Suppression of echo signals and the like

• US 7,062,040 B2
• Filed: 09/20/2002
• Issued: 06/13/2006
• Est. Priority Date: 09/20/2002
• Status: Expired due to Term

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Agere Systems Incorporated (Broadcom, Inc.)
• Inventors: Faller, Christof
• Primary Examiner(s): Harold, Jefferey F.
• Application Number:
  US10/251,404
• Publication Number:
  US 20040057574A1
• Time in Patent Office: 1,362 Days
• Field of Search: 379/406.1, 379/406.03, 379/406.06, 379/406.08, 379/406.12, 379/406.14, 379/406.01, 379/406.11, 379/406.15
• US Class Current: 379/406.11
• CPC Class Codes: H04M 9/082 using echo cancellers echo cancellers per se H04B3/23

Abstract

In a microphone signal, the signal component corresponding to, e.g., echo is suppressed using an echo control scheme that estimates the spectral envelope of the echo signal, without having to estimate the waveform for the echo signal. In one embodiment, the input signal (to be applied to a loudspeaker) and the microphone signal are spectrally decomposed into multiple subbands, where echo suppression processing is independently performed on each subband. The echo control of the present invention can be implemented with substantially reduced (1) computational complexity and (2) phase sensitivity, as compared to traditional acoustic echo cancellation, in which the waveform for the echo signal is estimated.

Claims

1. A method for processing audio signals, comprising the steps of:

(a) receiving an input signal, wherein the input signal is applied to a speaker;

(b) receiving a microphone signal generated by a microphone;

(c) estimating the spectral envelope of the input signal;

(d) estimating the spectral envelope of the microphone signal;

(e) applying the estimated input signal envelope to a filter to generate an estimate of the spectral envelope of a signal component in the microphone signal, wherein the signal component corresponds to sound generated by the speaker in response to the input signal and picked up by the microphone; and

(f) suppressing the signal component in the microphone signal based on the estimated signal component envelope and the estimated microphone signal envelope to generate an output signal.

2. The method of claim 1, wherein the output signal is generated without estimating the waveform for the signal component.

3. The method of claim 1, wherein:

the estimated input signal envelope is generated by estimating one or more power levels of the input signal;

the estimated microphone signal envelope is generated by estimating one or more power levels of the microphone signal and

the estimated signal component envelope is generated by estimating one or more power levels of the signal component.

4. The method of claim 1, wherein:

step (c) comprises the steps of;

(1) dividing the input signal into a plurality of frequency subbands; and

(2) estimating the power level of the signal component in each of a first set of one or more frequency subbands;

step (d) comprises the steps of;

(1) dividing the microphone signal into a plurality of frequency subbands; and

(2) estimating the power level of the microphone signal in each of the first set of one or more frequency subbands;

step (e) comprises, for each frequency subband in the first set, the step of applying the estimated input signal power for the frequency subband to a corresponding subband filter to generate an estimated signal component power for the frequency subband;

step (f) comprises the steps of;

(1) suppressing the signal component in each frequency subband in the first set based on the corresponding estimated signal component power and the corresponding estimated microphone signal power to generate a component-suppressed subband signal for each frequency subband in the first set; and

(2) generating the output signal from the one or more component-suppressed subband signals.

5. The method of claim 4, wherein:

step (c) further comprises the step of estimating the waveform for the signal component in each of a second set of one or more frequency subbands of the microphone signal;

step (f) further comprises the step of performing acoustic echo cancellation (AEC) processing on each frequency subband in the second set to generate one or more AEC-processed subband signals, wherein AEC processing reduces the subband signal component by adaptively estimating an impulse response of the subband signal component in time; and

step (f) comprises the step of generating the output signal from the one or more component-suppressed subband signals and the one or more AEC-processed subband signals.

6. The method of claim 5, wherein both AEC processing and the suppression processing of step (f)(1) are applied to at least one frequency subband shared by the first and second sets.

7. The method of claim 6, wherein the AEC processing is applied prior to the suppression processing of step (f)(1) in each shared frequency subband.

8. The method of claim 4, wherein:

step (c)(1) comprises the step of applying the input signal to a filterbank to divide the input signal into the plurality of frequency subbands; and

step (d)(1) comprises the step of applying the microphone signal to a filterbank to divide the microphone signal into the plurality of frequency subbands; and

step (f)(2) comprises the step of applying each component-suppressed subband signal to an inverse filterbank to generate the output signal.

9. The method of claim 1, wherein the signal component comprises an echo signal.

10. The method of claim 1, wherein the filter is a linear filter.

11. The method of claim 1, wherein the filter is adaptively updated based on the estimated input signal envelope and the estimated microphone signal envelope.

12. The method of claim 1, wherein the output signal is generated without estimating the echo path impulse response in order to reduce echo.

13. An apparatus for processing audio signals, comprising:

(a) a first operator configured to estimate the spectral envelope of an input signal, wherein the input signal is also applied to a speaker;

(b) a second operator configured to estimate the spectral envelope of a microphone signal generated by a microphone;

(c) an estimation filter configured to estimate, based on the estimated input signal envelope, the spectral envelope of a signal component in the microphone signal, wherein the signal component corresponds to sound generated by the speaker in response to the input signal and picked up by the microphone; and

(d) a suppression module configured to suppress the signal component in the microphone signal based on the estimated component signal envelope and the estimated microphone signal envelope to generate an output signal.

14. The apparatus of claim 13, further comprising:

(a) a first filterbank configured to divide the input signal into a plurality of input subband signals corresponding to a plurality of frequency subbands;

(b) a second filterbank configured to divide the microphone signal into a plurality of microphone subband signals corresponding to the plurality of frequency subbands, wherein;

the first operator is configured to estimate the power level of each input subband signal;

the second operator is configured to estimate the power level of each microphone subband signal;

the estimation filter is configured to estimate, based on each estimated input subband signal power, a power level of a subband signal component in the microphone signal; and

the suppression module is configured to suppress the subband signal component in each microphone subband signal based on the estimated subband component signal power and the estimated subband microphone signal power; and

(c) an inverse filterbank configured to generate the output signal from the component-suppressed microphone subband signals.

15. The apparatus of claim 14, wherein the estimation filter and the echo suppression module function independently for each frequency subband.

16. The apparatus of claim 13, wherein the output signal is generated without estimating the waveform for the signal component.

17. The apparatus of claim 13, wherein the signal component comprises an echo signal.

18. The apparatus of claim 13, wherein the estimation filter is a linear filter.

19. The apparatus of claim 13, further comprising an estimation filter controller configured to adaptively update the estimation filter based on the estimated input and microphone signal envelopes.

20. The apparatus of claim 13, wherein the output signal is generated without estimating the echo path impulse response in order to reduce echo.