JP5831910B2 - Sound field recording / reproducing apparatus, method, and program - Google Patents

Description

  The present invention relates to a wave field synthesis technique for collecting a sound signal with a microphone installed in a certain sound field and reproducing the sound field with a speaker using the sound signal.

  For example, Non-Patent Document 1 discloses a wave field synthesis technique for collecting a signal with a microphone array installed in a certain sound field and reproducing the sound field with the speaker array using the signal. Technologies are known.

Shoichi Koyama, 3 others, “Spatial-Time Frequency Domain Signal Conversion Method for Sound Field Recording and Reproduction”, Proceedings of the Acoustical Society of Japan, September 2011, p. 635-636

  In the technique described in Non-Patent Document 1, it is necessary to use a planar array in order to reproduce the image including the vertical direction. In addition, when a small number of arrays are used in the vertical direction, the reproduction error becomes large, and there is a possibility that the vertical feeling cannot be reproduced.

  An object of the present invention is to provide a sound field sound collecting / reproducing apparatus, method, and program capable of reproducing a sound field with higher accuracy than before.

In order to solve the above problems, sound field sound collecting and reproducing apparatus according to an aspect of the invention, at least four microphones are arranged in the circumferential surface of the rigid baffle cylindrical radius R _m of the first space The axial direction of the baffle is the z-axis direction, the circumferential direction of the baffle is the φ direction, j is the imaginary unit, ω is the frequency, c is the speed of sound, k = ω / c, and n is the φ direction Where k _{z, l} is the wave number in the z-axis direction, l is its index, H _n ⁽¹⁾ (・) is the nth-order first-class Hankel function, and w _nl is based on n, l Assuming that the weight is fixed, at least four speakers are arranged on the peripheral surface of the virtual cylinder having the radius R _{s in} the second space different from the first space, and are generated based on the signal collected by the microphone. Filtering is applied by applying the filter F ~ _nl (ω) defined by the following equation to the spatio-temporal frequency domain signal P ~ _nl (ω). A conversion filter unit that generates post-processing signals D to _nl (ω);

A spatial frequency inverse transform unit that transforms the filtered signal D ~ _nl (ω) into a frequency domain signal by inverse Fourier transform of space, and a frequency domain signal that is transformed into a time domain signal by inverse Fourier transform, and the transformed time A frequency inverse transform unit that outputs the region signal to the speaker.

Other sound field sound collecting and reproducing apparatus according to an aspect of the invention, the at least four microphones are arranged in the circumferential surface of the rigid baffle cylindrical radius R _m of the first space, the baffle cylinder Is the z axis direction, the circumferential direction of the baffle cylinder is the φ direction, j is the imaginary unit, ω is the frequency, c is the speed of sound, k = ω / c, and n is the order in the φ direction. , K _{z, l} is the wave number in the z-axis direction, l is its index, H _n ⁽¹⁾ (・) is the nth-order first-class Hankel function, and w _nl is a weight determined based on n, l Assuming that at least four speakers are arranged on the peripheral surface of the virtual cylinder having the radius R _{s in} the second space different from the first space, the signals collected by the microphone array are converted into frequency domain signals by Fourier transform. Frequency domain signal is converted into spatio-temporal frequency by Fourier transform of space and frequency converter to convert A spatial frequency transformation unit for converting the frequency signal P ~ _nl (ω), with respect to spatio-temporal frequency domain signal P ~ _nl (ω) and apply a filter F ~ _nl (ω) defined by the following equation filtering A conversion filter unit that generates a post-signal _D˜nl (ω).

  Even if the number of microphones and speakers constituting the microphone array and the speaker array is small, the microphone array and the speaker array are cylindrical, with a dense element arrangement in the left-right direction, and a coarse element arrangement in the vertical direction. A feeling of up and down can be reproduced. Therefore, the sound field can be reproduced with higher accuracy than before.

