JP2018191255A - Sound collecting device, method thereof, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sound collecting device and the like, capable of fetching a target sound with higher accuracy than the conventional one using an MVDR method.SOLUTION: A sound collecting device calculates a spatial correlation matrix in each frequency by using a microphone signal of an N channel frequency region, calculates an estimation value of a noise power included in the estimation value of intensity of an incoming wave from K directions from a spatial correlation matrix and noise power contained in each microphone signal, calculates an estimation value kof an arrival direction of a target sound, calculates the estimation value of a correlation matrix of the target sound and a non-target sound by using a matrix A(f)formed by K vectors a(f) and an N×N unit matrix and a matrix V(f,l) obtained by setting to 0 all elements other than an element of (k,k) of a diagonal matrix setting the estimation value of the intensity and the estimation value of the noise power as diagonal components, calculates a filter coefficient vector by using the estimation value of the correlation matrix, adapts the filter coefficient vector to the microphone signal, and calculates an output signal.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a sound collection device using a beam forming technique for forming a beam using a plurality of microphones, a method thereof, and a program.

  Multiple microphones are installed in the sound field to acquire multi-channel microphone signals, and target voices and sounds (hereinafter also referred to as target sounds) are cleared, noise and other sounds (hereinafter referred to as non-target sounds). In recent years, there has been an increasing need for a technology for removing as much as possible. For this purpose, beam forming technology for forming a beam by using a plurality of microphones has been researched and developed in recent years.

In the beam forming technique, each microphone signal y _n (t) picked up by N microphones 91-n (where n = 1, 2,..., N) is filtered by a filtering unit 92-n as shown in FIG. Apply. Note that t is an index indicating time. Next, the sum of the output values of the filtering unit 92-n is taken in the adding unit 93. The obtained sum is output as an output signal z (t) of the sound collecting device. With such a configuration, the noise can be greatly reduced and the target sound can be extracted more clearly. As a method for obtaining such a beam forming filter, a minimum variance distortionless response method (MVDR method) is often used (see Non-Patent Document 1).

Habets, E., Benesty, J., Cohen, I., Gannot, S., Dmochowski, J., "New Insights Into the MVDR Beamformer in Room Acoustics", IEEE Transactions on Audio, Speech, and Language Processing, 18, 1, pp. 158-170, 2010.

  In order to use the MVDR method, it is necessary to appropriately estimate the correlation characteristics of sounds other than the target sound (non-target sound) and the transfer characteristics from the sound source position of the target sound to each microphone. However, the components derived from the target sound and the component derived from the non-target sound are mixed in the plurality of microphone signals, and a desired correlation matrix and transfer characteristic cannot be extracted as they are. Therefore, with the MVDR method alone, the target sound cannot be clearly extracted from only the microphone signal.

  Therefore, the present invention has an object to provide a sound collecting apparatus, a method, and a program for estimating a correlation matrix from microphone signals in which target sound and non-target sound are mixed and extracting the target sound using the MVDR method. And

In order to solve the above-described problems, according to one aspect of the present invention, the sound collection device is configured such that N and K are each an integer of 2 or more, and n = 1, 2,..., N, k = 1 , 2,..., K, and N-channel frequency domain microphone signal Y _n (f, l) is used to calculate the spatial correlation matrix R (f, l) for each frequency, and the spatial correlation matrix R (f, l ) To obtain the estimated value p _k (f, l) of the intensity of incoming waves from K directions and the estimated noise power q _n (f, l) included in each microphone signal Y _n (f, l) An incoming wave decomposition unit, a target sound determination unit that obtains an estimated value k _t of the arrival direction of the target sound, and a microphone array when a plane wave of amplitude 1 arrives at a microphone array composed of N microphones from the kth direction The vector of output signals is a _k (f), and the matrix A (f) ^H = [a ₁ (f) a ₂ (f) consisting of K vectors a _k (f) and N × N identity matrix I _N … A _K (f) I _N ] and the intensity estimate p _k (f, l) Matrix V _s (f) where all elements other than (k _t , k _t ) elements of the diagonal matrix V (f, l) whose diagonal component is the estimated noise power q _n (f, l) are zero , l) and a correlation matrix synthesizer that calculates an estimated value R ^ _T (f, l) of the target sound correlation matrix and an estimated value R ^ _NT (f, l) of the non-target sound correlation matrix, , Find the filter coefficient vector h (f, l) using the correlation matrix estimates R ^ _T (f, l) and R ^ _NT (f, l), and filter coefficients into the microphone signal Y _n (f, l) An array filtering unit that applies the vector h (f, l) and obtains the output signal z (f, l).

