JP2015082745A - Microphone arrangement determination device and microphone arrangement determination method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To introduce quantity for evaluating correlation between transfer characteristics in each device pattern, and to determine a device configuration on the basis of this.SOLUTION: A transfer characteristic calculation part 1 calculates transfer characteristics between a microphone and a sound source about a microphone arrangement candidate consisting of the combination of all of the installation positions, angles and directivity of a plurality of microphones. A transfer characteristic correlation calculation part 2 calculates correlation between transfer characteristics about the microphone arrangement candidate related to the predetermined number of microphones. A frequency averaging part 3 frequency-averages the correlation between the transfer characteristics, and fixes the arrangement of the microphone sets related to the microphone arrangement candidate in which the correlation between the transfer characteristics is low. An optimal arrangement selection part 4 outputs the microphone arrangement candidate in which the correlation between the transfer characteristics is the lowest as the optimal microphone arrangement on the basis of the correlation between the transfer characteristics.

Description

  The present invention relates to a beam forming technique using an array device composed of a plurality of microphones and speakers. In particular, the present invention relates to a beam forming technique based on diffuse sensing that has clarified the optimum property of the transfer characteristic between a microphone and a sound source.

  Non-Patent Documents 1 and 2 are known as prior arts of speech enhancement technology based on diffusion sensing using a microphone array. In Non-Patent Documents 1 and 2, a pseudo diffuse sound field is generated by a reflecting structure, and a microphone array is installed therein to realize diffuse sensing.

K. Niwa, S. Sakauchi, K. Furuya, M. Okamoto, and Y. Haneda, "Diffused sensing for sharp directivity microphone array", ICASSP 2012, 2012, pp. 225-228 K. Niwa, Y. Hioka, K. Furuya, and Y. Haneda, "Telescopic microphone array using reflector for segregating target source from noises in same direction", ICASSP 2012, 2012, pp. 5457-5460

  However, in the related art, in order to generate a pseudo diffuse sound field by installing the reflecting structure in the vicinity of the microphone array, the scale of the device tends to increase. This is because the correlation between the volume of the reflecting structure and the reverberation time is high, and the longer the reverberation time, the closer to the diffuse sound field. For example, in Non-Patent Document 2, a reflective structure having a size that can fit in a sphere having a diameter of 1 meter is constructed. However, the device scale is often limited in advance depending on the application to be applied. If the volume is limited, there is a problem that the correlation of transfer characteristics increases and the directivity deteriorates.

  In order to reduce the correlation between transfer characteristics in a limited volume of the device, it is necessary to devise the device configuration. For example, devise the shape of the reflective structure, attach a mechanism that induces diffusion in the reflective structure, attach a movable part to the microphone or the reflective structure, and move it according to the state of the sound field. It is considered that this can be realized by means such as using an array in which different microphones are combined. However, since the devices considered in the above implementation are considered to have a considerable number of patterns, it is difficult to determine which device pattern is appropriate.

  An object of the present invention is to provide a technique for introducing a quantity for evaluating the correlation of transfer characteristics in each device pattern and determining the microphone arrangement based on the introduced quantity.

  In order to solve the above-described problem, a microphone arrangement determination apparatus according to an aspect of the present invention includes a transfer characteristic calculation unit, a transfer characteristic correlation calculation unit, a frequency average unit, and an optimum arrangement selection unit. The transfer characteristic calculation unit calculates transfer characteristics between the microphone and the sound source for microphone arrangement candidates including all combinations of installation positions, angles, and directivities of a plurality of microphones. The transfer characteristic correlation calculation unit calculates a correlation between transfer characteristics for microphone arrangement candidates related to a predetermined number of microphone sets selected from a plurality of microphones if there is no microphone set with a fixed arrangement, and a microphone set with a fixed arrangement. If there is, a correlation between transfer characteristics is calculated for a microphone arrangement candidate for a microphone set obtained by adding a predetermined number of microphones whose arrangement is not fixed to a microphone set whose arrangement is fixed. The frequency averaging unit averages the correlation between the transfer characteristics, and fixes the arrangement of the microphone sets related to the microphone arrangement candidates having low correlation of the transfer characteristics. The optimum arrangement selection unit outputs a microphone arrangement candidate having the lowest correlation of the transfer characteristics as the optimum microphone arrangement based on the correlation between the transfer characteristics.

  According to the present invention, a microphone arrangement that can most reduce the correlation between transfer characteristics can be derived from among microphone arrangement candidates that can analytically calculate the transfer characteristics.

FIG. 1 is a diagram for explaining arrangement candidates for a microphone and a sound source. FIG. 2 is a diagram illustrating a functional configuration of the microphone arrangement determining apparatus according to the first embodiment. FIG. 3 is a diagram illustrating a processing procedure of the microphone arrangement determination method according to the first embodiment. FIG. 4 is a diagram illustrating a functional configuration of the microphone arrangement determining apparatus according to the second embodiment. FIG. 5 is a diagram illustrating a processing procedure of the microphone arrangement determination method according to the second embodiment.

