Echo canceller for non-linear circuits

• US 5,796,819 A
• Filed: 07/24/1996
• Issued: 08/18/1998
• Est. Priority Date: 07/24/1996
• Status: Expired due to Term

• Current Assignee: Ericsson, Inc. (Telefonaktiebolaget LM Ericsson)
• Original Assignee: Ericsson, Inc. (Telefonaktiebolaget LM Ericsson)
• Inventors: Romesburg, Eric Douglas
• Primary Examiner(s): Hong, Harry S.
• Assistant Examiner(s): SAINT SURIN, JACQUES M
• Application Number:
  US08/685,495
• Time in Patent Office: 755 Days
• Field of Search: 379/410, 379/411, 379/406, 379/402, 379/390, 379/391, 379/407, 379/408, 379/409, 379/388, 379/389, 364/724.19, 370/291, 381/71.4, 381/71.11, 381/73.1, 381/93
• US Class Current: 379/406.09
• CPC Class Codes: H04B 7/015 Reducing echo effects
• CPC Class Codes: H04M 9/082 using echo cancellers echo cancellers per se H04B3/23

Abstract

A system and a method for echo suppression which use two or more microphones for beamforming of an echo signal corresponding to an incoming speech signal that is output by a loudspeaker. The outputs of the microphones are appropriately filtered and then linearly combined to cancel the echo signal, even if the incoming speech signal has been non-linearly distorted. Also, the microphones may be selectively positioned or pointed to allow the linear cancellation of ambient noise.

Claims

1. In an audio circuit comprising a microphone and a loudspeaker, said microphone detecting a speech signal from a near-end user, said loudspeaker receiving a far-end speech signal and generating a corresponding echo signal that is also detected by said microphone, a method of cancelling the echo signal comprising the steps of:

providing at least one other microphone in said audio circuit for detecting said echo signal, said at least one other microphone also detecting said near-end speech signal;

estimating a plurality of acoustic transfer functions in a plurality of adaptive filters each having a plurality of coefficients, at least one of said adaptive filters using the far-end speech signal as a reference signal for adjusting its coefficients;

filtering the outputs of one or more of said microphones using the coefficients of one or more of said adaptive filters; and

combining the filtered and any unfiltered microphone outputs so as to substantially cancel said echo signal while substantially preserving said near-end speech signal.

2. The method of claim 1 wherein said filtering step comprises filtering at least one microphone output in a fixed filter whose coefficients are copied from one or more of said adaptive filters.

3. The method of claim 1 wherein said filtering step comprises filtering at least one microphone output in an adaptive filter whose coefficients are adapted using the far-end speech signal as a reference signal.

4. The method of claim 1 wherein said filtering step comprises filtering at least one microphone output in an adaptive filter which uses the output of another adaptive filter as a reference signal.

5. The method of claim 1 wherein said audio circuit includes first and second microphones, the echo signal from said loudspeaker to said first and second microphones having acoustic transfer functions H₁ and H₂, respectively, and wherein the method comprises the steps of:

estimating H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating H₂ in a second adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said second adaptive filter;

filtering the output of said first microphone in a first fixed filter using the coefficients of said second adaptive filter;

filtering the output of said second microphone in a second fixed filter using the coefficients of said first adaptive filter; and

subtracting the output of said second fixed filter from the output of said first fixed filter.

6. The method of claim 5 further comprising the step of positioning said first and second microphones relative to said near-end user and said loudspeaker such that said first microphone receives a substantially higher level of said near-end speech signal than said second microphone, and said second microphone receives a substantially higher level of said echo signal than said first microphone.

7. The method of claim 5 further comprising the step of positioning said first and second microphones substantially equidistant from said loudspeaker so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

8. The method of claim 1 wherein said audio circuit includes first and second microphones, the echo signal from said loudspeaker to said first and second microphones having acoustic transfer functions H₁ and H₂, respectively, and wherein the method comprises the steps of:

estimating H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating H₂ in a second adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said second adaptive filter;

filtering the output of said second microphone in a fixed filter which estimates H₁ /H₂ using the coefficients of said first and second adaptive filters; and

subtracting the output of said fixed filter from the output of said first microphone.

