CN103916730B - A kind of sound field focusing method and system that can improve tonequality - Google Patents

Abstract

The present invention relates to a kind of sound field focusing method that can improve tonequality, including：Loudspeaker array is laid, sets the control point of area pellucida and dark space；Wherein, the area pellucida is the region with high acoustic energy, and the dark space is the region with energy in a low voice；Set worst contrast focusing performance parameter and the object vector in area pellucida on each Frequency point；According to the worst contrast focusing performance parameter on each Frequency point and the object vector in area pellucida, in conjunction with acoustic energy compared with control and least square method criterion, amplitude and phase place that the loudspeaker array on each Frequency point is input into multiple weight coefficient vector is adjusted respectively；Multiple weight coefficient vector component frequency on each Frequency point in each loudspeaker channel is responded, to frequency response adding window, and inverse Fourier transform is carried out, obtains the time domain impulse response signals of each passage.

Description

A kind of sound field focusing method and system that can improve tonequality

Technical field

The present invention relates to field of acoustics, more particularly to a kind of sound field focusing method and system that can improve tonequality.

Background technology

With growth in the living standard, such as mobile phone, notebook, the personal consumption such as multimedia player electronic product are being given birth to Apply in work more and more extensive.User generally listens to the sound of such product by earphone or loudspeaker, but this can be present Some problems, wearing earphone for a long time can not only make user feel under the weather, also can impaired user hearing, and adopt loudspeaker then Surrounding user can be bothered.Adopt the personal audio system of loudspeaker array can be with effectively solving problems.Personal audio system Most of acoustic energy can be assembled the region comprising user, so as to the private listening space of personalization is built for user.Individual In people's audio system, the most frequently used controlling mechanism is acoustic energy compared with control method.By adjusting the defeated of each passage of loudspeaker array Enter signal phase and amplitude, sound field compared with control method can cause the acoustic energy comprising user area (referred to as area pellucida) and other The ratio of the acoustic energy of non-associated users region (referred to as dark space) is maximum, so as to reach the effect of focusing acoustic field.

But acoustic energy compared with control method of the prior art only focuses on contrast focusing performance, does not focus on area pellucida Sound quality problem, this decline may the impression of user's audition, and on the other hand, in some cases, contrast focusing performance is more than one After determining threshold value, impression impact is focused on less to the subjective of people, and sound quality problem now should be more paid close attention to, to lift use further Family listen to impression.

For focusing performance being only focused in acoustic energy compared with control method of the prior art be not concerned with tonequality performance Problem, needs to find preferable balance of the more simple and effective method to obtain tonequality and contrast focusing performance.

Content of the invention

It is an object of the invention to overcome in existing acoustic energy compared with control method of the prior art only focusing on focusing Performance and be not concerned with the problem of tonequality performance, so as to provide a kind of sound field focusing method that can improve tonequality.

To achieve these goals, the invention provides a kind of sound field focusing method that can improve tonequality, including：

Step 1), lay loudspeaker array, set the control point of area pellucida and dark space；Wherein, the area pellucida be with loud The region of energy, the dark space are the region with energy in a low voice；

Step 2), worst contrast focusing performance parameter and the object vector in area pellucida set on each Frequency point；

Step 3), according to step 2) worst contrast focusing performance parameter on each Frequency point for obtaining and in area pellucida Object vector, in conjunction with acoustic energy compared with control and least square method criterion, adjusts the loudspeaker array on each Frequency point respectively The amplitude and phase place of the multiple weight coefficient vector of input；

Step 4), by step 3) multiple weight coefficient vector composition on each Frequency point in each loudspeaker channel for obtaining Frequency response, to frequency response adding window, and carries out inverse Fourier transform, obtains the time domain impulse response signals of each passage.

In technique scheme, in step 1) in, the loudspeaker array that laid be linear array or circular array or with Machine array.

In technique scheme, in step 1) in, the region shape of the area pellucida or dark space is square or circular or line style.

