CN104936089B - A kind of multi-channel system compressing method - Google Patents

Abstract

A multi-channel system simplification method, assuming that the speaker group to be replaced and the replacement speaker group are located on the same spherical surface, the center of the sphere is the listening point, the position information of all speakers is obtained, and the replacement speaker group is calculated to replace a single speaker to be replaced The initial distribution coefficient of , get the signal of the replacement speaker group and delete the speaker to be replaced, and get the final signal of each speaker in the replacement speaker group by summing. The traditional method uses three loudspeakers to replace a loudspeaker to be replaced, which can simplify the 22.2 multi-channel system to 10-channel and 8-channel systems, but fails to fully maintain the physical properties of the reconstructed sound at the listening point. The present invention uses the replacement loudspeaker All the speakers in the group participate in the replacement, which can more fully maintain the physical properties of the sound at the listening point, and can obtain the optimal distribution coefficient for multiple speakers to replace one speaker, which is conducive to the improvement of the sound field effect after the multi-channel system is simplified.

Description

A Streamlined Method for Multi-channel System

technical field

The invention belongs to the field of acoustics, in particular to a method for simplifying a multi-channel system.

Background technique

With the development of 3D TV and 3D movie technology, 3D audio technology has become a research hotspot in the field of multimedia. Three-dimensional panning is a technique that uses several speakers in a three-dimensional sound field to create a virtual sound source. Among the three-dimensional panning methods, a vector based amplitude panning technique (vector based amplitude panning abbreviated as VBAP) is widely recognized. In three-dimensional VBAP, the virtual sound source is synthesized by three speakers, the starting point is at the listening point, and the unit vector ending at the position of the virtual sound source can be linearly represented by three other vectors (the starting point of each vector is at the listening point, and the ending point is at Each vector corresponds to the position of the loudspeaker, and the length is 1). After the coefficients represented by the three vectors are normalized, they are used as weight coefficients so that the signals of the virtual sound source are distributed to the three speakers corresponding to the representation vectors. If the number of speakers used to represent the virtual sound source is more than three, the VBAP technology divides the entire reconstructed speaker space into several subspaces in groups of three speakers, and distributes signals in each subspace according to VBAP.

5.1 multi-channel systems used to be very popular home theater sound systems. However, with the development of 3D video technology, higher requirements are placed on audio technology. Now multi-channel audio research focuses on more advanced systems with more channels, which can provide people with a better sense of immersion. For example, the 22.2 multi-channel system of NHK Laboratories has been used for UHD TV broadcasting. This advanced multi-channel system requires speakers to be placed in their own unique way to produce the best sound. Although 24 loudspeakers can be placed optimally in a theater, they can be cumbersome to place in a home application. "Downmixing" is a great way to reduce speaker channels in multi-channel systems. Downmixing from 5.1 to two-channel stereo or mono has been standardized by the ITU-R Recommendation and is used in some TV receivers. Although this downmixing method is very efficient, it is not suitable for any number of loudspeaker configurations. In order to make the downmix between multiple systems feasible, a new sound field reconstruction or conversion technology is urgently needed.

In 2011, Akio Ando of the Japan Broadcasting Association Laboratory proposed a new downmixing method, which uses three speakers on the same spherical surface to replace a speaker surrounded by a spherical triangle formed by the three speakers (these four speakers are located at On the same sphere, the center of the sphere is the listening point), to ensure that the sound pressure and the direction of the proton velocity at the listening point before and after replacement remain unchanged. This method provides a physical basis for the VBAP technology. Repeated use of this method can gradually change the 22.2 The multi-channel system is reduced to a 10-channel or 8-channel system (the two low-frequency channels are not processed), and the listening effect is better than the traditional down-mixing method. The physical properties of sound can be represented by both sound pressure and proton velocity. Sound pressure is a scalar with only magnitude, and proton velocity is a vector with magnitude and direction. Therefore, Akio Ando's method only guarantees the sound pressure before and after the streamlining and the direction of the proton velocity, ignoring the proton velocity, which will bring errors to the reconstruction of the 22.2 multi-channel system.

Contents of the invention

Aiming at the deficiencies of the prior art, the present invention provides a multi-channel system simplification method.