Hereinafter, embodiments of the present invention will be described. In the drawings used for the following description, constituent parts having the same function and steps for performing the same process are denoted by the same reference numerals, and redundant description is omitted. In the following explanations, the symbols used in the text “ ^→ (superscript right arrow)”, “^ (caret)”, “ ^- (superscript hyphen)”, etc. are directly above the previous character. It should be written immediately after the character due to restrictions on text notation. In the formula, these symbols are written in their original positions. “_ (Underscore)” used in the subscript indicates a subscript. For example, x _{y_z} represents that _yz is a subscript to x. Further, the processing performed for each element of a vector or matrix is applied to all elements of the vector or matrix unless otherwise specified.

  First, a speech enhancement technique based on conventional diffusion sensing described in Non-Patent Document 1 will be described.

[Modeling of observed signals]
Consider a situation where one target sound and K (≧ 1) noises are received using M (≧ 2) microphones. It aims at the directivity control which emphasizes the target sound in arbitrary positions in the presence of many noises. This objective can be achieved by suppressing K noise sources and emphasizing the target sound. The impulse responses between the m (m = 1,2, ..., M) th microphone and the target sound, and the k (k = 1,2, ..., K) th noise are a _m (i) and b _k , respectively. _{, m} (i). However, the impulse response length is L, and i = 0, 1,..., L−1. The impulse response length L may be determined experimentally based on the reverberation time determined by the scale and structure of the apparatus and the situation of the installed room. When the target sound is s (t) and the k-th noise source signal is n _k (t), the observed signal x _m (t) observed by the m-th microphone is modeled by the following equation.

Here, t represents a time index.

By performing a short-time Fourier transform on the observation signal x _m (t), the convolutional mixture of Equation (1) is approximated as an instantaneous mixture in the frequency domain as shown in the following equation.

Here, ω represents a frequency, and τ represents a frame index. For example, sampling is performed at 48 kHz, and the number of taps is 2048. X _m (ω, τ) represents the time frequency representation of the observed signal x _m (t), S (ω, τ) represents the time frequency representation of the sound source signal s (t) of the target sound, and N _k ( ω, τ) represents a time-frequency representation of the k-th noise source signal n _k (t). a _m (ω) represents a frequency characteristic between the target sound and the m-th microphone, and b _{k, m} (ω) represents a frequency characteristic between the k-th noise and the m-th microphone. Hereinafter, a _m (ω) and b _{k, m} (ω) are referred to as transfer characteristics. When Expression (2) is expressed in matrix form, the following expression is obtained.

And ^T represents transposition.

[Beam forming]
The output signal y (t) after beam forming is obtained by convolving an observation signal x _m (t) with a filter w _m (t) designed to emphasize the target sound as in the following equation.

Here, J represents the filter length and may be approximately the same as the impulse response length L. Y (ω, τ), which is a time frequency representation of y (t), is approximately obtained by the following equation.

Here, ^H represents the conjugate transpose, and the complex conjugate of W ^→ _m (ω) corresponds to the frequency response of w _m (j).

When the noise component included in the output signal Y (ω, τ) is written as Y _N (ω, τ), the power p _N (ω) in the following equation is defined as the power of the noise component.

Here, E _T represents a temporal expected value calculation. Assuming that the sound source signals are uncorrelated with each other, the power p _N (ω) can be calculated as shown in the following equation using only the transfer characteristic b ^→ _k (ω) and the filter W ^→ (ω).

In the field of array signal processing, various filter design methods have been described to minimize power p _N (ω). As a representative example, the delay sum method and the maximum likelihood method will be described (see Reference 1).
[Reference 1] Taichi Asano, “Sound Array Signal Processing-Sound Source Localization / Tracking and Separation”, Corona, 2011

In the delay sum method, the filter W ^→ _DS is designed to emphasize the direct sound of the target sound by the following equation.

Represents the array manifold vector of the direct sound of the target sound. The element h _m (ω) represents the transfer coefficient of the direct sound path from the target sound to the m-th microphone, where d _{m is} the distance between the target sound and the m-th microphone, c is the speed of sound, and j is the imaginary unit. Then, for example, it can be calculated by the following equation.

In the maximum likelihood method, the filter W ^→ _ML is designed to enhance the direct sound of the target sound and minimize the power p _N (ω) by the following equation.

Here, R (ω) represents a spatial correlation matrix of noise. For example, assuming that the sound source signals are uncorrelated, the noise spatial correlation matrix R (ω) is calculated using only the transfer characteristic b ^→ _k (ω) as shown in the following equation.

