JP2015231087A - Local acoustic reproduction device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To implement an accurate local reproduction by calculating a stable driving signal without complicated computation by analytically deriving a spatial filter function and further, to accurately implement a multi local acoustic reproduction by simultaneously executing multiple local reproductions while changing positions or widths of the local reproductions.SOLUTION: A local reproduction device (10) includes a linear speaker array (12) formed by disposing a plurality of speakers (14, 14, ..., 14) at fixed intervals on a straight line in a direction of an (x) axis. In a drive unit (16), a sound pressure (p) at a listening position of a predetermined distance yfrom the linear speaker array is modeled by a rectangular window of which the bright region is "1" and the dark region is "0", and spatial Fourier transformation in the direction of the (x) axis is performed. A position xof the bright region is changed as needed and a filter function (f) is calculated. A drive signal (d) of each of the speakers is generated from the special filter function (f) and a sound source (s), and each of the speakers is driven.

Description

  The present invention relates to a local sound reproducing apparatus and program, and in particular, a novel local sound reproducing for locally reproducing sound with an arbitrary width at an arbitrary position with respect to a linear speaker array composed of a large number of linearly arranged speakers. The present invention relates to a device and a program.

  There are roughly two types of local sound reproduction using a large number of speakers. One is a method based on beam forming for controlling the directivity of the speaker array, and is disclosed in Patent Document 1, for example. The other is a method based on multi-point control in which the sound pressure at the control point is directly controlled by each speaker. For example, Non-Patent Document 1 and Non-Patent Document 2 have recently proposed several methods. The method based on beamforming is a method for maximizing the directivity in the target direction, whereas the method based on multipoint control is a method for maximizing the acoustic energy difference between control points. Therefore, in the case of a system that forms an area where sound cannot be heard, the latter method has high performance.

JP 2012-147414 [H04R 3/12, 3/00, 1/40, 1/34]

J.-W. Choi and Y.-H. Kim, “Generation of an acoustically bright zone with an illuminated region using multiple sources,” J. Acoust.Soc. Am., Vol. 111, no. 4, pp. 1695 -1700, Apr. 2002. M. Shin, SQ Lee, FM Fazi, PA Nelson, D. Kim, S. Wang, KH Park, and J. Seo, “Maximization of acoustic energy difference between two spaces,” J. Acoust. Soc. Am., Vol 128, no. 1, pp. 121-131, July 2010.

  In Non-Patent Document 1 that employs the contrast maximization method, the inverse matrix operation of the spatial correlation matrix between each speaker and the control points is used. It becomes stable. In order to solve this problem, regularization of the spatial correlation matrix is necessary, but there are limits to the appropriate regularization parameters, and it is necessary to perform iterative calculation for each parameter in order to determine the parameters. .

  On the other hand, the energy difference maximization method (EDM) as in Non-Patent Document 2 is based on the calculation of eigenvectors, not the inverse of the spatial correlation matrix, so there is no problem of unstable solutions. The local sound reproduction is realized with higher accuracy than that of Non-Patent Document 1.

  However, in this method, since it is necessary to determine the tuning coefficient for each frequency, it is necessary to perform an iterative calculation for determining an appropriate coefficient as in the regularization problem.

  Therefore, a main object of the present invention is to provide a novel local sound reproducing device and program.

  Another object of the present invention is to provide a local sound reproducing device and a program that enable highly accurate local sound reproduction without requiring cumbersome calculations.

  The present invention employs the following configuration in order to solve the above problems. The reference numerals in parentheses, supplementary explanations, and the like indicate correspondence relationships with embodiments described later to help understanding of the present invention, and do not limit the present invention in any way.

  A first aspect of the present invention is a local sound reproducing apparatus using a linear speaker array in which a plurality of speakers are linearly arranged at intervals, and the sound pressure at a reproduction position is modeled by a rectangular window having an arbitrary width. Modeling means for obtaining a model, deriving means for deriving a spatial filter function of each of the plurality of speakers based on the sound pressure model, and generating a driving signal for the plurality of speakers based on the sound source and the spatial filter function, It is a local sound reproduction apparatus provided with the drive means which drives each speaker.