The technical solution of the present invention provides a multi-channel system simplification method, comprising the following steps,

Step 1, assume that the replacement speaker group and the speaker group to be replaced are located on the same spherical surface, the speaker group to be replaced includes m speakers Spv1, Spv2, ..., Spvm, and the original signals of the speakers Spv1, Spv2, ..., Spvm are all S, the replacement The loudspeaker group includes n loudspeakers Sp1, Sp2, ..., Spn, n<m, and the position of the center of the sphere is the listening point; obtain n loudspeakers Sp1, Sp2, ..., Spn to replace a single loudspeaker Spvh to be replaced. , h=1, 2,..., m, including the following sub-steps,

Step 101, obtaining the respective position information of m loudspeakers Spv1, Spv2, ..., Spvm, and the respective position information of n loudspeakers Sp1, Sp2, ..., Spn;

Step 102, calculate and determine the initial distribution coefficients w _h1 , w h2 , ..., w _hn of n loudspeakers Sp1, _Sp2 , ..., Spn, including

The construction equation is as follows,

in,

W _h = (w _h1 w _h2 ... w _hn ) ^T , T represents matrix transposition, w _hj represents the signal distribution coefficient of the replacement speaker Spj for the speaker Spvh to be replaced, j=1, 2,..., n;

E ₁ =(0 0 1) ^T ;

θ _vh represents the angle between the projection of the line between the position of the speaker Spvh to be replaced and the origin O on the XOY plane and the X axis, h=1, 2,..., m;

Indicates the angle between the line between the position of the speaker Spvh to be replaced and the origin O and the XOY plane, h=1, 2, ..., m;

θ _j represents the angle between the projection of the line between the position of the speaker Spj and the origin O on the XOY plane and the X axis, j=1, 2,..., n;

Represent the angle between the line between the position of the speaker Spj and the origin O and the XOY plane, j=1, 2,..., n; Step 103, multiply the signal of the single speaker Spvh to be replaced by the initial value obtained in step 102 respectively Partition coefficient

After w _h1 , w _h2 ,...,w _hn are assigned to corresponding speakers Sp1, Sp2,..., Spn, delete the speaker Spvh to be replaced, h=1, 2,..., m;

Step 2, calculate and obtain the respective signals of the speakers Sp1, Sp2, ..., Spn, including summing a series of corresponding signals obtained by distributing the loudspeaker Spj obtained in step 1, the final distribution signal calculation formula of the loudspeaker Spj is as follows:

Wherein, j=1, 2, . . . , n.

Also, n=10, m=22.

Compared with the down-mixing method proposed by AkioAndo in 2011, the m loudspeaker systems provided by the present invention can be simplified into n loudspeaker systems, which can only ensure that the sound pressure and the direction of the proton velocity at the listening point remain unchanged before and after replacement. The method of this patent not only ensures that the sound pressure at the listening point and the direction of the proton velocity remain unchanged before and after the replacement, but also ensures that the error of the proton velocity before and after the replacement is the smallest, which is conducive to improving the effect of synthesizing the virtual sound source. The invention can simplify the multi-channel system including any number of loudspeakers, and has good universality.

Description of drawings

Fig. 1 is the basic flowchart of the present invention.

Fig. 2 is a diagram showing the placement of speakers to be replaced according to the embodiment of the present invention.

Fig. 3 is a diagram showing placement positions of replacement loudspeakers according to the embodiment of the present invention.

Detailed ways

The present invention proposes a multi-channel system simplification method for multi-channel systems placed on the same spherical surface. The core technology of the method is to use any number of speakers to replace a speaker to be replaced. The following is combined with the accompanying drawings and specific implementation Examples illustrate the technical solution of the present invention in detail.

The goal of an embodiment is to reduce a multi-channel system comprising 22 loudspeakers Spv1, Spv2, ..., Spv22 to a multi-channel system comprising 10 loudspeakers Sp1, Sp2, ..., Sp10. It is assumed that 22 speakers Spv1, Spv2, ..., Spv22 are speaker groups to be replaced, and the original signal of each speaker is S, and 10 speakers Sp1, Sp2, ..., Sp10 are replacement speaker groups. Therefore, the present invention mainly uses 10 speakers to replace a speaker to be replaced, and gradually replaces 22 speakers Spv1, Spv2, . . . , Spv22 with Sp1, Sp2, . . . , Sp10. Both the replacement speaker group and the speaker group to be replaced are located on the same spherical surface, the radius of the spherical surface is 2 meters, and the position of the center of the spherical surface is the listening point.

