JP6345633B2 - Sound field reproducing apparatus and method - Google Patents

Description

  The present invention relates to a sound field reproduction technique, and more particularly to a spatial multi-zone sound field reproduction technique.

  Spatial multi-zone sound field reproduction is a technique for reproducing a plurality of independent sound fields in a separated space by driving a speaker array (see, for example, Non-Patent Document 1). In the prior art, in order to reproduce a sound field in a plurality of areas, a circular speaker array in which omnidirectional speakers are arranged in a circle at equal intervals so as to surround the reproduction area is used.

Y. J. Wu, et al. “Spatial Multizone Soundfield Reproduction: Theory and Design,” Speech and Audio Processing, IEEE Transactions on 19.6 (2011): 1711-1720.

  In order to accurately reproduce the sound field up to the wave number k within the circumference of the radius R by the conventional circular speaker array, a huge number of speakers of 2kR + 1 or more are required. An object of the present invention is to realize spatial multi-zone sound field reproduction with a smaller number of sound sources than in the prior art.

  A predetermined number of higher-order sound sources that reproduce an acoustic signal with a reproduction pattern, which is composed of any one of a plurality of independent radiation modes having different directivity patterns or a combination of at least a part of the plurality of radiation modes, are used. In the reproduction pattern, each weight of the radiation mode is controlled in order to generate a desired sound field in a plurality of spatial regions using the echo of the acoustic signal.

  In the present invention, spatial multi-zone sound field reproduction can be realized with a smaller number of sound sources than in the prior art by using echoes while controlling the respective weights of the radiation modes.

FIG. 1 is a conceptual diagram for explaining the configuration of the sound field reproducing apparatus of the embodiment. FIG. 2 is a block diagram illustrating the control unit of the embodiment. FIG. 3 is a geometric diagram relating to Theorem 1. FIG. 4 is a conceptual diagram for explaining the configuration of the coefficient setting device of the embodiment. FIG. 5 is a block diagram illustrating the setting unit of the embodiment.

Embodiments of the present invention will be described below.
〔Overview〕
A predetermined number of higher order sources are arranged in a reverberant environment (in a reverberant environment). Each higher-order sound source reproduces an acoustic signal with a “reproduction pattern” composed of any one of a plurality of independent radiation modes having different directivity patterns or a combination (linear sum) of at least a part of a plurality of radiation modes. It is a sound source. An example of a higher order sound source is a higher order loudspeaker (for example, reference 1 “M. Poletti, T. Betlehem,“ Design of a prototype variable directivity loudspeaker for improved surround sound reproduction in rooms, ”AES 52nd. International Conference, Guildford, UK, 2013, September 2-4 "). Here, by using the “reproduction pattern” in which the weight of each radiation mode is controlled in order to generate the desired sound field in a plurality of spatial regions using the echo of the acoustic signal, a smaller number than in the past It is possible to realize spatial multi-zone sound field playback with any sound source. That is, as shown in Equations (20) and (21) of Non-Patent Document 1, in Non-Patent Document 1, spatial multi-zone sound field reproduction is realized by controlling the filter coefficient of each speaker. On the other hand, in this embodiment, each of the radiation modes of the higher-order sound source is controlled. By controlling the radiation mode, it is possible to improve the degree of freedom of mirror image control due to reverberation and to realize spatial multi-zone sound field reproduction with a small number of sound sources.

The “reproduction pattern” is, for example, the filter weight for the radiation mode m (ν) at the wave number k of the higher-order sound source hs (u).

It is what. Here, hs (1),..., Hs (L) represent L high-order sound sources, L is a positive integer (for example, L is an integer of 2 or more), and m (1),. _V +1) represents 2N _V +1 radiation modes, N _V is a positive integer (N _V is an integer of 1 or more), u = 1,..., L, ν = 1,..., 2N _V +1 It is. H _νu (r, θ; k) is an acoustic transfer function at the wave number k of the radiation mode m (ν) from the higher-order sound source hs (u) to the global polar coordinate (r, θ), and r is the radius And θ is a declination. M ₀ (k) is a positive integer, m = −M ₀ (k),..., M ₀ (k), and α _mνu (k) is m of the acoustic transfer function H _νu (r, θ; k). A (k) is a (2M ₀ (k) +1) × L ′ matrix with α _mνu (k) as elements, and L ′ = L (2N _V +1).