In technique scheme, in described step 2) in, set the worst contrast focusing performance ginseng on each Frequency point Number is comprised the following steps：

Step 2-1), determine a Frequency point f, obtain loudspeaker array arrive the control point of area pellucida and dark space respectively in frequency f The transfer function matrix G at place_B,fAnd G_D,f, its size is respectively M_B× L and M_D× L, wherein M_BAnd M_DIt is area pellucida and dark space respectively Control point number, L are number of loudspeakers；

Step 2-2), according to step 2-1) loudspeaker array that obtains control point respectively to area pellucida and dark space is in frequency f The transfer function matrix G at place_B,fAnd G_D,f, calculate maximum F of the ratio of area pellucida and dark space acoustic energy_max,f；Wherein,

Wherein λ_max{ } is the eigenvalue of maximum of matrix, δ_fFor robustness parameter, it is single more than 0, I that its selection range is Bit matrix, w_fIt is the multiple weight coefficient vector of loudspeaker array input on Frequency point f；

Step 2-3), by step 2-2) in area pellucida and dark space acoustic energy ratio maximum F_max,fDetermine on Frequency point f Worst contrast focusing performance parameter J_min,fSpan, in the range from (- ∞, F_max,f].

In technique scheme, in described step 2-1) in, using the audio-frequency test such as B＆K PULSE instrument or pass through Modeling and simulating obtains loudspeaker array and arrives transfer function matrix G of the control point of area pellucida and dark space at frequency f respectively_B,fWith G_D,f.

In technique scheme, in described step 2) in, object vector d in the area pellucida on Frequency point f_fTable Show, its size is L × 1, equal and its value of each element in vector is arbitrary value.

In technique scheme, described step 3) comprise the following steps：

Step 3-1), utilize step 2) in set parameter F_max,fAnd d_f, in conjunction with acoustic energy compared with control and least square Method criterion, is listed below optimal problem：

Step 3-2), orderDeployment step 3-1) obtained by expression formula, obtain：

Wherein Re { } is the real part for taking element；

Step 3-3), be defined as follows：

Wherein Im { } is the imaginary part for taking element；

Step 3-4), by step 3-3) defined in expression formula substitute into step 3-2) obtained by expression formula, so as to will Step 3-2) in expression formula be converted to real number problem, and give up constant termObtain following expression：

s.t. x^TH₂x≤0 (4)

Step 3-5), by step 3-4) obtained by expression formula described by optimal problem Lagrangian dual problem For：

Wherein ζ (x, λ)=x^TH₁x-2y^Tx+λx^TH₂x；

Step 3-6), definitionBy step 3-5) described in dual problem be converted into as follows Form, and then solve x；

Step 3-7), by step 3-6) obtained by x value by step 3-3) described in expression formula be converted into finally Solution w_f.

In technique scheme, in described step 3-6) in, solution x is obtained using convex optimisation technique.

In technique scheme, in step 4) in, the window function is rectangular window or Hanning window or hamming window.

Present invention also offers a kind of sound field focusing system that can improve tonequality, including：

Loudspeaker array cloth amplification module, for laying loudspeaker array, sets the control point of area pellucida and dark space；Wherein, institute State the region that area is with high acoustic energy clearly, the dark space is the region with energy in a low voice；

The first setting module of parameter, for setting the worst contrast focusing performance parameter on each Frequency point and in area pellucida Object vector；

Weight coefficient vector adjusting module is answered in input, on each Frequency point obtained for setting module according at the beginning of parameter most Difference contrast focusing performance parameter and the object vector in area pellucida, in conjunction with acoustic energy compared with control and least square method criterion, point Amplitude and phase place that loudspeaker array each Frequency point on be input into multiple weight coefficient vector are not adjusted；

Time domain impulse response signals computing module, for being input into, by described, each that weight coefficient vector adjusting module is obtained again Multiple weight coefficient vector component frequency response on each Frequency point in loudspeaker channel, to frequency response adding window, and is carried out inverse Fourier transformation, obtains the time domain impulse response signals of each passage.

It is an advantage of the current invention that：

1st, the present invention adopts all square minimum error principle so that the frequency response in area pellucida is approached by goal-selling vector In the frequency response of preferable Dirac pulse, so as to improve the tonequality in area pellucida.

2nd, the present invention can be obtained between contrast focusing performance and tonequality by adjusting worst contrast focusing performance parameter One preferable performance.

