Multi-input channel and multi-output channel echo cancellation

• US 20060002546A1
• Filed: 06/10/2005
• Published: 01/05/2006
• Est. Priority Date: 06/30/2004
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Stokes, Jack W. III, Platt, John C.
• Granted Patent: US 7,925,007 B2
• Time in Patent Office: Days
• US Class Current: 379/406.60
• CPC Class Codes: H04M 9/082 using echo cancellers echo cancellers per se H04B3/23

Abstract

An echo cancellation technique that can process multi-input microphone signals with only a small increase in the overall CPU consumption compared to implementing the algorithm for a single channel microphone signal. Furthermore, the invention provides an architecture that provides for echo cancellation for multiple applications in parallel with only a small increase in CPU consumption compared to a single instance of echo cancellation with a single microphone input and multi-output channel playback.

Claims

1. A system for canceling acoustic echoes for multiple inputs, comprising:

a plurality of C microphone inputs;

a plurality of N speaker playback signals, each decomposed into a plurality M of subbands;

a plurality of adaptive filters, equal to the number of microphone inputs C times a number of subbands M, whose outputs are used to reduce echo in the plurality of microphone inputs, wherein each adaptive filter comprises a weight update and wherein each adaptive filter processes N speaker playback signals; and

a single gain computation module for each subband M employed to compute the gain vector to be used in computing the weight update of said plurality of adaptive filters.

2. The system of claim 1, wherein the plurality of adaptive filters comprises recursive least squares (RLS) adaptive filters.

3. The system of claim 1, wherein the plurality of adaptive filters comprises fast recursive least squares (FRLS) adaptive filters.

4. The system of claim 1, wherein the plurality of adaptive filters comprises normalized least mean squares (NLMS) adaptive filters.

5. The system of claim 1 wherein the gain computation is independent of the plurality of the microphone inputs.

6. A method for canceling acoustic echoes for multiple inputs, comprising:

decomposing each of a plurality N of speaker playback signals into a plurality M of subband signals;

computing a single gain vector for all N speaker signals of each subband;

inputting a plurality of microphone inputs C;

applying a plurality of adaptive filters to the plurality of subband signals and the plurality of C microphone inputs using said single gain for each subband;

reducing the echo in the plurality of C microphone inputs using the outputs of the plurality of adaptive filters.

7. The method of claim 6 wherein computing a single gain vector for all N speaker signals of each subband, comprises:

using a stored inverse correlation matrix and the plurality N of speaker playback signals each decomposed into a plurality M of subband signals to compute the gain vector; and

updating the inverse correlation matrix; and

storing the updated inverse correlation matrix to be used for subsequent gain vector computations.

8. The method of claim 6 wherein each of the plurality of adaptive filters is selected in a round robin sequence;

and an inverse correlation matrix of each selected adaptive filter is regularized.

9. The method as set forth in claim 8, wherein regularizing further comprises ensuring that the plurality of adaptive filters are stable.

10. The method of claim 6, wherein the plurality of adaptive filters comprises recursive least squares (RLS) adaptive filters.

11. The method of claim 8, wherein regularizing the inverse correlation matrix comprises:

inverting the inverse correlation matrix to produce a correlation matrix;

regularizing the correlation matrix to produce a regularized correlation matrix; and

obtaining a regularized inverse correlation matrix by inverting the regularized correlation matrix.

12. The method of claim 6, wherein applying a plurality of adaptive filters comprises:

adapting a weight vector for each of the plurality of adaptive filters;

de-aggregating the weight vector to produce a plurality of single-channel filter coefficient vectors such that there is one single-channel filter coefficient vector for each playback channel; and

applying the plurality of single-channel filter coefficient vectors to the plurality of subband signals.

13. The method of claim 6, wherein the captured signal comprises an acoustic signal.

14. The method of claim 6, wherein the captured signal comprises an electrical signal.

15. The method of claim 6, wherein the plurality of playback signals is a stereo signal.

16. A computer-readable medium having computer-executable instructions for performing the method recited in claim 6.

19. The computer-implemented method of claim 15, wherein the reduced echoes in the captured signal are computed by multiple applications and wherein several copies of the updated gain vector are sent to each of the applications.

17. A computer-implemented method for eliminating echoes of a plurality of playback signals in a captured signal, comprising:

decomposing a plurality of playback signals into subbands;

computing a gain vector for each subband using inverse correlation matrices, wherein the gain vector and the inverse correlation matrices are independent of the captured signal;

processing the subbands using an adaptive filter for each of the subbands using the gain vector computed for the subband; and

reducing echoes in the captured signal using the outputs of the adaptive filters.

18. The computer-implemented method of claim 17, wherein more than one captured signal is input.

20. The computer-implemented method of claim 17 wherein the Kalman vector and the inverse correlation matrices are only updated once for each output device.