In order to solve the above-described problem, according to another aspect of the present invention, the sound collection method is such that N and K are each an integer of 2 or more, and n = 1, 2,..., N, k = 1, 2,..., K, and the spatial correlation matrix R (f, l) is calculated for each frequency using the N-channel frequency domain microphone signal Y _n (f, l), and the spatial correlation matrix R (f, l, estimate of the intensity of the incoming wave from the K direction from _{l) p k (f, l} ) and the microphone signal Y _n (f, estimate q _n (f noise power contained in l), l) the A arriving wave decomposition step to be obtained, a target sound determination step to obtain an estimated value k _t of the arrival direction of the target sound, and a microphone array when a plane wave having an amplitude of 1 arrives at the microphone array composed of N microphones from the kth direction. Let a _k (f) be the vector consisting of the output signals of, and the matrix A (f) ^H = [a ₁ (f) a ₂ (f with K vectors a _k (f) and the N × N identity matrix I _N )… A _K (f) I _N ] and strength Elements other than (k _t , k _t ) elements of the diagonal matrix V (f, l) whose diagonal components are the estimated value p _k (f, l) and the noise power estimate q _n (f, l) Using the matrix V _s (f, l) with all zeros, the estimated value R ^ _T (f, l) of the correlation matrix of the target sound and the estimated value R ^ _NT (f, l) of the correlation matrix of the non-target sound l) and a filter coefficient vector h (f, l) using the correlation matrix estimation values R ^ _T (f, l) and R ^ _NT (f, l), and the microphone signal Applying a filter coefficient vector h (f, l) to Y _n (f, l) to obtain an output signal z (f, l).

  According to the present invention, there is an effect that the target sound can be extracted with higher accuracy than before by using the MVDR method.

The functional block diagram of the sound collection device which concerns on a prior art. The functional block diagram of the sound collection device which concerns on 1st embodiment and 2nd embodiment. The figure which shows the example of the processing flow of the sound collection apparatus which concerns on 1st embodiment and 2nd 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 description, the symbol “^” or the like used in the text should be described immediately above the character immediately before, but it is described immediately after the character due to restrictions on text notation. In the formula, these symbols are written in their original positions. 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 embodiment>
FIG. 2 is a functional block diagram of the sound collection device 100 according to the first embodiment, and FIG. 3 shows a processing flow thereof.

The sound collection device 100 of the present embodiment receives an output signal (microphone signal) y _n (t) of a microphone array composed of N microphones 91-n. For example, the microphone 91-n includes an omnidirectional microphone element. N is an integer of 2 or more, and n = 1, 2,. The sound collection device 100 of this embodiment extracts an estimated value R ^ _NT (f, l) of the correlation matrix of the non-target sound from the N-channel microphone signal y _n (t), and extracts the target sound component by the MVDR method. The extracted signal is output as the output signal z (t).

  The sound collection device 100 is configured by a computer including a CPU, a RAM, and a ROM that stores a program for executing the following processing, and is functionally configured as follows.

  The sound collection device 100 includes a noise / arrival wave decomposition unit 101, a target sound determination unit 103, a correlation matrix synthesis unit 104, an array filtering unit 105, a Fourier transform unit 107, and an inverse Fourier transform unit.

<Fourier transform unit 107>
The Fourier transform unit 107 receives the N-channel time domain microphone signal y _n (t) as input, and performs a short-time Fourier transform to the frequency domain microphone signal Y _n (f, l) every frame l (S107). ),Output. The conversion result at frequency f and frame l

It is vectorized and treated as follows. Note that the microphone signal y (f, l) composed of the microphone signal Y _n (f, l) in the frequency domain of the N channel is
y (f, l) = x (f, l) + v (f, l)
Thus, the multi-channel signal x (f, l) composed of the direct wave of the target sound and the multi-channel signal v (f, l) composed of the reflection and reverberation components and noise are decomposed.