3rd, the sound field focusing method for improving tonequality of the present invention can be widely applied for television set, the individual such as computer and mobile phone Audio area, by adjusting worst contrast focusing performance parameter, can obtain one relatively between contrast focusing performance and tonequality Good performance.

Description of the drawings

Fig. 1 is the present invention based on the sound field focusing side for improving tonequality with reference to acoustic energy compared with control and least square method The flow chart of method；

Fig. 2 is the schematic diagram of linear loudspeaker array in one embodiment；

Fig. 3 is that contrast of the inventive method with acoustic energy compared with control method of the prior art in different frequency point is focused on The schematic diagram of performance；

Fig. 4 (a) is the schematic diagram of impulse response of the inventive method at the center control point of area pellucida；

Fig. 4 (b) is impulse response of the acoustic energy compared with control method of the prior art at the center control point of area pellucida Schematic diagram.

Specific embodiment

In conjunction with accompanying drawing, the invention will be further described.

The basic thought of the present invention is default worst contrast focusing performance parameter, is ensureing that contrasting focusing performance is not less than this In the case of parameter, using default object vector so that the frequency response in area pellucida approaches preferable Dirac-delta pulse Frequency response, so as to improve acoustical quality, and can be focused on by adjusting worst contrast focusing performance parameter and obtaining contrast Balance between performance and tonequality.

With reference to Fig. 1, the present invention is gathered based on the sound field for improving tonequality with reference to acoustic energy compared with control and least square method Burnt method is comprised the following steps：

Step 1), lay loudspeaker array, set the control point of area pellucida and dark space；Wherein, the area pellucida be with loud The region of energy, the dark space are the region with energy in a low voice；

Step 2), worst contrast focusing performance parameter and the object vector in area pellucida set on each Frequency point；

Step 3), combine acoustic energy compared with control and least square method criterion, adjust raising one's voice on each Frequency point respectively The amplitude and phase place of the multiple weight coefficient vector of device array input；

Step 4), by step 3) multiple weight coefficient vector composition on each Frequency point in each loudspeaker channel for obtaining Frequency response, to frequency response adding window, and carries out inverse Fourier transform, obtains the time domain impulse response signals of each passage.

Below each step in the inventive method is described further.

In step 1) in, the loudspeaker array for being laid is linear array or circular array, or random array.Institute The region shape for stating area or dark space clearly can be square or circular, or line style.

In described step 2) in, the worst contrast focusing performance parameter set on each Frequency point is comprised the following steps：

Step 2-1), determine a Frequency point f, obtain using the audio-frequency test such as B＆K PULSE instrument or by modeling and simulating Loudspeaker array arrives transfer function matrix G of the control point of area pellucida and dark space at frequency f respectively_B,fAnd G_D,f, its size divides Wei not M_B× L and M_D× L, wherein M_BAnd M_DIt is the control point number of area pellucida and dark space respectively, L is number of loudspeakers；

Step 2-2), according to step 2-1) loudspeaker array that obtains control point respectively to area pellucida and dark space is in frequency f The transfer function matrix G at place_B,fAnd G_D,f, calculate maximum F of the ratio of area pellucida and dark space acoustic energy_max,f.Wherein,

Wherein λ_max{ } is the eigenvalue of maximum of matrix, δ_fFor robustness parameter, it is single more than 0, I that its selection range is Bit matrix, w_fIt is the multiple weight coefficient vector of loudspeaker array input on Frequency point f；

Step 2-3), by step 2-2) in area pellucida and dark space acoustic energy ratio maximum F_max,fDetermine on Frequency point f Worst contrast focusing performance parameter J_min,fSpan, in the range from (- ∞, F_max,f].

In described step 2) in, object vector d in the area pellucida on Frequency point f_fRepresent, its size is L × 1, Equal and its value of each element in vector can be arbitrary value.