It is.

, And the represents a transpose of [·] ^T is _{^{[·], β d m (}} k) is the wave number k in the polar region including a plurality of spatial regions where desired sound field is generated (r, theta) Of the frequency domain signal S ^d (r, θ; k) for m, and g (k) is a solution or an approximate solution of A (k) g (k) = β ^d (k).

α _mνu (k) can be obtained from the acoustic transfer function H _νu (r, θ; k), and H _νu (r, θ; k) can be calculated from the output from the higher-order sound source and the sound collection result of the microphone. . However, depending on the environment, the error of H _νu (r, θ; k) may increase. in this case,

_Filter H _νu (a, θ _q ; k) by

You may get Thereby, the influence of the estimation error of the acoustic transfer function can be reduced. However, H _νu (a, θ _q ; k) is based on the observed acoustic signal at the microphone MIC (q) (where q = 1,..., Q) arranged at the global polar coordinates (a, θ _q ). Is an acoustic transfer function at a wave number k of the radiation mode m (ν) from the higher-order sound source hs (u) to the polar coordinates (a, θ _q ). a is a moving radius, θ _q is a declination,

, I is an imaginary unit, H ′ _m (•) is a derivative of the first kind Hankel function of degree m, (•) ^* represents the complex conjugate of (•), and λ _eq is equivalent It is a regularization parameter, and | (•) | represents the absolute value of (•).

[First Embodiment]
A first embodiment will be described with reference to the drawings.
<Configuration>
As illustrated in FIG. 1, the sound field reproducing apparatus 1 of this embodiment includes a control unit 11 and L high-order sound sources (high-order speakers) arranged in a reverberation room (in a reverberation environment) surrounded by a wall surface 13. ) 12-u (where u = 1,..., L). As illustrated in FIG. 2, the control unit 11 includes a storage unit 111, a weight acquisition unit 112, a filtering unit 113-u, and a drive signal generation unit 114-u. The control unit 11 is, for example, a general-purpose or dedicated computer that includes a processor (hardware processor) such as a CPU (central processing unit) and a memory such as random-access memory (RAM) and read-only memory (ROM). It is configured by executing a predetermined program. The computer may include a single processor and memory, or may include a plurality of processors and memory. This program may be installed in a computer, or may be recorded in a ROM or the like in advance. In addition, some or all of the processing units are configured using an electronic circuit that realizes a processing function without using a program, instead of an electronic circuit (circuitry) that realizes a functional configuration by reading a program like a CPU. May be. In addition, an electronic circuit constituting one device may include a plurality of CPUs. 1 and 2 are examples of L = 5, this does not limit the present invention.

≪Geometric arrangement≫
Two-dimensional global (global) spatial regions, and S two-dimensional local (circular) desired spatial regions (circular regions) 14-1,. Assume a corresponding desired sound field. However, S is an integer of 2 or more. The radius and origin of the s-th (where s = 1,..., S) space region 14-s are expressed by R _z ^(s) and O _s , respectively. Here, O _s is located at polar coordinates (r ^(s0) , θ ^(s0) ) with respect to the global origin O. r ^(s0) and θ ^(s0) represent a radius vector and a declination angle, respectively. A local polar coordinate of an arbitrary observation point in the s-th spatial region 14-s is expressed as (R ^(s) , Ω ^(s) ). (R ^(s) , Ω ^(s) ) are polar coordinates centered on the local origin O _s , and R ^(s) and Ω ^(s) represent the radial and declination, respectively. Local polar coordinates (R ^(s) , Ω ^(s) ) are located at polar coordinates (r, θ) with respect to the global origin O. r and θ represent a radius vector and a declination angle, respectively. All the spatial regions 14-1,..., 14-S are present in a circular region having a radius R _P ≧ r with the global origin O as the center. R _P is all the spatial regions 14-1, ..., the radius of the circular area including the 14-S inwards. L pieces of high-order tone 12-u are arranged on the circumference of a radius R _S ≧ R _P from global origin O. The higher order sound sources 12-1, ..., 12-L may be arranged at equal intervals or may not be arranged at equal intervals. L is a positive integer. For example, L is an integer of 2 or more.