The functional block diagram which shows the example of a sound field sound collection reproducing | regenerating apparatus. The figure for demonstrating the example of arrangement | positioning of a microphone and a speaker. The figure for demonstrating the example of arrangement | positioning of a microphone and a speaker. The flowchart which shows the example of the sound field sound collection reproduction | regeneration method.

  Embodiments of the present invention will be described below with reference to the drawings.

As shown in FIG. 1, the sound field sound collecting and reproducing apparatus and method are composed of N _z × N _φ microphones arranged on the peripheral surface of a cylindrical rigid baffle having a radius R _{m in} the first space. And a speaker array composed of N _z × N _φ speakers arranged on the circumferential surface of a virtual cylinder having a radius R _{s in} a second space different from the first space, The sound field of the first space formed by the sound generated by the sound source So in the second space is reproduced in the second space. In FIG. 1, the sound source So reproduced in the second space is expressed as a sound source So ′. The axial direction of the cylinder is the z-axis direction, and the circumferential direction is φ. The first space and the second space are different from each other. In this embodiment, the number of microphones arranged in the first space is the same as the number of speakers arranged in the second space.

As shown in FIG. 3, in each of the N _z number of circles of the peripheral surface of the baffle is cylindrical, N _phi number of microphones are arranged at regular intervals. N _z and N _φ are predetermined integers of 2 or more. That is, by arranging two microphones in each of the two circles on the circumferential surface of the cylinder, at least four microphones are disposed on the circumferential surface of the cylinder.

The N _z circles on the circumferential surface of the cylinder are positioned at z _c intervals, for example, with z _c as a predetermined distance. Further, the N _φ microphones arranged in the same circle are positioned at intervals of φ _c degrees, with φ _c being a predetermined angle.

The microphones do not need to be arranged at exactly equal intervals as long as they are arranged at approximately equal intervals. That is, z _c and φ _c that are intervals between adjacent microphones do not have to be exactly the same value, and may be almost the same value.

The speaker is also arranged in the same manner as the microphone. That is, as shown in FIG. 3, in each of the N _z number of circular circumferential surface of the cylindrical, N _phi number of speakers are arranged at regular intervals. N _z and N _φ are predetermined integers of 2 or more. At least four speakers are arranged on the circumferential surface of the cylinder.

The N _z circles on the circumferential surface of the cylinder are positioned at z _c intervals, for example, with z _c as a predetermined distance. Further, N _phi number of speakers arranged on the same circle, the phi _c as a predetermined angle are located at intervals of phi _c of.

As long as the speakers are arranged at almost equal intervals, the speakers need not be arranged at exactly equal intervals. That is, z _c and φ _c that are the distances between adjacent speakers need not be exactly the same value, and may be substantially the same value.

The radius R _m of the cylinder in which the microphone is arranged is, for example, 25 to 50 cm. Further, the radius R _s of the cylinder in which the speaker is arranged is, for example, about 1.5 m. R _m ≦ R _s or R _s ≦ R _m may be satisfied, but the accuracy of sound field reproduction is improved when R _m ≦ R _s .

  The microphone is arranged toward the outside of the circumferential surface of the cylinder. Further, the speaker is arranged toward the inside of the circumferential surface of the cylinder.

  The speaker may be disposed in the air in the second space in an acoustically transparent state, or may be disposed in the second space in a state that is not acoustically transparent. The acoustically transparent state is a state in which the same transmission characteristic as that of the second space where no speaker is arranged is maintained. For example, the speaker is placed in the air in the second space by being hung with a thread or fixed with a thin rod. The speaker may be arranged on a cylindrical rigid baffle as in the microphone.

The position of the microphone Mi-j in the first space is (R _m , φ _{m, i} , z _{m, j} ) [i = 1,2, ..., N _φ , j = 1,2, ..., N _z ] Express. The position of the speaker Si-j in the second space is (R _s , φ _{s, i} , z _{s, j} ) [i = 1,2, ..., N _φ , j = 1,2, ..., N _z ] Express.