Described step 3) specifically include following steps：

Step 3-1), utilize step 2) in set parameter F_max,fAnd d_f, in conjunction with acoustic energy compared with control and least square Method criterion, is listed below optimal problem：

Step 3-2), orderDeployment step 3-1) obtained by expression formula, obtain：

Wherein Re { } is the real part for taking element；

Step 3-3), be defined as follows：

Wherein Im { } is the imaginary part for taking element；

Step 3-4), by step 3-3) defined in expression formula substitute into step 3-2) obtained by expression formula, so as to will Step 3-2) in expression formula be converted to real number problem, and give up constant termObtain following expression：

s.t. x^TH₂x≤0 (4)

Step 3-5), by step 3-4) obtained by expression formula described by optimal problem Lagrangian dual problem For：

Wherein ζ (x, λ)=x^TH₁x-2y^Tx+λx^TH₂x；

Step 3-6), definitionBy step 3-5) described in dual problem be converted into as follows Form, and then solve x.

This problem is a positive semidefinite optimization problem, can obtain solution x by the technology of such as convex optimization etc.

Step 3-7), by step 3-6) obtained by x value by step 3-3) described in expression formula be converted into finally Solution w_f.

In step 4) in, the window function can be rectangular window, or Hanning window or hamming window.

In order to be better understood from the present invention, the present invention is made to retouch in detail further with specific embodiment below in conjunction with the accompanying drawings State：

In the present embodiment, as shown in Fig. 2 in Acoustical Inst., Chinese Academy of Sciences's whole elimination room, placing one and linearly raising Sound device array, loudspeaker array are made up of 8 moving-coil type favour prestige loudspeakers, and spacing is 12cm, and loudspeaker unit peak power is 15W, 8 Ω of rated impedance；Area pellucida and dark space are in each 45 degree of directions in left and right on loudspeaker array perpendicular bisector, apart from loudspeaker array Between distance be all 1m, and with loudspeaker array in the same level；The control point of area pellucida and dark space is all by 5 yuan of microphone arrays Row definition, at intervals of 8cm, microphone is 4189 microphones of B＆K company of Denmark；Experiment select hardware unit be power amplifier, PULSE measuring instrument and PC.

The specific implementation process of the present embodiment includes the following steps：

(1) systematic sampling rate f_s8kHz is set to, length N of time domain impulse response signals is set to 800, the calculating frequency of selection Point is 10Hz to 3920Hz, and frequency interval is 10Hz.Measurement loudspeaker array arrive area pellucida and secretly respectively in each Frequency point first The transfer function matrix at area control point, this can pass through using PULSE instrument direct measurement loudspeaker array to each control point Frequency receptance function, then by frequency response import PC platform on Matlab software processed, can a step obtain all frequencies The transfer function matrix G of rate_B,fAnd G_D,f.

(2) secondly, worst contrast focusing performance parameter J of all Frequency points is set_minIt is all 10, sets at Frequency point f Object vector is

(3) it is all 0.001 to be set in all Frequency point robustness parameters, according to reference to acoustic energy compared with control and a most young waiter in a wineshop or an inn Multiplication criterion calculates multiple weight coefficient vector w of optimum loudspeaker array input_f, frequency response is then combined into, and square is added to which Shape window does the time domain impulse response signals that inverse Fourier transform obtains respective channel.

Fig. 3 gives the contrast of the inventive method and acoustic energy compared with control method of the prior art in different frequency point Focusing performance, it can be seen that on selected Frequency point, the contrast focusing performance of although the inventive method declines, but It is can to ensure more than default worst contrast focusing performance 10log10 (J_min,f), from the acoustical testing being hereinafter previously mentioned It can be seen that, although contrast focusing performance declines, but tonequality is improved.

Fig. 4 sets forth the inventive method (Fig. 4 (a)) and acoustic energy compared with control method (Fig. 4 of the prior art (b)) impulse response at the center control point of area pellucida.Relatively two width figures can be seen that acoustic energy controlling party of the prior art Method oscillation amplitude is big, and conditions of streaking is serious, and is more approached preferably using the impulse response obtained by the inventive method Dirac-delta pulse, therefore greatly can improve tonequality, and this is improved by sacrificing the obtainable tonequality of contrast focusing.