In the first invention, the local sound reproducing device (10: reference numeral exemplifying a corresponding part in the embodiment, hereinafter the same) separates a plurality of speakers (14 ₁ , 14 ₂ ,..., 14 _N ) from each other. A linear speaker array (12) arranged linearly with (Δx) therebetween is used. For example, the modeling means (16) models the sound pressure (p) at the reproduction position (y _b ) parallel to the linear speaker array (12) with a rectangular window having an arbitrary width (l _b ), thereby generating a sound pressure model ( For example, the number [9]) is obtained. Based on the sound pressure model (p), the derivation means (16) analyzes the spatial filter function (f _i : for example, number [12]) of each of the plurality of speakers by using a method such as spatial Fourier transform. Derived automatically. Then, the drive means (16) generates a drive signal (d _i ) for the plurality of speakers (14 _i ) based on the sound source (s) and the spatial filter function (f _i ), and the drive signal (d _1). , _d 2, ..., each of the speakers _{(14 1 d N), 14 2, ...} , to drive the 14 _N).

  According to the first invention, since the spatial filter function can be derived analytically, no cumbersome calculation is required, and it is possible to perform local sound reproduction by setting the bright region and the dark region with high accuracy.

  A second invention is dependent on the first invention, wherein the modeling means obtains a sound pressure model modeled by a plurality of rectangular windows at different positions, and the derivation means obtains a sound pressure model modeled by the plurality of rectangular windows. Is a local sound reproduction device for deriving a spatial filter function according to

  In the second invention, the modeling means (16) obtains a sound pressure model from a plurality of rectangular windows at different positions, and the derivation means (16) analytically derives a spatial filter function according to each sound pressure model.

  According to the second invention, so-called multi-local sound reproduction with different bright areas becomes possible.

  A third invention is a local sound reproduction program executed by a computer of a local sound reproduction device using a linear speaker array in which a plurality of speakers are linearly arranged at intervals, and the local sound reproduction program is a computer. Modeling means for obtaining a sound pressure model by modeling the sound pressure at a reproduction position with a rectangular window of an arbitrary width, and a derivation means for deriving a spatial filter function (f) of each of a plurality of speakers based on the sound pressure model, And a local sound reproduction program that generates drive signals for a plurality of speakers based on a sound source and a spatial filter function, and causes the speakers to function as drive means for driving each speaker.

  In the third invention, the same effect as in the first invention can be expected.

  A fourth invention is dependent on the third invention, wherein the modeling means obtains a sound pressure model modeled by a plurality of different rectangular windows, and the derivation means has a sound pressure modeled by the plurality of rectangular windows. A local sound reproduction program for deriving a spatial filter function according to a model.

  In the fourth invention, the same effect as in the second invention can be expected.

  A fifth invention is a local sound reproducing device that reproduces sound locally at an arbitrary position at an arbitrary position with respect to the linear speaker array using a linear speaker array composed of a large number of speakers arranged linearly. Width defining means for defining the width of sound that can be heard in parallel with the speakers of the linear speaker array, position defining means for defining the position at which sound can be heard at a position parallel to the arrangement of the speakers of the linear speaker array, and width defining And a speaker driving means for driving each speaker of the linear speaker array by generating a drive signal by the spatial filter function set by the means and the position defining means and the sound source.

In the fifth invention, the local sound reproducing device (10) includes a linear speaker array (12) in which a plurality of speakers (14 ₁ , 14 ₂ ,..., 14 _N ) are linearly arranged with an interval (Δx) therebetween. Is used. The width defining means (16) defines, for example, a width (l _b ) in which sound can be heard in parallel with the arrangement of the speakers (x direction) of the linear speaker array, in which a bright region is to be set. The position defining means (16) defines, for example, a position (x _b ) where sound can be heard in parallel with the arrangement (x direction) of the speakers of the linear speaker array where a bright region should be set. The speaker driving means (16) generates a drive signal (d _i ) based on the spatial filter function (f _i ) set by the width defining means and the position defining means and the sound source (s) given from, for example, the sound source unit (18). And each speaker is driven by the drive signal.

  According to the fifth invention, since the spatial filter function can be derived analytically by defining the width and the position, no complicated calculation is required, and the bright region and the dark region are set with high accuracy, and the local acoustic wave is determined. Playback can be performed.

  A sixth invention is a local sound reproducing device according to the fifth invention, further comprising distance defining means for defining a distance from the linear speaker array, wherein the spatial filter function is further set by the distance defining means.

In the sixth invention, the distance defining means (16) defines the distance (y _b ) from the linear speaker array in the direction orthogonal to the linear speaker array. The spatial filter function is further set by the distance defining means, and the drive signal is generated by the spatial filter function and the sound source.