  As shown in FIG. 1, the sound field sound collecting and reproducing apparatus includes, for example, a frequency conversion unit 1, a spatial frequency conversion unit 2, a conversion filter unit 3, a spatial frequency reverse conversion unit 4, a frequency reverse conversion unit 5, and a window function unit 6. The processing of each step illustrated in FIG. 4 is performed.

First microphone M1-1 arranged in the _{space, M2-1, ..., MN φ -N} z generates a picked-up by a time-domain signal a sound emitted by the sound source S of the first space . The generated signal is sent to the frequency converter 1. A signal at time t in the time domain picked up by the microphone Mi-j of (R _m , φ _{m, i} , z _{m, j} ) is expressed as p _ij (t).

The frequency converter 1 converts the signal p _ij (t) collected by the microphones M1-1, M2-1,..., MN _φ −N _z into a frequency domain signal P _ij (ω) by Fourier transform (step). S1). The generated frequency domain signal P _ij (ω) is sent to the spatial frequency converter 2. ω is a frequency. For example, the frequency domain signal P _ij (ω) is generated by short-time discrete Fourier transform. Of course, the frequency domain signal P _ij (ω) may be generated by other existing methods. Further, the frequency domain signal P _ij (ω) may be generated using a method such as overlap add. When the input signal is long or when the signal is continuously input as in real time processing, the processing is performed for each frame such as every 10 ms. For example, the frequency domain signal P _ij (ω) is defined as follows. J in the argument of the function exp is an imaginary unit.

Spatial frequency transformation unit 2, a frequency-domain signal P _ij by the Fourier transform of the spatial (omega) into a space-time frequency domain signal P ~ _nl (ω) (Step S2). The spatio-temporal frequency domain signal P ~ _nl (ω) is calculated for each ω. The converted spatio-temporal frequency domain signal P _{n n} (ω) is sent to the conversion filter unit 3. Specifically, the spatial frequency converter 2 calculates _P˜nl (ω) defined by the equation (1).

k _{z, l} is the wave number in the z-axis direction, l is its index, and n is the order in the φ direction. The wave number is a so-called spatial frequency or angular spectrum. Expression (1) is an example of the transformation into the spatio-temporal frequency domain, and the spatial Fourier transform may be performed by other methods.

The transform filter unit 3 applies the filter F ~ _nl (ω) defined by the equation (2) to the spatio-temporal frequency domain signal P ~ _nl (ω) to obtain the filtered signal D ~ _nl (ω). Generate (step S3). The filtered signal _D˜nl (ω) is transmitted to the spatial frequency inverse transform unit 4.

In Equation (2), c is the speed of sound, and k = ω / c is the wave number. Further, w _nl is a weight determined as follows based on n and l, for example. In the following formulas, n _c is the cut-off value of a predetermined value n. k _c is a predetermined value and is a cutoff value of k _z . α _n and α _z are predetermined values, for example, 0.05. Of course, other weight functions may be used as w _nl .

H _n ⁽¹⁾ (•) is an n-th order Hankel function of the first kind. The nth-order first kind Hankel function H _n ⁽¹⁾ (x) is defined as follows using the first kind Bessel function J _n (x) and the second kind Bessel function Y _n (x).

The spatial frequency inverse transform unit 4 transforms the filtered signal _D˜nl (ω) into a frequency domain signal D _ij (ω) by inverse spatial Fourier transform (step S4). The converted frequency domain signal D _ij (ω) is sent to the frequency inverse transform unit 5. Specifically, the spatial frequency inverse transform unit 4 calculates the frequency domain signal D _ij (ω) defined by the equation (3).

The frequency inverse transform unit 5 transforms the frequency domain signal D _ij (ω) into a time domain signal P ^d _ij (t) by inverse Fourier transform (step S5). The time domain signal P ^d _ij (t) obtained for each frame by the inverse Fourier transform is appropriately shifted to obtain a linear sum to be a continuous time domain signal. For the inverse Fourier transform, an existing method such as a short-time discrete inverse Fourier transform may be used. The time domain signal P ^d _ij (t) is sent to the window function unit 6.

