NOISE-REDUCING DIRECTIONAL MICROPHONE ARRAY

• US 20090175466A1
• Filed: 03/09/2007
• Published: 07/09/2009
• Est. Priority Date: 02/05/2002
• Status: Active Grant

• Current Assignee: mh acoustics LLC
• Original Assignee: mh acoustics LLC
• Inventors: Gaensler, Tomas Fritz, Elko, Gary W., Meyer, Jens M.
• Granted Patent: US 8,942,387 B2
• Time in Patent Office: Days
• US Class Current: 381/94.200
• CPC Class Codes: G10L 2021/02166 Microphone arrays; Beamforming
• CPC Class Codes: G10L 21/0216 characterised by the method used for estimating noise
• CPC Class Codes: G10L 21/0264 characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques
• CPC Class Codes: H04R 1/326 for microphones H04R1/34 and H04R1/40 take precedence
• CPC Class Codes: H04R 2410/01 Noise reduction using microphones having different directional characteristics
• CPC Class Codes: H04R 2410/07 Mechanical or electrical reduction of wind noise generated by wind passing a microphone
• CPC Class Codes: H04R 2430/20 Processing of the output signals of the acoustic transducers of an array for obtaining a desired directivity characteristic H04R2203/12 takes precedence
• CPC Class Codes: H04R 2430/21 Direction finding using differential microphone array [DMA]
• CPC Class Codes: H04R 2430/23 Direction finding using a sum-delay beam-former
• CPC Class Codes: H04R 25/407 Circuits for combining signals of a plurality of transducers
• CPC Class Codes: H04R 3/005 for combining the signals of two or more microphones specially adapted for hearing aids H04R25/407
• CPC Class Codes: H04R 3/04 for correcting frequency response

Abstract

In one embodiment, a directional microphone array having (at least) two microphones generates forward and backward cardioid signals from two (e.g., omnidirectional) microphone signals. An adaptation factor is applied to the backward cardioid signal, and the resulting adjusted backward cardioid signal is subtracted from the forward cardioid signal to generate a (first-order) output audio signal corresponding to a beampattern having no nulls for negative values of the adaptation factor. After low-pass filtering, spatial noise suppression can be applied to the output audio signal. Microphone arrays having one (or more) additional microphones can be designed to generate second- (or higher-) order output audio signals.

Claims

1. A method for processing audio signals, comprising:

(a) generating first and second cardioid signals (e.g., 1509(1) and 1509(2)) from first and second microphone signals (e.g., 1503(1) and 1503(2));

(b) generating a first adaptation factor (e.g., 1511);

(c) applying the first adaptation factor to the second cardioid signal to generate an adapted second cardioid signal (e.g., 1513); and

(d) combining the first cardioid signal and the adapted second cardioid signal to generate a first output audio signal (e.g., 1515) corresponding to a first beampattern having no nulls for at least one value of the first adaptation factor.

2. The invention of claim 1, wherein:

the first cardioid signal is a forward cardioid signal;

the second cardioid signal is a backward cardioid signal;

the adapted backward cardioid signal is subtracted from the forward cardioid signal to generate the first output audio signal; and

the first beampattern has no nulls for negative values of the first adaptation factor.

3. The invention of claim 2, wherein the first beampattern has a null for non-negative values of the first adaptation factor.

4. The invention of claim 2, further comprising:

(e) determining whether a nearfield source is present based on the forward and backward cardioid signals.

5. The invention of claim 4, wherein the nearfield source is determined to be present if a power level of the forward cardioid signal exceeds a power level of the backward cardioid signal by a specified threshold level.

6. The invention of claim 4, wherein the nearfield source is determined to be present based on a comparison of different linear combinations of the forward and backward cardioid signals.

7. The invention of claim 4, wherein the nearfield source is determined to be present based on a comparison of different linear combinations of the first and second microphone signals.

8. The invention of claim 1, wherein the first adaptation factor is generated based on the second cardioid signal and the first output audio signal.

9. The invention of claim 8, wherein the first adaptation factor is updated according to:

β

_t+1=β

_t+2μ

yc_B,

wherein;

β

_t is the first adaptation factor at time t;

β

_t+1 is the first adaptation factor at time t+1;

μ

is an update step-size;

y is the first output audio signal; and

c_B is the second cardioid signal.