≪ Desired sound field≫
A desired sound field in the two-dimensional space region 14-s is expressed as a two-dimensional Green function (point source 3 in the two-dimensional plane, x _{source (s)} = (r _s cos θ _s , r _s sin θ _s ) ^T It is defined as a sound field caused by a linear sound source in a dimensional space. This sound source is called a virtual sound source. However, an arbitrary polar coordinate with respect to the local origin O _s is expressed as (r _s , θ _s ). r _s and θ _s represent a radius vector and a declination angle, respectively. (•) ^T represents transposition of (•). That is, the purpose of this embodiment is to use higher-order sound sources 12-1,..., 12-L, and the sound field caused by the virtual sound source located at x _{source (s)} To reproduce. A desired sound field in the spatial domain 14-s is represented by the following time frequency domain signal S ^{d (s)} (x _{d (s)} ; k).

However, k is a wave number, i is an imaginary unit, and || (•) || represents a norm of (•). H _n (·) is a first type Hankel function of order n. _{xd (s)} is an arbitrary observation point in the spatial region 14-s. d (s) is an index representing a desired sound field in the spatial region 14-s. When x _{d (s)} is expressed in polar coordinates with respect to the local origin O _s , it becomes (R ^(s) , Ω ^(s) ), and when expressed in polar coordinates with respect to the global origin O, it becomes (r, θ). By multiplying S ^{d (s)} (x _{d (s)} ; k) by a coefficient S (k), a sound field by an arbitrary acoustic signal can be expressed.

≪Cylinder harmonic function expansion of desired sound field≫
x _{d (s)} is represented by polar coordinates (R ^(s) , Ω ^(s) ) with respect to the local origin O _s . In this case, the time-frequency domain signal S ^{d (s)} (R ^(s) , Ω ^(s) ; k) representing the desired sound field in the spatial domain 14-s is expressed as follows by cylindrical harmonic function expansion. The

However, J _m (•) is a Bessel function of degree m, and α _m ^{d (s)} (k) uniquely represents S ^{d (s)} (R ^(s) , Ω ^(s) ; k). Expansion coefficient (sound field coefficient) for m, and e is the Napier number (the base of natural logarithm). Expression (2) is a Fourier series expansion, which can represent an arbitrary two-dimensional sound field derived from an arbitrary number of cylindrical waves and plane waves. In addition, “ ^{d (s)} ” of “α _m ^{d (s)} (k)” should be written directly above “m”, but “α _m ^{d (s)} (k” is restricted due to the limitation of the description. ) ”. The same notation may be used for other symbols.

Equation (2) has an infinite number of orthogonal modes. However, due to the nature of the Bessel function and the fact that the sound field is limited in the spatial region where all sound sources are outside, this series expansion can be discontinued by a finite number of orthogonal modes. In this case, S ^{d (s)} (R ^(s) , Ω ^(s) ; k) can be approximated as follows.

However, M _s (k) is a positive integer. In this case, S ^{d (s)} (R ^(s) , Ω ^(s) ; k) is expressed by at least 2M _s (k) +1 modes. When M _s (k) = ceil (keR _z ^(s) / 2), the truncation error is 16.1% or less. However, ceil (•) is a ceiling function of (•).

≪Equivalent global sound field≫
An equivalent global sound field composed of the sound fields of the S spatial regions 14-1,..., S (multi-zone) described above is defined. That is, the sound field reproduction problem of the plurality of spatial regions 14-1,..., S is reduced to reproduction of a global desired sound field over the entire area. A time-frequency domain signal S ^d (r, θ; k) representing a desired global sound field is approximated by cylindrical harmonic function expansion as follows.

Here, β _m ^d (k) is an expansion coefficient (sound field coefficient) for m for uniquely expressing S ^d (r, θ; k). In this case, S ^d (r, θ; k) is expressed by at least 2M ₀ (k) +1 modes. For example, M ₀ (k) = ceil (keR _P / 2). If all multizone falls within a circular region of radius R _P, in order to reproduce the sound field in all multi-zone mode limitation by M 0 _(k) is
M ₀ (k) ≧ (M ₁ (k) + M ₂ (k) +... + M _S (k)) (5)
It is necessary to satisfy.