9. The method of claim 8 further comprising the step of positioning said first and second microphones relative to said near-end user and said loudspeaker such that said first microphone receives a substantially higher level of said near-end speech signal than said second microphone, and said second microphone receives a substantially higher level of said echo signal than said first microphone.

10. The method of claim 8 further comprising the step of positioning said first and second microphones substantially equidistant from said loudspeaker so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

11. The method of claim 1 wherein said audio circuit includes first and second microphones, the echo signal from said loudspeaker to said first and second microphones having acoustic transfer functions H₁ and H₂, respectively, and wherein the method comprises the steps of:

estimating 1/H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating 1/H₂ in a second adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said second adaptive filter;

filtering the output of said first microphone in said first adaptive filter;

filtering the output of said second microphone in said second adaptive filter; and

subtracting the output of said second adaptive filter from the output of said first adaptive filter.

12. The method of claim 11 further comprising the step of positioning said first and second microphones relative to said near-end user and said loudspeaker such that said first microphone receives a substantially higher level of said near-end speech signal than said second microphone, and said second microphone receives a substantially higher level of said echo signal than said first microphone.

13. The method of claim 11 further comprising the step of positioning said first and second microphones substantially equidistant from said loudspeaker so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

14. The method of claim 1 wherein said audio circuit includes first and second microphones, the echo signal from said loudspeaker to said first and second microphones having acoustic transfer functions H₁ and H₂, respectively, and wherein the method comprises the steps of:

estimating H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating 1/H₂ in a second adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said second adaptive filter;

filtering the output of said second microphone in said second adaptive filter;

filtering the output of said second adaptive filter in a fixed filter using the coefficients of said first adaptive filter; and

subtracting the output of said fixed filter from the output of said first microphone.

15. The method of claim 14 further comprising the step of positioning said first and second microphones relative to said near-end user and said loudspeaker such that said first microphone receives a substantially higher level of said near-end speech signal than said second microphone, and said second microphone receives a substantially higher level of said echo signal than said first microphone.

16. The method of claim 14 further comprising the step of positioning said first and second microphones substantially equidistant from said loudspeaker so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

17. The method of claim 1 wherein said audio circuit includes first and second microphones, the echo signal from said loudspeaker to said first and second microphones having acoustic transfer functions H₁ and H₂, respectively, and wherein the method comprises the steps of:

estimating H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating H₁ /H₂ in a second adaptive filter using the output of said first adaptive filter as a reference signal for adapting the coefficients of said second adaptive filter;

filtering the output of said second microphone in said second adaptive filter; and

subtracting the output of said second adaptive filter from the output of said first microphone.

18. The method of claim 17 further comprising the step of positioning said first and second microphones relative to said user and said loudspeaker such that said first microphone receives a substantially higher level of said near-end speech signal than said second microphone, and said second microphone receives a substantially higher level of said echo signal than said first microphone.

19. The method of claim 17 further comprising the step of positioning said first and second microphones substantially equidistant from said loudspeaker so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

20. The method of claim 1 wherein said audio circuit includes first, second and third microphones, the echo signal from said loudspeaker to said first, second and third microphones having acoustic transfer functions H₁, H₂ and H₅, respectively, and wherein the method comprises the steps of:

estimating H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating H₂ in a second adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said second adaptive filter;

estimating H₅ in a third adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said third adaptive filter;

filtering the output of said first microphone in a first pair of fixed filters using the coefficients of said second and third adaptive filters, respectively;

filtering the output of said second microphone in a second pair of fixed filters using the coefficients of said first and third adaptive filters, respectively;

filtering the output of said third microphone in a third pair of fixed filters using the coefficients of said first and second adaptive filters, respectively;

multiplying the output of said first pair of fixed filters by a constant (c) in a first multiplier, where 0≦

c≦

1;

multiplying the output of said second pair of fixed filters by a constant (1-c) in a second multiplier; and

subtracting the output of said third pair of fixed filters from the outputs of said first and second multipliers.

