CN1700306A - Adaptive valley point noise reduction method and system - Google Patents

Abstract

The invention discloses a system and method of sub strip impulse dip point noise reduction. It uses first access sampling signal and second access sampling signal which are outputted by two non-directing microphones separately minus second access sampling signal and first access sampling signal which have been time-delayed to obtain forward directing signal and rear directing signal; it then decomposes it into many fore-and-after sub strip signals; impulse filter does multiplication with the corresponding strip rear sub strip signal and impulse filter factor and outputs the adjust signal of rear sub strip; accumulator uses each strip's foreword sub strip signal minus rear sub strip adjust signal and outputs it; at last synthesis filter set accumulates the outputs of accumulator and outputs the signals. The dip spots of different strip can align several noise sources at the same time to assure controlling the noise of output signals.

Description

A kind of self-adaptation valley point noise-reduction method and system

Technical field

The present invention relates to a kind of voice de-noising method, more particularly, the present invention relates to a kind of voice de-noising method of adaptive pointing.

Background technology

There is the hearing device (or instrument) of directive property to utilize echo signal source and noise signal source to realize the enhancing of echo signal and the inhibition of noise signal usually from the characteristics of different directions.At present this directive acoustical signal pickup apparatus has two kinds of patterns: a kind of have a fixed directional, comprise that single directional microphone and two microphones are to (dipole) etc., so-called fixed directional just is meant at fixed-direction and strengthens and decay, can not change along with the orientation of target sound source and noise source and change; Another kind of its directive property is adaptive, can come the Adjustment System parameter according to the variation in the orientation of target sound source and noise source, and with the enhancing that dynamically realizes echo signal and the inhibition of noise signal, promptly directive property is variable.

Fig. 1 is the system of the fixed directional of two common microphones compositions, d represents distance between two microphones among the figure, c is the velocity of sound, θ is that the echo signal source is to the angle between microphone and the two microphone lines, signal comes then θ=0 ° directly over the microphone of a passage, if same signal comes then θ=180 ° under the microphone of two passages.Regulate time-delay τ and can produce several outputs, different figure under corresponding the polar coordinate system with different directive property.For example, get respectively when retardation coefficient and to do τ=0,0.32 (d/c) and (d/c) time, 1kHz is a sound source, output is three kinds of shapes shown in corresponding diagram 2A, 2B, the 2C respectively.Thick line is represented the normalized energy of the signal that array receives among the figure on all directions.It is variant that system has the signal energy that directive property that is to say that system receives in different directions.The valley point is exactly the direction of received energy minimum.

Various figures have different valley points, and corresponding system can suppress the noise source of different directions.In different noise circumstances, only need to make the direction of the valley point of system to quasi-noise by adjusting, just can suppress noise preferably.Yet the environment that becomes and move during for noise source, the system of this fixed directional just can not meet the demands.

At Fa-Long Luo, Jun Yang, Senior Member, Chaslav Pavlovic, andArye Nehorai, " the Adaptive Null-Forming Scheme in Digital HearingAids " that is shown, IEEE Trans.Signal Processing, vol.50, no.7, disclose a kind of existing adaptive pointing microphone system among the JULY 2002., please refer to the system chart shown in Fig. 3, the signal that a passage (front microphone) and two passages (back microphone) are received is respectively f (n) and b (n), corresponding d/c is the time delay of two passages, W is the coefficient of each band-adaptive wave filter, and a (n) is the output of sef-adapting filter, and z (n) is the output of total system.

This system can come the Adjustment System parameter according to the variation in the orientation of target sound source and noise source, with the enhancing that dynamically realizes echo signal and the inhibition of noise signal.But during some is used, such as osophone, mobile phone, PDA, meeting room Mike etc., their environment of applications are normally Protean, noise source is very complicated, they may be in different directions, and be distributed in the different frequency section, for example, at meeting room, the fan noise of projector is arranged, it almost is a kind of single-frequency noise, traffic noise from is outside window arranged, they are some low-frequency noises, and these two kinds of noises make just to be difficult to simultaneously relatively good with these two kinds of noise removing in this way from different directions for Mike.