During specific implementation, the method of the present invention can use computer software technology to realize the automatic operation process. The embodiment uses 10 loudspeakers to replace a loudspeaker to be replaced and gradually realizes the simplification process of the 22.2 multi-channel system. Referring to FIG. 1, the process provided by the embodiment includes the following steps:

Step 1, assuming that m speakers Spv1, Spv2, ..., Spvm are located on the same spherical surface, the original signals of the speakers Spv1, Spv2, ..., Spvm are all S, and the speakers Spv1, Spv2, ..., Spvm are the speaker groups to be replaced; n Speakers Sp1, Sp2, ..., Spn are replacement speaker groups (n<m), which are used to simplify speaker groups Spv1, Spv2, ..., Spvm to be replaced, and speakers Sp1, Sp2, ..., Spn can be connected with speakers Spv1, Spv2, ... 1. The positions of some speakers in Spvm are the same, or they can be completely different. Obtain n loudspeakers Sp1, Sp2, ..., Spn to replace a loudspeaker Spvh (h=1, 2, ..., m) to be replaced by distributing the obtained signal, the replacement loudspeaker group and the loudspeaker group to be replaced are all located on the same spherical surface, and the spherical surface The center of the ball is the listening point. During specific implementation, those skilled in the art can determine the values of n and m according to specific situations, and the implementation methods are the same.

The embodiment obtains the signals distributed by replacing a single speaker to be replaced by 10 speakers Sp1, Sp2, . . . , Sp10. Include the following sub-steps,

Step 101, obtaining the respective position information of 22 loudspeakers Spv1, Spv2, ..., Spv22 to be replaced, and the respective position information of 10 replacement loudspeakers Sp1, Sp2, ..., Sp10;

Let the listening point be the coordinate origin O to set up a three-dimensional rectangular coordinate system XYZ, the present invention adopts the polar coordinate form, such as the coordinates of point A (ρ _A , θ _A , ), ρ _A represents the distance between point A and the origin of the coordinates, θ _A represents the angle between the projection of the line between point A and origin O on the XOY plane and the X axis, Indicates the angle between the line between point A and origin O and the XOY plane. Assume that the coordinates of a single speaker Spvh (h=1, 2, ..., 22) to be replaced are (ρ, θ _vh , ), the coordinates of the 10 replacement speakers Sp1, Sp2, ..., Sp10 are (ρ,θ ₁ , ), (ρ,θ ₂ , ), (ρ,θ ₃ , ), (ρ,θ ₄ , ), (ρ,θ ₅ , ), (ρ,θ ₆ , ), (ρ,θ ₇ , ), (ρ,θ ₈ , ), (ρ,θ ₉ , ), (ρ,θ ₁₀ , ).

Assume that in this embodiment, the hollow dots indicate the location of the listening point, and the solid dots indicate the location of the speaker. The replacement speakers Sp1, Sp2, ..., Sp10 are respectively located at points on the surface of the ball O, see Figure 3, the speakers are mainly distributed at high angles Elevation=45°, 0°, -30°, and the coordinates are: Sp1(2,0° ,90°), Sp2(2,0°,45°), Sp3(2,90°,45°), Sp4(2,180°,45°), Sp5(2,0°,0°), Sp6(2 ,60°,0°), Sp7(2,120°,0°), Sp8(2,180°,0°), Sp9(2,270°,0°), Sp10(2,90°,-30°), speakers to be replaced Groups Spv1, Spv2, ..., Spv22 are respectively located at points on the surface of the ball O, see Figure 2, the coordinates are: Spv1(2,0°,90°), Spv2(2,0°,45°), Spv3(2, 45°,45°), Spv4(2,90°,45°), Spv5(2,135°,45°), Spv6(2,180°,45°), Spv7(2,225°,45°), Spv8(2,270°, 45°), Spv9(2,315°,45°), Spv10(2,0°,0°), Spv11(2,30°,0°), Spv12(2,60°,0°), Spv13(2, 90°,0°), Spv14(2,120°,0°), Spv15(2,150°,0°), Spv16(2,180°,0°), Spv17(2,225°,0°), Spv18(2,270°,0° ), Spv19(2,315°,0°), Spv20(2,45°,-30°), Spv21(2,90°,-30°), Spv22(2,135°,-30°).

Step 102, calculate and determine the initial allocation coefficients of the 10 replacement speakers Sp1, Sp2, . . . , Sp10.

According to the sound pressure magnitude and proton velocity direction produced by a single loudspeaker Spvh (h=1, 2, ..., m) to be replaced at the listening point and the sound pressure magnitude and proton velocity produced by n replacement loudspeakers at the listening point The directions are strictly equal, and the error between the proton velocity generated by a single speaker Spvh (h=1, 2, ..., m) at the listening point and the proton velocity generated by n replacement speakers at the listening point is as small as possible Based on the principle, the present invention proposes an initial distribution coefficient calculation scheme:

The embodiment calculates the initial allocation coefficients w _h1 , w _h2 , ..., w _h10 (h=1, 2, ... ,twenty two).