21. The method of claim 20 wherein the value of (c) is adjusted so as to minimize noise detected by said microphones.

22. The method of claim 20 wherein the value of (c) is adjusted so as to vary the sensitivity of said microphones to said near-end speech signal.

23. The method of claim 1 wherein said audio circuit includes first, second and third microphones, the echo signal from said loudspeaker to said first, second and third microphones having acoustic transfer functions H₁, H₂ and H₅, respectively, and wherein the method comprises the steps of:

estimating H₁ in a first adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said first adaptive filter;

estimating H₂ in a second adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said second adaptive filter;

estimating H₅ in a third adaptive filter using said far-end speech signal as a reference signal for adapting the coefficients of said third adaptive filter;

filtering the output of said first microphone in a first fixed filter using the coefficients of said second adaptive filter;

filtering the output of said second microphone in a second fixed filter using the coefficients of said first adaptive filter;

multiplying the output of said first fixed filter by a constant (c) in a first multiplier, where 0≦

c≦

1;

multiplying the output of said second fixed filter by a constant (1-c) in a second multiplier;

adding the outputs of said first and second fixed filters in an adder;

filtering the output of said adder in a third fixed filter using the coefficients of said third adaptive filter;

filtering the output of said third microphone in a pair of fixed filters using the coefficients of said first and second adaptive filters, respectively; and

subtracting the output of said pair of fixed filters from the output of said third fixed filter.

24. The method of claim 23 wherein the value of (c) is adjusted so as to minimize noise detected by said microphones.

25. The method of claim 23 wherein the value of (c) is adjusted so as to vary the sensitivity of said microphones to said near-end speech signal.

26. The method of claim 1 wherein said audio circuit is part of a hands-free accessory for a wireless telephone.

27. The method of claim 1 wherein said adaptive filters use a Gradient Descent or Least Squares Estimation (LSE) algorithm to estimate said acoustic transfer functions.

28. The method of claim 1 wherein the far-end speech signal is distorted in said loudspeaker and/or in an amplifier or a digital-to-analog converter (DAC) connected to said loudspeaker.

29. The method of claim 1 wherein said adaptive filters are finite impulse response (FIR) filters.

30. A circuit for cancelling an echo signal from an audio source comprising:

a first microphone for detecting said echo signal along a first path defined by a first acoustic transfer function H₁ ;

a second microphone for detecting said echo signal along a second path defined by a second acoustic transfer function H₂ ;

a first adaptive filter for estimating H₁ ;

a second adaptive filter for estimating H₂ ;

a first fixed filter for filtering the output of said first microphone using the estimate of H₂ ;

a second fixed filter for filtering the output of said second microphone using the estimate of H₁ ; and

means for subtracting the output of said second fixed filter from the output of said first fixed filter so as to cancel said echo signal.

31. The circuit of claim 30 wherein the coefficients of said first fixed filter are copied from the coefficients of said second adaptive filter, and wherein the coefficients of said second fixed filter are copied from the coefficients of said first adaptive filter.

32. The circuit of claim 30 wherein said first and second microphones are positioned substantially equidistant from said audio source so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

33. The circuit of claim 30 wherein said echo signal is generated from an incoming signal received by said audio source, and wherein said first and second adaptive filters use said incoming signal as a reference signal for estimating H₁ and H₂, respectively.

34. The circuit of claim 33 wherein said echo signal is a distorted version of said incoming signal.

35. A circuit for cancelling an echo signal from an audio source comprising:

a first microphone for detecting said echo signal along a first path defined by a first acoustic transfer function H₁ ;

a second microphone for detecting said echo signal along a second path defined by a second acoustic transfer function H₂ ;

a first adaptive filter for estimating H₁ ;

a second adaptive filter for estimating H₂ ;

a fixed filter for filtering the output of said second microphone using an acoustic transfer function H₁ /H₂ based on the estimates of H₁ and H₂ ; and

means for subtracting the output of said fixed filter from the output of said first microphone so as to cancel said echo signal.