《Spatial harmonic coefficient conversion》
Consider a sound field in an area where no sound source exists. O _1, O ₂ as shown in FIG. 3, is the origin of the two coordinate systems, they have the axis of the same direction, which moves the O ₁ a known conversion is assumed to be O ₂ . When the polar coordinates (r ⁽¹⁾ , θ ⁽¹⁾ ) in the coordinate system with O ₁ as the origin are represented by the polar coordinates in the coordinate system with O ₂ as the origin, (r ⁽²⁾ , θ ⁽²⁾ ) and Become. Here, the polar coordinates of O ₂ with respect to O ₁ (r ⁽¹²⁾ , θ ⁽¹²⁾ ) are represented. Also, {α _m ⁽¹⁾ (k)} and {α _m ⁽²⁾ (k)} denote the sound field in the region where no sound source exists in the two coordinate systems with the origins as O ₁ and O ₂ , respectively. It is assumed that this is a set of expansion coefficients for m for uniquely expressing. In this case, the following theorems 1 and 2 hold.

[Theorem 1]
α _m ⁽¹⁾ (k) and α _m ⁽²⁾ (k) are related by:

However,

Where “*” represents discrete convolution at mode order m. T _m ⁽²¹⁾ represents a conversion operator from the origin O ₂ to O ₁ , and T _m ⁽¹²⁾ represents a conversion operator from the origin O ₁ to O ₂ .

[Theorem 2]
The polar coordinates of O ₁ with respect to O ₂ are expressed as (r ⁽²¹⁾ , θ ⁽²¹⁾ ),

And applying the theorem of Spatial Harmonic Coefficient Translation, the following relation can be derived.

Here, T _m ^(0s) is a conversion operator from a global coordinate system having O as the origin to a local coordinate system having O _s in the sth spatial region 14- _s as the origin.

《Search for global sound field coefficient》
Coefficient conversion from α _m ^{d (s)} (k) to β _m ^d (k) is performed. The convolution of equation (9) is written as a linear sum as follows.

When this equation (10) is written down for S = 1,..., S, these simultaneous equations are constructed and expressed in matrix form as follows.
α ^d (k) = T (k) β ^d (k) (11)
However, the following is satisfied.

Equation (11) can be modified as follows.
β ^d (k) = T (k) ⁺ α ^d (k) (14)
Where T (k) ⁺ = [T (k) ^H T (k)] ⁻¹ T (k) ^H is a Moore-Peron's pseudo-inverse matrix of T (k), and (·) ^H is (·) Represents the complex conjugate transpose of. Perform coefficient conversion from α _m ^{d (s)} (k) to β _m ^d (k) by obtaining a solution of β ^d (k) or an approximate solution of equation (14) using a least square method or the like. Can do.

《Multi-zone sound field reproduction in reverberation room with high-order sound source》
Sound field reproduction in a spatial multi-zone is performed by L high-order sound sources 12-1, ..., 12-L arranged in a reverberation room (under reverberation environment). Each higher-order sound source 12-s can generate a plurality of directivity patterns independent of each other. Considering a two-dimensional plane, a higher-order sound source 12-s of order N _V can generate 2N _V +1 independent radiation modes. They _{^{{e inθ: n = -N V}} , ..., N V} -phase mode with a polar response of the, {1, cos (nθ) ■, sin (nθ) ■: n = 1, ..., N V} It is represented by an amplitude mode having a polar response of or a linear sum of these polar responses. The reproduced sound field in the input signal S (k) = 1 generated by these L high-order sound sources 12-1,..., 12-L is the following time frequency domain signal S (r, θ; k). Can be written.

Here, G _νu (k) represents the filter weight applied to the radiation mode m (ν) of the wave number k of the higher-order sound source 12-s. However, ν = 1,..., 2N _V +1 is an index corresponding to each radiation mode. H _νu (r, θ; k) is an acoustic transfer function of wave number k in the radiation mode m (ν) from the higher-order sound source 12-s to the polar coordinates (r, θ).

