Self calibration of envelope-based acoustic echo cancellation
View Patent ↗Systems and methods for envelope-based acoustic echo cancellation in a communication device are provided. In exemplary embodiments, a primary acoustic signal is received via a microphone of the communication device, and a far-end signal is received via a receiver. Frequency analysis is performed on the primary acoustic signal and the far-end acoustic signal to obtain frequency sub-bands. An echo mask based on magnitude envelopes of the primary and far-end acoustic signals for each frequency sub-band is generated. A noise mask based on at least the primary acoustic signal for each frequency sub-band may also be generated. A combination of the echo mask and noise mask may then be applied to the primary acoustic signal to generate a masked signal. The masked signal is then output.
1. A method for acoustic echo cancellation, comprising:
receiving a primary acoustic signal via a primary microphone and a far-end signal via a receiver;
adapting a plurality of echo gains based on a coherence value between the primary acoustic signal and the far-end signal;
adjusting a speed of adaptation for each echo gain of the plurality of echo gains, the adjusting the speed of adaptation being based on detecting that the coherence value for a sub-band satisfies a threshold;
applying the plurality of the adapted echo gains to the far-end signal to generate a predicted echo signal;
configuring an echo mask based on the predicted echo signal and the primary acoustic signal;
applying the configured echo mask to the primary acoustic signal to generate a masked signal; and
outputting the masked signal.
2. The method of claim 1 , wherein the coherence value is measured in sub-bands of the primary acoustic signal and the far-end signal.
3. The method of claim 1 , wherein the configured echo mask is applied to selected sub-bands of the primary acoustic signal.
4. The method of claim 1 , wherein adapting a plurality of echo gains includes determining a normalized coherence between the primary acoustic signal and the far-end signal.
5. The method of claim 1 , further comprising performing frequency analysis on the primary acoustic signal and the far-end acoustic signal to obtain frequency sub-bands for the primary and far-end acoustic signals.
6. The method of claim 1 , wherein adjusting a speed of adaptation for each echo gain comprises updating each echo gain based on one or more adaptation control cues.
7. The method of claim 6 , wherein the one or more adaptation control cues comprise an adaptive far-end noise floor.
8. The method of claim 6 , wherein the one or more adaptation control cues comprise a noise floor.
9. The method of claim 6 , wherein the one or more adaptation control cues comprise an echo dominance cue associated with the coherence value.
10. The method of claim 1 , further comprising synthesizing the masked signal into time domain for outputting.
11. The method of claim 1 , wherein the echo mask is applied as a gain mask.
12. A method for acoustic echo cancellation, comprising:
receiving a primary acoustic signal via a primary microphone and a far-end signal via a receiver;
adapting an echo gain based on a coherence value between the primary acoustic signal and the far-end signal;
adjusting a speed of the echo gain adaptation based on a first ratio between the primary acoustic signal and the far-end signal;
applying the adapted echo gain to the far-end signal to generate a predicted echo signal;
configuring an echo mask based on the predicted echo signal and the primary acoustic signal;
applying the echo mask to the primary acoustic signal to generate a masked signal; and
outputting the masked signal.
13. The method of claim 12 , wherein the first ratio is a power ratio.
14. The method of claim 12 , wherein adjusting the speed of the echo gain adaptation includes detecting that a global coherence value satisfies a threshold.
15. The method of claim 12 , wherein adjusting the speed of the echo gain adaptation includes detecting that the coherence value for a sub-band satisfies a threshold.
16. The method of claim 12 , wherein adjusting the speed of the echo gain adaptation includes reducing the speed of the echo gain adaptation during doubletalk.
17. A method for acoustic echo cancellation, comprising:
receiving a primary acoustic signal via a primary microphone and a far-end signal via a receiver;
determining a coherence value between the primary acoustic signal and the far-end acoustic signal, wherein the coherence value is determined from a cochlea filter output;
adapting a plurality of echo gains based on the coherence value;
adjusting a speed of adaptation for each echo gain of the plurality of echo gains, the adjusting the speed of adaptation being based on detecting that the coherence value for a sub-band satisfies a threshold;
applying the plurality of the adapted echo gains to the far-end signal to generate a predicted echo signal;
configuring an echo mask based on the predicted echo signal and the primary acoustic signal;
applying the configured echo mask to the primary acoustic signal to generate a masked signal; and
outputting the masked signal.
18. The method of claim 17 , further comprising obtaining frequency sub-bands for the primary and far-end acoustic signals, wherein the coherence value is determined based on the cochlea filter output for all sub-bands.
19. The method of claim 17 , further comprising performing frequency analysis on the primary acoustic signal and the far-end acoustic signal to obtain frequency sub-bands for the primary acoustic signal and far-end acoustic signal.
20. The method of claim 17 , wherein the coherence value is determined based on a phase difference between the primary acoustic signal and the far-end acoustic signal.
21. The method of claim 17 , wherein the coherence value is determined based on an exponential function.
22. The method of claim 20 , wherein the phase difference is computed using decimation of the cochlea filter output.
23. A system for acoustic echo cancellation in a communication device, comprising:
acoustic sensors configured to receive a primary acoustic signal and a far-end acoustic signal;
an echo mask generator configured to:
generate an echo mask based on a predicted echo signal and the primary acoustic signal;
adapt a plurality of echo gains based on a coherence value between the primary acoustic signal and the far-end acoustic signal;
adjust an adaptation speed for each echo gain of the plurality of echo gains, the adjustment being based on detecting that the coherence value for a sub-band satisfies a threshold; and
apply the plurality of the adapted echo gains to the far-end signal to generate a predicted echo signal;
a modifier module configured to apply the generated echo mask to the primary acoustic signal to generate a masked signal; and
an output device configured to output the masked signal.
24. The system of claim 23 , wherein the echo mask generator is further configured to adjust the adaption speed of each echo gain of the plurality of echo gains based on a first ratio between the primary acoustic signal and the far-end acoustic signal.
25. The system of claim 23 , wherein the echo mask generator is further configured to determine the coherence value based on a phase difference between the primary acoustic signal and the far-end acoustic signal.