JP2020058085A - Sound pickup device - Google Patents

Abstract

To provide a sound pickup device capable of achieving directional sound collection in a low range even when a reflection part is small.SOLUTION: The sound pickup device includes: a reflector that has a rotational paraboloid; and P microphones (P is any integer greater than or equal to 2). Defining that each of M and N is an integer of 1 or more, and P≥M+N, M microphones are placed near the focus position of the rotational paraboloid, and N microphones are placed near the center position of the surface on the focal side of the rotational paraboloid. The diameter of a circle formed by the ends of the rotational paraboloid is about twice the distance between the focus position of the rotational paraboloid and the center position of the surface of the rotational paraboloid.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a technique for collecting a target direction sound (hereinafter, also referred to as a target sound) by arranging a plurality of microphones three-dimensionally.

  A technique is known in which a plurality of microphones are arranged at a plurality of different positions near a focal position of a reflector having a paraboloid of revolution, and a beamforming process is performed based on collected signals to emphasize only a target direction sound. (See Patent Document 1). Further, a beamforming filter is designed based on a transfer function (array manifold vector) from a target sound source position or a target sound source direction to each microphone and a spatial correlation matrix (noise spatial correlation matrix) of a noise signal received by the microphone. Is performed to enhance a target sound source and suppress noise (see Non-Patent Document 1). Further, in Patent Document 1, a transfer function from a target sound source to a microphone is measured in advance for an apparatus in which a plurality of microphones are installed near a focal position of a reflection unit, and a beamforming filter is designed using the transfer function. By doing so, it is shown that the target sound source can be more efficiently emphasized than when there is no reflective portion. In addition, by installing a plurality of microphones near the focal point, only the front direction of the reflector (the direction of the rotation axis of the paraboloid of revolution of the reflector and the direction of opening (the direction on the focal side)) Instead, it is possible to form directivity even in a direction deviated from the front direction, and it is possible to emphasize sound in an arbitrary direction.

JP 2015-198411 A

Futoshi Asada, "Sound Array Signal Processing-Localization / Tracking and Separation of Sound Sources", Corona, 2011.

There is a system that emphasizes a target sound by performing beamforming sound collection using a device in which a plurality of microphones are installed near a focal position of a reflection unit as disclosed in Patent Literature 1.
However, when the reflection part is small, there is a problem that directional sound collection cannot be performed in a low sound range. In particular, when the wavelength of the sound wave is longer than the diameter of the reflection part, the effect is remarkable. An example of the directivity characteristics of the reflecting section is shown in Reference 1.
(Reference 1) "Sound collection microphone", [online], [Searched on February 20, 2017], Internet <URL: http://www.kobayasi-riken.or.jp/news/No118/118_3. htm>

  It is an object of the present invention to provide a sound collection device that can realize directional sound collection in a low sound range even when a reflection unit is small.

  In order to solve the above problems, according to the first aspect of the present invention, even when the paraboloid of revolution constituting the sound pickup device is small, the wavelength is larger than the diameter of the paraboloid of revolution. This is for accurately emphasizing a long signal. The sound pickup device includes a reflector having a paraboloid of revolution, P is any integer of 2 or more, includes P microphones, M and N are each an integer of 1 or more, and P ≧ M + N, M microphones are arranged near the focal point of the paraboloid of revolution, N microphones are arranged near the center of the focal side surface of the paraboloid of revolution, and the end of the paraboloid of revolution The diameter of the circle formed by the portion is approximately twice the distance between the focal position of the paraboloid of revolution and the center position of the surface of the paraboloid of revolution.

  According to the present invention, it is possible to achieve the same level of directivity sound collection as in the past using a smaller reflector than before, or to use a reflector that is about the same size as the conventional There is an effect that directional sound collection can be realized in a lower sound range than that.

FIG. 3 is a front view showing an example of microphone arrangement. FIG. 4 is a side view showing an example of microphone arrangement. FIG. 2 is a perspective view showing an example of microphone arrangement. The figure for demonstrating the effect of a reflector. The figure for demonstrating the microphone arrangement | positioning for enhancing a target sound source efficiently. The figure for demonstrating the example of arrangement | positioning of a parabolic reflector and a microphone. The figure which shows the sensitivity characteristic of the sound collection device of a prior art. FIG. 4 is a diagram illustrating sensitivity characteristics of the sound collection device according to the first embodiment. FIG. 2 is a functional block diagram of the sound collection device according to the first embodiment. The figure showing the example of the processing flow of the sound collection device concerning a first embodiment. The figure which shows the example 1 of a microphone arrangement. The figure which shows the example of arrangement | positioning of the microphone with respect to a parabolic reflector. The figure which shows the example 2 of a microphone arrangement. The figure which shows the example of arrangement | positioning of the microphone with respect to a parabolic reflector. The figure which shows the example of the external appearance of arrangement | positioning of the microphone with respect to the parabolic reflector of 2nd Embodiment. The figure which shows the example of the external appearance of arrangement | positioning of the microphone with respect to the parabolic reflector of 3rd Embodiment.