Summary of the invention

The technical problem to be solved in the present invention provides a kind of self-adaptation valley point noise-reduction method, can effectively eliminate to be distributed in the noise that different frequency section and a plurality of noise sources on different azimuth are sent.The present invention also will provide a kind of system that realizes this method.

In order to solve the problems of the technologies described above, the invention provides a kind of self-adaptation valley point noise-reduction method, be applied to comprise the digital information processing system of two non-directive microphones, may further comprise the steps:

(a) described two microphones pick up respectively and obtain a channel sample signal and two channel sample signals;

(b) deduct the two channel sample signals that process is delayed time with a channel sample signal, obtain directive property signal forward, deduct two channel sample signals, obtain directive property signal backward with a channel sample signal through time-delay; Described delay time equals the product of two distances between the microphone divided by the velocity of sound and sample frequency;

(c) be decomposed into a plurality of subband signals forward and a plurality of subband signal backward with directive property signal backward in identical frequency band division mode forward with described;

(d) to each frequency band, with subband signal forward and backward the cross correlation function of subband signal determine the adaptive filter coefficient of this frequency band divided by the power of subband signal backward;

(e) on each frequency band, deduct the product of subband signal and this band-adaptive filter coefficient backward with subband signal forward, again all differences are added up the voice signal that obtains handling.

For the output consistance that guarantees microphone better, in described step (b) before, measure the transport function of two microphones earlier, obtain penalty coefficient, in described step (b), the signal that the signal of a described passage or two passages adopts original signal multiply by and obtains behind the described penalty coefficient participates in computing.

Self-adaptation provided by the invention valley point noise reduction system comprises two non-directive microphones, first and second totalizers, two chronotrons, be characterized in, also comprise two analysis filterbank, a plurality of sef-adapting filter, a plurality of secondary totalizer and a synthesis filter group, wherein:

Described two non-directive microphones are used to receive acoustical signal, export a channel sample signal and two channel sample signals respectively after sampling and analog to digital conversion;

Described two chronotrons are respectively applied for and will export behind a described passage and the two channel sample signal lags, and delay time equals distance between described two microphones divided by the product of the velocity of sound and sample frequency;

Described first adder is used for a channel sample signal of input is deducted the two channel sample signals that process is delayed time, and obtains directive property signal forward; Described second adder is used for two channel sample signals of input are deducted the channel sample signal that process is delayed time, and obtains directive property signal backward;

Described two analysis filterbank, be respectively applied for described forward, the directive property signal is decomposed into a plurality of subband signals forward and a plurality of subband signal backward by identical frequency band division mode backward;

Described a plurality of sef-adapting filter, the subband signal backward that is respectively applied for frequency band multiply by the auto adapted filtering coefficient, export backward subband and adjust signal, described auto adapted filtering coefficient equal this frequency band forward subband signal and backward the cross correlation function of subband signal divided by the power of subband signal backward;

Described a plurality of secondary totalizer, the subband signal forward that is respectively applied for each frequency band deducts subband adjustment signal and output backward;

Described synthesis filter group is used for the output of all secondary totalizers being added up the signal that obtains handling.

In order to obtain better effect, the phase differential of the transport function between described two non-directive microphones is preferably less than 5 degree, and the amplitude ratio is less than 2dB.

In order to obtain better effect, described two microphones distance apart is preferably less than 40mm.

Better for the output consistance that guarantees microphone, can also increase a compensator, be used for outputing to again behind the penalty coefficient that sampled signal with a described passage or two passages multiply by two microphones described first or second adder and chronotron.

As from the foregoing, the present invention is distributed under different frequency bands and the situation in different azimuth in several noise sources by divide band on frequency, and the valley point of different frequency section can be aimed at several noise sources simultaneously, guarantees that noise is inhibited in the output signal.