Improve tonequality to comment by PESQ (Perceptual evaluation of speech quality) fraction To estimate, the score value of PESQ is 0 to 4.5/, 0 represents worst, and 4.5 represent best, and the voice after process only has pure with primitive sound Postpone.Here the two speech signal samples rates for adopting are 8kHz, and the content for being respectively designated as e1 and e2, e1 and e2 is respectively " the birch canoe slid on the smooth planks " and " glue the sheet to the dark blue Background ", is read aloud by male voice.Table 1 gives the speech quality PESQ fraction produced at the center control point of area pellucida, from table As can be seen that acoustic energy compared with control method of the prior art has infringement to speech quality, and compared to of the prior art Acoustic energy compared with control method, the inventive method greatly improve speech quality, almost can revert to speech quality original The tonequality of voice.

The speech quality evaluation produced at 1 area pellucida center control point of table, using PESQ fraction

	The inventive method	Method of the prior art
			e1	4.3317	2.9238
e2	4.3179	2.9073

Although sample frequency is defined in the present embodiment for 8kHz, and select area pellucida and dark space for linear regions, this is only It is only that one of method provided by the present invention is illustrated, does not limit method provided by the present invention and be only suitable for people and speak audio frequency In the range of rate, or area pellucida, dark space can only select line style.It is true that method provided by the present invention can be extended to entire audible The broadband signal of acoustic frequency section simultaneously realizes acoustic energy contrast focusing.

Present invention also offers a kind of sound field focusing system that can improve tonequality with preceding method identical, including：

Loudspeaker array cloth amplification module, for laying loudspeaker array, sets the control point of area pellucida and dark space；Wherein, institute State the region that area is with high acoustic energy clearly, the dark space is the region with energy in a low voice；

The first setting module of parameter, for setting the worst contrast focusing performance parameter on each Frequency point and in area pellucida Object vector；

Weight coefficient vector adjusting module is answered in input, on each Frequency point obtained for setting module according at the beginning of parameter most Difference contrast focusing performance parameter and the object vector in area pellucida, in conjunction with acoustic energy compared with control and least square method criterion, point Amplitude and phase place that loudspeaker array each Frequency point on be input into multiple weight coefficient vector are not adjusted；

Time domain impulse response signals computing module, for being input into, by described, each that weight coefficient vector adjusting module is obtained again Multiple weight coefficient vector component frequency response on each Frequency point in loudspeaker channel, to frequency response adding window, and is carried out inverse Fourier transformation, obtains the time domain impulse response signals of each passage.

It should be noted last that, above example is only in order to illustrate technical scheme and unrestricted.Although ginseng The present invention is described in detail according to embodiment, it will be understood by those within the art that, the technical side to the present invention Case is modified or equivalent, and without departure from the spirit and scope of technical solution of the present invention, which all should be covered in the present invention Right in the middle of.

Claims (9)

1. a kind of sound field focusing method that can improve tonequality, including：

Step 1), lay loudspeaker array, set the control point of area pellucida and dark space；Wherein, the area pellucida be with high acoustic energy Region, the dark space is the region with energy in a low voice；

Step 2), worst contrast focusing performance parameter and the object vector in area pellucida set on each Frequency point；The step Further include：

Step 2-1), determine a Frequency point f, obtain loudspeaker array and arrive the control point of area pellucida and dark space respectively at frequency f Transfer function matrix G_B,fAnd G_D,f, its size is respectively M_B× L and M_D× L, wherein M_BAnd M_DIt is the control of area pellucida and dark space respectively Point number, L are number of loudspeakers；

Step 2-2), according to step 2-1) loudspeaker array that obtains control point respectively to area pellucida and dark space is at frequency f Transfer function matrix G_B,fAnd G_D,f, calculate maximum F of the ratio of area pellucida and dark space acoustic energy_max,f；Wherein,

$\begin{array}{c}{F}_{\max ,f}=\underset{w_f}{\max }\frac{M_D}{M_B}\frac{w_f^H{G}_{B,f}^H{G}_{B,f}{w}_f}{w_f^H(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)w_f}\\ =\frac{M_D}{M_B}{\mathrm{\λ}}_{\max }\left\{{(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)}^{-1}{G}_{B,f}^H{G}_{B,f}\right\}\end{array}$

Wherein λ_max{ } is the eigenvalue of maximum of matrix, δ_fFor robustness parameter, it is unit square more than 0, I that its selection range is Battle array, w_fIt is the multiple weight coefficient vector of loudspeaker array input on Frequency point f；