  Hereinafter, embodiments of the present invention will be described. In the drawings used in the following description, components having the same functions and steps for performing the same processing are denoted by the same reference numerals, and redundant description will be omitted. In the following description, the processing performed for each element of a vector or matrix is applied to all elements of the vector or matrix unless otherwise specified.

<Points of the first embodiment>
In the present embodiment, by arranging the microphones three-dimensionally, a sound arrival time difference between the microphones occurs. The directivity is improved using this time difference (phase difference). With such a configuration, sound can be collected up to a frequency at which a half wavelength is included in the distance between the microphone 20-n at the back and the microphone 20-m at the front.

  As shown in Patent Literature 1, there is a method of arranging a plurality of microphones in the vicinity of a focal position of a parabolic reflector in a planar manner to emphasize a target direction sound. However, there has not been proposed a method of collecting sound in the target direction by arranging it not only near the focal position but also at the bottom of the parabola away from the focal point, as proposed in the present embodiment.

  In the present embodiment, the microphone 20-n is disposed near the focal point of the reflector (hereinafter also referred to as a parabolic reflector) 10 having a paraboloid of revolution, and at the center of the focal-side surface of the parabolic reflector 10. . 1, 2, and 3 are a front view, a side view, and a perspective view, respectively, illustrating an arrangement example of microphones. With such a configuration, it is possible to accurately estimate a low-frequency phase difference that cannot be detected by the microphone 20-m disposed near the focal position of the parabolic reflector 10, and to achieve directional sound collection in a low-frequency range of the target direction sound. .

In the present embodiment, M microphones 20-m are arranged near the focal position of the parabolic reflector 10. Further, N microphones 20-n are arranged near the center position of the focal-side surface of the parabolic reflector 10. Hereinafter, the microphone 20-m is also called a focus microphone 20-m, and the microphone 20-n is also called a bottom microphone 20-n. _{m = 1 m, 2 m,} ..., M m, n = 1 n, 2 n, ..., a N _n. Subscripts m and n are indices indicating the focus microphone and the bottom microphone, respectively. With such a configuration, it is possible to collect directional sound including the bass range.

  In the conventional parabolic reflector, the effect of the reflector is significantly reduced at a frequency whose wavelength is longer than its diameter. The effect of the reflector is that when a microphone is installed near the focal point of the reflector, sharp directivity is formed at an angle close to the front direction of the reflector.

  At frequencies where the wavelength is shorter than the diameter of the reflector, the sound concentrates on a certain point depending on the direction of arrival of the sound, and the sound pressure increases at that point. That is, the effect of the reflector is high. FIG. 4 shows this state. The sound arriving from the front of the reflector is reflected by the reflector and collected at the focal point (point A) of the reflector, so that the sound pressure at the focal point of the reflector increases. Also, sound arriving from a direction slightly deviated from the front of the reflector is reflected by the reflector and gathers at a point B slightly deviated from the focal point of the reflector, so that the sound pressure at the point B increases. Since the point at which the sound pressure increases depends on the direction of arrival of the sound, a large amplitude difference occurs between the sounds observed by the plurality of microphones installed near the focal point depending on the direction of arrival of the sound. By performing beam forming processing mainly using the information of the amplitude difference, it is possible to realize sound collection in which sound in an arbitrary direction is emphasized.

  However, at frequencies where the wavelength is longer than the diameter of the reflector, the effect of the reflector is significantly reduced. Therefore, in a plurality of microphones installed near the focal position of the reflector, the amplitude difference due to the direction of sound arrival becomes smaller (compared to the case where the wavelength is shorter than the diameter of the reflector). Further, since the microphones are concentrated near the focal point, the phase difference due to the sound arrival direction is small. For this reason, a sharp directional characteristic cannot be formed even if the beam forming process is performed.

  In the present embodiment, the arrangement of the microphone is devised so that a sharp directional characteristic can be formed without using reflection of the reflector for a frequency having a wavelength longer than the diameter of the reflector. Further, from the viewpoint of cost, it is better that the number of microphones is as small as possible. Therefore, it is desirable that the number of additional microphones (bottom microphones) be as small as possible.

First, a microphone arrangement for efficiently emphasizing a target sound source will be described. As shown in FIG. 5, when the target sound source is on a straight line passing through the positions of the two microphones, the target sound source can be efficiently emphasized. Since the directional characteristics formed by beamforming using two microphones are symmetrical with respect to a line passing through the microphones, if the axial direction is to be enhanced, only one direction can be enhanced. On the other hand, if the direction other than the axial direction is a direction to be emphasized, there is a direction having the same sensitivity in a direction other than the direction to be emphasized (a direction having the same angle with the axis (a direction symmetric with respect to the axis)). As described above, when the microphones are arranged such that the target sound source is located on a straight line passing through the microphone, the target sound can be efficiently emphasized. Further, as described in Non-Patent Document 1, the microphone interval is preferably smaller than a half wavelength of the target frequency where spatial aliasing distortion does not occur, and the interval is preferably as small as possible so that the phase difference can be easily observed. Wider is better. That is, an interval of about a half wavelength of the target frequency is desirable.