《Mode matching》
In this embodiment, the desired sound field is decomposed into a cylindrical harmonic function, and the active modes (m = −M ₀ (k ),..., Mode matching for controlling the mode in the range of M ₀ (k). Specifically, G _ν u (k) (where u = 1,..., L so as to match the desired sound field mode within a range of m = −M ₀ (k),..., M ₀ (k). ). The advantage of the mode matching approach is that accurate sound field reproduction is guaranteed in the control region up to the spatial Nyquist frequency of the higher-order sound sources 12-1, ..., 12-L. In order to accurately reproduce the active mode of the sound field, the number of degrees of freedom, that is, the number of directivity patterns provided by the higher-order sound sources 12-1,..., 12-L L ′ = L (2N _V +1 ) Must be equal to or greater than the number of active modes (2M ₀ (k) +1). That is, it is desirable that L (2N _V +1) ≧ 2M ₀ (k) +1.

The acoustic transfer function of wave number k in the radiation mode m (ν) from each higher-order sound source 12-s to polar coordinates (r, θ) can be expressed as follows.

Here, α _mνu (k) is an expansion coefficient for m of the acoustic transfer function H _νu (r, θ; k) in the radiation mode m (ν) of the higher-order sound source 12-s.

From the equations (4), (15) and (16), the following relationship can be derived.

However, m = −M ₀ (k),..., M ₀ (k). This set of equations can be expressed as a matrix as follows.
A (k) g (k) = β ^d (k) (18)
However, A (k) is a matrix of (2M ₀ (k) +1) × L ′ having α _mνu (k) as an element,

It is. [(•)] _{ω, ι} represents the ω row ι column element of the matrix (•). As shown in Equation (16), A (k) can be determined by obtaining the acoustic transfer function H _νu (r, θ; k) by actual measurement.

It is. [·] ^T represents transposition of [·], β ^d _m (k) is a frequency region of wave number k in polar coordinates (r, θ) of a region including a plurality of spatial regions in which a desired sound field is generated This is the expansion coefficient for m of the signal S ^d (r, θ; k). As shown in equation (3) (13) (14), higher sound 12-1, ..., 12-L and the spatial regions 14-1, ..., be determined desired sound field 14-S beta ^d (K) can be defined. g (k) is a vector having G _νu (k) as an element,

It is. [(•)] _ω represents the ω-th element of the vector (•).

From the above, mode matching can be realized by solving equation (18) and obtaining g (k). In this embodiment, a solution or approximate solution of g (k) in Expression (18) is obtained. The square error J (k) between A (k) g (k) and β ^d (k) can be expressed as follows.

Where W (k) is [w _m (k)] _{M0 (k) + m + 1, M0 (k) + m + 1} = w _m (k) (where m = −M ₀ (k),..., M ₀ (k) ) Defined by (2M ₀ (k) +1) rows (2M ₀ (k) +1) columns. However, the subscript “M0 (k)” represents “M ₀ (k)”. Also,

It is. w _m (k) represents the magnitude of the relative energy contribution of mode m to the control region. G (k) corresponding to the minimum value of the square error J (k) (least square error) is as follows.

Here, η is a Tikhonov regularization parameter, and I is a unit matrix of (2M ₀ (k) +1) × (2M ₀ (k) +1). The result of equation (20) is taken as the solution or approximate solution of g (k) in equation (18).

<Operation>
G (k) obtained as described above is stored in the storage unit 111 of the control unit 11 (FIG. 2). The input signal S (k) is also stored in the storage unit 111. The weight acquisition unit 112 acquires G _νu (k) from g (k) read from the storage unit 111, and sends G _νu (k) to the filtering unit 113-u (where u = 1,..., L). . The filtering unit 113-u reads the input signal S (k) from the storage unit 111, and inputs a filter whose filter weight is G _νu (k) for the radiation mode m (ν) at the wave number k of the higher-order sound source 12-u. Apply to the signal S (k) to obtain the output signal S _νu (k). The processing of the filtering unit 113-u may be performed in the time domain or may be performed in the time frequency domain. Each higher-order sound source 12-u is formed by, for example, D _u speakers sp _u (χ) (provided that χ∈ [0, D _u −1]) arranged annularly at equal intervals on the outer periphery of a cylindrical baffle. It can be constituted by an annular array (for example, Reference 1). In this case, the weight of the acoustic signal of each speaker sp _u (χ) that radiates from the higher-order sound source 12-u in the radiation mode m (ν) is W _{u, χ} = (1 / D _u ) e ^{−iνφ (χ)}
It becomes. However, χ∈ [0, D _u −1], and φ (χ) is an angle φ (χ) = 2χπ / D _u on the two-dimensional plane with respect to the center of the annular array forming the higher-order sound source 12- _u. It is. In this example, the output signal S _νu (k) is as follows.
S _νu (k) = G _νu (k) W _{u, χ} S (k)

