 Забыли?

?

# JP2008048294

код для вставкиСкачать
```Patent Translate
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2008048294
PROBLEM TO BE SOLVED: To provide a directional array microphone which has the same
directivity as the treble to the bass even if the array size is small, and has the single directivity of
a narrower beam width. SOLUTION: A first two-dimensional FIR filter is formed by connecting a
microphone array 2 arranged along an x-axis and a y-axis and first and second two-dimensional
FIR filters 12, 13 in cascade. At 12, the directivity corresponding to the y-axis direction is
obtained, and at the second two-dimensional FIR filter 13, the directivity corresponding to the xaxis direction is obtained. In these two-dimensional FIR filters 12 and 13, when the amplitude
characteristic is represented on a two-dimensional frequency plane consisting of a time
frequency axis and a space frequency axis, the passband in the cross section parallel to the space
frequency axis is It is formed from a physical area to a part of the physical area adjacent to the
non-physical area, and is configured to have a larger amplitude in the non-physical area than in
the physical area. As a result, the directivity characteristics can be broadened. [Selected figure]
Figure 1
Directional array microphone and directional array speaker

The present invention relates to a directional array microphone and a directional array speaker.

As a method of realizing a directional microphone that receives only sound waves coming from a
specific direction, there is a directional array microphone that uses a microphone array in which
04-05-2019
1
a plurality of microphones are arranged.
Also, as a directional speaker that emits sound waves only in a specific direction, a directional
array speaker using a speaker array in which a plurality of speakers are arranged is known.

In the directional array microphone, only the component of the sound wave coming from a
specific direction is extracted by the digital filter from the acoustic signal input to each of the
microphones constituting the microphone array, whereby the directional characteristic is
obtained. In the directional array speaker, the sound wave having directional characteristics can
be output by individually controlling the amplitude and phase of the acoustic signal output from
each of the speakers constituting the speaker array by the digital filter. Hereinafter, sound waves
in a specific direction extracted by a directional array microphone and sound waves in a specific
direction emitted from a directional array speaker are collectively referred to as an acoustic
beam. Moreover, electroacoustic transducers, such as a microphone and a speaker, may be
generically called a transducer.

When designing a digital filter used in such a directional array microphone and directional array
speaker, time and the arrangement position of each transducer are used as variables rather than
displaying the acoustic signal for each transducer. The method (refer nonpatent literature 1) of
displaying as a two-dimensional acoustic signal is used. In this method, the frequency spectrum
obtained by performing two-dimensional Fourier transform on a two-dimensional acoustic signal
is on a straight line according to the traveling direction of the acoustic beam in a twodimensional frequency plane represented by the time frequency axis and the space frequency
axis. It is used to appear in Therefore, the characteristics of the digital filter can be represented
by a fan filter having fan-shaped characteristics on a two-dimensional frequency plane.

Specifically, as a method of designing a target amplitude characteristic of a fan filter, the time
frequency axis of the two-dimensional frequency plane is divided, and the divided amplitude
04-05-2019
2
characteristic at the cross section in the spatial frequency axis direction for each divided time
frequency is There is known a method (refer to Non-Patent Document 2) of designing to be an
amplitude characteristic of a filter (Dolph-Chebyshev filter). In this method, the Chebyshev
polynomial is associated with the amplitude characteristic in the cross section in the spatial
frequency axis direction so that the passband has a single peak and the stopband has a plurality
of equal amplitude ripples. As a result, the amplitude of the stop band equal ripple can be
reduced compared to the pass band amplitude. From the obtained discrete data of the target
amplitude characteristics, the filter coefficients of the two-dimensional FIR filter can be obtained
by two-dimensional discrete Fourier inverse transform (see Non-Patent Document 3).

In the directional microphone and the directional speaker, it is desirable that the same directional
characteristic be realized in a wide band from the high range to the low range. However, in the
fan filter, the passband in the spatial frequency axial cross section in the bass region is narrower
than that in the treble region, so if you try to obtain the same amplitude and directivity as the
trebleband in the bass region, you will have sharp peaks You need to have a passband. However,
since it is difficult to obtain a passband having such a sharp peak, in the prior art, in order to
maintain the amplitude as high as that of the high passband to the narrow passband, the
passband in the low passband is low. The passband characteristics with steep peaks have been
realized by extending the passband to make the directivity wider by the bass range, or increasing
the number of transducers to increase the size of the array (Non-patent literature) 2).