In this embodiment, as shown in FIG. 6, one or more (three in the example of FIG. 6) microphones 20-m are installed near the focal position of the parabolic reflector 10, and the center of the focal-side surface of the parabolic reflector 10 is set. By installing one microphone 20-n at the position, for a frequency having a wavelength longer than the diameter of the parabolic reflector 10, the microphone 20-n disposed at the center of the surface of the parabolic reflector 10 and the focal position Beamforming processing is performed by mainly utilizing the arrival time difference (phase difference) of the sound with the nearby microphone 20-m to efficiently emphasize the target sound. For example, if the front direction of the parabolic reflector 10 (solid line in FIG. 6) of the target sound direction, two of the installed microphone 20-1 _m to the microphone 20-n and the point A, which is installed in the point D Mainly used for beamforming. If a direction deviated from the front (dashed line in FIG. 6) of the target sound direction, mainly using two of the installed microphone 20-3 _m to the microphone 20-n and the point C which is installed in the point D Perform beamforming. As described above, in addition to the vicinity of the focal position of the parabolic reflector 10, by installing one microphone 20-n at the center position of the surface of the parabolic reflector 10, a frequency longer than the diameter of the parabolic reflector 10 can be obtained. On the other hand, directivity can be efficiently formed in a target sound direction by beamforming mainly using a phase difference.

  Since the microphone interval is desirably a half wavelength of the target frequency, the distance from the focal point (point A) of the parabolic reflector 10 to the center of the parabolic surface is about 0.25 to 1 times the diameter of the parabolic reflector 10. Yes, especially about half is desirable. Thus, the shape of the parabolic reflector 10 may be designed.

An example in which the directivity is actually formed is shown. FIG. 7 shows sensitivity characteristics in each direction when the microphone 20-m is installed only near the focal position of the parabolic reflector 10 having a diameter of 30 cm and beamforming is performed so as to emphasize the 0-degree direction. FIG. 8 shows microphones 20-m and 20-n installed near the focal point of the parabolic reflector 10 having a diameter of 30 cm and the center position of the surface of the parabolic reflector 10 when beamforming is performed so as to emphasize the 0-degree direction. This is a sensitivity characteristic in each direction. Looking at the result at 1 kHz where the wavelength is longer than the diameter of 30 cm, the sensitivity difference between the 0 degree direction and the 80 degree direction when the microphone 20-m is installed only near the focal position is 10 dB, whereas the parabolic reflector is It can be seen that the sensitivity difference is as large as 15 dB when the microphone 20-n is also installed at the center of the surface of the surface 10. By arranging the microphone 20-n also at the center position of the surface of the parabolic reflector 10, it is possible to form a sharper directivity for a frequency having a wavelength longer than the diameter of the parabolic reflector 10. Become.

<Sound collection device 100 according to first embodiment>
9 and 10 show a functional block diagram and a processing flow of the sound collection device 100 according to the first embodiment.
The sound pickup device 100 includes P microphones 20-p and the parabolic reflector 10, and further includes a CPU, a RAM, and a computer including a ROM recording a program for executing the following processing. . This computer is functionally configured as follows. The sound collection device 100 includes an AD conversion unit 120, a frequency domain conversion unit 130, a filtering unit 160, a time domain conversion unit 170, a filter calculation unit 150, and a transfer characteristic storage unit 140.

<Parabolic reflector 10>
The parabolic reflector 10 has a paraboloid of revolution. The paraboloid of revolution has a shape, a material, and a size capable of reflecting a sound wave, and forms a focal point. In this embodiment, the parabolic reflector 10 is a rigid body having a paraboloid of revolution. The diameter of the circle formed by the edge of the paraboloid of revolution is about 0.25 to 1 or more times the largest wavelength width among the wavelength widths to be handled, and particularly preferably about half a wavelength (0.5 times) or more. For example, when the wavelength width handled by the wavelength of the sound wave is 0.01 to 1 m, it is desirable that the diameter of the circle formed by the edge of the paraboloid of revolution is about 0.5 m or more. The material of the parabolic reflector 10 is desirably a material that easily reflects sound waves (in other words, a material having a high reflection coefficient), and is preferably a hard material. Therefore, in the present embodiment, a parabolic shaped rigid body having a hard and large area is used as the parabolic reflector 10.

