Low complexity echo compensation system
View Patent ↗An echo reduction system includes a signal analysis filter that converts an input into sub-band signals. A signal down-sampling circuit down-samples the sub-band signals at a first down-sampling rate. An echo analysis filter converts a loudspeaker signal into echo sub-band signal that are further processed by an echo down-sampling circuit. The circuit down-samples the echo sub-band signals at a second down-sampling rate to generate down-sampled echo sub-band signals. An echo compensation filter folds the down-sampled echo sub-band signals with an estimated impulse response of a loudspeaker-room-input system. A second echo down-sampling circuit down-samples the folded down-sampled echo sub-band signals at a third down-sampling rate to generate estimated echo sub-band signals. The first down-sampling rate is equal to the product of the second and third down-sampling rates.
1. A method for reducing echoes in an input signal containing an audio signal and an echo signal, the echo signal generated by a loudspeaker signal in a loudspeaker-room-microphone system, the method comprising:
converting a portion of the input signal to microphone sub-band signals;
down-sampling the microphone sub-band signals by a first down-sampling rate;
converting the loudspeaker signal to echo sub-band signals;
down-sampling the echo sub-band signals by a second down-sampling rate to generate down-sampled echo sub-band signals;
folding the down-sampled echo sub-band signals with an estimated impulse response of the loudspeaker-room-microphone system to generate the folded down-sampled echo sub-band signals;
down-sampling the folded down-sampled echo sub-band signals by a third down-sampling rate to generate estimated echo sub-band signals, where the first down-sampling rate is equal to the product of the second and third down-sampling rates
subtracting for each sub-band, the estimated echo sub-band signal from the corresponding microphone sub-band signal to obtain error sub-band signals; and
up-sampling and synthesizing the error sub-band signals to generate an echo compensated audio signal.
2. The method according to claim 1 , where generating the estimated echo sub-band signal comprises adapting filter coefficients of an echo compensation filter based on the stored down-sampled echo sub-band signals at a rate equal to the first down-sampling rate.
3. The method according to claim 2 , where the filter coefficients are adapted using a normalized least mean square process.
4. The method according to claim 1 , where the second down-sampling rate is equal to an integer value ranging from 32 to 128.
5. The method according to claim 1 , where the third down-sampling rate is equal to an integer value ranging from 2 to 4.
6. The method according to claim 1 , where the first down-sampling rate is equal to the product of the second down-sampling rate and the third down-sampling rate.
7. The method according to claim 1 , where converting the microphone signal to microphone sub-band signals comprises filtering the signals using a bank of square-root Hann window filters.
8. The method according to claim 1 , where converting the loudspeaker signal to echo sub-band signals comprises filtering the signals using a bank of square-root Hann window filters.
9. The method according to claim 1 , where synthesizing the error sub-band signals comprises filtering the signals using a bank of square-root Hann window filters.
10. The method according to claim 1 , where the error sub-band signals are up-sampled at rate equal to the first down-sampling rate.
11. The method according to claim 1 , where converting the microphone signal to microphone sub-band signals comprises filtering the signals using a bank of Hann window filters, each Hann window filter raised to the power of a first rational number, and where synthesizing the error sub-band signals comprises filtering the signals using a bank of Hann window filters, each Hann window filter raised to the power of a second rational number.
12. The method according to claim 11 , where the sum of the first and second rational numbers is equal to 1.
13. The method according to claim 12 , where the first rational number ranges from 0.50 to 0.95 and the second rational number ranges from 0.05 to 0.50.
14. A method for reducing echoes in a plurality of input signals, each in put signal corresponding to a channel, and containing an audio signal and an echo signal, the echo signals generated by a loudspeaker signal in a loudspeaker-room-microphone system, the method comprising:
converting at least a portion of each input signal to respective microphone sub-band signals;
down-sampling the microphone sub-band signals for each microphone by a first down-sampling rate;
converting the loudspeaker signal to echo sub-band signals for each channel;
down-sampling the echo sub-band signals for each channel by a second down-sampling rate to generate down-sampled echo sub-band signals, the second down-sampling rate less than the first down-sampling rate;
folding the down-sampled echo sub-band signals for each channel with an estimated impulse response of the loudspeaker-room-microphone system to generate the folded down-sampled echo sub-band signals;
down-sampling the folded down-sampled echo sub-band signals for each channel by a third down-sampling rate to generate estimated echo sub-band signals, where the first down-sampling rate is equal to the product of the second and third down-sampling rates;
for each sub-band within each channel, subtracting the estimated echo sub-band signal from the corresponding microphone sub-band signal to obtain error sub-band signals for each channel;
beamforming the error sub-band signals for each channel to generate beamformed error sub-band signals; and
up-sampling and synthesizing the error sub-band signals to generate an echo compensated audio signal.
15. An echo reduction system for reducing echoes in a microphone signal containing an audio signal and an echo signal, the echo signal generated by a loudspeaker signal in a loudspeaker-room-microphone system, the system comprising:
a signal analysis filter bank configured to convert at least a portion of a microphone signal to microphone sub-band signals;
a signal down-sampling circuit configured to down-sample the microphone sub-band signals by a first down-sampling rate
an echo analysis filter bank configured to convert a loudspeaker signal to a plurality of echo sub-band signals;
a first echo down-sampling circuit configured to down-sample the echo sub-band signals by a second down-sampling rate to generate down-sampled echo sub-band signals, the second down-sampling rate less than the first down-sampling rate;
an echo compensation filter configured to fold the down-sampled echo sub-band signals with an estimated impulse response of the loudspeaker-mom-microphone system to generate the folded down-sampled echo sub-band signals; and
a second echo down-sampling circuit configured to down-sample the folded down-sampled echo sub-band signals by a third down-sampling rate to generate estimated echo sub-band signals, where the first down-sampling rate is equal to the product of the second and third down-sampling rates.
16. The system of claim 15 further comprising:
a combining circuit configured to subtract for each sub-band, the estimated echo sub-band signal from the corresponding microphone sub-band signal to obtain error sub-band signals;
an up-sampling circuit configured to up-sample the error sub-band signals by an up-sampling rate equal to the first down-sampling rate; and
a synthesis filter bank configured to synthesize the up-sampled error sub-band signals to generate an echo compensated audio signal.
17. The system of claim 15 further comprising at least one of a residual echo suppression filter and a noise reduction filter configured to filter the error sub-band signals.
18. The system of claim 15 , where the signal analysis filter bank, comprises a plurality of square-root Hann window filters.
19. The system of claim 15 , where the echo analysis filter bank comprises a plurality of square-root Hann window filters.
20. The system of claim 16 , where the synthesis filter bank comprises a plurality of square-root Hann window filters.
21. The system of claim 16 , where the up-sampling circuit up-samples the error sub-band signals at rate equal to the first down-sampling rate.
22. The system of claim 16 , where the signal analysis filter bank comprises a plurality of Hann window filters, each Hann window filter raised to the power of a first rational number, and the echo analysis filter bank comprises a plurality of Hann window filters, each Hann window filter raised to the power of a second rational number.
23. The system of claim 22 , where the sum of the first and second rational numbers is equal to 1.
24. The method according to claim 23 , where the first rational number ranges from 0.50 to 0.95 and the second rational number ranges from 0.05 to 0.50.
25. The system of claim 15 , where the signal analysis filter bank, comprises a plurality of window-based filters.
26. The system of claim 25 , where the plurality of window-based filters provides an output value of zero when an input value is outside a pre-defined interval.