  According to the sixth aspect, since the distance is defined by the distance defining means, the accuracy becomes higher.

  According to a seventh aspect of the present invention, there is provided a computer for a local sound reproducing apparatus that reproduces sound locally at an arbitrary position at an arbitrary position with respect to the linear speaker array using a linear speaker array including a plurality of speakers arranged in a straight line. A local sound reproduction program executed by the computer according to claim 1, wherein the local sound reproduction program is configured to control the computer with respect to the arrangement of the speakers of the linear speaker array; The position defining means for defining the position where sound can be heard at the parallel position, and the spatial filter function set by the width defining means and the position defining means and the sound source generate drive signals to drive each speaker of the linear speaker array. This is a local sound reproduction program that functions as a speaker driving means.

  In the seventh invention, the same effect as in the fifth invention can be expected.

  The eighth invention is dependent on the seventh invention, and further causes the computer to function as a distance defining means for defining a distance from the linear speaker array, and the drive signal is supplied by the distance defining means in addition to the spatial filter function. A local sound reproduction program to be set.

  In the eighth invention, the same effect as in the sixth invention can be expected.

  According to the present invention, since the spatial filter function can be derived analytically, highly accurate local sound reproduction can be performed without troublesome calculations.

  The above object, other objects, features and advantages of the present invention will be described with reference to the drawings. It will become more apparent from the detailed description of the following examples.

FIG. 1 is a schematic diagram showing the configuration of a local sound reproducing apparatus according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing a state of sound field reproduction using an infinite linear secondary sound source based on a spectrum division method (SDM). FIG. 3 is a schematic diagram showing local sound reproduction using an infinite linear secondary sound source by wave number domain spatial filtering. FIG. 4 is a schematic diagram showing the configuration of a local sound reproducing device according to another embodiment of the present invention. FIG. 5 is a schematic diagram showing a rectangular window for calculating a drive signal when a finite number of linear speaker arrays is used. FIG. 6 is a schematic diagram showing the arrangement of speakers and local sound reproduction control (bright area and dark area) in EDM. FIG. 7 is a graph showing the results of spatio-temporal average sound pressure levels in computer simulation in comparison with EDM. FIG. 8 is a graph showing the results of BDR in a computer simulation in comparison with EDM, showing the results of 500 Hz, 1 kHz, and 2 kHz in order from the top, in which both the left side is EDM and the right side is the example. FIG. 9 is a graph showing the results of time average sound pressure levels in computer simulation in comparison with EDM, showing the results of 3 kHz, 4 kHz, and 5 kHz in order from the top, all of which are EDM on the left side and the examples on the right side. Is.

FIG. 1 is a schematic diagram showing a local sound reproducing apparatus according to an embodiment of the present invention. The local sound reproducing device 10 of this embodiment uses a linear speaker array 12 in which a large number of speakers are arranged linearly. The linear speaker array 12 has a plurality (N = 64 in the embodiment) of speakers 14 ₁ , 14 ₂ ,..., 14 _N arranged linearly at a predetermined interval, for example, every 5 cm. As an example, it has a diameter of 1 inch (2.5 cm). However, these numerical values can be changed as appropriate.

For example, the drive device (computer) 16 formed of a computer has drive signals d ₁ , d ₂ ,..., D based on a sound source s (ω) given from a sound source unit 18 such as an appropriate storage medium (memory, HDD, optical disk, network, etc.). _N are created, and the speakers 14 ₁ , 14 ₂ ,..., 14 _N are driven by these drive signals.

  In the local sound reproduction to which the present invention is directed, “acoustic” means sound, music, and other sound signals related to sound, and the sound signal is supplied from the sound source 18 to the driving device 16.

  However, in this embodiment, of the listeners P1, P2, P3,... Existing in front of the linear speaker array 12, as an example, only the listener P2 hears the sound source s and cannot hear other listeners. Local sound reproduction. Here, the region where the sound source s can be heard is expressed as “Bright” in FIG. 1, and the region that cannot be heard is expressed as “Dark”. In the bright area, all listeners can listen to the sound source regardless of whether one or more listeners exist in the area.

In order to form such bright areas and dark areas, in this embodiment, the driving device, that is, the computer 16 calculates a spatial filter function F _i for each speaker 14 as will be described later. By multiplying the sound source s (ω), a drive signal d _i for each speaker is generated. Then, the driving signal _{d 1,} d 2 for the speakers produced, ..., the speakers ₁₄ 1 _{d N,} 14 2, ..., to drive the 14 _N.