<Microphone 20-p>
Sound is collected using the P microphones 20-p (s1), and an analog signal (sound pickup signal) is output to the AD converter 120. Note that, among the P microphones 20-p, the M microphones 20-m are arranged near the focal position of the parabolic reflector 10, and the N microphones 20-n are disposed on the focal-side surface of the parabolic reflector 10. It is arranged near the center position. However, P is any integer of 2 or more, M and N are each any integers of 1 or more, P = M + N, p = 1,2, ..., P, and m _{= 1 m, 2 m, ...} , M m, n = 1 n, 2 n, ..., a n _n. The microphone 20-m and the microphone 20-m 'are arranged at different positions near the focal position formed by the parabolic reflector 10. Here, m ′ is one of 1, 2,..., M, and m ≠ m ′. Further, for example, the microphone 20-n and the microphone 20-n 'are arranged at different positions near the center position of the focal-side surface of the parabolic reflector 10. Here, n ′ is one of 1, 2,..., N, and n ≠ n ′.

<Position of microphone 20-p with respect to parabolic reflector 10>
P microphones 20-p may be arranged such that the determinant det (R) of the noise spatial correlation matrix R is maximized. However, the spatial correlation matrix R (ω) of the noise is calculated using only the transfer characteristic b ^→ _k (ω) of the noise as in the following equation (see Non-Patent Document 1).

P microphones 20-p are arranged to increase the SN ratio. At this time, the M microphones 20-m are arranged near the focal position. Further, the M microphones 20-m are arranged at different positions so that the inter-channel correlation is low. For example, as described in Patent Literature 1, the M microphones 20-m may be arranged so that the correlation between sound waves converted into electric signals by the M microphones 20-m is low. . Further, N microphones 20-n are arranged near the center position of the focal-side surface of the parabolic reflector 10.

(Arrangement example 1)
FIG. 11 shows the conditions for microphone arrangement. The direction of the central axis of the parabolic reflector 10 is described as a coordinate axis y. A plane passing through the intersection of the central axis of the parabolic reflector 10 and the paraboloid of revolution of the parabolic reflector 10 and orthogonal to the central axis of the parabolic reflector 10 is defined as y = H1. A plane passing through the focal position of the parabolic reflector 10 and orthogonal to the central axis of the parabolic reflector 10 is defined as y = H2. However, in FIG. 11, if the direction of the parabolic reflector 10 as viewed from the focal point is the positive direction on the y-axis, then H1> H2.

  In this arrangement example, the microphones 20-p are arranged so as to satisfy the following.

  The microphone 20-m is arbitrarily arranged in a region on the y = H2 plane and at a radius r2 from the focal position (a hatched region in FIG. 11, a circular region when viewed from the front). However, the radius r2 is set to be equal to or less than the radius r1 of the parabolic reflector 10. For example, the shaded area in FIG. 11 corresponds to the above-described “near the focal position”. The expression “near the focal position” includes the focal position itself, and the microphone 20-m may be arranged at the focal position.

  When a plane separated by a distance D from the y = H2 plane toward the y = H1 plane is defined as y = H2 + D (where H1> H2 + D), the microphone 20-n is positioned inside the parabolic reflector 10 and y = H1. N microphones 20-n are arbitrarily arranged in a region (the matte region in FIG. 11) sandwiched between the plane and the y = H2 + D plane. For example, the satin area in FIG. 11 corresponds to “near the center position of the focal-side surface” of the parabolic reflector 10. However, the plane passing through the edge of the parabolic reflector 10 and orthogonal to the central axis of the parabolic reflector 10 (the plane closest to the y = H1 plane when a plurality of planes passing through the parabolic edge can be defined) is defined as y = When the plane is H3, the distance D is set so that the plane y = H2 + D is located between the plane y = H1 and the plane y = H3 (H3 ≦ H2 + D <H1). By setting the distance D in this manner, the N microphones 20-n are arranged between the plane y = H2 + D and the focal-side surface of the parabolic reflector 10. The distance D is 0.25 to 1 times the diameter of the parabolic reflector 10, and is preferably about 0.5 times.

  FIG. 12 shows an arrangement example of the microphones 20-p with respect to the parabolic reflector 10.

  In this embodiment, the microphone 20-m is supported by a support 191 having a shape that does not easily block waves arriving at the parabolic reflector 10.

  In FIG. 12, the support portion 191 is a structure including a plane y = H2 located near the focal position formed by the parabolic reflector 10, and has a surface for holding M microphones 20-m. A plurality of holes are formed. For example, M ′ (M ′> M) holes are formed, and each microphone 20-m is embedded in any of the M ′ holes. In this embodiment, the support 191 includes a mesh member 191A and a support member 191B. The support member 191B is a rod-shaped structure that forms each side extending from the top of the square pyramid whose top is the bottom of the parabolic reflector 10 having a parabolic shape. A net-like mesh member 191A is provided on the same plane as the bottom surface of the regular quadrangular pyramid. The support member 191B couples the parabolic reflector 10 and the mesh member 191A, and fixes the mesh member 191A to the parabolic reflector 10. The mesh member 191A is a structure including a surface located near a focal position formed by the parabolic reflector 10. Further, the mesh member 191A has a plurality of holes for holding the microphone 20-m. In other words, the M microphones 20-m can be embedded in any of the meshes of the mesh-like mesh member 191A.