10. The invention of claim 9 wherein the first adaptation factor is limited to values from −

1 to +1, inclusive.

11. The invention of claim 9, further comprising the steps of:

determining whether a nearfield source is present; and

decreasing the update step-size μ

to reduce adaptation speed for generating the first output audio signal, if the nearfield source is determined to be present.

12. The invention of claim 1, wherein:

the first and second microphone signals are generated by two omnidirectional microphones; and

each cardioid signal is generated by subtracting a delayed version of one microphone signal from another microphone signal.

13. The invention of claim 1, further comprising the step of low-pass filtering the first output audio signal.

14. The invention of claim 1, further comprising:

(e) applying noise suppression processing to the first output audio signal to generate a noise-suppressed output audio signal.

15. The invention of claim 14, wherein the noise suppression processing is controlled based on the first adaptation factor.

16. The invention of claim 14, wherein step (e) comprises:

(1) generating a difference-signal power based on the first and second microphone signals;

(2) generating a sum-signal power based on first and second microphone signals;

(3) generating a power ratio based on the difference-signal power and the sum-signal power;

(4) generating a suppression value based on the power ratio; and

(5) applying the noise suppression processing to the first output audio signal based on the suppression value to generate the noise-suppressed output audio signal.

17. The invention of claim 16, wherein the suppression processing is based on both the power ratio and the first adaptation factor.

18. The invention of claim 16, wherein step (b) comprises generating the first adaptation factor based on the power ratio.

19. The invention of claim 18, wherein:

if the power ratio is above a specified threshold, then the first adaptation factor is set equal to a specified value; and

if the power ratio is below the specified threshold, then the first adaptation factor is based on the second cardioid signal and the first output audio signal.

20. The invention of claim 19, wherein the specified value implies that the first beampattern is omnidirectional.

21. The invention of claim 16, wherein the difference-signal power and the sum-signal power are generated from the first and second microphone signals.

22. The invention of claim 16, wherein:

the first and second microphone signals are applied to a plurality of time-domain band-pass filters to generate a power ratio value for each band-pass section;

a cutoff frequency is selected based on the plurality of power ratio values; and

the first output audio signal is high-pass filtered based on the selected cutoff frequency.

23. The invention of claim 16, wherein the difference-signal power and the sum-signal power are generated by differencing and summing the first and second cardioid signals.

24. The invention of claim 16, wherein step (e) is implemented in a subband domain to generate a suppression level for each subband.

25. The invention of claim 1, wherein steps (b), (c), and (d) are implemented in a subband domain.

26. The invention of claim 25, wherein:

step (a) is implemented in a time domain to generate time-domain first and second cardioid signals; and

the time-domain first and second cardioid signals are applied to a subband filterbank to generate subband-domain first and second cardioid signals for steps (b), (c), and (d).

27. The invention of claim 25, wherein:

the first and second microphone signals are applied to a subband filterbank to generate subband-domain microphone signals; and

step (a) is implemented in the subband domain to generate subband-domain first and second cardioid signals for steps (b), (c), and (d).

28. The invention of claim 1, wherein step (a) comprises filtering at least one of the first and second microphone signals based on a first weight factor prior to generating the first and second cardioid signals.

29. The invention of claim 28, wherein the first weight factor is generated by:

(1) selecting one microphone signal as a reference signal and another microphone signal as a calibrated signal;

(2) determining an envelope level for each of the first and second microphone signals;

(3) applying a calibration weight factor to the envelope level of the calibrated signal to generate an adjusted calibration-signal envelope level;

(4) updating the calibration weight factor to decrease a difference between the envelope level of the reference signal and the adjusted calibration-signal envelope level; and

(5) applying the updated calibration weight factor to a first low-pass filter to generate the first weight factor for the filtering of step (a).

30. The invention of claim 29, further comprising the step of applying the updated calibration weight factor to a second low-pass filter to generate a second weight factor for use in reducing noise in the first output audio signal, wherein the first low-pass filter has a cutoff frequency lower than a cutoff frequency of the second low-pass filter.