<Noise / arrival wave decomposition unit 101>
The noise / arrival wave decomposition unit 101 receives the microphone signal y (f, l) in the frequency domain, and uses the microphone signal y (f, l) in the frequency f and the frame l, and the spatial correlation matrix R (f, l l) is calculated. For example, it is calculated by the following formula.
R (f, l) = E [y (f, l) y (f, l) ^H ] (2)

However, E [] means taking the expected value. Y (f, l) ^H is a vector obtained by transposing y (f, l) and taking a complex conjugate. In actual processing, a short-time average is usually used instead of E []. Then, from the spatial correlation matrix R (f, l), the estimation value p _k (f, l) of the intensity of the incoming wave from K directions and the estimation of the noise power included in each microphone signal Y _n (f, l) A value q _n (f, l) is obtained (S101), and a diagonal matrix V (f, l) having p _k (f, l) and q _n (f, l) as diagonal components is output. Here, k is an index of arrival direction, and K direction is assumed as a possible arrival direction of plane waves, and k = 1, 2,. Therefore, the diagonal matrix V (f, l) is expressed as follows.

K> N. As an estimation method of the estimated value p _k (f, l) of the intensity and the estimated value q _n (f, l) of the noise power, for example, the method of Reference Document 1 can be used.
(Reference 1) P. Stoica, P. Babu, and J. Li, "SPICE A sparse covariance-based estimation method for array processing", IEEE Transactions on signal processing, vol. 59, no. 2, 2011, 629- 638.

In this method, a K direction (> N) is assumed in advance as a plane wave arrival direction. When a plane wave having an amplitude of 1 arrives at the microphone array at the frequency f from the kth direction, the response (output signal) of each microphone is expressed as a _k (f) = [a _{k, 1} (f) a _{k, 2} ( f)… a _{k, N} (f)] Let ^T. a _{k, n} (f) represents the response (output signal) of the n-th microphone to the plane wave having an amplitude of 1 coming from the k-th direction at the frequency f. Note that a _k (f) is obtained in advance prior to sound collection. However, a _k (f) may be obtained in advance by experiment (actual measurement) or simulation, or a theoretical value obtained by calculation may be used. A matrix consisting of K response vectors a _k (f) and an N × N identity matrix I _N
A (f) ^H = [a ₁ (f) a ₂ (f)… a _K (f) I _N ] (3)
In Reference Document 1,
R (f, l) = A (f) ^H V (f, l) A (f) (4)
The matrix R (f, l) is decomposed into a product of a matrix A (f) ^H , a diagonal matrix V (f, l) and a matrix A (f). By this decomposition, an estimated value p _k (f, l) of the plane wave intensity from the kth direction included in the diagonal matrix V (f, l) and an estimated value of the noise power of the nth microphone 91-n. q _n (f, l) is obtained. Actually, the above decomposition is
|| (A (f) ^H V (f, l) A (f)) ^-1/2 (R (f, l) -A (f) ^H V (f, l) A (f)) R (f , l) ^-1/2 || ² (5)
Corresponds to finding a diagonal matrix V (f, l) that minimizes. In this equation (5), || x || means to take the Frobenius norm of the matrix x.

<Target sound determination unit 103>
The target sound determination unit 103 obtains an estimated value k _t of the arrival direction of the target sound (S103) and outputs it. For example, the target sound determination unit 103 receives the diagonal matrix V (f, l) as an input, and estimates the intensity p _k (f, l) of the arrival directions k included in the diagonal matrix V (f, l). with the highest intensity in the direction it is determined that arrival direction of the target sound (S103), and outputs the determination result (the estimated value of the arrival direction) k _t. In this example, the target sound determination unit 103 obtains an estimated value k _t of the target sound arrival direction using the estimated value p _k (f, l) of the intensity of the band 100 to 500 Hz where the sound power is concentrated. In this band, the intensity of each direction of arrival k is

become. In this example, f ₀ corresponds to 100 Hz and f ₁ corresponds to 500 Hz. k that maximizes b (k, l) is determined as the arrival direction k _t of the target sound in frame l. An incoming wave from direction k _t is regarded as a target sound, and an incoming wave from directions other than k _t is regarded as a non-target sound.