  The microphone 20-n is disposed near the center position (bottom) of the surface on the focal side of the parabolic reflector 10, in this example, on a fixing member of the support member 191B.

(Arrangement example 2)
FIG. 13 shows other microphone arrangement conditions. In this arrangement example, the microphones 20-p are arranged so as to satisfy the following.
The microphone 20-m is arranged at an arbitrary position in a region (shaded area in FIG. 13, spherical) within a radius r2 from the focal position. For example, the shaded area in FIG. 13 corresponds to the vicinity of the focal position.

  The microphone 20-n is located inside the parabolic reflector 10 and in a region (three-dimensional space formed by the parabolic reflector 10) between the planes of y = H1 and y = H3, and is separated from the focal point by a distance D or more. Arrange them arbitrarily in the area (the satin portion in FIG. 13). However, r2 is a size of about 1/4 of the radius r1 of the parabolic reflector 10. The distance D is about half the diameter of the parabola.

  In FIG. 14, the support portion 191 is a rod-shaped structure that couples the parabolic reflector 10 and the microphone 20-m.

  The shape of the support portion 191 is not limited to the above-described shape, and may be any shape as long as it is difficult to block the wave arriving at the parabolic reflector 10 and can support the microphone 20-m. . The microphone 20-m is supported by the support 191 and is arranged near the focal position.

<AD converter 120>
The AD converter 120 converts the P analog signals collected by the P microphones 20-p into digital signals x ^→ (t) = [x ₁ (t),..., X _P (t)] ^T. (S2), and outputs the result to the frequency domain transform unit 130. t represents a discrete time index.

<Frequency domain transforming section 130>
The frequency domain conversion unit 130 receives the digital signal x ^→ (t) = [x ₁ (t),..., X _P (t)] ^T output from the AD conversion unit 120, and outputs Q samples for each channel. It is stored in a buffer and generates a digital signal x ^→ (τ) = [x ^→ ₁ (τ),..., X ^→ _P (τ)] ^T in frame units. τ is the index of the frame number. x ^→ _p (τ) = [x _p ((τ−1) Q + 1),..., x _p (τQ)] (1 ≦ p ≦ P). Q depends on the sampling frequency, but in the case of 48 kHz sampling, around 2048 points is appropriate. Next, the frequency domain conversion unit 130 converts the digital signal x ^→ (τ) of each frame into a frequency domain signal X ^→ (ω, τ) = [X ₁ (ω, τ),..., X _P (ω, τ) )] Convert to ^T (s3) and output. ω is an index of a discrete frequency. One of the methods for converting a time-domain signal into a frequency-domain signal is a fast discrete Fourier transform, but is not limited thereto, and another method for converting into a frequency-domain signal may be used. The frequency domain signal X ^→ (ω, τ) is output for each frequency ω and each frame τ.

<Transfer characteristic storage unit 140>
The transfer characteristic storage unit 140 stores the transfer characteristics A ^→ (ω) = [a ^→ (ω), b ^→ ₁ (ω),..., B ^→ _K (ω)] measured in advance using the sound collection device 100. Remember. a ^→ (ω) = [a ₁ (ω), ..., a _P (ω)] ^T is the transfer characteristic at the frequency ω between the target sound and the P microphones, in other words, a ^→ (ω ) = [A ₁ (ω),..., A _P (ω)] ^T is the transfer characteristic of the target sound to each microphone included in the microphone array at the frequency ω. k = 1,2, ..., is a K, K is the number of ^{_{noise, b k → (ω) =}} [b k1 (ω), ..., b kP (ω)] and ^T, noise k and P this The transfer characteristic at the frequency ω between the microphone and the other microphones, that is, b _k ^→ (ω) = [b _k1 (ω),..., B _kP (ω)] ^T is transmitted to each microphone included in the microphone array. Of the noise k at the frequency ω. Note that the transfer characteristic A ^→ (ω) may be prepared in advance by a theoretical formula or a simulation, instead of performing the preliminary measurement.

<Filter calculation unit 150>
The filter calculation unit 150 extracts the transfer characteristic A ^→ (ω) from the transfer characteristic storage unit 140, calculates the filter W ^→ (ω), and outputs the calculated filter to the filtering unit 160. For example, a filter W ^→ (ω) used for signal processing for suppressing an acoustic signal from a specific position or direction is calculated.