FIG. 2 is a schematic diagram showing sound field reproduction by SDM as a premise of the embodiment in FIG. However, although the embodiment of FIG. 1 uses a linear speaker array 12 composed of a finite number of speakers 14 ₁ , 14 ₂ ,..., 14 _N , in the following description, infinite linear linear speakers (infinite linear Consider sound field reproduction assuming a secondary sound source.

  However, for SDM, reference 1 (J. Ahrens, and S. Spors, “Sound field reproduction using planar and linear arrays of loudspeakers,” IEEE Trans. Audio, Speech, Lang. Process., Vol. 18, no. 8, pp. 2038-2050, Nov. 2010.).

The infinite linear secondary sound source is assumed to be on the x axis (y = 0), and the z axis is omitted. The frequency is f, and the angular frequency and omega = 2 [pi] f, infinitely linear secondary position of the sound source _x 0 = _[x, ^0] a drive signal in the ^T d (x0, ω) and when the position x = [x, y] The sound pressure p (x, ω) at ^T is given by the number [1].

  In each of the following expressions, the function before Fourier transformation is expressed in lower case letters, and the function after Fourier transformation is expressed in upper case letters instead of using “^ (hat)”.

here,

Is a three-dimensional Green function indicating the characteristic of transmitting sound from an infinite linear secondary sound source, j = √−1, k = ω / c indicates the wave number, and c indicates the speed of sound. However, reflection etc. are not considered here.

The number [1] indicates the sound pressure p obtained by summing the sound pressures from the speakers at the sound receiving position y _ref in front of the infinite linear secondary sound source. When the number [1] is spatial Fourier transformed in the x-axis direction, From the convolution theorem, Equation 3 is obtained.

Here, k _x is the spatial frequency in the x-axis direction, D (k _x , ω) is the Fourier transform of the drive signal, and G (k _x , y, ω) is the three-dimensional green function g _3D (x−x0, ω). Is represented by the number [4].

  However, for spatial Fourier transform, see Reference 2 (E. G. Williams, Fourier Acoustics: Sound Radiation and Nearfield Acoustic Holography, London: Academic Press, 1999.).

  By performing the spatial Fourier transform, the component amount of the wave from each direction at the sound receiving position can be obtained.

H ₀ ⁽²⁾ and k ₀ indicate a 0th-order second-type Hankel function and a 0th-order modified Bessel function, respectively.

Therefore, as shown in FIG. 2, when the spatial Fourier transform of the sound pressure at the sound receiving position (y = y _ref ) continuous in parallel with the x axis is P (k _x , y _ref , ω), in SDM, the wave number region The driving signal D (k _x , ω) of the infinite linear secondary sound source at can be analytically obtained from the number [3] as the number [5]. That is, the sound field from the infinite linear secondary sound source based on SDM at the sound receiving position y _ref is obtained by the number [5]. That is, it is already known in the above-mentioned Reference 1 that the drive signal D is obtained by Fourier transforming the sound pressure p and the transfer function g in SDM as shown in Equation [5].

  Based on the above, next, a method for forming a bright region and a dark region by the drive device, that is, the computer 16 as in the embodiment of FIG.

When the filter of each speaker for reproducing the sound source s (ω) only in a certain region is set to f (x ₀ , ω) using the infinite linear secondary sound source shown in FIG. The signal d (x ₀ , ω) can be described as the number [6].

However, the sound source s supplied from the sound source unit 18 shown in FIG. 1 may be an arbitrary sound source and assumes a monaural sound source.

In this embodiment, the optimal filter f (x ₀ , ω) is designed in the equation [6], thereby realizing local sound reproduction of the sound source s (ω) as in FIG. 1 embodiment.

Since the filter does not depend on the sound source, s (ω) = 1. Therefore, the spatial Fourier transform of the sound field at the sound receiving position y = y _ref by the filter f (x ₀ , ω) becomes the number [7] by the number [3].

Therefore, similarly to the number [5], the spatial filter F (k _x , ω) in the wave number domain is the number [8]. The spatial filter F shown by this number [8] shows the result of spatial Fourier transform of the filter function f for each speaker.