As a method of designing a two-dimensional fan filter in which the same directivity characteristic
is further expanded to the low frequency range without increasing the array size, there is a
method of actively using the non-physical region of the two-dimensional frequency plane (NonPatent Document 4) , 5). Here, the non-physical region is a region in which a value obtained by
dividing the spatial frequency by the time frequency exceeds the reciprocal of the speed of sound
in a two-dimensional frequency plane, and physically refers to a region where the spectrum of
the acoustic signal can not be distributed. .

04-05-2019
3
In the design method described in Non-Patent Document 4, the stop band ripple is designed to
have a large amplitude in the non-physical region of the two-dimensional frequency plane using
the design method of Parks & McClellan. The width of the passband becomes narrower as the
amplitude of the stopband ripple increases, so that an amplitude characteristic having a steep
peak passband can be realized, and as a result, the directivity characteristic identical to that of
the high tone range can be realized to a more bass range. be able to.

In the design method described in Non-Patent Document 5, when designing the amplitude
characteristic in the cross section in the spatial frequency direction on the two-dimensional
frequency plane, the passband is from one non-physical region to the non-physical region among
The Chebyshev polynomial is made to correspond to the amplitude characteristic in the spatial
frequency axis direction so that it is formed over a partial region adjacent to the physical region
and has a larger amplitude in the non-physical region than in the physical region. By giving the
maximum amplitude of the passband in the non-physical region, the steep passband amplitude
characteristics can be realized in the physical region adjacent to the non-physical region, so the
directivity characteristics identical to the high-tone region up to the bass region Can be given.

Kiyoshi Nishikawa, "Two-dimensional domain analysis of beamforming", Transactions of the
Institute of Electronics and Communication Engineers, The Institute of Electronics, Information
and Communication Engineers, September, 1994, Volume J77-A, No. 9, p. 1304-1306, K.
Matsumoto, K. Nishikawa, "A Design Method of Directional Array Loudspeakers with a Constant
Sidelobe Amount," Technical Report of IEICE, The Institute of Electronics, Information and
Communication Engineers, October 2004, EA 2004-74, p. 13-18 Sei Nishikawa, et al., "Design
method of two-dimensional FIR fan filter for broadband beam forming by two-dimensional
Fourier series approximation", Transactions of the Institute of Electronics and Communication
Engineers, The Institute of Electronics, Information and Communication Engineers, Dec. 2000,
J83-A, No. 12, p. 1357-1367 Mitsuru Ota, Kiyoshi Nishikawa, "Wideband design of directional
array loudspeakers with a fixed sidelobe amount", IEICE Technical Report, The Institute of
Electronics, Information and Communication Engineers, October 2005, EA 2005-50, p. 7-12 N.
Nagakawa, K. Nishikawa, "Wideband design of a single sided array microphone with a fixed
sidelobe amount", IEICE Technical Report, The Institute of Electronics, Information and
Communication Engineers, October 2005, EA 2005-49, p. 1-6

04-05-2019
4
As described above, in the directional array microphone and the directional array speaker, there
is a technical problem in that the amplitude and the directivity in the high range are similarly
obtained in the low range.
It is possible to achieve the same directivity in a wide band by arranging a large number of
transducers in a wide range, but since the system becomes large, the same wide band from bass
to treble band with the smallest possible array size It is desirable that directional characteristics
be obtained. In this respect, widening of the directional characteristics can be achieved by
adopting a design method that positively utilizes the non-physical region in the two-dimensional
frequency plane as in Non-Patent Documents 4 and 5 described above.

However, since the acoustic characteristics in the case of arranging the transducers along a
straight line become rotational symmetry characteristics around the straight line in which the
transducers are arranged, the direction of the acoustic beam in the design method described in
Non-Patent Document 4 described above Is in the direction intersecting the straight line in which
the transducers are arranged, and is rotationally symmetric about the straight line. For example,
if the direction of the acoustic beam is orthogonal to the array direction of the transducers, the
wave front of the acoustic wave will be formed to extend in an arc when viewed in a cross section
perpendicular to the array direction of the transducers. Therefore, in this design method, the
direction of travel of the acoustic beam does not have so-called single directivity characteristics
distributed around the central axis of the beam.