In classical array signal processing as described in Reference 1, it has been considered how to arrange the intervals between microphones. However, the correlation between microphones is often high except for specific frequencies. The following two are known as typical problems. The first is that in a low frequency band with a long wavelength, the correlation between transfer characteristics tends to be high, so that narrow directivity control is difficult. Second, in a high-frequency band with a short wavelength, spatial aliasing that emphasizes sounds other than a specific target sound occurs unless microphones are arranged at intervals of half a wavelength or less. From the above two points, it has been difficult to reduce the power p _N (ω) over a wide band.

[Diffusion sensing]
In Non-Patent Document 1, in order to reduce the power p _N (ω) over a wide band, the nature of the transfer characteristic should be examined, and the basic theory of diffusion sensing is summarized.

  The concept of diffuse sensing is to make the transfer characteristic uncorrelated over a wide band by “physical modulation of the transfer characteristic” as shown in the following equation.

Here, the physical modulation of the transfer characteristic means any physical means for changing the nature of the transfer characteristic itself, for example, (1) a reflective structure is installed near the microphone, (2) reflection Complicate the structure of the structure, (3) Install an object that induces diffraction, such as a sphere (rigid body), (4) Change the placement of the microphone, (5) Directivity of the microphone itself (omnidirectional, simple (6) Install reflection / diffraction in the vicinity of the directional microphone, (7) Change the material such as the walls and floor of the recording room, etc. It is done. The method proposed in Non-Patent Document 1 is a method in which a sound field (diffused sound field) in which reflected sound is repeated is generated repeatedly, and a microphone array is installed therein. For example, if a reflection structure having a shape surrounding the microphone array is made and only one surface is opened, the sound arriving in the reflection structure is automatically reflected and a pseudo diffuse sound field is generated.

  The reason why transfer characteristics are uncorrelated when a microphone array is installed in a diffuse sound field will be briefly described. It is known that the correlation γ (ω) in the diffuse sound field is calculated by the following equation, where γ (ω) is the correlation between the transfer characteristics.

Here, E _S represents the spatial expectation operation, p ^→ represents the position vector between microphones.

If the distance || p ^→ || between the microphones is sufficiently wide, the expected value of the correlation γ (ω) between the transfer characteristics in the diffuse sound field is zero.

  In order to utilize this characteristic, in the prior art, a pseudo diffuse sound field is physically generated by a reflecting structure, and a microphone array is installed therein (see Non-Patent Documents 1 and 2).

In addition, in order to reduce the power p _N (ω), filter design methods using transfer characteristics prepared by prior simulation and measurement have been studied. Simply put, only the target sound has been emphasized, but the control based on diffuse sensing is designed to emphasize the transfer characteristic itself.

When the delay sum method is used as a base, the filter W ^→ _DS1 (ω) is designed by replacing the array manifold vector h ^→ (ω) with the target sound transfer characteristic a ^→ (ω) as shown in the following equation. it can.

In this case, it is necessary to prepare the transfer characteristic a ^→ (ω) in advance by simulation or actual measurement.

Further, when the maximum likelihood method is used as a base, the filter W ^→ _DS2 (ω) can be designed by the following equation.

Similarly in this case, it is necessary to prepare the transfer characteristic a ^→ (ω) and the spatial correlation matrix R (ω) in advance by simulation or actual measurement. When using a physical means such as those mentioned above to generate a pseudo diffuse sound field and pick up the sound, it is expected that the transfer characteristics are naturally uncorrelated, so the power p _N ( ω) could be reduced over a wide band.

<First embodiment>
The purpose of this embodiment is to reduce the correlation between transfer characteristics by some physical means. In the prior art, it has been practiced to reduce the correlation between transfer characteristics by arranging a microphone in a sound field diffused using a reflective structure. However, it has been a problem to prepare transfer characteristics in advance by simulation or actual measurement.

  In this embodiment, elements such as (1) installation position, (2) angle, and (3) directivity of microphone elements that can be analytically modeled on transfer characteristics analytically in advance are variables, and a large number of combinations of each variable. Among these arrangement candidates, a microphone arrangement candidate that most likely reduces the correlation between the transfer characteristics is searched, and a microphone array device that implements an optimum microphone arrangement can be configured.

Specifically, the number of microphones constituting the microphone array device is M, the number of candidate positions of each microphone is P ₁ , the angle of each microphone (including the combination of horizontal angle and elevation angle) is P ₂ , and the direction of each microphone If the sex type is P ₃ , there can be a K (= P ₁ × P ₂ × P ₃ ) pattern as a microphone arrangement method. An optimum microphone arrangement is determined by evaluating a correlation between transfer characteristics in an L (= _K C _M ) pattern microphone arrangement candidate in which _M microphones are arranged with respect to a K pattern microphone arrangement candidate.