In order to perform local sound reproduction in the multipoint control method, the sound pressure at a place where sound can be heard is set to “1”, and the place where sound cannot be heard is set to “0”. Therefore, in this embodiment, as shown in FIG. 3, x = 0 in order to realize the local sound reproduction of the width l _b at a distance of y = y _b mainly, y = y sound in _b pressure p (x , Y _y ) are modeled as follows using a rectangular window of sound pressure distribution where “1” is heard and “0” is not heard, and the computer 16 obtains a sound pressure model shown in Equation [9]. . In that sense, it can be said that the computer 16 that calculates the sound pressure model according to the number [9] functions as modeling means.

  Since this rectangular window does not depend on the frequency, ω can be omitted.

Then, the spatial Fourier transform (see Reference 2) of the rectangular window p (x, y _b ) in the x-axis direction is as shown in the equation [10].

Furthermore, in order to realize local sound reproduction at an arbitrary position _xb , the number [10] becomes the number [11] according to the shift theorem (see Reference 2).

Therefore, if the number [11] is substituted into the number [8], the local reproduction filter F (k _x , ω) by the infinite linear secondary sound source in the wave number domain can be analytically derived as the number [12].

This spatial filtering, an infinite linear secondary source local sound reproduction areas of any width l _b can be realized any position [x _b, y b] to ^T using. Thus, the driving device or computer 16 (in the embodiment, the arrangement direction of the speaker 14 of the linear loudspeaker array 12) direction of infinite linear secondary source by setting the position x _b and window width l _b in the space Deriving a filter. Therefore, it can be said that the computer 16 functions as a position defining means and a width defining means. Further, the driving device, that is, the computer 16 uses the distance (y _b ) from the linear speaker array 12 to derive a spatial filter. Therefore, the computer 16 also functions as a distance defining means.

Although the sound source s = 1 is assumed as the premise of the number [7], in order to actually drive the speakers 14 ₁ , 14 ₂ ,..., 14 _N in FIG. The drive signal d is generated according to the equation [13] using the spatial filter function f and the sound source s derived in step [13].
[Equation 13]
d _i = s _i f _i
The drive device or computer 16, the number [13] drive signal generated for each speaker in _{d 1,} d 2, ..., by _{d N,} a speaker _{14 1,} 14 2, ..., to drive the 14 _N. Therefore, the driving device, that is, the computer 16 functions as driving means.

Up to several [12], an infinite linear secondary sound source was assumed for the calculation. However, as shown in FIG. 1, local sound reproduction using the mathematical expression [12] is limited to N speakers (64 in the embodiment). It is realized by using the linear speaker array 12 in which 14 ₁ , 14 ₂ ,..., 14 _N are arranged on a straight line. For that purpose, it is necessary to discretize the number [12] and cut it off with a finite length.

  Here, if the number of channels (number of speakers) of the linear speaker array 12 is N and the element spacing is Δx, the array length is L = (N−1) Δx. The discrete inverse Fourier transform (IDFT) of the number [12] is obtained by the number [14].

Here, x = nΔx, k _x = 2πm = / NΔx, and −N / 2 ≦ n ≦ N / 2-1.

  In proximity acoustic holography (reference 2), in order to reduce errors due to truncation of the aperture length of the array, “0” is virtually inserted at both ends of the aperture length of the sound pressure recording surface, and more than twice the aperture length. A DFT is calculated using points of 2N or more using a surface of length. Similarly to this method, in the embodiment of FIG. 1, as shown in FIG. 5, the length of the rectangular window of the sound pressure distribution is set to 2L which is twice the array length, and the number [14] is obtained by IDFT using 2N points. ] Is calculated.

  Of the obtained 2N-channel drive signals, the center N-channel portion is used as the actual drive signal. In the embodiment of FIG. 1, the drive signal d is the central 64 channels of the 128-channel drive signals. In this way, the actual drive signal d is calculated by performing a discrete inverse Fourier transform on the number [12].

  In the embodiment of FIG. 1, the sound from the linear speaker array 12 is reproduced only to the listener P2 (in the region thereof) and cannot be heard by other listeners. That is, one bright region for the listener P2 was formed. Thus, one sound source can be reproduced only for the listener P2.