The window function unit 6 generates a post-window function time domain signal d _ij (t) by multiplying the time domain signal P ^d _ij (t) by the window function (step S6). After the window function time domain signal d _ij (t) is the speaker Si-j, S2-1, ..., it is sent to the SN φ -N _z.

For example, a so-called tukey window function w _ij defined by the following equation is used as the window function. N _tpr is the number of points to which the taper is applied, and is an integer between 1 and N _φ , N _z . Of course, other window functions may be used.

The speaker arrays S1-1, S2-1,..., SN _φ -N _z reproduce sound based on the time domain signal _dij (t) after the window function. Specifically, with i = 1,..., _Nφ , j = 1,..., _Nz , the speaker Si-j reproduces sound based on the time domain signal _dij (t) after the window function. Thereby, the sound field of the first space can be reproduced in the second space.

  In this way, by making the microphone array and the speaker array cylindrical, even when the number of microphones and speakers constituting the microphone array and the speaker array is small, a dense element arrangement in the left-right direction is used. Can reproduce the vertical feeling as a rough element arrangement. Therefore, the sound field can be reproduced with higher accuracy than in the past.

Hereinafter, the reason why the filter _F˜nl (ω) is expressed as in Expression (2) will be described.

The position vector of the reproduction region is r '= (r, φ, z), the cylindrical surface where the secondary sound source is placed is S, and the position vector on S is r _s ' = (R _s , φ _s , z _s ). At frequency ω, the sound pressure at position r ′ in the reproduction region is P (r ′, ω), and the secondary sound source signal at position r _s ′ is D (r _s ′, ω), from r _s ′ to r ′. Let G (r _s '-r', ω) be the transfer function of. At this time, the sound field synthesized in the reproduction region by the secondary sound source can be written as follows.

  This expression is nothing but a convolution operation on the variables φ and z of the functions D (•) and G (•). Therefore, from the convolution theorem, Equation (4) can be expressed as follows in the helical wave spectrum region.

Here, Expression (5) is developed for the case where the secondary sound source can be approximated as a monopole characteristic. At this time,

であるから、

Because

となる。ここで、

It becomes. here,

である。Hankel関数の加法定理より、

It is. From the addition theorem of the Hankel function,

であるから、

Because

となる。式（６）を式（５）に代入すると、

It becomes. Substituting equation (6) into equation (5),

となる。

It becomes.

The ideal sound field can be given from the sound pressure on a cylindrical rigid baffle using an omnidirectional microphone array or the like. If the sound pressure distribution on the cylindrical surface on the sound collection side is P (r _m ', ω), P (r', ω) is the incident sound field P _i (r ', ω) and the scattered sound field P _s (r ', ω).

となる。P_i(r’,ω)とP_s(r’,ω)をヘリカル波スペクトル領域で書くと、 From the boundary condition that the sound pressure gradient is 0 on the rigid baffle,

It becomes. If P _i (r ', ω) and P _s (r', ω) are written in the helical wave spectrum region,

となり、式（８）の境界条件より、以下の関係式が成り立つ。

From the boundary condition of equation (8), the following relational expression is established.

Therefore,

It becomes. Since the sound field to be reproduced is the incident sound field P _i (•), the ideal sound field can be written in the helical wave spectrum region as follows.

What is desired to be obtained here is a conversion formula from the cylindrical sound pressure distribution P (r _m ′, ω) to the drive signal D (r _s ′, ω) of the secondary sound source. When obtaining the conversion equation assuming that the secondary sound source has a monopole characteristic, equations (7) and (8) may be solved simultaneously.

When written in filter form, it looks like this:

である。 However,

It is.

[Modifications, etc.]
The number of microphones arranged in the first space and the number of speakers arranged in the second space may be different. When the number of microphones constituting the microphone array is larger than the number of speakers constituting the speaker array, the post-window function time domain signal _dij (t) may be thinned out. On the other hand, if the number of microphones constituting the microphone array is smaller than the number of speakers constituting the speaker array, interpolation may be performed by taking an average of the time domain signal _dij (t) after the window function. Good.