FIG. 1 shows an example of K pattern microphone arrangement candidates and N pattern sound source position candidates. Each of the M microphones is arranged at any one of P ₁ predetermined installation position candidates. Each microphone is placed toward any angle P ₂ types determined in advance. Each microphone has either the directivity of the P ₃ kinds of directivity predetermined. The sound source is arranged in one of N predetermined installation position candidates. In FIG. 1, an example of a microphone arrangement candidate is shown as M = 7 and P ₁ = 10, and an example of a sound source position candidate is shown as N = 6.

  With reference to FIG. 2, an example of a functional configuration of the microphone arrangement determination device according to the first embodiment will be described. The microphone arrangement determination device includes a transfer characteristic calculation unit 1, a transfer characteristic correlation calculation unit 2, a frequency averaging unit 3, an optimum arrangement selection unit 4, a microphone arrangement candidate storage unit 7, and a sound source position candidate storage unit 8. The microphone arrangement determination device is a special configuration in which a special program is read into a known or dedicated computer having a central processing unit (CPU), a main storage device (Random Access Memory, RAM), and the like. Device. For example, the microphone arrangement determination device executes each process under the control of the central processing unit. The data input to the microphone placement determination device and the data obtained in each process are stored in the main storage device, for example, and the data stored in the main storage device is read out as needed and used for other processing. Is done. Each storage unit included in the microphone placement determination device includes, for example, a main storage device such as a RAM (Random Access Memory), an auxiliary storage device configured by a semiconductor memory element such as a hard disk, an optical disc, or a flash memory, or It can be configured with middleware such as a relational database or key-value store. Each storage unit included in the microphone arrangement determination device is only required to be logically divided, and may be stored in one physical storage device.

The microphone placement candidate storage unit 7 has a K (= P ₁ × P ₂ × P ₃ ) pattern that combines all of P ₁ preset installation position candidates, P ₂ angles, and P ₃ types of directivity. Microphone placement candidates are stored.

  The sound source position candidate storage unit 8 stores N pattern sound source position candidates corresponding to predetermined N installation position candidates.

  With reference to FIG. 3, an example of the processing flow of the microphone arrangement determination method executed by the microphone arrangement determination apparatus according to the first embodiment will be described in the order of procedures actually performed.

In step S 1, the transfer characteristic calculation unit 1 receives the K pattern microphone arrangement candidates stored in the microphone arrangement candidate storage unit 7 and the N pattern sound source position candidates stored in the sound source position candidate storage unit 8 as inputs. The characteristic A (ω) = [A ₁ ^T (ω),..., A _K ^T (ω)] ^T is derived analytically. The transfer characteristic A (ω) is input to the transfer characteristic correlation calculation unit 2. Here, A _k (ω) = [A _{k, 1} (ω),..., A _{k, N} (ω)], and A _{k, n} (ω) is the kth microphone arrangement candidate and the nth sound source. It represents the transfer characteristics between position candidates. An example of a method for calculating transfer characteristics analytically will be described below.

q _{k, n} = r _{k, n} (sin (θ _{k, n} ) cos (φ _{k, n} ), relative position vector of the nth sound source position candidate when the position of the kth microphone placement candidate is the origin sin (θ _{k, n} ) sin (φ _{k, n} ), cos (θ _{k, n} )] ^T Where r _{k, n} represents the distance between the kth microphone placement candidate and the nth sound source position candidate, and θ _{k, n} is between the kth microphone placement candidate and the nth sound source position candidate. Φ _{k, n} represents the horizontal angle between the kth microphone arrangement candidate and the nth sound source position candidate. Θ _{k, n} ^(R) represents the elevation angle of the installation angle in the kth microphone arrangement candidate, and φ _{k, n} ^(R) represents the horizontal angle of the installation angle in the kth microphone arrangement candidate. Also, the installation angle vector of the _kth microphone placement candidate is q _k ^(R) = (sin (θ _{k, n} ^(R) ) cos (φ _{k, n} ^(R) ), sin (θ _{k, n} ^(R) ) sin (φ _{k, n} ^(R) ), cos (θ _{k, n} ^(R) )] ^T. At this time, the sound reception gain D _{k, n} between the kth microphone arrangement candidate and the nth sound source position candidate can be calculated as follows according to the type of directivity of the microphone.

(i) The sound reception gain D _{k, n} when using an omnidirectional microphone element is expressed by the following equation.

(ii) The sound reception gain D _{k, n} when a unidirectional microphone element is used is expressed by the following equation.

(iii) The sound reception gain D _{k, n} when a super-directional microphone element is used is expressed by the following equation.

The transfer characteristic on the premise of recording in a free sound field is calculated by the following equation using the sound reception gain D _{k, n} calculated by any one of the above methods.

Up to this point, we did not model the transfer characteristics including the effect of reflected waves.For example, when a microphone array is installed at a fixed position in a space such as a room or a car, the transfer characteristics including the reflected waves are not transmitted. It is better to model the characteristics. Although there are various methods for modeling the reflected wave, for example, modeling by a mirror image method is given (see Reference 2).
[Reference 2] JB Allen and DA Berkley, “Image method for efficiently simulating small-room acoustics”, The Journal of the Acoustical Society of America, vol. 65, no. 4, pp. 943-950, 1979.