On the other hand, in the embodiment of FIG. 4, two sound source units 181 and 182 are provided, and two sound sources s ₁ (ω) and sound sources s ₂ (ω) are supplied from the sound source units 181 and 182 to the drive device, that is, the computer 16. To give. On the other hand, according to the method described below, a first bright area for the listener P2 and a second bright area for the listener P1 are set. Then, for example, the English sound source s ₁ (ω) is reproduced in the first bright region, and for example, the Japanese sound source s ₂ (ω) is reproduced in the second bright region. Therefore, this embodiment can be applied to, for example, a personal audio system, a multilingual simultaneous guide at an art museum or the Olympics, and other virtual reality technologies as in the embodiment of FIG.

Moreover, in the platform of the station where the limited express train and the ordinary train stop in common, there are many cases where the position where the expected express train rider is lined up and the ordinary train are different. In such a case, the embodiment shown in FIG. For example, the first bright area is set for the person P1 who rides on the limited express train, and the announcement (sound source s ₁ ) for the express train is made to flow, and at the same time, the second line is set for the person P2 who rides the regular train. By setting the bright area and playing the announcement for the ordinary train (sound source s ₂ ), it is possible for each prospective rider to get on the correct train without erroneously boarding.

In this way, in order to realize multi-local acoustic reproduction in which M sound sources s _i (ω) are reproduced in M different bright regions, as shown in FIG. The drive signal is obtained by the number [15] by superimposing the filter f _i at the sound reproduction position and the combination s _i (ω) f _i (x ₀ , ω) of the corresponding sound source s _i . That is, in the embodiment of FIG. 4, the driving device or computer 16, the sound receiving position y = y _b, in two different positions in the arrangement direction, that is the x direction of the speaker 14, and sets a rectangular window, each rectangular window A drive signal d _i is generated by the spatial filter function derived from the equation [12] and the sound sources s ₁ and s ₂ from the sound source units 181 and 182.

In other words, as the number [15], in the embodiment shown in FIG. 4, the driving device or computer 16, the sound source _{s i} and the speakers ₁₄ 1 by the spatial filter _{f i} of the speaker 14 _i, 14 2, ..., of 14 _N Drive signals d ₁ , d ₂ ,..., D _N are generated to drive each speaker.

  However, in the above description, a plurality of sound sources are reproduced locally in a plurality of bright areas. However, even when one sound source is reproduced locally in a plurality of bright areas, the embodiment shown in FIG. Of course, it can be used. That is, the embodiment shown in FIG. 4 is an embodiment in which a plurality of bright areas are individually set.

Also, in FIG. 4, the two bright areas are depicted as being set at different listening positions from the linear speaker array 12, but this is merely for illustration purposes, and in fact the two bright areas are the same. distance y = Note that is to be set on the y _b.

  Here, the effectiveness of the present invention is verified by computer simulation. The local sound reproduction accuracy is compared with EDM, which is one of the conventional methods.

In EDM, an N-channel speaker is used to control a K-channel control point. When each speaker position is x _sp and the control point position is x _co , the spatial correlation matrix between each speaker and the control point is

Is calculated as here,

It is.

In EDM, the spatial correlation matrix between each speaker and the control point of the local sound reproduction region is r _b (ω), the spatial correlation matrix between each speaker and the control point of the silent region is r _d (ω), and the filter of each speaker f (ω) is

As the eigenvector corresponding to the largest eigenvalue of. Here, α is a tuning coefficient, and is set for each angular frequency ω.

The accuracy of local sound reproduction assuming a three-dimensional free sound field is evaluated by computer simulation. Each speaker is assumed to be an omnidirectional point sound source. The simulation conditions and the speaker and local sound reproduction control line arrangement are shown in Table 1 and FIG. 6, respectively. The speaker array is arranged on the x axis with x = 0 as the center. x <0 is the bright region (x _b , p = 1), and x> 0 is a dark region (x _d , p = 0). In EDM, control points are set on y = y _b (= 2 m) in parallel with the speaker, 32 channels in the region of x <0 are set as control points x _b for local sound reproduction, and the remaining half of the 32 channels are silent. The number [17] is calculated as the control point _xd of the control region.

The tuning coefficient was set so as to maximize the energy difference between the _xb region and the _xd region by preliminary examination. When a linear speaker array was used, the maximum was achieved at α = 0.9999 regardless of the frequency.

For example, in the method of FIG. 1 embodiment, the length of the window function of the number [12] is 2L, and the number [14] is calculated by IDFT of 2N = 128 points. Of the obtained 128 channels of d (x ₀ , ω), the central 64 channels are used as the actual drive signals.