For example, the filter W ^→ (ω) can be designed by the same method as in the related art. For example, <1> a filter design method based on the SN ratio maximization criterion described in Reference Document 2, <2> a filter design method based on Power Inversion, <3> one or more blind spots (noise Filter design method with minimum variance distortion-free response method with constraint on the direction in which gain is suppressed), <4> filter design method with delay-and-sum beam forming, and <5> maximum likelihood method The filter W ^→ (ω) can be designed by a filter design method, <6> AMNOR (Adaptive Microphone-array for noise reduction) method, or the like.
[Reference 2] International Publication No. WO2012 / 086864 pamphlet

For example, when the delay-and-sum method is used as a base, a filter W ^→ _DS1 (ω) is calculated by the following equation.

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

For example, in the case of a filter design method using a minimum variance distortionless response method having one or more blind spots as constraints, a filter W ^→ _DS3 (ω) is calculated by the following equation.

Here, f _S (ω) and f _k (ω) represent the pass characteristics at the frequency ω regarding the target sound and the noise k (k = 1, 2,..., K), respectively. For example, if the transfer characteristic a ^→ (omega) is the direction transfer characteristic a which depends on θ ^{→ (ω,} θ) as it is prepared in advance, the transfer characteristic a ^→ (ω, θ) with the filter W ^→ ( omega, calculates a theta), the filtering unit 160, perform the signal processing in a specific direction theta _s. If the transfer characteristic a ^→ (ω) can be prepared in advance as the transfer characteristic a ^→ (ω, θ, D) depending on the direction θ and the distance D, the transfer characteristic a ^→ (ω, θ, D) using the filter ^{W → (ω, θ, D} ) to calculate the, in the filtering unit 160, perform signal processing at a particular position (position specified by the particular direction theta _s and the distance D _H).

  With such a configuration, for a frequency having a wavelength shorter than the diameter of the parabolic reflector 10, beamforming is performed mainly using the M microphones 20-m, and the diameter of the parabolic reflector 10 is reduced. For a frequency having a wavelength equal to or larger than the diameter, a beamforming filter is designed so that beamforming is performed using the M microphones 20-m and the other microphones 20-n. In the unit 160, the above-described beam forming is performed.

<Filtering unit 160>
Filtering section 160 receives filter W ^→ (ω) from filter calculating section 150 in advance, receives frequency domain signal X ^→ (ω, τ), and for each frame τ, frequency domain signal X ^→ (ω, τ) = [X ₁ (ω, τ),..., X _P (ω, τ)] Apply a filter W ^→ (ω) to ^{T as shown} in the following equation (s4), and output signal Outputs Y (ω, τ).

For example, the filtering unit 160 makes the sound pickup characteristics of the sound signals emitted from at least a plurality of positions or directions in space different based on the sound pickup signal by the microphone 20-p and the sound pickup signal by the microphone 20-p '. Anything should do. "Different sound pickup characteristics" means, for example, that a sound signal emitted at a specific position is locally picked up so that an acoustic signal emitted at another position is not picked up as much as possible, or conversely, a specific signal is picked up. This means that an acoustic signal emitted at a position is suppressed (silenced) and only an acoustic signal emitted at another position is collected.

<Time domain converter 170>
The time domain conversion unit 170 converts the output signal Y (ω, τ) of each frequency ω∈Ω of the τth frame into the time domain (s5), and converts the frame unit time domain signal y (τ) of the τth frame. Then, the obtained frame unit time domain signal y (τ) is connected in the order of frame number index to output a time domain signal y (t). The method of converting the frequency domain signal into the time domain signal is an inverse transform corresponding to the transform method used in the processing of s3, for example, a fast discrete inverse Fourier transform.

<Effect>
With such a configuration, the idea of reducing the inter-channel correlation in Patent Document 1 can be inherited, the impulse response length can be shortened, and the SN ratio at the time of sound reception can be increased. Therefore, waves arriving from various directions (and information indicated by the waves) can be analyzed stably at the same time with high spatial resolution. For example, it is possible to perform sound collection in which components in the unnecessary sound direction are suppressed as compared with the related art, and the target direction sound including the low range is further emphasized. In particular, the directivity in the frequency range where the wavelength of the sound wave is longer than the diameter of the parabolic reflector can be made sharper than before. In the present embodiment, the filter W ^→ (ω) is calculated in advance. However, after the sound source position and the microphone arrangement are determined according to the calculation processing capability of the sound pickup device 100, the filter calculation unit 150 May be configured to calculate the filter W ^→ (ω) for each filter.

<Modification>
In the present embodiment, the directivity of the microphone is not mentioned, but by using a mixture of microphones having various directivities, the correlation between the transfer characteristics may be reduced and the correlation may be reduced. For example, the directivity of the microphone is not limited, but microphones having various directivities such as omnidirectional, unidirectional, bidirectional, and hypercardioid are used in combination. If the electro-acoustic transducers having different directivities are arranged at the same position, the transfer characteristics with the same control point will be different. For example, if an omnidirectional microphone and a unidirectional microphone are arranged at the same position, the transfer characteristics between the control point and the omnidirectional microphone and the transmission characteristics between the control point and the unidirectional microphone The transfer characteristics between them are different. Therefore, under this condition, the correlation between the transfer characteristics is further reduced by utilizing the change in the transfer characteristics due to the difference in directivity, and the correlation is eliminated. In other words, by making the directional characteristics of at least one of the plurality of microphones different from the directional characteristics of the other microphone, decorrelation is achieved.