The sound pressure p _vh produced by a single speaker Spvh to be replaced at the listening point is:

The sound pressure produced by 10 replacement speakers Sp1, Sp2, ..., Sp10 at the listening point is _ph :

in:

ρ represents the distance between the position of a single speaker Spvh to be replaced and the coordinate origin O;

G represents the proportional coefficient between the sound pressure of the speaker at a unit distance from the speaker and the sound pressure generated by the speaker;

e is a mathematical constant;

i is the imaginary unit;

k is the wave number, f is the sound signal frequency;

c is the speed of sound in air;

w _hj represents the signal distribution coefficient of the replacement loudspeaker Spj for the loudspeaker Spvh to be replaced, j=1,2,...,10, h=1,2,...,22;

s(ω) represents the Fourier transform of the speaker input signal.

Calculated by formula (1) and formula (2):

w _h1 +w _h2 +…+w _h10 ＝1 (3)

The proton velocity u _vh produced by a single speaker Spvh to be replaced at the listening point is:

in:

ρ represents the distance between the position of a single speaker Spvh to be replaced and the coordinate origin O;

θ _vh represents the angle between the projective projection of the single speaker Spvh to be replaced and the origin O on the XOY plane and the X axis;

Represent the angle between the line between the position of the single speaker Spvh to be replaced and the origin O and the XOY plane;

G represents the proportional coefficient between the sound pressure of the speaker at a unit distance from the speaker and the sound pressure generated by the speaker;

e is a mathematical constant;

i is the imaginary unit;

k is the wave number, f is the sound signal frequency;

c is the speed of sound in air;

λ is the air density;

s(ω) represents the Fourier transform of the speaker input signal.

Proton velocity produced by 10 replacement speakers Sp1, Sp2, ..., Sp10 at the listening point for:

W _h ＝(w _h1 w _h2 … w _h10 ) ^T

in:

θ _j represents the angle between the projection of the line between the position of the replacement speaker Spj and the origin O on the XOY plane and the X axis, j=1, 2, ..., 10;

Indicates the angle between the line between the position of the replacement speaker Spj and the origin O and the XOY plane, j=1, 2, ..., 10;

w _hj respectively represent the signal distribution coefficients of the replacement speakers Spj, j=1, 2, . . . , 10, h=1, 2, . . . , 22.

From (4) and (5) are equal to get:

In the formula (6), the first line and the second line are respectively divided by the third line to get:

Equation (7) ensures that the direction of the proton velocity generated by the speaker Spvh to be replaced at the listening point is equal to the direction of the proton velocity generated by the 10 replacement speakers Sp1, Sp2, . . . , Sp10 at the listening point. From equation (7), we can get:

Combine (8) and (3) to get:

in:

The error E(w _h1 ,w _h2 ,…,w _h10 ) of the proton velocity at the listening point is:

For a given signal s(ω), the position information of 10 replacement speakers Sp1, Sp2, ..., Sp10 and the position information of a single speaker Spvh (h=1, 2, ..., 22) to be replaced, as a constant. Therefore, to minimize the error of the proton velocity, that is, to minimize the value of the formula _E2 .

Then using 10 replacement speakers Sp1, Sp2, ..., Sp10 to replace a single speaker Spvh (h=1, 2, ..., 22) to be replaced is equivalent to solving:

in,

W _h ＝(w _h1 w _h2 … w _hn ) ^T , T means matrix transposition;

E ₁ =(0 0 1) ^T ;

θ _vh represents the angle between the projection of the line between the position of the speaker Spvh to be replaced and the origin O on the XOY plane and the X axis, h=1, 2, ..., m;

Represent the angle between the line between the position of the speaker Spvh to be replaced and the origin O and the XOY plane, h=1,

2,...,m;

θ _j represents the angle between the projection of the line between the position of the replacement speaker Spj and the origin O on the XOY plane and the X axis, j=1, 2, ..., n;

Indicates the angle between the line connecting the position of the replacement speaker Spj and the origin O and the XOY plane, j=1, 2, . . . , n.

Equation (12) can be solved by an existing mature algorithm, and this embodiment uses a trust domain algorithm to solve it. This equation can also be used for other values of n and m.

In this embodiment, according to formula (12), it can be obtained that 10 replacement speakers Sp1, Sp2, ..., Sp10 are used to replace a single speaker Spvh (h=1, 2, ..., 22) to be replaced respectively to obtain 10 replacement speakers Sp1, Sp2 , ..., a series of initial partition coefficients of Sp10, as shown in Table 1.