36. The circuit of claim 35 wherein the coefficients of said first and second adaptive filters are copied for use by said fixed filter.

37. The circuit of claim 35 wherein said first and second microphones are positioned substantially equidistant from said audio source so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

38. The circuit of claim 35 wherein said echo signal is generated from an incoming signal received by said audio source, and wherein said first and second adaptive filters use said incoming signal as a reference signal for estimating H₁ and H₂, respectively.

39. The circuit of claim 38 wherein said echo signal is a distorted version of said incoming signal.

40. A circuit for cancelling an echo signal from an audio source comprising:

a first microphone for detecting said echo signal along a first path defined by a first acoustic transfer function H₁ ;

a second microphone for detecting said echo signal along a second path defined by a second acoustic transfer function H₂ ;

a first adaptive filter for estimating 1/H₁ and for filtering the output of said first microphone;

a second adaptive filter for estimating 1/H₂ and for filtering the output of said second microphone; and

means for subtracting the output of said second adaptive filter from the output of said first adaptive filter so as to cancel said echo signal.

41. The circuit of claim 40 wherein said first and second microphones are positioned substantially equidistant from said audio source so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

42. The circuit of claim 40 wherein said echo signal is generated from an incoming signal received by said audio source, and wherein said first and second adaptive filters use said incoming signal as a reference signal for estimating 1/H₁ and 1/H₂, respectively.

43. The circuit of claim 42 wherein said echo signal is a distorted version of said incoming signal.

44. A circuit for cancelling an echo signal from an audio source comprising:

a first microphone for detecting said echo signal along a first path defined by a first acoustic transfer function H₁ ;

a second microphone for detecting said echo signal along a second path defined by a second acoustic transfer function H₂ ;

a first adaptive filter for estimating H₁ ;

a second adaptive filter for estimating 1/H₂ and for filtering the output of said second microphone;

a fixed filter for filtering the output of said second adaptive filter using the estimate of H₁ ; and

means for subtracting the output of said fixed filter from the output of said first microphone so as to cancel said echo signal.

45. The circuit of claim 44 wherein the coefficients of said fixed filter are copied from the coefficients of said first adaptive filter.

46. The circuit of claim 44 wherein said first and second microphones are positioned substantially equidistant from said audio source so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

47. The circuit of claim 44 wherein said echo signal is generated from an incoming signal received by said audio source, and wherein said first and second adaptive filters use said incoming signal as a reference signal for estimating H₁ and 1/H₂, respectively.

48. The circuit of claim 47 wherein said echo signal is a distorted version of said incoming signal.

49. A circuit for cancelling an echo signal from an audio source comprising:

a first microphone for detecting said echo signal along a first path defined by a first acoustic transfer function H₁ ;

a second microphone for detecting said echo signal along a second path defined by a second acoustic transfer function H₂ ;

a first adaptive filter for estimating H₁ ;

a second adaptive filter for estimating H₁ /H₂ and for filtering the output of said second microphone; and

means for subtracting the output of said second adaptive filter from the output of said first microphone so as to cancel said echo signal.

50. The circuit of claim 49 wherein said first and second microphones are positioned substantially equidistant from said audio source so as to also suppress ambient noise which is substantially equally received by said first and second microphones.

51. The circuit of claim 49 wherein said echo signal is generated from an incoming signal received by said audio source, said first adaptive filters uses said incoming signal as a reference signal for estimating H₁, and said second adaptive filter uses the output of said first adaptive filter for estimating H₁ /H₂.

52. The circuit of claim 51 wherein said echo signal is a distorted version of said incoming signal.

53. In a speech communications system including a loudspeaker which generates an echo signal from an incoming signal, an echo cancellation circuit comprising:

a plurality of microphones for detecting said echo signal from said loudspeaker;

means for estimating a plurality of acoustic transfer functions using the incoming signal as a reference signal;

means for filtering one or more of the outputs of said microphones using one or more of the estimated acoustic transfer functions; and

means for combining the filtered outputs of the microphones so as to cancel said echo signal.