  In the present embodiment, of the P microphones 20-p included in the sound pickup device 100, the M microphones 20-m are arranged near the focal position of the parabolic reflector 10, and the N microphones 20-n are parabolic. It is arranged near the center position of the surface on the focal side of the mold reflector 10. That is, P = M + N. However, the microphone holon may be arranged in a portion other than the vicinity of the focal position or the central position of the surface on the focal side. That is, P> M + N may be satisfied.

In the present embodiment, the filtering unit 160, for each frequency Omega∈omega, the frequency domain signal X ^→ (ω, τ) = a _{[X 1 (ω, τ)} , ..., X P (ω, τ)] T, The filter W ^→ (ω) is applied to output the output signal Y (ω, τ). For a frequency whose wavelength is shorter than the diameter of the parabolic reflector 10, the M microphones 20-m are output. A configuration in which beamforming is performed by using only one of them may be adopted. For example, let Th be the index of the frequency corresponding to the diameter of the parabolic reflector 10, and if ω <Th, the frequency domain signal X ^→ (ω, τ) = [X ₁ (ω, τ),. , _XP (ω, τ)] ^T is obtained. When ω ≧ Th, the frequency domain signal X ^→ _m (ω, τ) corresponding to the M analog signals collected by the M microphones 20-m is obtained. = [X _{1_m} (ω, τ),…, X _{M_m} (ω, τ)] Apply filter W ^→ _m (ω) only to ^T , and output signal Y _m (ω, τ) = [Y _{1_m} ( ω, τ),…, Y _{M_m} (ω, τ)] ^T is output. When ω ≧ Th, the filter calculation unit 150 calculates only the filter W ^→ _m (ω). The transfer characteristic storage section 140 only needs to store only transfer characteristics necessary for calculating the filter W ^→ _m (ω) when ω ≧ Th. For example, when ω ≧ Th, the transfer characteristic a ^→ (ω) = [a _{1_m} (ω),..., A _{M_m} (ω)] ^T at the frequency ω between the target sound and the M microphones and the noise k transfer characteristics b _k at a frequency between the M microphones ^{ω → (ω) = [b} k_m (ω), ..., b kM_m (ω)] is stored and ^T. With such a configuration, the same effect as in the first embodiment can be obtained.

In the present embodiment, the filtering process is performed in the frequency domain. However, the filtering process may be performed in the time domain. For example, as a configuration in which time-domain filter coefficients are convolved in the time domain for each channel p with _P digital signals x ₁ (t),..., X _P (t) in the time domain, and the convolution results of all channels are added. Is also good. In this case, the frequency domain transform section 130 and the time domain transform 170 need not be provided. Further, the transfer characteristic storage unit 140 stores the transfer characteristics in the time domain measured using the sound pickup device 100 in advance, and the filter calculation unit 150 calculates the filter in the time domain.

<Second embodiment>
The following description focuses on the differences from the first embodiment.

The sound collection device according to the present embodiment also includes the P microphones 20-p and the parabolic reflector 10. However, among the P microphones 20-p, the M microphones 20-m are arranged near the focal position of the parabolic reflector 10, and the L microphones 20-1 are located at the center axis of the parabolic reflector 10. It is arranged at a position away from the front. However, P is any integer of 2 or more, M and L are each any integers of 1 or more, P = M + L, p = 1, 2,..., P, and m = 1 _m , 2 _m ,..., M _m , l = 1 _l , 2 _l ,..., L _l , and the subscript l indicates that the microphone is located away from the front. It is an index. With such a configuration, it is possible to collect directional sound including the bass range as in the first embodiment. FIG. 15 shows an example of the appearance of the arrangement. The same effect as in the first embodiment can be obtained by placing the microphone 20-1 on the front side of the focal position.

  However, in the case of the present embodiment, there is no restriction on the distance between the focal position and the microphone placed in front of the focal position, so that there is an advantage that the microphone can be placed at an arbitrary distance. For example, the L microphones 20-1 are arranged at a distance of 0.25 to 1 times the wavelength of the frequency at which the effect is desired to be obtained. In particular, it is preferable to arrange the L microphones 20-1 so that the distance is 0.5 times. That is, the microphone 20-1 may be installed at a position about half the diameter of the parabolic reflector 10 where the frequency and the wavelength at which the effect of the reflector is low coincide with each other.