<Correlation matrix synthesis unit 104>
The correlation matrix synthesis unit 104 receives the diagonal matrix V (f, l) and the estimated direction of arrival k _t and inputs the matrix A (f) ^H = [a ₁ (f) a ₂ (f)... A _K ( f) I _N ] and a matrix V _s (f, l) (diagonal matrix V (f, l)) with all elements other than (k _t , k _t ) elements of the diagonal matrix V (f, l) set to 0 This is a matrix in which all elements other than the elements of (k _t , k _t ) in K p _k (f, l) included in l) are set to 0. From the arrival direction (estimated value) of the target sound Using the matrix V _s (f, l)) that leaves only the intensity (estimated value) of the incoming sound and sets the other elements to 0, the estimated value R ^ _T (f, l) of the target sound correlation matrix ) And the non-target sound correlation matrix estimate R ^ _NT (f, l) is obtained (S104) and output. As described above, a _k (f) is obtained in advance prior to sound collection.

For example, the correlation matrix synthesis unit 104 obtains the estimated value R ^ _T (f, l) of the correlation matrix of the target sound and the estimated value R ^ _NT (f, l) of the correlation matrix of the non-target sound by the following equations. .
R ^ _T (f, l) = A (f) ^H V _s (f, l) A (f)
R ^ _NT (f, l) = A (f) ^H (V (f, l) -V _s (f, l)) A (f) (7)

<Array Filtering Unit 105>
The array filtering unit 105 performs frequency domain microphone signal y (f, l), target sound correlation matrix estimate R ^ _T (f, l), and non-target sound correlation matrix estimate R ^ _NT (f, l, l) is an input.

First, the filter coefficient vector (N-dimensional complex vector) h () using the estimated value R ^ _T (f, l) of the target sound correlation matrix and the estimated value R ^ _NT (f, l) of the non-target sound correlation matrix f, l). For example, a filter coefficient vector is obtained by using the MVDR method based on the estimated correlation matrix estimate R ^ _NT (f, l). The MVDR method solves the following constrained optimization problem and obtains its filter coefficient vector h (f, l).

Here, g ^{(d) and} (f) are vectors composed of frequency transfer characteristics of the direct path from the sound source position of the target sound to each microphone. g ₁ ^{(d) and} (f) are frequency transfer characteristics of the direct path from the sound source position of the target sound to the reference microphone 91-1. In this example, the microphone 91-1 is used as the reference, but any of the other microphones 91-2 to 91-N may be used as the reference.

This constraint condition can be rewritten using the sound source signal S (f, l) (target sound) and the signal component X ₁ ^(d) (f, l) that directly reaches the microphone 91-1 from the sound source of the target sound. it can. Note that X ₁ ^(d) (f, l) = g ₁ ^(d) (f) S (f, l).

Multiply S (f, l) X ₁ ^{(d) *} (f, l) from the right to the above constraint condition expression to obtain the expected value. However, the superscript * means taking a complex conjugate. The rewritten constraint is
h ^H (f, l) E [x ^(d) (f, l) X ₁ ^{(d) *} (f, l)] = E [X ₁ ^(d) (f, l) X ₁ ^{(d) *} ( f, l)] (9)
become. Here, x ^(d) (f, l) is a vector of signal components that directly reach each microphone from the target sound source.

Here, E [] on the left side of Equation (9) is the first vertical vector of the estimated value R ^ _T (f, l) of the correlation matrix of the target sound. E [] on the right side is the (1,1) element of the estimated value R ^ _T (f, l) of the target sound correlation matrix. The target sound source signal S (f, l) and the frequency transfer characteristics g ^(d) (f) from the target sound source to each microphone are unknown. However, the statistical procedure for taking the expected value allows the new constraint coefficient to be obtained from the estimated value R ^ _T (f, l) of the correlation matrix of the target sound.

The array filtering unit 105 applies the obtained filter coefficient vector h (f, l) to the microphone signal y (f, l) (see the following equation) to obtain the output signal z (f, l) (S105) and outputs it. To do.
z (f, l) = h ^H (f, l) y (f, l) (15)
With such a configuration, the component of the frequency f of the target sound can be extracted.

<Inverse Fourier Transform Unit 108>
The inverse Fourier transform unit 108 receives the frequency domain output signal z (f, l) as input, and performs a short time inverse Fourier transform on the processing results at all frequencies (S108) to obtain the time domain output signal z (t). ,Output.

<Effect>
With the above configuration, it is possible to estimate each correlation matrix from a microphone signal in which target sound and non-target sound are mixed, and to extract the target sound using the MVDR method.