  Each unit constituting the sound field sound collecting / reproducing device may be provided in either the sound collecting device arranged in the first space or the reproducing device arranged in the second space. In other words, the processes of the frequency conversion unit 1, the spatial frequency conversion unit 2, the conversion filter unit 3, the spatial frequency reverse conversion unit 4, the frequency reverse conversion unit 5, and the window function unit 6 are arranged in the first space. It may be executed by a sound collecting device or may be executed by a reproducing device arranged in the second space. The signal generated by the sound collection device is transmitted to the reproduction device.

  The positions of the first space and the second space are not limited to those shown in FIG. The first space and the second space may be adjacent to each other or separated from each other. Also, the orientation of the first space and the second space may be any.

  The window function processing by the window function unit 6 may be performed at any stage or in multiple stages. That is, the window function unit 6 is provided between the microphone array and the frequency conversion unit 1, between the frequency conversion unit 1 and the spatial frequency conversion unit 2, between the spatial frequency conversion unit 2 and the conversion filter unit 3, and between the conversion filter unit. 3 and the spatial frequency inverse transform unit 4, and at least one between the spatial frequency inverse transform unit 4 and the frequency inverse transform unit 5. When the window function processing is performed on the signal input to each section, each section of the sound field sound collecting / reproducing apparatus performs the window function processing in the same manner as described above instead of the input signal. The processed signal is processed.

Further, the window function unit 6 may not be provided. In this case, with i = 1,..., _Nφ , j = 1,..., _Nz , the speaker Si-j reproduces the sound based on the time domain signal P ^d _ij (t).

  As long as the sound field sound collecting / reproducing apparatus includes the conversion filter unit 3, the sound field collecting / reproducing device may not include other units. For example, the sound field sound collecting / reproducing apparatus may include a transform filter unit 3, a spatial frequency inverse transform unit 4, and a frequency inverse transform unit 5. Further, the sound field sound collecting / reproducing apparatus may include a frequency conversion unit 1, a spatial frequency conversion unit 2, and a conversion filter unit 3.

  You may perform the process of the frequency converter 1 and the process of the spatial frequency converter 2 simultaneously. Similarly, the process of the spatial frequency inverse transform unit 4 and the process of the frequency inverse transform unit 5 may be performed simultaneously. Further, the spatial frequency conversion unit 2 and the spatial frequency inverse conversion unit 4 may be interchanged.

  The sound field sound collecting / reproducing apparatus can be realized by a computer. In this case, the processing content of each part of this apparatus is described by a program. Then, by executing this program on a computer, each unit in this apparatus is realized on the computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. In this embodiment, these apparatuses are configured by executing a predetermined program on a computer. However, at least a part of these processing contents may be realized by hardware.

  The present invention is not limited to the above-described embodiment, and can be modified as appropriate without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 Frequency conversion part 2 Spatial frequency conversion part 3 Conversion filter part 4 Spatial frequency reverse conversion part 5 Frequency reverse conversion part 6 Window function part

Claims (6)

At least four microphones and is disposed on the peripheral surface of the rigid baffle cylindrical radius R _m of the first space, the axial direction of the baffle and the z-axis direction, the direction of the circumferential direction of the baffle φ Where j is the imaginary unit, ω is the frequency, c is the speed of sound, k = ω / c, n is the order in the φ direction, k _{z, l} is the wave number in the z-axis direction, and l is its index. , H _n ⁽¹⁾ (·) is an n-th order first-class Hankel function, w _nl is a weight determined based on n, l, and at least four speakers are different from the first space. As arranged on the circumference of the virtual cylinder of radius R _s of
Filtered signal D by applying filter F ~ _nl (ω) defined by the following equation to spatio-temporal frequency domain signal P ~ _nl (ω) generated based on the signal collected by the microphone ~ Conversion filter part that generates _nl (ω);