Table 1 Signal Initial Distribution Coefficient

Step 103, the signal of a single speaker Spvh (h=1, 2, ..., 22) to be replaced is multiplied by the initial distribution coefficients w _h1 , w _h2 , ..., w _h10 obtained in step 102, and then distributed to the corresponding speakers Sp1, Sp2, ..., Sp10, delete the speaker to be replaced Spvh (h=1, 2, ..., 22);

In this embodiment, the signal of the speaker Spv1 to be replaced is multiplied by the initial distribution coefficient 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 in Table 1 to obtain the speakers Sp1, Sp2, ... , the initial distribution signal of Sp10, and delete the speaker Spv1 to be replaced. Similar operations are performed on the signals of the remaining speakers to be replaced in turn.

Step 2, calculate the respective signals of the speakers Sp1, Sp2, ..., Spn, including summing a series of corresponding signals obtained from the distribution of the speakers Spj obtained in step 1, the final distribution signal calculation formula of the speakers Spj is Wherein, j=1, 2, . . . , n.

The embodiment calculates and obtains the signal of replacing loudspeaker Sp1, Sp2, ..., Sp10 respectively, and the final distribution signal calculation formula of replacing loudspeaker Spj (j=1, 2, ..., 10) is as follows:

In this embodiment, the initial allocation coefficients of the replacement speaker Sp1 obtained during the 22 replacements of the speakers Spv1, Spv2, ..., Spv22 to be replaced are respectively: 1, 0, 0, 0, 0, 0, 0.1464, 0.5, 0.1464 , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, as shown in the second column of Table 1, according to formula (13), the final distribution signal of the replacement speaker Sp1 is _Sf1 = 1.7928S. Similar calculations can be performed to obtain the final signal for the remaining replacement speakers.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which the present invention belongs can make various modifications or supplements to the described specific embodiments or adopt similar methods to replace them, but they will not deviate from the spirit of the present invention or go beyond the definition of the appended claims range.

Claims (2)

1. A multi-channel system reduction method, characterized by: comprises the following steps of (a) carrying out,

step 1, setting a replacement loudspeaker group and a to-be-replaced loudspeaker group to be located on the same spherical surface, wherein the to-be-replaced loudspeaker group comprises m loudspeakers Spv1, Spv2, … and Spvm, original signals of the loudspeakers Spv1, Spv2, … and Spvm are S, the replacement loudspeaker group comprises n loudspeakers Sp1, Sp2, … and Spn, n is less than m, and the spherical center position of the spherical surface is a listening point; obtaining the signals assigned by the n loudspeakers Sp1, Sp2, …, Spn to replace the single loudspeaker to be replaced Spvh, h being 1, 2, …, m, comprising the sub-steps of,

step 101, obtaining the position information of m loudspeakers Spv1, Spv2, … and Spvm respectively, and the position information of n loudspeakers Sp1, Sp2, … and Spn respectively;

102, calculating and determining initial distribution coefficients w of the n loudspeakers Sp1, Sp2, … and Spn_h1,w_h2,…,w_hnThe method comprises the following steps of constructing an equation,

s.t.LW_h＝E₁

w_h1,w_h2,…w_hn≥0

wherein,

W_h＝(w_h1w_h2… w_hn)^Tt denotes the matrix transposition, w_hjA signal distribution coefficient representing the replacement speaker Spj for the speaker Spvh to be replaced, j being 1, 2, …, n;

E₁＝(0 0 1)^T；

θ_vhan included angle between the projection of the XOY plane and the X axis, which is a connecting line between the position of the speaker Spvh to be replaced and the origin O, is represented, and h is 1, 2, …, m;

representing the angle between the line between the position of the loudspeaker Spvh to be replaced and the origin O and the XOY plane,h＝1，2，…，m；

θ_jan angle between the projection of the line between the position of the speaker Spj and the origin O on the XOY plane and the X axis is represented, j is 1, 2, …, n;

an included angle between a connecting line between the position of the speaker Spj and the origin O and the XOY plane is represented, and j is 1, 2, …, n; step 103, multiplying the signals of the individual speakers Spvh to be replaced by the initial distribution coefficients w obtained in step 102 respectively_h1,w_h2,…,w_hnThen, the signals are distributed to corresponding loudspeakers Sp1, Sp2, … and Spn, and the loudspeaker Spvh, h being 1, 2, … and m to be replaced is deleted;

step 2, calculating to obtain signals of the speakers Sp1, Sp2, … and Spn respectively, including summing a series of corresponding signals obtained by distributing the speakers Spj obtained in step 1, wherein a final distribution signal calculation formula of the speakers Spj is as follows:

wherein j is 1, 2, …, n.

2. The multi-channel system reduction method of claim 1, wherein: n is 10 and m is 22.