  In addition, the sound source does not necessarily have to be located at the positions of the N pattern sound source position candidates. By giving a high weight to the position where the sound source is likely to exist in advance, the spatial position of the observed signal predicted in the sound field Characteristics may be modeled.

In step S2, the transfer characteristic correlation calculation section 2, as an input a transfer characteristic A (ω), L (= K C M) correlation C between transfer characteristics of the microphone arrangement candidates Each pattern (ω) = [C ₁ ( ω), ..., C _L (ω)] ^T is calculated. The index of the microphone in the i (i = 1,..., L) th microphone arrangement candidate is represented as {σ ₁ (i),..., σ _M (i)} ∈σ (i). Here, σ _m (i) is a function that outputs a corresponding value among integers of 1 or more and K or less.

There are various methods for calculating the correlation between transfer characteristics. (I) The method using the power average C _i ⁽¹⁾ (ω) of the correlation of transfer characteristics, (ii) The channel capacity C _i ⁽²⁾ (ω) There are four methods: (iii) a method using the condition number C _i ⁽³⁾ (ω), and (iv) a method using the determinant C _i ⁽⁴⁾ (ω).

(i) First, the calculation method of the power average C _i ⁽¹⁾ (ω) of the correlation of the transfer characteristics is shown below. The correlation power between the transfer characteristics is calculated and averaged by the following equation for all combinations of control points.

Here, β _n is a weight for the nth sound source position, and is set to a larger value as the position where the sound source is more likely to exist. If there is no prior information regarding the position of the sound source, β _n = 1 may be set. The value of the lower correlation of the transfer characteristic C _i ^{(1) (ω)} becomes small, between fully transfer characteristic becomes ^{_{C i (1) (ω)}} = 0 in the case of uncorrelated.

(ii) Next, a method using the channel capacity will be described. Channel capacity is a measure often used in wireless multiple-input and multiple-output (MIMO) systems, and is the maximum amount of information that can be sent on the transmission path when the sound source and microphone are considered as the transmission path. Yes (see Reference 3).
[Reference 3] GJ Foschini et al., “On limits of wireless communications in a fading environment when using multi-element antennas”, Wireless Personal Communications, 1998, vol. 6, no. 3, pp. 311-335

The channel capacity C _i ⁽²⁾ (ω) can be calculated by the following equation.

Here, P _SNR (ω) is an average S / N ratio between the sound source signal and sensor noise. Λ _{i, m} (ω) is the m-th eigenvalue of the spatial correlation matrix R _i (ω) expected in the microphone placement candidate by the i-th M microphones, and Λ _{i, 1} (ω) ≧. Aligned as ≧ Λ _{i, M} (ω) ≧ 0. Assuming that the sound source signals are uncorrelated with each other, the spatial correlation matrix R _i (ω) can be approximately calculated by the following equation using the analyzed transfer characteristics.

The lower the transfer characteristic correlation, the larger the value of C _i ⁽²⁾ (ω). When the transfer characteristics are completely uncorrelated, the eigenvalue is smoothed as Λ _{i, 1} (ω) ≒… ≒ Λ _{i, M} (ω), so the trace of the spatial correlation matrix R _i (ω) is constant. Under such conditions, the channel capacity C _i ⁽²⁾ (ω) is maximized.

(iii) Next, a method using the condition number C _i ⁽³⁾ (ω) will be described. The condition number in the nth microphone is calculated by the ratio of the maximum eigenvalue and the minimum eigenvalue of the spatial correlation matrix R _i (ω) as shown in the following equation.

The lower the transfer characteristic correlation, the smaller the value of C _i ⁽³⁾ (ω). When the transfer characteristics are completely uncorrelated, C _i ⁽³⁾ (ω) = 1.

(iv) Finally, a method using the determinant C _i ⁽⁴⁾ (ω) is shown. The determinant is one evaluation function used to evaluate the smoothness of the eigenvalue distribution.

The lower the transfer characteristic correlation, the larger the value of C _i ⁽⁴⁾ (ω). If the transfer characteristics are completely uncorrelated, C _i ⁽⁴⁾ (ω) = 1.

In step S3, the frequency averaging unit 3 correlates the transfer characteristics C (ω) = [C ₁ (ω),..., C _L (calculated by any one of the four methods described above. ω)] ^T is input, and the correlation C ⁻ = [C ⁻ ₁ ,..., C ⁻ _L ] ^T between the transfer characteristics averaged by the following equation is output.

Here, Ω is a set of frequency indexes to be averaged, and | Ω | represents the total number. G (ω) represents a weight for each frequency. If it is assumed that the voice is white, there is no problem even if g (ω) = 1.