<Modification>
In the present embodiment, the L microphones 20-1 are arranged at positions separated from the central axis of the parabolic reflector 10 in the front direction, but are not necessarily arranged “on the central axis”. The L microphones 20-1 are arranged on the front side of the focal position of the parabolic reflector 10, and the distance from the focal position of the parabolic reflector 10 is 0.25 to 1 times the diameter of the parabolic reflector 10. What is necessary is just to arrange | position in the "near the point on the central axis" of the reflector 10. In particular, it is preferable to dispose it at “near a point on the central axis” where the distance from the focal position becomes 0.5 times the diameter. It is sufficient that the M microphones 20-m and the L microphones 20-1 are separated to such an extent that the low-frequency phase difference can be accurately estimated. “Nearby” in “nearby” means a range in which such an effect can be obtained. When a point on the central axis is O, for example, “near a point on the central axis” is a half-wavelength or less of the center frequency of the center frequency of the target sound frequency band, within a sphere around the point O. Area. In this case, `` near a point on the central axis '' is an area within a sphere centered on the point O having a radius of at most a half wavelength of the center frequency of the frequency band, and a minimum point O (radius is zero). ). The expression “near a point on the central axis” includes the point itself on the central axis, and the microphone 20-1 may be arranged at a point on the central axis.

<Third embodiment>
The following description focuses on the differences from the first embodiment.

  The sound collection device according to the present embodiment also includes the P microphones 20-p and the parabolic reflector 10. However, among the P microphones 20-p, among the P microphones 20-p, the M microphones 20-m are arranged near the focal position of the parabolic reflector 10, and the N microphones 20-p are arranged. n is disposed near the center of the focal-side surface of the parabolic reflector 10, and the L microphones 20-1 are disposed at positions away from the central axis of the parabolic reflector 10 in the front direction. Here, P is any integer of 3 or more, M, N, and L are each any integers of 1 or more, and P = M + N + L. The arrangement of the microphones is as described in the first embodiment, the second embodiment, and their modifications.

  With such a configuration, it is possible to collect directional sound including the bass range. FIG. 16 shows an example of the appearance of the arrangement. By placing the microphone both on the bottom part of the parabolic reflector 10 and on the front side of the focal position, it is possible to obtain higher effects than in the first embodiment and the second embodiment. As in the first embodiment and the second embodiment, the distance between the focal point and the microphone at the bottom of the parabola, and the distance between the focal point and the microphone installed in front of the focal point are smaller than the diameter of the parabolic reflector 10. It is about 0.25 to 1 times, and especially about half is desirable.

<Other modifications>
The present invention is not limited to the above embodiments and modifications. For example, the above-described various processes may be executed not only in chronological order as described, but also in parallel or individually according to the processing capability of the device that executes the processes or as necessary. In addition, changes can be made as appropriate without departing from the spirit of the present invention.

<Program and recording medium>
Further, various processing functions in each device described in the above embodiment and the modified examples may be realized by a computer. In this case, the processing content of the function that each device should have is described by a program. By executing this program on a computer, various processing functions of the above-described devices are realized on the computer.

  A program describing this processing content can be recorded on a computer-readable recording medium. As a 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 distribution of the program is performed by, for example, selling, transferring, lending, or the like, a portable recording medium such as a DVD or a CD-ROM on which the program is recorded. Further, the program may be stored in a storage device of a server computer, and the program may be distributed by 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 unit. Then, when executing the processing, the computer reads the program stored in its own storage unit and executes the processing according to the read program. As another embodiment of the program, a computer may directly read the program 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 sequentially performed. A configuration in which the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes a processing function only by executing an instruction and acquiring a result without transferring a program from the server computer to the computer. It may be. It should be noted that the program includes information to be used for processing by the computer and which is similar to the program (such as data that is not a direct command to the computer but has properties that define the processing of the computer).

  Further, each device is configured by executing a predetermined program on a computer, but at least a part of the processing contents may be realized by hardware.

Claims (2)

Even if the paraboloid of revolution constituting the sound pickup device is small, a sound pickup device for accurately emphasizing a signal whose wavelength is longer than the diameter of the paraboloid of revolution,
A reflector having a paraboloid of revolution,
P is any integer of 2 or more, including P microphones,
M and N are each an integer of 1 or more, and P ≧ M + N, M microphones are disposed near the focal position of the paraboloid of revolution, and N microphones are located on the paraboloid of revolution. It is located near the center position of the focal side surface,
The diameter of a circle formed by the end of the paraboloid of revolution is approximately twice the distance between the focal position of the paraboloid of revolution and the center of the surface of the paraboloid of revolution.
Sound pickup device. The sound pickup device according to claim 1,
M is any integer of 2 or more,
Based on a sound pickup signal by a certain microphone and a sound pickup signal by another microphone, including a filtering unit that makes sound pickup characteristics of sound signals emitted from at least a plurality of positions or directions in space different.
Sound pickup device.

Images