Description of drawings

Fig. 1 is the system of the fixed directional of two common microphones compositions;

Fig. 2 A, 2B, 2C are in the system of Fig. 1, the different synoptic diagram that postpone to produce the different directive property that the 1kHz acoustical signal is formed;

Fig. 3 is the system chart of existing adaptive pointing microphone;

Fig. 4 is the functional block diagram of embodiment of the invention self-adaptation valley point noise reduction system.

Embodiment

Fig. 4 is the functional block diagram of embodiment of the invention self-adaptation valley point noise reduction system, as shown in the figure, two non-directive microphone (not shown)s receive acoustical signal, after sampling and analog to digital conversion sampled signal f (n) and b (n) are outputed to a passage and two passages respectively, f (n) and b (n) are respectively through chronotron 11 time-delay τ=d/cf _sBack (f _sBe sample frequency), and b (n), f (n) signal subtract each other on first, second totalizer 12 and obtain forward and directive property signal x (n) backward and y (n), again signal x (n) and y (n) are sent into analysis filterbank 13 respectively and carry out frequency division, obtain a plurality of subband signals forward and subband signal x (n backward, f) and y (n, f), again with each subband signal y (n backward, f) send into the sef-adapting filter 14 of this frequency band respectively, W _iIt is the coefficient of each band-adaptive wave filter, a (n, f) be that sef-adapting filter is exported subband adjustment signal backward, the x of subband signal forward (n with each frequency band, f) and backward subband is adjusted signal a (n, f) after each secondary totalizer is subtracted each other, comprehensively just obtained the output signal z (n) of system again through synthesis filter group 15.According to the analysis of Fig. 1 as can be known, the polar coordinates figure of the directive property correspondence of x (n) is the valley point at 180 ° heart shape diagram, and the corresponding polar coordinates figure of same y (n) is the valley point at 0 ° heart shape diagram.

Valley point and the coefficient W (n) of the output z of system (n) have following relation:

$W(n,f)=-\frac{\sin \left(\mathrm{\πf}\frac{d}{c}(1+\cos \left({\mathrm{\θ}}_{\mathrm{null}}\right))\right)}{\sin \left(\mathrm{\πf}\frac{d}{c}(1-\cos \left({\mathrm{\θ}}_{\mathrm{null}}\right))\right)}---\left(1.1\right)$

Wherein, f is the frequency of signal, and d represents distance between two microphones, and c is the velocity of sound, and θ is that the echo signal source is to the angle between microphone and the two microphone lines, θ _NullIt is the angle value of valley point correspondence.

Formula (1.1) has shown valley point and adaptation coefficient W, and (n, relation f) is to the dependence of signal frequency.Carry out the branch band by the signal that x (n) and these two of y (n) is had fixed directional by frequency domain, in each frequency band, carry out Adaptive Identification respectively, obtain adaptive filter coefficient W (n separately, f), (n f) changes the time-delay of each frequency band, just can change the polar coordinates figure of output correspondence to regulate W, make the valley point of total system aim at noise source, realize reducing the purpose of noise.

To specify below adaptation coefficient W (n, computing method f) suppose that s (n) and i (n) are useful signal (voice signal) and the noise signals that enters in the signal of two passages, voice signal on ° direction of θ=0, τ _iBe the delay that enters respectively between the noise signal of two passages,, can obtain following various according to the system of Fig. 4:

f(n)＝s(n)+i(n-τ _i) (1.2)

$b\left(n\right)=s(n-\frac{d}{c})+i\left(n\right)---\left(1.3\right)$

$x\left(n\right)=f\left(n\right)-b(n-\frac{d}{c})---\left(1.4\right)$

$y\left(n\right)=f(n-\frac{d}{c})-b\left(n\right)---\left(1.5\right)$

a(n，f)＝W(n，f)y(n，f) (1.6)

$z\left(n\right)=x\left(n\right)-\underset{f=0}{\overset{f=N}{\mathrm{\Σ}}}a(n,f)---\left(1.7\right)$

With formula (1.2) and (1.3) substitution formula (1.5), can obtain after the decomposition:

$y(n,f)=$ $(n-\mathrm{\τ}-\frac{d}{c},f)-i(n,f)---\left(1.8\right)$

Can see that directive property signal y (n) only contains noise component backward, in present specification, the directive property signal promptly is the valley point sound source that aims at the mark backward, only contains the pickoff signals of noise component.These characteristics show that reducing system's output power is equivalent to the power that reduces the noise component in the output power, because voice signal and noise signal are uncorrelated.The expectation of the power of the output z of system (n) is as follows

${=R}_{\mathrm{xx}}(n,f)-2W(n,f)R_{\mathrm{xy}}(n,f)+W^2(n,f)R_{\mathrm{yy}}(n,f)---\left(1.9\right)$

R wherein _Xx(n, f) and R _Yy(n, f) be respectively x (n, f) and y (n, power f), R _Xy(n, f) be x (n, f) and y (n, cross correlation function f).Formula (1.9) is that (n, quadratic expression f) has and only have a minimum value W about W _Opt, making its minimum, can obtain:

$W_{\mathrm{opt}}(n,f)=\frac{R_{\mathrm{xy}}(n,f)}{R_{\mathrm{yy}}(n,f)}---\left(1.10\right)$

According to above analysis, (n f), just can guarantee output signal z (n, f) direction of subtend noise source in the valley point of each frequency band to utilize formula (1.10) to calculate the coefficient W of each sef-adapting filter in the present embodiment system.

Accordingly, in the embodiment of the invention method,, determine delay time τ with the product of the distance between two microphones divided by the velocity of sound and sample frequency;

Be formulated promptly: $\mathrm{\τ}=\frac{d}{{\mathrm{cf}}_s},$ D represents distance between two microphones, and c is the velocity of sound, f _sBe sample frequency;

Processing procedure to acoustical signal may further comprise the steps:

Step 110, two microphones pick up respectively and obtain a channel sample signal f (n) and two channel sample signal b (n);

Step 120, deduct two channel sample signals of process time-delay with a channel sample signal, obtain directive property signal forward, be x (n)=f (n)-b (n-τ), deduct two channel sample signals with a channel sample signal through time-delay, obtain directive property signal backward, i.e. y (n)=f (n-τ)-b (n);

Step 130, will be forward with directive property signal x (n) and y (n) backward with identical frequency band division mode be decomposed into a plurality of x of subband signal forward (n, f) with a plurality of y of subband signal backward (n, f);

Step 140, to each frequency band, with subband signal x forward (n, f) and backward (n, cross correlation function f) obtain the adaptive filter coefficient of this frequency band divided by the power of subband signal backward to subband signal y;

Be formulated promptly: $W(n,f)=\frac{R_{\mathrm{xy}}(n,f)}{R_{\mathrm{yy}}(n,f)};$

Step 150 on each frequency band, deducts the product of subband signal and this band-adaptive filter coefficient backward with subband signal forward, again with all difference additions, the voice signal that obtains exporting.

Be formulated promptly: $z\left(n\right)=\underset{f=0}{\overset{N}{\mathrm{\Σ}}}(x(n,f)-y(n,f)*W(n,f))$

In sum, by on frequency, dividing band can guarantee in the valley point of each frequency band of output signal all directions of subtend noise source, be distributed in different frequency bands and in the situation of different azimuth in several noise sources, the valley point of different frequency section can be aimed at several noise sources simultaneously, guarantees that noise is inhibited in the output signal.Be example still with background technology institute act situation, should the present invention be arranged after, can make the valley point of low-frequency band aim at window, and projector is aimed in the valley point of the frequency band at fan noise place, therefore can take into account the noise source of both direction simultaneously.

On the basis of basic scheme of the present invention, various conversion can also be arranged:

For example, when two microphone consistance are not so good, can compensate the output consistance that improves two microphones.Measure earlier the transport function of two microphones, obtain penalty coefficient C, earlier with two channel sample signal input offset devices, two channel sample signal: b ' after being compensated (n)=Cb (n), be entered into first adder and corresponding chronotron again; In the step 120 of embodiment method, then adopt after the compensation two channel sample calculated signals forward, directive property signal backward, i.e. x (n)=f (n)-b ' (n-τ), (n-τ)-b ' (n) for y (n)=f.It also is the same that one channel sample signal is compensated.