The output signal S _νu (k) or a combination (linear sum) thereof is sent to the drive signal generation unit 114-1, and the drive signal generation unit 114-1 outputs the drive signal S corresponding to the output signal S _νu (k) or a combination thereof. ' _νu (k) or a combination thereof is generated and output to the higher-order sound source 12-u. The higher order sound source 12-u radiates an acoustic signal having a reproduction pattern corresponding to the drive signal S ′ _νu (k) or a combination thereof. Thereby, a desired sound field is generated in the desired spatial region 14-1,..., 14-S, and spatial multi-zone sound field reproduction is realized. In this embodiment, each of the radiation modes m (ν) (where ν = 1,..., 2N _V +1) of the higher-order sound source 12-u (where u = 1,..., L) is controlled by G _νu (k). To do. This can improve the flexibility of the control of the mirror image due to reverberation, the number of high-order tone 12-u of order N _V required to reproduce an accurate sound field into a plurality of regions in the reverberation room (order N _V) L can be reduced to a maximum of N _V + 1 / compared with Non-Patent Document 1.

[Second Embodiment]
As described in the first embodiment, in order to obtain a solution or approximate solution of g (k) in Expression (18), A (k) needs to be specified, and A (k) is specified accurately. For this purpose, it is necessary to accurately measure the acoustic transfer function H _νu (r, θ; k). The acoustic transfer function H _νu (r, θ; k) can be measured by observing the acoustic signal radiated from the higher-order sound source 12-u with a microphone array. However, it is not easy to accurately measure the acoustic transfer function H _νu (r, θ; k). Microphone mismatch, sensor noise, spatial aliasing, and imperfect mode equivalence each contribute to the error in the coefficients of the acoustic transfer function. In this embodiment, a method for reducing the influence of measurement error of the acoustic transfer function and improving the accuracy of A (k) will be described.

<Configuration>
As illustrated in FIG. 1, the coefficient setting device 2 according to the present embodiment includes a setting unit 21 and L high-order sound sources 12-u (provided that the reverberation room is surrounded by a wall surface 13). , U = 1,..., L) and a microphone 24-q (where q = 1,..., Q). Q is a positive integer, for example, Q ≧ 2. As illustrated in FIG. 2, the setting unit 21 includes a control unit 211, an acoustic transfer function estimation unit 212, and a coefficient calculation unit 213. The setting unit 21 is configured by, for example, the above-described general-purpose or dedicated computer executing a predetermined program. Each microphone 24-q is embedded in the baffle with the cylindrical bottom surface of radius a ≦ R _P around the global origin O. The baffle may be made of a hard material or a soft material. The microphone 24-q is arranged at polar coordinates (a, θ _q ) with respect to the origin O. 4 and 5 are examples of L = 5 and Q = 6, this does not limit the present invention.

The wave number of the radiation mode m (ν) from the higher-order sound source 12-u to the polar coordinates (a, θ _q ) obtained based on the observed acoustic signal with the microphone 24-q (where q = 1,..., Q). The acoustic transfer function H _νu (a, θ _q ; k) at k is as follows.

Here, H ′ _m (•) is a derivative (derivative) of the first kind Hankel function of order m. Based on the Wronskian relation, equation (21) can be transformed into equation (22).

here

Represents the strength of the mode in the cylindrical portion described above.

^Multiplying both sides of equation (22) by e ^−imθ and integrating, the following relationship is derived.

Furthermore, equation (23) can be modified as follows.