On the other hand, in the design method described in Non-Patent Document 5 described above,
the passband of the sound wave is a cross section parallel to the spatial frequency axis of the
two-dimensional frequency plane. As it is set in the area adjacent to the area, the central axis of
the acoustic beam coincides with the alignment direction of the linearly arranged transducers,
and the acoustic beam is emitted conically around the axis along which the transducers are
arranged. It will be. Therefore, in the design method described in Non-Patent Document 5, socalled single directivity characteristics are realized, but as in Non-Patent Document 4 described
above, an acoustic beam is formed in a direction intersecting the array direction of the
04-05-2019
5
transducers. In the case of the same array size, there is a drawback that the beam width of the
acoustic beam is wider than in the case of

The object of the present invention is to provide a directional array microphone and a directional
array speaker having the same directivity as the treble to the bass even if the array size is small
and having the uni-directional characteristic with a narrower beam width. It is to do.

According to the present invention, there is provided a microphone array comprising a plurality
of microphones arranged along a first plane and a second direction crossing each other on the
same plane, and a microphone comprising a plurality of microphones arranged along the first
direction. A first two-dimensional digital filter, provided for each column, for filtering the signals
output from the plurality of microphones constituting each of the microphone columns so as to
have directivity corresponding to the first direction And a second two-dimensional digital filter
that filters the signals output from the plurality of first two-dimensional digital filters so as to
have directivity corresponding to the second direction, The first and second two-dimensional
digital filters have their amplitude characteristics as a two-dimensional frequency consisting of a
time frequency axis and a space frequency axis. When expressed on a plane, the passband in a
cross section parallel to the spatial frequency axis is formed from one non-physical area to a part
of the physical area adjacent to the one non-physical area, and from the physical area Is a
directional array microphone characterized in that it is configured to have a large amplitude in
the non-physical region.

Further, according to the present invention, each of the microphones constituting the
microphone array is held by holding means capable of changing an angle formed by the first
direction and the second direction.

Further, according to the present invention, there is provided a speaker array comprising a
plurality of speakers arranged along a first plane and a second direction crossing each other on a
same plane, a plurality of first two-dimensional digital filters, and a second two-dimensional A
directional array speaker including a digital filter, wherein the second two-dimensional digital
filter filters an input signal so as to have directivity corresponding to the second direction.
Outputting the filtered signal to each first two-dimensional digital filter, wherein the first twodimensional digital filter is provided for each of the speaker trains including a plurality of
04-05-2019
6
speakers arranged along the first direction. And each signal output from the second twodimensional digital filter is filtered to have directivity corresponding to the first direction. When
the first and second two-dimensional digital filters are represented on a two-dimensional
frequency plane consisting of a time frequency axis and a space frequency axis, the first and
second two-dimensional digital filters have cross sections parallel to the space frequency axis.
The passband is formed to extend from one non-physical area to a part of the physical area
adjacent to the one non-physical area, and configured to have a larger amplitude in the nonphysical area than in the physical area. It is a directional array speaker to be characterized.

Further, according to the present invention, each of the speakers constituting the speaker array is
held by holding means capable of changing an angle formed by the first direction and the second
direction.

According to the present invention, the directional array microphone has a microphone array
consisting of a plurality of microphones arranged along the first and second directions crossing
each other on the same plane, and is received by the microphone array The directivity is given to
the acoustic signal by filtering the first and second two-dimensional digital filters.

Here, the passbands of the first and second two-dimensional digital filters in a cross section
parallel to the spatial frequency axis on the two-dimensional frequency plane are adjacent to one
non-physical region from the one non-physical region Therefore, an acoustic beam centered on
the first direction is formed by the filtering process by the first two-dimensional digital filter
provided for each microphone array along the first direction. Further, the filtering process by the
second two-dimensional digital filter to which the output from the first two-dimensional digital
filter is supplied forms an acoustic beam centered on the second direction.
As a result, the acoustic beam finally formed by the directional microphone is an acoustic beam
formed by the filtering process by the first two-dimensional digital filter and an acoustic beam
formed by the filtering process by the second two-dimensional digital filter. It is formed in the
common area with the beam.
04-05-2019
7
Therefore, even if only a wide acoustic beam can be formed only by the microphone array along
one of the first and second directions, the common part of the characteristics in both directions
becomes the characteristic of the final acoustic beam. A narrow beam width can be obtained.

Furthermore, compared to the case where the microphones are arranged along a straight line as
in the prior art, the present invention realizes a narrow acoustic beam by arranging the
microphones in a two-dimensional manner on the same plane. A smaller and smaller array size
directional array microphone can be realized.