In addition, in order to obtain better effect, should select reasonable two the non-directive microphones of consistance in the system, so-called high conformity is exactly the phase differential of transport function between two microphones less (| φ |≤5 °), amplitude is more straight (| A _Max-A _Min|＜2dB); Two microphones distance apart preferably satisfies d＜40mm.15 of analysis filterbank 13 and synthesis filter groups are not limited to certain type.

Claims (6)

1, a kind of self-adaptation valley point noise-reduction method is applied to comprise the digital information processing system of two non-directive microphones, may further comprise the steps:

(a) described two microphones pick up respectively and obtain a channel sample signal and two channel sample signals;

(b) deduct the two channel sample signals that process is delayed time with a channel sample signal, obtain directive property signal forward, deduct two channel sample signals, obtain directive property signal backward with a channel sample signal through time-delay; Described delay time equals the product of two distances between the microphone divided by the velocity of sound and sample frequency;

It is characterized in that, in step (b) afterwards, also comprise:

(c) be decomposed into a plurality of subband signals forward and a plurality of subband signal backward with directive property signal backward in identical frequency band division mode forward with described;

(d) to each frequency band, with subband signal forward and backward the cross correlation function of subband signal determine the adaptive filter coefficient of this frequency band divided by the power of subband signal backward;

(e) on each frequency band, deduct the product of subband signal and this band-adaptive filter coefficient backward with subband signal forward, again all differences are added up the voice signal that obtains handling.

2, self-adaptation as claimed in claim 1 or 2 valley point noise-reduction method, it is characterized in that, in described step (b) before, measure the transport function of two microphones earlier, obtain penalty coefficient, in described step (b), the sampled signal that the sampled signal of a described passage or two passages adopts original signal multiply by and obtains behind the described penalty coefficient participates in computing.

3, a kind of self-adaptation valley point noise reduction system comprises two non-directive microphones, first and second totalizers and two chronotrons, wherein:

Described two non-directive microphones are used to receive acoustical signal, sample and analog to digital conversion after export a channel sample signal and two channel sample signals respectively;

Described two chronotrons are respectively applied for and will export behind a described passage and the two channel sample signal lags, and delay time equals distance between described two microphones divided by the product of the velocity of sound and sample frequency;

Described first adder is used for a channel sample signal of input is deducted the two channel sample signals that process is delayed time, and obtains directive property signal forward; Described second adder is used for two channel sample signals of input are deducted the channel sample signal that process is delayed time, and obtains directive property signal backward;

It is characterized in that, also comprise two analysis filterbank, a plurality of sef-adapting filter, a plurality of secondary totalizer and a synthesis filter group, wherein:

Described two analysis filterbank, be respectively applied for described forward, the directive property signal is decomposed into a plurality of subband signals forward and a plurality of subband signal backward by identical frequency band division mode backward;

Described a plurality of sef-adapting filter, the subband signal backward that is respectively applied for frequency band multiply by the auto adapted filtering coefficient, export backward subband and adjust signal, described auto adapted filtering coefficient equal this frequency band forward subband signal and backward the cross correlation function of subband signal divided by the power of subband signal backward;

Described a plurality of secondary totalizer, the subband signal forward that is respectively applied for each frequency band deducts subband adjustment signal and output backward;

Described synthesis filter group is used for the output of all secondary totalizers being added up the signal that obtains handling.

4, self-adaptation as claimed in claim 3 valley point noise reduction system is characterized in that, the phase differential of the transport function between described two non-directive microphones is less than 5 degree, and amplitude is than less than 2dB.

5, self-adaptation as claimed in claim 3 valley point noise reduction system is characterized in that described two microphones distance apart is less than 40mm.

6, self-adaptation as claimed in claim 3 valley point noise reduction system, it is characterized in that, also comprise a compensator, be used for outputing to again behind the penalty coefficient that sampled signal with a described passage or two passages multiply by two microphones described first or second adder and chronotron.

Images