Method of capturing constant echo path information in a full duplex speakerphone
View Patent ↗A method of improving convergence of an echo canceller in a full duplex speakerphone, wherein the echo canceller includes LEC (Line Echo Canceller) and AEC (Acoustic Echo Canceller) portions, comprising the steps of capturing coefficients during operation, storing the coefficients, and utilizing the stored coefficients as default values during start-up of a subsequent call. According to the method of the present invention, the amount of echo cancellation necessary to provoke a save of the coefficients decreases with time when it is not achieved by the system despite the presence of a reference signal (i.e. speakerphone signal).
1. A method of operating an echo canceller which receives a reference signal and converges to an estimated echo signal of a received input signal via feedback of an error signal, said echo canceller being characterized by predetermined filter coefficients when converged, said method comprising the steps of:
a) detecting power levels of said received input signal and said error signal;
b) calculating an Echo Return Loss Enhancement value based on said power levels of said received input signal and said error signal, wherein said Echo Return Loss Enhancement value is expressed in decibels as
ERLE(dB)=10 log 10 [Power(ReceivedSignal)/Power(ErrorSignal)]; and
c) in the event said Echo Return Loss Enhancement value exceeds a threshold then storing said predetermined filter coefficients for subsequent use by said echo canceller, increasing said threshold by a predetermined factor and re-executing steps a) to c), and if said Echo Return Loss Enhancement value does not exceed the threshold, determining whether voice is present in said input signal and if so reducing said threshold by a predetermined factor and re-executing steps a) to c).
2. The method of claim 1 , wherein said threshold is reduced only in the event said threshold exceeds a predetermined amount.
3. A use of the method according to claim 2 in a full duplex speakerphone, wherein said power levels of said received input signal and said error signal are detected during a call in progress, and said predetermined filter coefficients are stored for use by said echo canceller during calls subsequent to said call in progress.
4. A use of the method according to claim 3 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
5. A use of the method according to claim 2 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
6. The method of claim 1 , wherein said threshold is increased by approximately 3 dB in the event said Echo Return Loss Enhancement value exceeds said threshold.
7. A use of the method according to claim 6 in a full duplex speakerphone, wherein said power levels of said received input signal and said error signal are detected during a call in progress, and said predetermined filter coefficients are stored for use by said echo canceller during calls subsequent to said call in progress.
8. A use of the method according to claim 7 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
9. A use of the method according to claim 6 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
10. The method of claim 1 , wherein said threshold is decreased by approximately 3 dB every approximately 5 seconds in the event said Echo Return Loss Enhancement value does not exceed said threshold and voice is present in said input signal.
11. A use of the method according to claim 10 in a full duplex speakerphone, wherein said power levels of said received input signal and said error signal are detected during a call in progress, and said predetermined filter coefficients are stored for use by said echo canceller during calls subsequent to said call in progress.
12. A use of the method according to claim 11 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
13. A use of the method according to claim 10 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
14. The method of claim 1 , wherein said threshold is at least approximately 21 dB.
15. A use of the method according to claim 14 in a full duplex speakerphone, wherein said power levels of said received input signal and said error signal are detected during a call in progress, and said predetermined filter coefficients are stored for use by said echo canceller during calls subsequent to said call in progress.
16. A use of the method according to claim 15 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
17. A use of the method according to claim 14 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
18. A use of the method according to claim 1 in a full duplex speakerphone, wherein said power levels of said received input signal and said error signal are detected during a call in progress, and said predetermined filter coefficients are stored for use by said echo canceller during calls subsequent to said call in progress.
19. A use of the method according to claim 18 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
20. A use of the method according to claim 1 in a conferencing system having multiple microphones in respective look directions under control of a beamformer, wherein said power levels of said received input signal and said error signal are detected for each of said look directions, and said predetermined filter coefficients are stored for use by said echo canceller in response to changes in said look directions.
21. The method of claim 1 , wherein the threshold is maintained above a minimum value that is greater than zero.