54. The circuit of claim 53 wherein:

said microphones comprise a first, second and third microphones for detecting said echo signal along first, second and third paths, respectively, defined by first, second and third acoustic transfer functions H₁, H₂ and H₅, respectively;

said estimating means comprises first, second and third adaptive filters for estimating H₁, H₂ and H₅, respectively;

said filtering means comprises;

a first pair of fixed filters for filtering the output of said first microphone using the estimates of H₂ and H₅, respectively;

a second pair of fixed filters for filtering the output of said second microphone using the estimates of H₁ and H₅, respectively; and

a third pair of fixed filters for filtering the output of said third microphone using the estimate of H₁ and H₂, respectively; and

said combining means comprises;

a first multiplier for multiplying the output of said first pair of fixed filters by a constant (c), where 0≦

c≦

1;

a second multiplier for multiplying the output of said second pair of fixed filters by a constant (1-c); and

a subtractor for subtracting the output of said third pair of fixed filters from the outputs of said first and second multipliers.

55. The circuit of claim 54 wherein the value of (c) is adjusted so as to minimize noise detected by said microphones.

56. The circuit of claim 54 wherein the value of (c) is adjusted so as to vary the sensitivity of said microphones to a user signal.

57. The circuit of claim 53 wherein:

said microphones comprise a first, second and third microphones for detecting said echo signal along first, second and third paths, respectively, defined by first, second and third acoustic transfer functions H₁, H₂ and H₅, respectively;

said estimating means comprises first, second and third adaptive filters for estimating H₁, H₂ and H₅, respectively; and

said filtering means and said combining means comprise;

a first fixed filter for filtering the output of said first microphone using the estimate of H₂ ;

a second fixed filter for filtering the output of said second microphone using the estimate of H₁ ;

a first multiplier for multiplying the output of said first fixed filter by a constant (c), where 0≦

c≦

1;

a second multiplier for multiplying the output of said second fixed filter by a constant (1-c);

an adder for adding the outputs of said first and second multipliers;

a third fixed filter for filtering the output of said adder using the estimate of H₅ ;

a pair of fixed filters for filtering the output of said third microphone using the estimates of H₁ and H₂, respectively; and

a subtractor for subtracting the output of said pair of fixed filters from the output of said third fixed filter.

58. The circuit of claim 57 wherein the value of (c) is adjusted so as to minimize noise detected by said microphones.

59. The circuit of claim 57 wherein the value of (c) is adjusted so as to vary the sensitivity of said microphones to a user signal.

60. The circuit of claim 53 wherein said system comprises a hands-free accessory for a wireless telephone.

61. The circuit of claim 53 wherein said incoming signal is non-linearly distorted in said loudspeaker or in an amplifier or a digital-to-analog converter (DAC) connected to said loudspeaker.

62. The circuit of claim 53 wherein said estimating means comprises a plurality of adaptive filters.

63. The circuit of claim 62 wherein said adaptive filters use a Gradient Descent or Least Squares Estimation (LSE) algorithm to estimate said acoustic transfer functions.

64. The circuit of claim 62 wherein said filtering means comprises a plurality of fixed filters each having coefficients which are copied from one or more of said adaptive filters.

65. The circuit of claim 64 wherein each of said adaptive filters and each of said fixed filters is a finite impulse response (FIR) filter.

66. The circuit of claim 53 wherein said combining means comprises:

means for multiplying one or more of the filtered outputs of said microphones by a corresponding weight function; and

means for adding the filtered and/or weighted outputs of said microphones.

67. The circuit of claim 66 wherein each weight function is selected so as to minimize noise detected by said microphones.

68. The circuit of claim 66 wherein each weight function is selected so as to vary the sensitivity of said microphones to a user signal.