In step S4, the optimum arrangement selection unit 4 receives the correlation C ⁻ = [C ⁻ ₁ ,..., C ⁻ _L ] ^T between the frequency-averaged transfer characteristics, and the microphone having the lowest transfer characteristic correlation. {Σ ₁ (i _opt ), ..., σ _M (i _opt )} ∈σ (i _opt ) are output.

For example, when the power average C _j ⁽¹⁾ (ω) or the condition number C _j ⁽³⁾ (ω) is used as the correlation between the transfer characteristics, the index i _opt of the microphone combination is selected by the following equation: .

In addition, for example, when the channel capacity C _j ⁽²⁾ (ω) or the determinant C _j ⁽⁴⁾ (ω) is used as the correlation between the transfer characteristics, the index i _opt of the microphone combination is Select.

  As described above, the microphone arrangement determining apparatus according to the first embodiment has a plurality of arrangement candidates including combinations of variables, with variables such as (1) installation position, (2) angle, and (3) directivity of the microphone element. Since the transfer characteristic is derived analytically, it is not necessary to prepare the transfer characteristic in advance by simulation or actual measurement. By calculating the correlation between the transfer characteristics based on the transfer characteristics thus derived, it is possible to determine the microphone arrangement with the lowest correlation of the transfer characteristics.

<Second embodiment>
In the first embodiment, the processing flow for determining the optimum microphone arrangement to reduce the correlation between the transfer characteristics based on the analytically calculated transfer characteristics has been described. However, if a large number of candidates for the number of microphones and the arrangement method are prepared, the number of microphone arrangement candidate patterns becomes a large value, and the calculation amount may be enormous. In the second embodiment, a description will be given of a microphone arrangement determination apparatus that reduces the amount of calculation in the optimum arrangement calculation of microphones.

  The basic idea for reducing the amount of calculation is to narrow down candidates while sequentially selecting microphones that contribute to reducing the correlation of transfer characteristics. As a result, the number of patterns for calculating the correlation of the transfer characteristics is reduced, and the overall calculation amount can be reduced.

  With reference to FIG. 4, an example of a functional configuration of the microphone arrangement determination device according to the second embodiment will be described. The microphone arrangement determination device according to the second embodiment includes a transfer characteristic calculation unit 1, an optimum arrangement selection unit 4, a microphone arrangement candidate storage unit 7, and a sound source position candidate storage unit 8 as in the first embodiment. The microphone arrangement determining apparatus according to the second embodiment further includes a transfer characteristic correlation calculating unit 5 and a frequency averaging unit 6.

  With reference to FIG. 5, an example of the processing flow of the microphone arrangement determination method executed by the microphone arrangement determination apparatus according to the second embodiment will be described in the order of procedures actually performed.

In step S <b> 1, as in the first embodiment, the transfer characteristic calculation unit 1 performs the K pattern microphone arrangement candidate stored in the microphone arrangement candidate storage unit 7 and the N pattern sound source stored in the sound source position candidate storage unit 8. Using the position candidates as inputs, transfer characteristics A (ω) = [A ₁ ^T (ω),..., A _K ^T (ω)] ^T are analytically derived. Since the method for calculating the transfer characteristic is the same as in the first embodiment, a detailed description thereof is omitted here.

  In step S5a, the transfer characteristic correlation calculation unit 5 selects P combinations of ρ microphones that contribute to a reduction in correlation between transfer characteristics from among K pattern microphone arrangement candidates. When there are microphones whose arrangement is already fixed, P sets of combinations of ρ microphones are selected from the microphones whose arrangement is not yet fixed. The combination of microphones is sometimes called a microphone set. ρ is an arbitrary integer of 2 or more. The process of the placement candidate selection unit 5 is repeatedly executed, but the value of ρ may be a different value each time it is executed. In the following description, a case where ρ = 2 is described as an example. P is an arbitrary integer of 2 or more. The value of P is not limited, but a smaller value is preferable because it corresponds to a reduction in calculation amount as the value is smaller.

In step S5b, the transfer characteristic correlation calculation unit 5 receives the transfer characteristic A (ω) as an input, and the correlation C (ω) = [C ₁ between the transfer characteristics in each of the microphone placement candidates of the L ₂ (= _K C _P ) pattern. (ω), ..., C _{L_2} (ω)] ^T is calculated. The index of the p (p = 1,..., P) th selected microphone is represented as {σ ₁ (p),..., σ _ρ (p)} ∈σ (p). The transfer characteristic between the j (j = 1,..., L ₂ ) th ρ microphone combination candidate and the nth sound source position is _expressed as α _{j, n} ⁽²⁾ (ω) = [A _{σ_1 (j) , n} (ω),..., A _{σ_ρ (j), n} (ω)] ^T. α _{j, n} ⁽²⁾ (ω) is calculated from the input transfer characteristic A (ω).