Also, the passbands of the first and second two-dimensional digital filters in a cross section
parallel to the spatial frequency axis on the two-dimensional frequency plane are formed so that
the amplitude in the non-physical region is larger than that in the physical region. The amplitude
characteristics of the passband of the region adjacent to the non-physical region in the physical
Therefore, it is possible to realize the same directivity as the high tone range up to the low tone
range.

Further, according to the present invention, the angle formed by the first direction and the
second direction can be changed by the holding means for holding the microphones constituting
the microphone array.
As described above, the acoustic beam according to the present invention is formed in the
common region of the acoustic beam whose central direction is the first direction and the
acoustic beam whose central direction is the second direction.
Therefore, the smaller the angle between the central direction of the acoustic beam along the
first direction and the central direction of the acoustic beam along the second direction, the
04-05-2019
8
wider the common area of the two acoustic beams, and thus the wider the beam width. Can be
realized.
Conversely, the larger the angle between the central direction of the acoustic beam along the first
direction and the central direction of the acoustic beam along the second direction, the narrower
the common area of the two acoustic beams, and thus the narrower the beam width. Can be
realized.
By changing the angle between the first direction and the second direction in this manner, the
width of the acoustic beam can be changed.

Further, according to the present invention, the directional array speaker has a speaker array
composed of a plurality of speakers arranged along the first and second directions crossing each
other on the same plane, and the signal inputted from the outside is On the other hand,
directivity is given to the sound wave radiated from the speaker array by performing the filtering
process by the first and second two-dimensional digital filters.

Here, the passbands of the first and second two-dimensional digital filters in a cross section
parallel to the spatial frequency axis on the two-dimensional frequency plane are from one nonphysical region to the non-physical region of the physical region. It is formed over a part of
Thus, as with the directional array microphone described above, the acoustic beam formed by the
directional array speaker is an acoustic beam having a central axis along the first direction
formed by the first two-dimensional digital filter; Formed in a common area with an acoustic
beam with a central axis along the second direction formed by the second two-dimensional
digital filter, thus achieving a narrower beam width, even with small and small array sizes can do.

Also, as in the case of the directional array microphone described above, the passbands of the
04-05-2019
9
first and second two-dimensional digital filters in a cross section parallel to the spatial frequency
axis of the two-dimensional frequency plane are in the non-physical region more than the
physical region. Because the amplitude characteristic of the pass band of the region adjacent to
the non-physical region in the physical region is steep, the directivity characteristic can be

Further, according to the present invention, the angle formed by the first direction and the
second direction can be changed by the holding means for holding the speakers constituting the
speaker array, and therefore, as in the case of the microphone array, , The width of the acoustic
beam can be changed.

FIG. 1 is a block diagram showing the configuration of a directional array microphone 1
according to an embodiment of the present invention.
The directional array microphone 1 includes a microphone array 2 including a plurality of
microphones 10 arranged along both the y-axis direction as the first direction and the x-axis
direction as the second direction, and the y-axis direction It includes a first two-dimensional FIR
(finite impulse response) filter 12 and a second two-dimensional FIR filter 13 provided for each
microphone row 11.

In FIG. 1, in the microphone array 2, NA + 1 (10 is an even number equal to or more than 2)
microphones 10 are arranged at a predetermined interval DA along the x-axis direction which is
the second direction. In the first direction, along the y-axis direction, NB + 1 (NB is an even
number of 2 or more) microphones 10 are arranged at a constant interval DB from each other,
and a total of (NA + 1) × (NB + 1) Microphones 10 are arranged in a grid.
In FIG. 1, the case of the minimum configuration of NA = NB = 2 is illustrated. Here, the
microphone array 11 refers to a group of NB + 1 (NB is an even number of 2 or more)
microphones 10 along the y-axis direction which is the first direction. The microphone array 2
04-05-2019
10
shown in FIG. 1 includes NA + 1 microphone rows 11.

In the present embodiment, the y-axis direction which is the first direction and the x-axis
direction which is the second direction are orthogonal to each other, and the microphone 10 is
disposed on the xy plane. Each microphone 10 is assumed to have the same shape and the same
omnidirectional acoustic characteristics. Further, in the present embodiment, the case where the
odd number of microphones 10 are arrayed in the x-axis and y-axis directions is exemplified by
setting the NA and NB to an even number of 2 or more. The even number of microphones 10
may be arranged in each direction, with H being an odd number of 3 or more. In this case, the
microphones 10 are not arranged on the symmetry axis of the microphone array 2. Further, one
of NA and NB is an even number of 2 or more, the other is an odd number of 3 or more, an odd
number of microphones 10 are arranged in one direction, and an even number of microphones
10 are arranged in the other direction. It is also good.

