reducer are implemented by a digital signal processor.

8. The apparatus of claim 1, wherein the apparatus is implemented in a hands-free communication device.

9. The apparatus of claim 1, wherein the apparatus is implemented in a mobile communication device.

10. The apparatus of claim 1, wherein the noise reducer utilizes a noise cancellation algorithm.

11. An apparatus for processing a signal comprising:

a means for receiving an audio signal, the audio signal including a voice signal component and a noise signal component;

a means adapted for detecting the voice signal component;

a means adapted for detecting the noise signal component;

a means for canceling a first echo component in the voice signal component to generate a voice signal component with reduced echo comprising determining a near end speech talk state, a far-end speech talk state, a no speech talk state, or both a near-end and far-end speech talk state;

a means for canceling a second echo component in the noise signal component to generate a noise signal component with reduced echo comprising determining the near end speech talk state, the far-end speech talk state, the no speech talk state, or both the near-end and far-end speech talk state;

a means for reducing noise in the audio signal using the voice signal component with reduced echo and the noise signal component with reduced echo to generate an optimized voice signal with reduced noise;

a first voice activity detecting means for receiving the voice signal component with reduced echo, the noise signal component with reduced echo, and the optimized voice signal with reduced noise, and for outputting a first voice/no voice signal to the noise reducer; and

a second voice activity detecting means for receiving a far end audio signal and outputting a second voice/no voice signal to the means adapted for detecting the voice signal component and the means adapted for detecting the noise signal component.

12. A communication device for use in a bi-directional communications system comprising:

a microphone array for receiving a near end audio signal, the audio signal including a voice signal component and a noise signal component;

a speaker for broadcasting to a near end user of the communication device a far-end audio signal which is generated by a far end user and received at the communication device; and

a signal processing circuit for attenuating an echo signal generated by the speaker detected by the microphone array and for attenuating background noise detected by the microphone array, wherein the signal processing circuit comprises a voice beamformer adapted to detect the voice signal component, a noise beamformer adapted to detect a noise signal component, a first echo controller coupled to the output of the voice beamformer wherein the first echo controller determines a near end speech talk state, a far-end speech talk state, a no speech talk state, or both a near-end and far-end speech talk state, wherein a first echo controller output comprises a voice signal with reduced echo, a second echo controller coupled to the output of the noise beamformer, wherein the second echo controller determines the near end speech talk state, the far-end speech talk state, the no speech talk state, or both the near-end and far-end speech talk state, and a noise reducer coupled to the output of the first echo controller and the second echo controller, wherein the second echo controller comprises;

an adaptive filter for generating an approximated echo signal in the audio signal; and

a subtractor for subtracting the approximated echo signal from an output of the noise beamformer, wherein the output of the second echo controller is a noise signal with reduced echo, and wherein the noise reducer utilizes the noise signal with reduced echo and the voice signal with reduced echo in a noise reduction process to generate an output signal with reduced noise;

a first voice activity detector, wherein the first activity detector receives a voice beamformer output, noise beamformer output, and the output signal with reduced noise from the noise reducer, and the first voice activity detector outputs a first voice/no voice signal to the noise reducer; and

a second voice activity detector, wherein the second activity detector receives a far end audio signal, and the second voice activity detector outputs a second voice/no voice signal to the voice beamformer and the nose beamformer.

13. The apparatus of claim 12, wherein the first echo controller comprises:

a first adaptive filter with adaptive weights for generating a first approximated echo signal in the audio signal; and

a first subtractor for subtracting the first approximated echo signal from an output of the voice beamformer, wherein the output of the first echo controller is a voice signal with reduced echo.

14. The apparatus of claim 12, wherein the signal processing circuit further comprises an analog to digital converter for processing the audio signal to generate a digitized signal.

15. The apparatus of claim 12, wherein the signal processing circuit further comprises a double talk detector that receives the far end audio signal and is coupled to the output of the first voice activity detector and the second voice activity detector, wherein the double talk detector determines when the near end user and far end user are simultaneously speaking, and wherein the output of the double talk detector is provided to the first and second echo controller and the noise reducer.

16. The apparatus of claim 12, wherein the signal processing circuit is implemented by a digital signal processor.

17. The apparatus of claim 12, wherein the noise reducer utilizes a noise cancellation algorithm.

18. A method for processing a signal to reduce undesired noise, the method comprising:

receiving an audio signal with a microphone array, the audio signal comprising one or more components;

providing the audio signal to a voice processing path comprising a voice beamform processor having an input coupled to the microphone array and adapting the voice processing path to detect voice signals;

providing the audio signal to a noise processing path comprising a noise beamform processor having an input coupled to the microphone array and adapting the noise processing path to detect noise signals;

canceling an acoustic echo component in the audio signal with a first echo controller coupled to the voice processing path comprising determining a near end speech talk state, a far-end speech talk state, a no speech talk state, or both a near-end and far-end speech talk state;

outputting a voice signal with reduced echo from the first echo controller;

canceling an acoustic echo component in the audio signal with a second echo controller coupled to the noise processing path, comprising;

determining the near end speech talk state, the far-end speech talk state, the no speech talk state, or both the near-end and far-end speech talk state;

generating an approximated echo signal in the audio signal; and

subtracting the approximated echo signal from a noise signal component on the noise processing path;

outputting a noise signal with reduced echo from the second echo controller;

processing the voice signal with reduced echo and the noise signal with reduced echo with a noise reducer coupled to the output of the first echo controller and the second echo controller to generate an output signal with reduced noise;

inputting to a first voice activity detector a voice beamform processor output, a noise beamform processor output, and the output signal with reduced noise;

outputting from the first voice activity detector a first voice/no voice signal to the noise reducer;

inputting to a second voice activity detector a far end audio signal; and

outputting from the second activity detector a second voice/no voice signal to the beamform voice processor and beamform noise processor.

19. The method of claim 18, wherein adapting the voice processing path to detect voice signals comprises:

weighting the audio signal with adaptive complex weighting coefficients selected to form one or more beams or nulls to detect voice signals.

20. The method of claim 18, wherein adapting the noise processing path to detect noise signals comprises:

weighting the audio signal with adaptive complex weighting coefficients selected to form one or more beams or nulls to detect noise signals.

21. The method of claim 18, wherein canceling an acoustic echo component in the audio signal with a first echo controller comprises:

generating a first approximated echo signal in the audio signal; and

subtracting the first approximated echo signal from an output of the first echo controller, wherein the output of the first echo controller is a voice signal with reduced echo.

22. The method of claim 18, further comprising processing the audio signal to generate a digitized signal.

23. The method of claim 18, further comprising determining whether the audio signal contains a voice signal corresponding to a near end user.

24. The method of claim 18, further comprising determining when a near end user and a far end user are simultaneously speaking.