<Modification>
In the present embodiment, the diagonal matrix V (f, l) is output from the noise / arrival wave decomposition unit 101 to the target sound determination unit 103, but the intensity included in the diagonal matrix V (f, l) is estimated. Only the value p _k (f, l) may be output. In short, it is only necessary that the target sound determination unit 103 can determine the direction of arrival of the target sound.

  In the present embodiment, the sound collection device 100 includes a Fourier transform unit 107 and an inverse Fourier transform unit 108. However, the Fourier transform unit 107 and the inverse Fourier transform unit 108 are separate devices, and the sound collection device 100 has a frequency. A microphone signal y (f, l) in the region may be input, or an output signal z (f, l) in the frequency region may be output.

<Second embodiment>
A description will be given centering on differences from the first embodiment.

  FIG. 2 is a functional block diagram of the sound collection device 200 according to the second embodiment, and FIG. 3 shows a processing flow thereof.

  The sound collection device 200 includes a noise / arrival wave decomposition unit 101, an intensity correction unit 202 (indicated by a broken line in FIG. 2), a target sound determination unit 103, a correlation matrix synthesis unit 204, an array filtering unit 105, a Fourier transform unit 107, and An inverse Fourier transform unit 108 is included.

<Intensity correction unit 202>
The intensity correction unit 202 receives the diagonal matrix V (f, l) as an input, the total signal power of the spatial correlation matrix R (f, l), and the matrix A (f) ^H = [a ₁ (f) a ₂ ( f)... a _K (f) I _N ] and the total signal power obtained from the diagonal matrix V (f, l), a correction coefficient β (f, l) is obtained (S202, indicated by a broken line in FIG. 3). ),Output.

  For example, the intensity correction unit 202 obtains a correction coefficient β (f, l) by the following equation.

Tr () is a function that takes a matrix trace. For example, the spatial correlation matrix R (f, l) may be the one calculated by the noise / arrival wave decomposition unit 101, and the matrix A (f) ^H is obtained in advance prior to sound collection by the method described in the first embodiment. You can use the ones you have left.

<Correlation matrix synthesis unit 204>
Correlation matrix synthesis section 204 receives as input spatial correlation matrix R (f, l), correction coefficient β (f, l), diagonal matrix V (f, l) and arrival direction estimation value k _t . Spatial correlation matrix R (f, l), correction coefficient β (f, l), matrix A (f) ^H = [a ₁ (f) a ₂ (f)… a _K (f) I _N ] and a matrix V _s (f, l) in which all elements other than (k _t , k _t ) elements of the diagonal matrix V (f, l) are zeroed Estimated value R ^ _NT (f, l) is obtained (S204) and output. For example, the estimated value R ^ _NT (f, l) of the correlation matrix is obtained by the following equation.
R ^ _NT (f, l) = R (f, l) -β (f, l) A (f) ^H V _s (f, l) A (f)
Note that the estimated value R ^ _T (f, l) of the correlation matrix of the target sound can be obtained by the same method as in the first embodiment.

<Effect>
By setting it as such a structure, the effect similar to 1st embodiment can be acquired. Furthermore, the correlation matrix of the non-target sound can be obtained better by using the correction coefficient.

<Other variations>
The present invention is not limited to the above-described embodiments and modifications. For example, the various processes 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 processes. In addition, it can change suitably in the range which does not deviate from the meaning of this invention.

<Program and recording medium>
In addition, various processing functions in each device described in the above embodiments and modifications may be realized by a computer. In that case, 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. Further, 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 storage unit. When executing the process, this computer reads the program stored in its own storage unit and executes the process according to the read program. As another embodiment of this program, a computer may read a program directly from a portable recording medium and execute processing according to the program. Further, each time a program is transferred from the server computer to the computer, 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 includes information provided for processing by the electronic computer and equivalent 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 addition, although each device is configured by executing a predetermined program on a computer, at least a part of these processing contents may be realized by hardware.