Here, B _m ⁻¹ (ka) is a mode equivalent filter. In this expression, aliasing errors are ignored on the assumption that an array is constituted by a sufficient number of microphones 24-q. The maximum mode order M ₀ (k) of H _νu (a, θ _q ; k) that can be measured by a circular microphone array composed of microphones 24-1,..., 24-Q is suppressed from above by Q and M ₀ (k) ≦ (Q−1) / 2.

However, λ _eq is an equivalent regularization parameter, and | (•) | represents the absolute value of (•). This regularization limits the maximum gain of the mode equivalent filter to − (6 + ₁₀ log ₁₀ λ _eq ) dB. As a result, extreme amplification of microphone noise is prevented.

<Operation>
First, the acoustic transfer function H _νu (a, θ _q ; k) is measured. When _measuring H _νu (a, θ _q ; k), the control unit 211 of the setting unit 21 (FIG. 5) sets the wave number in the mode m (ν) to the drive signal generation unit 114-u and the acoustic transfer function estimation unit 212. Send k test signals t _νu (k). The drive signal generation unit 114-u generates a drive signal t ′ _νu (k) corresponding to the test signal t _νu (k) and sends it to the higher-order sound source 12-u. The higher order sound source 12-u emits an acoustic signal corresponding to the drive signal t ′ _νu (k). The observation signal observed by the microphone 24-q is sent to the acoustic transfer function estimation unit 212. The acoustic transfer function estimation unit 212 calculates and outputs an acoustic transfer function H _νu (a, θ _q ; k) using the test signal t _νu (k) and the observation signal. The coefficient calculation unit 213 receives H _νu (a, θ _q ; k) as input, generates α _mνu (k) according to the equation (24), and outputs it. The output α _mνu (k) is used as an element of A (k) described in the first embodiment.

In the present embodiment, extreme amplification of microphone noise is prevented by using the mode equivalent filter B _m ⁻¹ (ka) of Expression (25), and the influence of measurement error of the acoustic transfer function H _νu (a, θ _q ; k). And the accuracy of A (k) can be improved.

[Modifications, etc.]
The present invention is not limited to the embodiment described above. For example, the processes of the control unit and the setting unit described above are not only executed in time series according to the description, but may also be executed in parallel or individually as required by the processing capability of the apparatus that executes the process. Needless to say, other modifications are possible without departing from the spirit of the present invention.

  When the configuration of the control unit and the setting unit is realized by a computer, the processing contents of the functions that these should have are described by a program. By executing this program on a computer, these processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. An example of a computer-readable recording medium is a non-transitory recording medium. Examples of such a recording medium are a magnetic recording device, an optical disk, a magneto-optical recording medium, a semiconductor memory, and the like.

  This program is distributed, for example, by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Furthermore, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its own storage device. When executing the process, this computer reads a program stored in its own recording device and executes a process according to the read program. As another execution form of the program, the computer may read the program directly from the portable recording medium and execute processing according to the program, and each time the program is transferred from the server computer to the computer. The processing according to the received program may be executed sequentially. The above-described processing may be executed by a so-called ASP (Application Service Provider) type service that realizes a processing function only by an execution instruction and result acquisition without transferring a program from the server computer to the computer. Good.

  In the above embodiment, the processing functions of the apparatus are realized by executing a predetermined program on a computer. However, at least a part of these processing functions may be realized by hardware.

1 Sound field reproduction device 2 Coefficient setting device

Claims (6)

A predetermined number of higher-order sound sources for reproducing an acoustic signal in a reproduction pattern, comprising any one of a plurality of independent radiation modes having different directivity patterns or a combination of at least a part of the plurality of radiation modes,
The sound field reproducing apparatus, wherein the weight of each of the radiation modes is controlled so that the reproduction pattern generates a desired sound field in a plurality of spatial regions using the echo of an acoustic signal. The sound field reproducing apparatus according to claim 1,
The predetermined high-order sound sources are L high-order sound sources hs (1),..., Hs (L), L is a positive integer, and the plurality of radiation modes are 2N _V +1 radiation modes. m (1), ..., m (2N _V +1), N _V is a positive integer, u = 1, ..., L, ν = 1, ..., 2N _V +1,
H _νu (r, θ; k) is an acoustic transfer function at the wave number k of the radiation mode m (ν) from the higher-order sound source hs (u) to the global polar coordinate (r, θ), and r , Θ is a declination, M ₀ (k) is a positive integer, m = −M ₀ (k),..., M ₀ (k), and α _mνu (k) is the acoustic transmission. The expansion coefficient of the function H _νu (r, θ; k) with respect to m, and A (k) is a matrix of (2M ₀ (k) +1) × L ′ with α _mνu (k) as elements, and L '= L (2N _V +1),