A spatial frequency inverse transform unit for transforming the filtered signal D to _nl (ω) into a frequency domain signal by inverse Fourier transform of the space;
A frequency inverse transform unit that transforms the frequency domain signal into a time domain signal by inverse Fourier transform and outputs the transformed time domain signal to the speaker;
Sound field collection and playback device including And at least four microphones are arranged in the circumferential surface of the rigid baffle cylindrical radius R _m of the first space, the axial direction of the baffle and the z-axis direction, a circumferential direction of the cylinder of the baffle φ direction, j as imaginary unit, ω as frequency, c as sound velocity, k = ω / c, n as order in φ direction, k _{z, l} as wave number in z axis direction, l as its H _n ⁽¹⁾ (·) is an n-th order first-class Hankel function, w _nl is a weight determined based on n, l, and at least four speakers are different from the first space. As arranged on the circumference of the virtual cylinder with radius R _s of the space of
A frequency converter that converts signals collected by the microphone array into frequency domain signals by Fourier transform;
A spatial frequency converter that converts the frequency domain signal into a spatio-temporal frequency domain signal P to _nl (ω) by Fourier transform of space;
A transform filter unit that generates a filtered signal D to _nl (ω) by applying a filter F to _nl (ω) defined by the following equation to the spatio-temporal frequency domain signal P to _nl (ω):

Sound field collection and playback device including In the sound field sound collecting and reproducing device according to claim 1 or 2,
At least one of the spatio-temporal frequency domain signals P to _nm (ω), P to _n (ω), the frequency domain signal, and the time domain signal converted by the frequency inverse transform unit is a window by a predetermined window function. A signal that has undergone function processing.
Sound field recording and playback device. At least four microphones and is disposed on the peripheral surface of the rigid baffle cylindrical radius R _m of the first space, the axial direction of the baffle and the z-axis direction, the direction of the circumferential direction of the baffle φ Where j is the imaginary unit, ω is the frequency, c is the speed of sound, k = ω / c, n is the order in the φ direction, k _{z, l} is the wave number in the z-axis direction, and l is its index. , H _n ⁽¹⁾ (·) is an n-th order first-class Hankel function, w _nl is a weight determined based on n, l, and at least four speakers are different from the first space. As arranged on the circumference of the virtual cylinder of radius R _s of
The conversion filter unit applies the filters F to _nl (ω) defined by the following equations to the spatio-temporal frequency domain signals P to _nl (ω) generated based on the signals collected by the microphone. A transform filter step for generating a filtered signal D ~ _nl (ω);

The spatial frequency inverse transform unit performs the above-mentioned filtered signal D ~ _nl (
a spatial frequency inverse transform step for transforming (ω) into a frequency domain signal;
A frequency inverse transform unit that transforms the frequency domain signal into a time domain signal by inverse Fourier transform, and outputs the transformed time domain signal to the speaker;
Sound field collection and playback method including At least four microphones and is disposed on the peripheral surface of the rigid baffle cylindrical radius R _m of the first space, the axial direction of the baffle and the z-axis direction, the direction of the circumferential direction of the baffle φ Where j is the imaginary unit, ω is the frequency, c is the speed of sound, k = ω / c, n is the order in the φ direction, k _{z, l} is the wave number in the z-axis direction, and l is its index. , H _n ⁽¹⁾ (·) is an n-th order first-class Hankel function, w _nl is a weight determined based on n, l, and at least four speakers are different from the first space. As arranged on the circumference of the virtual cylinder of radius R _s of
A frequency converting step in which the frequency converting unit converts the signal collected by the microphone array into a frequency domain signal by Fourier transform;
A spatial frequency transforming step, wherein the spatial frequency transforming unit transforms the frequency domain signal into a spatio-temporal frequency domain signal P to _nl (ω) by Fourier transform of the space;
The transform filter unit generates a filtered signal D ~ _nl (ω) by applying a filter F ~ _nl (ω) defined by the following equation to the spatio-temporal frequency domain signal P ~ _nl (ω) A transform filter step;

Sound field collection and playback method including   A sound field recording / reproducing program for causing a computer to function as each unit of the sound field sound collecting / reproducing apparatus according to any one of claims 1 to 3.

Images