The sound waves 14 reaching the microphones 10 constituting the microphone array 2 can be
approximated by plane waves. The traveling direction of the sound wave, which is a plane wave,
is perpendicular to the wavefront. Here, in the xy plane shown in FIG. 1, an angle φ between the
traveling direction of the plane wave and the y axis is defined as the incident angle of the sound
wave. The incident angle when the sound wave is incident from the positive side to the negative
side of the y axis along the y axis direction is φ = 0 °, and the incident angle when the sound
wave is incident from the positive direction side of the x axis is represented by positive The case
where a sound wave is incident from the negative direction side of the x axis is represented as
negative. In the present specification, when the incident angle of the sound wave to the
microphone 10 is represented, it is represented as in FIG.

The first two-dimensional FIR filter 12 is provided corresponding to each of the microphone
arrays 11 in NA + 1. In each first two-dimensional FIR filter 12, acoustic signals received by the
plurality of microphones 10 constituting the microphone array 11 corresponding to the first twodimensional FIR filter 12 are A / D (analog / digital) ) Converted to a digital signal by a converter
04-05-2019
11
(not shown) and then input.

The signal filtered by each first two-dimensional FIR filter 12 is input to the second twodimensional FIR filter 13. Then, the second two-dimensional FIR filter 13 performs filtering to
obtain an acoustic signal from the sound wave 14 that has arrived from a specific direction. Here,
although the FIR filter is used as the digital filter in this embodiment, an infinite impulse
response (IIR) filter may be used. Signal processing in digital filters such as FIR filters and IIR
filters are performed as operations by a microprocessor.

The first and second two-dimensional FIR filters 12 and 13 can be regarded as a configuration of
parallel sums of one-dimensional FIR filters, respectively. FIG. 2 is a block diagram showing the
configuration of a two-dimensional FIR filter 20 used as the first and second two-dimensional FIR
filters 12 and 13 in the directional array microphone 1 shown in FIG.

The two-dimensional FIR filter 20 includes an NC + 1 one-dimensional FIR filter 22 for filtering a
digital signal input from NC + 1 (NC is an even number of 2 or more) input terminals 21, and
each one-dimensional FIR filter 22. And an adder 23 for adding the output.

When the two-dimensional FIR filter 20 is used as the first two-dimensional FIR filter 12 shown
in FIG. 1, NC = NB, and the number of one-dimensional digital filters 22 is equal to the number of
microphones 10 constituting each microphone row 11. Is equal to NB + 1.
The outputs from the microphones 10 constituting each microphone row 11 are input to the onedimensional FIR filters 22 corresponding to the respective microphones 10, and the outputs from
the one-dimensional FIR filters 22 are input to the adder 23 and added. Result is output.
04-05-2019
12

When the two-dimensional FIR filter 20 is used as the second two-dimensional FIR filter 13
shown in FIG. 1, NC = NA, and the number of one-dimensional digital filters 22 is the same as
that of each first two-dimensional FIR filter 12. It is equal to NA + 1 which is the number. The
output of each first two-dimensional FIR filter 12 is input to the corresponding one-dimensional
FIR filter 22 via the input terminal 21, and the output from each one-dimensional FIR filter 22 is

Hereinafter, the configuration and function of the directional array microphone 1 shown in FIG. 1
will be described in more detail. First, as shown below, the directional array microphone 1 can be
regarded as a configuration in which a linear directional array microphone in which the
microphones are linearly arranged is cascaded in two stages.

FIGS. 3 and 4 are block diagrams showing the configuration of linear directional array
microphones 15a and 15b (hereinafter referred to as "linear array microphones") in which the
microphones 10 are linearly arranged. The linear array microphone 15a shown in FIG. 3 includes
NB + 1 (NB is an even number of 2 or more) microphones 10 and a first two-dimensional FIR
filter 12, which are arranged along a straight line at intervals DB. The linear array microphone
15b shown in FIG. 4 includes NA + 1 pieces (NA is an even number of 2 or more) of microphones
10 arranged along a straight line at intervals DA and a second two-dimensional FIR filter 13. The
arrangement direction of the microphones 10 is different between the linear array microphone
15a of FIG. 3 and the linear array microphone 15b of FIG. 4, and the microphones 10 are arrayed
in the y-axis direction in FIG. The microphones 10 are arranged in the x-axis direction.