As in the first embodiment, the method for calculating the correlation between the transfer characteristics is as follows: (i) a method using the power average C _i ⁽¹⁾ (ω) of the correlation of the transfer characteristics, and (ii) a channel capacity C _i ^{(2 )} Method using (ω), (iii) Method using condition number C _i ⁽³⁾ (ω), (iv) Method using determinant C _i ⁽⁴⁾ (ω) It is sufficient to calculate using

(i) First, the calculation method of the power average C _i ⁽¹⁾ (ω) of the correlation of the transfer characteristics is shown below.

The value of the lower correlation of the transfer characteristic C _i ^{(1) (ω)} becomes small, between fully transfer characteristic becomes ^{_{C i (1) (ω)}} = 0 in the case of uncorrelated.

(ii) Next, a calculation method of the channel capacity C _i ⁽²⁾ (ω) is shown below.

Here, P _SNR (ω) is an average S / N ratio between the sound source signal and sensor noise. Also, Λ _{j, m} ⁽²⁾ (ω) is the m-th eigenvalue of the spatial correlation matrix R _j ⁽²⁾ (ω) expected in the microphone placement candidate by the j-th ρ microphone set, and Λ _{j , 1} ⁽²⁾ (ω) ≧… ≧ Λ _{j, ρ} ⁽²⁾ (ω) ≧ 0. Assuming that the sound source signals are uncorrelated with each other, the spatial correlation matrix R _j ⁽²⁾ (ω) can be approximately calculated by the following equation using the analyzed transfer characteristics.

The lower the transfer characteristic correlation, the larger the value of C _i ⁽²⁾ (ω). When the transfer characteristics are completely uncorrelated, the eigenvalues become smooth as Λ _{i, 1} ⁽²⁾ (ω) ≈… ≈Λ _{i, ρ} ⁽²⁾ (ω), so the spatial correlation matrix R _i ⁽²⁾ Trace capacity of (ω) The channel capacity C _i ⁽²⁾ (ω) is maximized under constant conditions.

(iii) Next, the calculation method of the condition number C _i ⁽³⁾ (ω) is shown below.

The lower the transfer characteristic correlation, the smaller the value of C _i ⁽³⁾ (ω). When the transfer characteristics are completely uncorrelated, C _i ⁽³⁾ (ω) = 1.

(iv) Finally, the calculation method of the determinant C _i ⁽⁴⁾ (ω) is shown below.

The lower the transfer characteristic correlation, the larger the value of C _i ⁽⁴⁾ (ω). If the transfer characteristics are completely uncorrelated, C _i ⁽⁴⁾ (ω) = 1.

In step S6a, the frequency averaging unit 6 correlates the transfer characteristics C (ω) = [C ₁ (ω),..., C _{L_2} ( ω)] ^T is input, and a correlation C ⁻ = [C ⁻ ₁ ,..., C ⁻ _L ^{— 2} ] ^T between frequency-averaged transfer characteristics by the following equation is output.

In step S6b, the frequency averaging unit 6, the correlation C between the transfer characteristic ^- based on, P number of microphones combination candidate correlation of the transfer characteristic decreases _{{σ 1 (p), ...} , σ ρ (p)} (P = 1, ..., P) is output. For example, when the power average C _j ⁽¹⁾ (ω) or the condition number C _j ⁽³⁾ (ω) is used as the correlation between the transfer characteristics, P microphone arrangement candidates are selected in ascending order of C ⁻ _j. select. As the correlation between the transfer characteristic, for example, in the case of using the communication channel capacity C _j ^{(2) (ω)} and the determinant C _j ^{(4) (ω)} is, C ^- _j a descending order into P microphone arrangement candidates Select.

In step S6c, the frequency averaging unit 6 fixes the combination of P microphones σ ₁ (p),..., Σ _ρ (p) selected in the above process. Subsequently, in step S6d, the frequency averaging unit 6 determines whether the arrangement of M microphones has been determined. If the arrangement of M or more microphones has been determined, the process proceeds to the next step S4. Proceed with the process. If the number of microphones whose arrangement has been determined is less than M, the process returns to step S5a. Thereby, the processes of the transfer characteristic correlation calculation unit 5 and the frequency averaging unit 6 are repeatedly executed until a combination of M microphones is determined. With the above processing, in the combination of M microphones, the correlation C ⁻ = [C ⁻ ₁ ,..., C ⁻ _P ] ^T between the frequency averaged transfer characteristics corresponding to the microphone arrangement candidates of the P pattern is calculated be able to.

In step S4, the optimum arrangement selecting unit 4 receives the correlation C ⁻ = [C ⁻ ₁ ,..., C ⁻ _P ] ^T between the transfer characteristics, and the combination of the microphones {σ ₁ (p _opt ), ..., σ _M (p _opt )} ∈σ (p _opt ) is output.

For example, when the power average C _j ⁽¹⁾ (ω) or the condition number C _j ⁽³⁾ (ω) is used as the correlation between the transfer characteristics, the microphone combination index p _opt is selected by the following equation: .