Claims (5)

N and K are each an integer of 2 or more, n = 1, 2,..., N, k = 1, 2,..., K, and a microphone signal Y _n (f, l) in the N-channel frequency domain Is used to calculate a spatial correlation matrix R (f, l) for each frequency, and an estimated value p _k (f, l) of the intensity of incoming waves from K directions from the spatial correlation matrix R (f, l) And an incoming wave decomposition unit for obtaining an estimated value q _n (f, l) of the noise power included in each microphone signal Y _n (f, l),
A target sound determination unit for obtaining an estimated value k _{t of the} direction of arrival of the target sound;
The vector consisting of the output signal of the microphone array when a plane wave of amplitude 1 arrives at the microphone array consisting of N microphones from the k-th direction is defined as a _k (f), and K vectors a _k (f) and N A matrix A (f) ^H = [a ₁ (f) a ₂ (f) ... a _K (f) I _N ] consisting of × N unit matrix I _N and the estimated value p _k (f, l) of the intensity A matrix V _s (all elements other than (k _t , k _t ) elements of the diagonal matrix V (f, l) having the noise power estimate q _n (f, l) as a diagonal component are set to 0. Correlation matrix synthesizer that calculates the target sound correlation matrix estimate R ^ _T (f, l) and non-target sound correlation matrix estimate R ^ _NT (f, l) When,
A filter coefficient vector h (f, l) is obtained using the estimated values R ^ _T (f, l) and R ^ _NT (f, l) of the correlation matrix, and the microphone signal Y _n (f, l) An array filtering unit that applies a filter coefficient vector h (f, l) to obtain an output signal z (f, l),
Sound collection device. The sound collection device according to claim 1,
The total signal power tr (R (f, l)) of the spatial correlation matrix R (f, l) and the matrix A (f) ^H = [a ₁ (f) a ₂ (f) ... a _K (f) I _N ] and the total signal power tr (A (f) ^H V (f, l) A (f)) obtained from the diagonal matrix V (f, l), the correction coefficient β (f, l) Including the required intensity correction unit,
In the correlation matrix synthesis unit, the spatial correlation matrix R (f, l), the correction coefficient β (f, l), the matrix A (f) ^H, and the matrix V _s (f, l) To obtain the estimated value R ^ _NT (f, l) of the correlation matrix,
Sound collection device. N and K are each an integer of 2 or more, n = 1, 2,..., N, k = 1, 2,..., K, and a microphone signal Y _n (f, l) in the N-channel frequency domain Is used to calculate a spatial correlation matrix R (f, l) for each frequency, and an estimated value p _k (f, l) of the intensity of incoming waves from K directions from the spatial correlation matrix R (f, l) And an incoming wave decomposition step for obtaining an estimated value q _n (f, l) of the noise power included in each microphone signal Y _n (f, l),
A target sound determination step for obtaining an estimated value k _{t of the} direction of arrival of the target sound;
The vector consisting of the output signal of the microphone array when a plane wave of amplitude 1 arrives at the microphone array consisting of N microphones from the k-th direction is defined as a _k (f), and K vectors a _k (f) and N A matrix A (f) ^H = [a ₁ (f) a ₂ (f) ... a _K (f) I _N ] consisting of × N unit matrix I _N and the estimated value p _k (f, l) of the intensity A matrix V _s (all elements other than (k _t , k _t ) elements of the diagonal matrix V (f, l) having the noise power estimate q _n (f, l) as a diagonal component are set to 0. Correlation matrix synthesis step to obtain the target sound correlation matrix estimate R ^ _T (f, l) and the non-target sound correlation matrix estimate R ^ _NT (f, l) When,
A filter coefficient vector h (f, l) is obtained using the estimated values R ^ _T (f, l) and R ^ _NT (f, l) of the correlation matrix, and the microphone signal Y _n (f, l) Applying a filter coefficient vector h (f, l) to determine an output signal z (f, l), and
Sound collection method. The sound collection method according to claim 3,
The total signal power tr (R (f, l)) of the spatial correlation matrix R (f, l) and the matrix A (f) ^H = [a ₁ (f) a ₂ (f) ... a _K (f) I _N ] and the total signal power tr (A (f) ^H V (f, l) A (f)) obtained from the diagonal matrix V (f, l), the correction coefficient β (f, l) Including a desired intensity correction step,
In the correlation matrix synthesis step, the spatial correlation matrix R (f, l), the correction coefficient β (f, l), the matrix A (f) ^H, and the matrix V _s (f, l) To obtain the estimated value R ^ _NT (f, l) of the correlation matrix,
Sound collection method.   A program for causing a computer to function as the sound collecting device according to claim 1.

Images