Step 2-3), by step 2-2) in area pellucida and dark space acoustic energy ratio maximum F_max,fDetermine on Frequency point f most Difference contrast focusing performance parameter J_min,fSpan, in the range from (- ∞, F_max,f]；

Step 3), according to step 2) worst contrast focusing performance parameter on each Frequency point for obtaining and the target in area pellucida Vector, in conjunction with acoustic energy compared with control and least square method criterion, adjusts the loudspeaker array input on each Frequency point respectively The amplitude and phase place of multiple weight coefficient vector；

Step 4), by step 3) multiple weight coefficient vector component frequency on each Frequency point in each loudspeaker channel for obtaining Response, to frequency response adding window, and carries out inverse Fourier transform, obtains the time domain impulse response signals of each passage.

2. the sound field focusing method that tonequality can be improved according to claim 1, it is characterised in that in step 1) in, institute The loudspeaker array for laying is linear array or circular array or random array.

3. the sound field focusing method that tonequality can be improved according to claim 1, it is characterised in that in step 1) in, institute The region shape for stating area or dark space clearly is square or circular or line style.

4. the sound field focusing method that tonequality can be improved according to claim 1, it is characterised in that in described step 2- 1), in, using the audio-frequency test such as B＆K PULSE instrument or loudspeaker array is obtained by modeling and simulating arrive area pellucida and secretly respectively Transfer function matrix G of the control point in area at frequency f_B,fAnd G_D,f.

5. the sound field focusing method that tonequality can be improved according to claim 1, it is characterised in that in described step 2) In, object vector d in the area pellucida on Frequency point f_fRepresent, its size is L × 1, each element in vector is equal And its value is arbitrary value.

6. the sound field focusing method that tonequality can be improved according to claim 5, it is characterised in that described step 3) bag Include following steps：

Step 3-1), utilize step 2) in set parameter F_max,fAnd d_f, accurate in conjunction with acoustic energy compared with control and least square method Then, optimal problem is listed below：

$\begin{array}{cc}\underset{w_f}{min}& \left|\right|G_{B,f}{w}_f-d_f|{|}^2\\ s.t.& \frac{M_D}{M_B}\frac{w^H{G}_{B,f}^H{G}_{B,f}w}{w^H(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)w}\≥J_{\min ,f}\end{array}---\left(1\right)$

Step 3-2), orderDeployment step 3-1) obtained by expression formula, obtain：

$\begin{array}{cc}\underset{w_f}{\min }& w_f^H{G}_{B,f}^H{G}_{B,f}{w}_f-2\mathrm{Re}\left\{w_f^H{G}_{B,f}^H{d}_f\right\}+d_f^H{d}_f\\ s.t.& w_f^H\[J_0(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)-G_{B,f}^H{G}_{B,f}\]w_f\≤0\end{array}---\left(2\right)$

Wherein Re { } is the real part for taking element；

Step 3-3), be defined as follows：

$\begin{array}{c}\begin{array}{ccc}x=\left[\begin{array}{c}\mathrm{Re}\left\{w_f\right\}\\ \mathrm{Im}\left\{w_f\right\}\end{array}\right]& y=\left[\begin{array}{c}\mathrm{Re}\left\{G_{B,f}^H{d}_f\right\}\\ \mathrm{Im}\left\{G_{B,f}^H{d}_f\right\}\end{array}\right]& S=J_0(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)-G_{B,f}^H{G}_{B,f}\end{array}\\ \begin{array}{cc}{H}_1=\left[\begin{array}{cc}\mathrm{Re}\left\{G_{B,f}^H{G}_{B,f}\right\}& -\mathrm{Im}\left\{G_{B,f}^H{G}_{B,f}\right\}\\ \mathrm{Im}\left\{G_{B,f}^H{G}_{B,f}\right\}& \mathrm{Re}\left\{G_{B,f}^H{G}_{B,f}\right\}\end{array}\right]& H_2=\left[\begin{array}{cc}\mathrm{Re}\left\{S\right\}& -\mathrm{Im}\left\{S\right\}\\ \mathrm{Im}\left\{S\right\}& \mathrm{Re}\left\{S\right\}\end{array}\right]\end{array}\end{array}---\left(3\right)$

