Bi-magnitude processing framework for nonlinear echo cancellation in mobile devices

Patent metadata

Field	Value
Publication number	US-10045137-B2
Application number	US-201715639263-A
Country	US
Kind code	B2
Filing date	Jun 30, 2017
Priority date	Jun 30, 2016
Publication date	Aug 7, 2018
Grant date	Aug 7, 2018

Abstract

Techniques of performing acoustic echo cancellation involve providing a bi-magnitude filtering operation that performs a first filtering operation when a magnitude of an incoming audio signal to be output from a loudspeaker is less than a specified threshold and a second filtering operation when the magnitude of the incoming audio signal is greater than the threshold. The first filtering operation may take the form of a convolution between the incoming audio signal and a first impulse response function. The second filtering operation may take the form of a convolution between a nonlinear function of the incoming audio signal and a second impulse response function. For such a convolution, the bi-magnitude filtering operation involves providing, as the incoming audio signal, samples of the incoming audio signal over a specified window of time. The first and second impulse response functions may be determined from an input signal input into a microphone.

First claim

What is claimed is:

1. A method, comprising:
receiving, by processing circuitry configured to reduce acoustic echo in an audio system including a loudspeaker and a microphone, at the loudspeaker of the audio system, an audio signal from a source location remote from the audio system;
performing, by the processing circuitry, a comparison operation on a magnitude of the audio signal and a threshold magnitude to produce a comparison result;
in response to the comparison result indicating that the magnitude of the audio signal is less than the threshold magnitude:
performing, by the processing circuitry, a first filtering operation on an input signal into the microphone of the audio system to produce a first filtered input signal; and
transmitting, by the processing circuitry, the first filtered input signal to the source location;

in response to the comparison result indicating that the magnitude of the audio signal is greater than the threshold magnitude:
performing, by the processing circuitry, a second filtering operation on the input signal into the microphone of the audio system to produce a second filtered input signal, the second filtered input signal being different from the first filtered input signal; and
transmitting, by the processing circuitry, the second filtered input signal to the source location.

2. The method as in claim 1 , wherein performing the first filtering operation includes:
sampling the audio signal over a specified window of time to produce a windowed audio signal;
generating a first impulse response function based on the input signal input into the microphone; and
generating a convolution of the windowed audio signal and the first impulse response function to produce a first filtered incoming signal.

3. The method as in claim 2 , wherein generating the first impulse response function based on the input signal input into the microphone includes producing, as the first impulse response function, an impulse response function that optimizes a power of a residual signal, the residual signal being equal to a difference between the input signal input into the microphone and a convolution of the windowed audio signal and the generic impulse response function, and
wherein transmitting the first filtered input signal to the source location includes sending, as the first filtered input signal, a power-optimized residual signal, the power-optimized residual signal being equal to the difference between the input signal input into the microphone and the convolution of the windowed audio signal and the first impulse response function.

4. The method as in claim 2 , wherein the first impulse response function is a weighted sum of harmonics, each of the harmonics having a frequency equal to a multiple of a fundamental frequency.

5. The method as in claim 2 , wherein performing the first filtering operation further includes:
after a specified amount of time after generating the first impulse response function, generating another, first impulse response function.

6. The method as in claim 2 ,
wherein the windowed audio signal includes a plurality of samples of the audio signal, each of the plurality of the samples being a value of the audio signal at a time that occurs within the specified window of time, and
wherein performing the comparison operation on the magnitude of the audio signal and a threshold magnitude includes:
generating an absolute value of each of the plurality of samples of the audio signal to produce a plurality of absolute values; and
producing, as the magnitude of the audio signal, the largest of the plurality of absolute values.

7. The method as in claim 2 , wherein performing the second filtering operation includes:
generating a nonlinear function of the windowed audio signal;
generating a second impulse response function based on the input signal input into the microphone; and
generating a convolution of (i) the nonlinear function of the windowed audio signal and (ii) the first impulse response function to produce a first filtered incoming signal.

8. The method as in claim 7 , wherein generating the nonlinear function of the windowed audio signal includes squaring the magnitude of the windowed audio signal.

9. The method as in claim 7 , wherein generating the second impulse response function based on the input signal input into the microphone includes producing, as the second impulse response function, a generic impulse response function that optimizes a power of a residual signal, the residual signal being equal to a difference between the input signal input into the microphone and a convolution of the nonlinear function of the windowed audio signal and the generic impulse response function, and
wherein transmitting the second filtered input signal to the source location includes sending, as the second filtered input signal, a power-optimized residual signal, the power-optimized residual signal being equal to the difference between the input signal input into the microphone and the convolution of the nonlinear function of the windowed audio signal and the second impulse response function.

10. The method as in claim 7 , wherein performing the first filtering operation includes:
generating a first nonlinear function of the windowed audio signal, the first nonlinear function being different from the nonlinear function;
generating a first impulse response function based on the input signal input into the microphone; and
generating a convolution of (i) the first nonlinear function of the windowed audio signal and (ii) the first impulse response function to produce a first filtered incoming signal.

11. The method as in claim 6 , wherein performing the second filtering operation further includes:
after a specified amount of time after generating the second impulse response function, generating another, third response function.

12. The method as in claim 1 , wherein the method further comprises:
in response to the comparison result indicating that the magnitude of the audio signal is greater than the threshold magnitude and greater than a second threshold magnitude:
performing, by the processing circuitry, a third filtering operation on the input signal into the microphone of the audio system to produce a second filtered input signal, the third filtered input signal being different from the first filtered input signal and the second filtered input signal; and
transmitting, by the processing circuitry, the third filtered input signal to the source location.

13. A computer program product comprising a nontransitory storage medium, the computer program product including code that, when executed by processing circuitry configured to reduce acoustic echo in an audio system including a loudspeaker and a microphone, causes the processing circuitry to perform a method, the method comprising:
receiving, at the loudspeaker of the audio system, an audio signal from a source location remote from the audio system;
performing a comparison operation on a magnitude of the audio signal and a threshold magnitude to produce a comparison result;
in response to the comparison result indicating that the magnitude of the audio signal is less than the threshold magnitude:
performing a first filtering operation on an input signal into the microphone of the audio system to produce a first filtered input signal; and
transmitting the first filtered input signal to the source location;

in response to the comparison result indicating that the magnitude of the audio signal is greater than the threshold magnitude:
performing a second filtering operation on the input signal into the microphone of the audio system to p

Assignees

• Google Inc
• Google Llc

Inventors

• Skoglund Jan
• Huang Yiteng
• Luebs Alejandro

Classifications

• G01N29/4427

  with stored values, e.g. threshold values · CPC title

• H04M9/082 Primary

  using echo cancellers (echo cancellers per se H04B3/23) · CPC title

• H04B3/237

  using two adaptive filters, e.g. for near end and for end echo cancelling · CPC title

• H04R29/001 Primary

  for loudspeakers (H04R29/007 takes precedence) · CPC title

• G10L2021/02082

  the noise being echo, reverberation of the speech · CPC title