As a correlation between transfer characteristics, for example, when channel capacity C _j ⁽²⁾ (ω) or determinant C _j ⁽⁴⁾ (ω) is used, the index p _opt of the combination of microphones is Select.

  As described above, the microphone arrangement determining apparatus according to the second embodiment calculates the correlation between the transfer characteristics in order to determine the optimum microphone arrangement while sequentially selecting the microphones that contribute to reducing the correlation of the transfer characteristics. The number of patterns to be reduced can be reduced, and the overall calculation amount can be reduced.

  The present invention is not limited to the above-described embodiment, and it goes without saying that modifications can be made as appropriate without departing from the spirit of the present invention. The various processes described in the above-described embodiments are not only executed in time series according to the order described, but may be executed in parallel or individually as required by the processing capability of the apparatus that executes the processes.

[Program, recording medium]
When various processing functions in each device described in the above embodiment are realized by a computer, the processing contents of the functions that each device should have are described by a program. Then, by executing this program on a computer, various processing functions in each of the above devices are realized on the computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. As the computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.

  The program is distributed 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, the computer reads a program stored in its own recording medium and executes a process according to the read program. As another execution form of the program, the computer may directly read the program from a portable recording medium and execute processing according to the program, and the program is transferred from the server computer to the computer. Each time, the processing according to the received program may be executed sequentially. Also, the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes the processing function only by the execution instruction and result acquisition. It is good. Note that the program in this embodiment includes information that is used for processing by an electronic computer and that conforms to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).

  In this embodiment, the present apparatus is configured by executing a predetermined program on a computer. However, at least a part of these processing contents may be realized by hardware.

DESCRIPTION OF SYMBOLS 1 Transfer characteristic calculation part 2, 5 Transfer characteristic correlation calculation part 3, 6 Frequency averaging part 4 Optimal arrangement selection part 7 Microphone arrangement candidate storage part 8 Sound source arrangement candidate storage part

Claims (5)

A transfer characteristic calculation unit that calculates transfer characteristics between the microphone and the sound source for microphone placement candidates composed of all combinations of installation positions, angles, and directivities of a plurality of microphones;
If there is no microphone set with a fixed arrangement, the correlation between the transfer characteristics is calculated for the microphone arrangement candidates for a predetermined number of microphone sets selected from the plurality of microphones, and if there is a microphone set with a fixed arrangement, A transfer characteristic correlation calculation unit for calculating a correlation between the transfer characteristics for the microphone arrangement candidate related to the microphone set obtained by adding a predetermined number of microphones whose arrangement is not fixed to the microphone set whose arrangement is fixed;
A frequency averaging unit that averages the correlation between the transfer characteristics, and fixes the arrangement of the microphone set related to the microphone arrangement candidate with low transfer characteristic correlation; and
An optimum arrangement selection unit that outputs the microphone arrangement candidate having the lowest correlation of the transfer characteristics based on the correlation between the transfer characteristics as an optimum microphone arrangement;
A microphone arrangement determination device including: The microphone arrangement determining apparatus according to claim 1, wherein
The microphone arrangement determining apparatus, wherein the transfer characteristic correlation calculation unit calculates a determinant representing a smoothness of an eigenvalue distribution of a spatial correlation matrix calculated from the transfer characteristic as a correlation between the transfer characteristics. A transfer characteristic calculating step for calculating a transfer characteristic between the microphone and the sound source for a microphone arrangement candidate including all combinations of installation positions, angles, and directivities of a plurality of microphones;
A first transfer characteristic correlation calculating step in which a transfer characteristic correlation calculating unit calculates a correlation between the transfer characteristics for the microphone arrangement candidates related to a predetermined number of microphone sets selected from the plurality of microphones;
A frequency averaging unit that frequency-averages the correlation between the transfer characteristics and fixes the arrangement of the microphone set with low transfer characteristic correlation; and
A second transmission in which the transfer characteristic correlation calculation unit calculates a correlation between the transfer characteristics for the microphone arrangement candidates related to a microphone set in which a predetermined number of microphones whose arrangement is not fixed is added to a microphone set whose arrangement is fixed. A characteristic correlation calculation step;
An optimal placement selection unit that outputs the microphone placement candidate having the lowest correlation in transfer characteristics based on the correlation between the transfer characteristics as an optimum microphone placement; and
A microphone placement determination method including: A microphone arrangement determination method according to claim 3,
After performing the first transfer characteristic correlation calculating step, the frequency averaging step and the second transfer characteristic correlation until the number of microphones included in the microphone set having the fixed arrangement becomes equal to or greater than the number of the plurality of microphones. A microphone placement determination method that repeatedly executes calculation steps.   The program for functioning a computer as a microphone arrangement | positioning determination apparatus of Claim 1 or 2.

Images