Dynamic Beamformer Processing for Acoustic Echo Cancellation in Systems with High Acoustic Coupling

• US 20120243698A1
• Filed: 03/22/2012
• Published: 09/27/2012
• Est. Priority Date: 03/22/2011
• Status: Active Grant

• Current Assignee: mh acoustics LLC
• Original Assignee: mh acoustics LLC
• Inventors: Elko, Gary W., Gaensler, Tomas F., Diethorn, Eric J., Meyer, Jens M.
• Granted Patent: US 8,942,382 B2
• Time in Patent Office: Days
• US Class Current: 381/66
• CPC Class Codes: H04M 9/082 using echo cancellers echo cancellers per se H04B3/23

Abstract

Near-end equipment for a communication channel with far-end equipment. The near-end equipment includes at least one loudspeaker, at least two microphones, a beamformer, and an echo canceller. The communication channel may be in one of a number of communication states including Near-End Only state, Far-End Only state, and Double-Talk state. In one embodiment, when the echo canceller determines that the communication channel is in either the Far-End Only state or the Double-Talk state, the beamformer is configured to generate a nearfield beampattern signal that directs a null towards a loudspeaker. When the echo canceller detects the Near-End Only state, the beamformer is configured to generate a farfield beampattern signal that optimizes reception of acoustic signals from the near-end audio source. Using different beamformer processing for different communication states allows echo cancellation processing to be more successful at reducing echo in the signal transmitted to the far-end equipment.

Claims

1. Near-end equipment for a near-end audio source communicating with a far-end equipment via an audio communication channel, the near-end equipment comprising:

a beamformer configured to receive two or more microphone signals generated by two or more local microphones and generate one or more beampattern signals; and

an echo canceller configured to (i) receive the one or more beampattern signals from the beamformer and an incoming signal from the far-end equipment that is applied to one or more local loudspeakers, (ii) determine an active communication state for the audio communication channel, (iii) select a beampattern for the beamformer based on the determined active communication state, and (iv) apply echo-cancellation processing to the beampattern signal corresponding to the selected beampattern for the beamformer, to generate an outgoing signal to be transmitted to the far-end equipment.

2. The invention of claim 1, wherein:

possible active communication states comprise a near-end only talking state, a far-end only talking state, and a double-talk state;

the one or more beampatterns comprise a first beampattern and a second beampattern different from the first beampattern;

when the echo canceller determines that the active communication state is the far-end only talking state or the double-talk state, the echo canceller selects the first beampattern; and

when the echo canceller determines that the active communication state is the near-end only talking state, the echo canceller selects the second beampattern.

3. The invention of claim 2, wherein:

the first beampattern has at least one null corresponding to the one or more local loudspeakers; and

the second beampattern optimizes reception of acoustic signals from the near-end audio source.

4. The invention of claim 1, wherein:

the beamformer simultaneously generates at least two beampattern signals corresponding to at least two different beampatterns; and

the echo canceller selects one of the at least two beampattern signals as the beampattern signal corresponding to the selected beampattern.

5. The invention of claim 1, wherein:

the beamformer is configured to generate a first beampattern signal for a first communication state; and

the echo canceller is configured to (i) detect a change from the first communication state to a second communication state different from the first communication state and (ii) control the beamformer to be re-configured to generate a second beampattern signal for the second communication state and different from the first beampattern signal.

6. The invention of claim 1, wherein the near-end equipment further comprises the one or more local loudspeakers and the two or more local microphones.

7. The invention of claim 6, wherein at least one of the local microphones is (i) located inside housing of one of the local loudspeakers and (ii) configured to be used to reduce audio coupling between the one or more local loudspeakers and the two or more local microphones during at least one determined active communication state.

8. The invention of claim 1, wherein the beamformer is configured to concurrently generate different beampattern signals for two or more different frequency bands of the microphone signals.

9. The invention of claim 8, wherein the beamformer is configured to concurrently generate:

a first beampattern signal for a first frequency band of the microphone signals having a null directed towards a first loudspeaker; and

a second, different beampattern signal for a second, different frequency band of the microphone signals having a null directed towards a second, different loudspeaker.

10. The invention of claim 1, wherein:

the one or more local loudspeakers comprises at least a first local loudspeaker and a second local loudspeaker; and

the echo canceller is configured to generate a beampattern signal having a first null directed towards the first local loudspeaker and a second null directed towards the second local loudspeaker.

11. The invention of claim 1, wherein the echo canceller is further configured to control the one or more local loudspeakers based on the determined active communication state.

12. The invention of claim 11, wherein:

there are two or more local loudspeakers; and

the echo canceller is configured to control the two or more local loudspeakers to generate a loudspeaker beampattern based on the determined active communication state.

13. The invention of claim 12, wherein the echo canceller is configured to steer the loudspeaker beampattern away from the two or more local microphones for at least one determined active communication state.