[•] ^T represents the transpose of [•]

Β ^d _m (k) is a frequency domain signal S ^d (r, θ; wavenumber k) in polar coordinates (r, θ) of an area including the plurality of spatial areas in which the desired sound field is generated. k) is the expansion coefficient for m, and g (k) is a solution or approximate solution of A (k) g (k) = β ^d (k),
The reproduction pattern is a filter weight for the radiation mode m (ν) at the wave number k of the higher-order sound source hs (u).

The sound field reproduction device which is the pattern. The sound field reproducing device according to claim 2,
H _νu (a, θ _q ; k) is obtained based on the acoustic signal observed by the microphone MIC (q) (where q = 1,..., Q) arranged at the global polar coordinates (a, θ _q ). Is an acoustic transfer function at a wave number k of the radiation mode m (ν) from the higher-order sound source hs (u) to the polar coordinates (a, θ _q ), a is a radial, and θ _q is a declination Yes,

I is an imaginary unit, H ′ _m (·) is a derivative of the first-class Hankel function of degree m,

(•) ^* represents the complex conjugate of (•), λ _eq is the equivalent regularization parameter, | (•) | represents the absolute value of (•),

A sound field reproduction device. A predetermined number of high-order sound sources reproduces an acoustic signal with a reproduction pattern consisting of any one of a plurality of independent radiation modes having different directivity patterns or a combination of at least some of the plurality of radiation modes,
The sound reproduction method according to claim 1, wherein the weight of each of the radiation modes is controlled so that the reproduction pattern generates a desired sound field in a plurality of spatial regions using the echo of an acoustic signal. The sound field reproduction method according to claim 4,
The predetermined high-order sound sources are L high-order sound sources hs (1),..., Hs (L), L is a positive integer, and the plurality of radiation modes are 2N _V +1 radiation modes. m (1), ..., m (2N _V +1), N _V is a positive integer, u = 1, ..., L, ν = 1, ..., 2N _V +1,
H _νu (r, θ; k) is an acoustic transfer function at the wave number k of the radiation mode m (ν) from the higher-order sound source hs (u) to the global polar coordinate (r, θ), and r , Θ is a declination, M ₀ (k) is a positive integer, m = −M ₀ (k),..., M ₀ (k), and α _mνu (k) is the acoustic transmission. The expansion coefficient of the function H _νu (r, θ; k) with respect to m, and A (k) is a matrix of (2M ₀ (k) +1) × L ′ with α _mνu (k) as elements, and L '= L (2N _V +1),

[•] ^T represents the transpose of [•]

Β ^d _m (k) is a frequency domain signal S ^d (r, θ; wavenumber k) in polar coordinates (r, θ) of an area including the plurality of spatial areas in which the desired sound field is generated. k) is the expansion coefficient for m, and g (k) is a solution or approximate solution of A (k) g (k) = β ^d (k),
The reproduction pattern is a filter weight for the radiation mode m (ν) at the wave number k of the higher-order sound source hs (u).

The sound field reproduction method that is the pattern. The sound field reproduction method according to claim 5,
H _νu (a, θ _q ; k) is obtained based on the acoustic signal observed by the microphone MIC (q) (where q = 1,..., Q) arranged at the global polar coordinates (a, θ _q ). Is an acoustic transfer function at a wave number k of the radiation mode m (ν) from the higher-order sound source hs (u) to the polar coordinates (a, θ _q ), a is a radial, and θ _q is a declination Yes,

I is an imaginary unit, H ′ _m (·) is a derivative of the first-class Hankel function of degree m,

(·) ^* Is the complex conjugate of (·), λ _eq is the equivalent regularization parameter, | (·) | represents the absolute value of (·),

A sound field reproduction method.