31. The invention of claim 30, further comprising the step of applying the updated calibration weight factor to a third low-pass filter to generate a third weight factor for use in detecting presence of any of wind noise, thermal noise, and circuit noise in the first and second microphone signals, wherein the second low-pass filter has a cutoff frequency lower than a cutoff frequency of the third low-pass filter.

32. The invention of claim 29, further comprising:

(6) determining whether any of wind noise, thermal noise, and circuit noise are present in the first and second microphone signals; and

(7) determining whether a nearfield source is present, wherein updating of the first weight factor based on the updated calibration weight factor is suspended if any of the wind noise, the thermal noise, and the circuit noise are determined to be present or if the nearfield source is determined to be present.

33. The invention of claim 1, wherein:

the first output audio signal is a first-order signal; and

further comprising;

(e) generating third and fourth cardioid signals (e.g., C_F2 and C_B2 of FIG. 20) from one of the first and second microphone signals (e.g., p₂) and a third microphone signal (e.g., p₃);

(f) generating a second adaptation factor (e.g., β

₁);

(g) applying the second adaptation factor to the fourth cardioid signal to generate an adapted fourth cardioid signal;

(h) combining the third cardioid signal and the adapted fourth cardioid signal to generate a second, first-order output audio signal corresponding to a second beampattern having no nulls for at least one value of the second adaptation factor; and

(i) combining the first output audio signal and the second output audio signal to form a second-order output audio signal corresponding to a third beampattern having no nulls for at least one value of the first adaptation factor and at least one value of the second adaptation factor.

34. The invention of claim 33, wherein the first adaptation factor is substantially equal to the second adaptation factor.

35. The invention of claim 33, wherein step (i) comprises:

(1) generating first and second second-order cardioid signals from the first and second first-order output audio signals;

(2) generating a third adaptation factor (e.g., β

₂);

(3) applying the third adaptation factor to the first second-order cardioid signal to generate an adapted first second-order cardioid signal;

(4) combining the second second-order cardioid signal and the adapted first second-order cardioid signal to generate the second-order output audio signal.

36. The invention of claim 35, wherein the first, second, and third adaptation factors are adapted together.

37. The invention of claim 33, wherein the first, second, and third microphone signals are generated by a one-dimensional array of three omnidirectional microphones.

38. The invention of claim 1, further comprise:

(e) determining whether any of wind noise, thermal noise, and circuit noise are present, wherein the generation of the first adaptation factor depends on whether any of the wind noise, the thermal noise, and the circuit noise are determined to be present.

39. The invention of claim 38, wherein:

if the wind noise, the thermal noise, and the circuit noise are determined not to be present, then the first adaptation factor is set equal to a specified value; and

if any of the wind noise, the thermal noise, and the circuit noise are determined to be present, then the first adaptation factor is adaptively generated based on the second cardioid signal and the first output audio signal.

40. An audio system for processing audio signals, comprising:

(a) means for generating first and second cardioid signals from first and second microphone signals;

(b) an adaptation block adapted to generate a first adaptation factor;

(c) a multiplication node adapted to apply the first adaptation factor to the second cardioid signal to generate an adapted second cardioid signal; and

(d) a combiner adapted to combine the first cardioid signal and the adapted second cardioid signal to generate a first output audio signal corresponding to a first beampattern having no nulls for at least one value of the first adaptation factor.

41. The invention of claim 40, wherein:

the first and second microphone signals are first and second omnidirectional microphone signals;

the first cardioid signal is a forward cardioid signal;

the second cardioid signal is a backward cardioid signal; and

means (a) comprises;

a first delay block adapted to delay the first omnidirectional microphone signal;

a second delay block adapted to delay the second omnidirectional microphone signal;

a first subtraction node adapted to generate the forward cardioid signal based on a difference between the first omnidirectional microphone signal and the delayed second omnidirectional microphone signal; and

a second subtraction node adapted to generate the backward cardioid signal based on a difference between the second omnidirectional microphone signal and the delayed first omnidirectional microphone signal; and

the combiner node is a third subtraction node adapted to generate the first output audio signal based on a difference between the forward cardioid signal and the adapted backward cardioid signal.