  In order to evaluate the accuracy depending on the position of the local sound reproduction, the sound pressure level PSPL (x) at the position x is defined as the number [19].

Here, p (x, ω) is the sound pressure at the position x and the angular frequency ω.

Further, in order to evaluate the sound pressure level between the local sound reproduction region _xb and the silence control region _xd , a bright region to dark region ratio (BDR) is defined as a number [20]. .

Each region is centered at y = _{y b} as shown in FIG. 5, the width 0.4 m, length was calculated as L / 2 = 1.6m.

  The result of the sound pressure level, which is a simulation result, is shown in FIG. 7, and the result of BDR is shown in FIGS. From these results, it can be seen that the local sound reproduction can be realized with higher accuracy than the EDM particularly in the case of 1000 Hz or more. Further, from the results of FIG. 8 and FIG. 9, the width of local sound reproduction becomes narrower as the frequency becomes higher in EDM, whereas in the embodiment, local sound reproduction of a set width can be realized without depending on the frequency. Can be confirmed.

  As described above, the set local sound reproduction position can be effectively controlled by the spatial filtering according to the embodiment.

  It should be noted that the specific numerical values such as the dimensions mentioned above are merely examples, and can be appropriately changed according to the needs of product specifications and the like.

10 ... local sound reproducing apparatus 12 ... linear loudspeaker array _{14 1, 14 2, ...,} 14 N ... speaker 16 ... driving apparatus (computer)

Claims (8)

A local sound reproducing apparatus using a linear speaker array in which a plurality of speakers are linearly arranged at intervals from each other,
Modeling means for obtaining a sound pressure model by modeling a sound pressure at a reproduction position with a rectangular window of an arbitrary width;
Deriving means for deriving a spatial filter function of each of the plurality of speakers based on the sound pressure model, and generating a drive signal for the plurality of speakers based on a sound source and the spatial filter function, A local sound reproducing device comprising driving means for driving.   2. The modeling means obtains a sound pressure model modeled by a plurality of rectangular windows at different positions, and the derivation means derives the spatial filter function according to a sound pressure model modeled by the plurality of rectangular windows. The local sound reproduction apparatus as described. A local sound reproduction program that is executed by a computer of a local sound reproduction device that uses a linear speaker array in which a plurality of speakers are arranged linearly at intervals, the local sound reproduction program comprising:
Modeling means for obtaining a sound pressure model by modeling a sound pressure at a reproduction position with a rectangular window of an arbitrary width;
Deriving means for deriving a spatial filter function of each of the plurality of speakers based on the sound pressure model, and generating a drive signal for the plurality of speakers based on a sound source and the spatial filter function, A local sound reproduction program that functions as a driving means for driving.   4. The modeling means obtains a sound pressure model modeled by a plurality of rectangular windows at different positions, and the derivation means derives the spatial filter function according to a sound pressure model modeled by the plurality of rectangular windows. The local sound reproduction program described. A local sound reproducing device for reproducing sound locally at an arbitrary position at an arbitrary position with respect to the linear speaker array using a linear speaker array composed of a large number of speakers arranged in a straight line,
Width defining means for defining a width in which sound can be heard in parallel with the speakers of the linear speaker array;
Position defining means for defining a position where sound can be heard at a position parallel to the arrangement of the speakers of the linear speaker array;
A local sound reproducing apparatus comprising: speaker driving means for generating a drive signal by a spatial filter function set by the width defining means and the position defining means and a sound source to drive each speaker of the linear speaker array. A distance defining means for defining a distance from the linear speaker array;
The local sound reproduction device according to claim 5, wherein the spatial filter function is further set by the distance defining means. Local sound executed by a computer of a local sound reproduction device that reproduces sound locally at an arbitrary position and at an arbitrary position with respect to the linear speaker array using a linear speaker array composed of a plurality of speakers arranged in a straight line. A reproduction program, wherein the local sound reproduction program causes the computer to
A width defining means for defining a width in which sound can be heard in parallel with the speakers of the linear speaker array;
A position defining means for defining a position where sound can be heard at a position parallel to the speaker arrangement of the linear speaker array, and a drive signal is generated by a spatial filter function and a sound source set by the width defining means and the position defining means. A local sound reproduction program for causing a speaker driving means for driving each speaker of the linear speaker array. Further, the computer is caused to function as a distance defining means for defining a distance from the linear speaker array,
The local sound reproduction program according to claim 7, wherein the spatial filter function is further set by the distance defining means.