Wherein Im { } is the imaginary part for taking element；

Step 3-4), by step 3-3) defined in expression formula substitute into step 3-2) obtained by expression formula, so as to by step Expression formula in 3-2) is converted to real number problem, and gives up constant termObtain following expression：

$\begin{array}{cc}\underset{x}{\min }& x^T{H}_{1}x-2y^{T}x\\ s.t.& x^T{H}_{2}x\≤0\end{array}---\left(4\right)$

Step 3-5), by step 3-4) obtained by expression formula described by the Lagrangian dual problem of optimal problem be：

Wherein

Step 3-6), definitionBy step 3-5) described in dual problem be converted into following form, And then solve x；

$\begin{array}{cc}\underset{\mathrm{\λ}}{max}& \mathrm{\γ}\\ s.t.& \begin{array}{c}\mathrm{\λ}\≥0\\ \left[\begin{array}{cc}{H}_1+{\mathrm{\λH}}_2& -y\\ -y^T& -\mathrm{\γ}\end{array}\right]\≥0\end{array}\end{array}---\left(6\right)$

Step 3-7), by step 3-6) obtained by x value by step 3-3) described in expression formula be converted into last solution w_f.

7. the sound field focusing method that tonequality can be improved according to claim 6, it is characterised in that in described step 3- 6), in, solution x is obtained using convex optimisation technique.

8. the sound field focusing method that tonequality can be improved according to claim 1, it is characterised in that in step 4) in, institute Window function is stated for rectangular window or Hanning window or hamming window.

9. a kind of sound field focusing system that can improve tonequality, it is characterised in that include：

Loudspeaker array cloth amplification module, for laying loudspeaker array, sets the control point of area pellucida and dark space；Wherein, stated clearly Area is the region with high acoustic energy, and the dark space is the region with energy in a low voice；

The first setting module of parameter, for setting worst contrast focusing performance parameter and the target in area pellucida on each Frequency point Vector；Specifically include：

First, determine a Frequency point f, obtain loudspeaker array and arrive transmission of the control point of area pellucida and dark space at frequency f respectively Jacobian matrix G_B,fAnd G_D,f, its size is respectively M_B× L and M_D× L, wherein M_BAnd M_DIt is the control point of area pellucida and dark space respectively Number, L are number of loudspeakers；

Then, transfer function matrix G of the control point of area pellucida and dark space at frequency f is arrived respectively according to loudspeaker array_B,fWith G_D,f, calculate maximum F of the ratio of area pellucida and dark space acoustic energy_max,f；Wherein,

$\begin{array}{c}{F}_{\max ,f}=\underset{w_f}{\max }\frac{M_D}{M_B}\frac{w_f^H{G}_{B,f}^H{G}_{B,f}{w}_f}{w_f^H(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)w_f}\\ =\frac{M_D}{M_B}{\mathrm{\λ}}_{\max }\left\{{(G_{D,f}^H{G}_{D,f}+{\mathrm{\δ}}_{f}I)}^{-1}{G}_{B,f}^H{G}_{B,f}\right\}\end{array}$

Wherein λ_max{ } is the eigenvalue of maximum of matrix, δ_fFor robustness parameter, it is unit square more than 0, I that its selection range is Battle array, w_fIt is the multiple weight coefficient vector of loudspeaker array input on Frequency point f；

Then, by area and maximum F of the ratio of dark space acoustic energy_max,fDetermine the worst contrast focusing performance ginseng on Frequency point f Number J_min,fSpan, in the range from (- ∞, F_max,f]；

The multiple weight coefficient vector adjusting module of input, for according to worst right on parameter each Frequency point that just setting module is obtained Object vector than focusing performance parameter and in area pellucida, in conjunction with acoustic energy compared with control and least square method criterion, is adjusted respectively Loudspeaker array on each Frequency point whole is input into the amplitude of multiple weight coefficient vector and phase place；

Time domain impulse response signals computing module, for multiple for input weight coefficient vector adjusting module is obtained each raise one's voice Multiple weight coefficient vector component frequency response on each Frequency point on device passage, to frequency response adding window, and is carried out in inverse Fu Leaf transformation, obtains the time domain impulse response signals of each passage.