In FIG. 3, the numbers of the microphones 10 arranged at the origin of the xy plane are 0, and
numbers 1, 2, 3, ... are sequentially assigned in the + y direction, and -1, -2 in the -y direction. , -
04-05-2019
13
3,... Therefore, in the linear array microphone 15a shown in FIG. 3, numbers from -NB / 2 to NB /
2 are assigned from -y direction to + y direction. Further, in FIG. 4, as the number 0 of the
microphone 10 disposed at the origin of the xy plane, the numbers are sequentially assigned 1, 2,
3, ... in the + x direction, and -1, -2 in the -x direction. , -3,... Therefore, in the linear array
microphone 15b shown in FIG. 4, the numbers from -NA / 2 to NA / 2 are assigned from the -x
direction toward the + x direction.

Comparing FIG. 3 and FIG. 4 with FIG. 1, the directional array microphone 1 shown in FIG. 1 is
different from the microphone 10 of the linear array microphone 15b shown in FIG. 4 in the
linear array microphone 15a shown in FIG. It has been arranged. The positional relationship of
each microphone 10 is such that the zeroth microphone 10 of the origin is located at the position
of each of the linear array microphones 15a shown in FIG. 3 at the position of each of the
microphones 10 constituting the linear array microphones 15b shown in FIG. Be placed.

As described above, the directional array microphone 1 shown in FIG. 1 has a configuration in
which the linear array microphones 15a and 15b are connected in a two-stage cascade, so that
the characteristics of the directional array microphone 1 are linear array microphones. It can
explain based on the characteristic of 15a and 15b. Therefore, the directivity characteristics of
the linear array microphones 15a and 15b will first be described below.

FIG. 5 is an explanatory diagram for illustrating how sound waves are received by a microphone
array consisting of microphones arranged at equal intervals along a straight line. In order to
distinguish from the two-dimensional array microphone array 2 of the present invention, the
linear array microphone array shown in FIG. In FIG. 5, N (N is a natural number) omnidirectional
microphones 10 ̶ n (n is an integer from 0 to N−1) are arranged at an interval d on the x axis
in the xy plane. Here, the reference code of the microphone disposed at the origin of the xy plane
is represented by 10_0, and the reference code of the microphone disposed at a position d away
from the origin in the + x direction is represented by 10_1, and similarly n from the origin to the
+ x direction It is assumed that reference symbols of microphones disposed at positions
04-05-2019
14
separated by a distance of xd are represented by 10_n.

In FIG. 5, assuming that the sound wave 14 at the sound velocity c approximated as a plane wave
is incident on the microphone array 16 as φ, the sound wave arriving at a certain microphone
10_n further advances by d · sin φ, the distance d in the −x direction To reach the microphone
10 ̶ n−1 arranged. In time, the sound wave 14 arrives at the microphone 10 ̶ n−1 arranged
at an interval d in the −x direction with a delay of d · (sin φ) / c.

FIG. 6 is a diagram showing a time change of an acoustic signal observed by each microphone 10
̶ n in the microphone row 16 shown in FIG. 5. In FIG. 6, the horizontal axis represents time t,
and the vertical axis represents the position of each microphone by the distance from the origin
of the x axis as in FIG. The acoustic characteristic of the microphone 10_n disposed at the
position where the distance from the origin is nd (n is an integer from 0 to N-1) has a point on
the straight line of x = nd as the origin at each time t and the + x direction It is displayed as
positive. As described above, the acoustic characteristic of the microphone 10_n + 1 at the
position of x = (n + 1) d leads the acoustic characteristic of the microphone 10_n at the position
of x = nd by time d · (sin φ) / c, x = ( The acoustic characteristics of the microphone 10_n-1 at
the position n-1) d are delayed by the time d · (sin φ) / c, so the acoustic signal observed in the
entire microphone array 16 is xt as shown in FIG. When expressed as a two-dimensional acoustic
signal en (t) on a plane, the slope of the observed two-dimensional acoustic signal en (t) is −c /
sin φ.

FIG. 7 is a diagram in which a spectrum of an acoustic signal obtained by performing a twodimensional Fourier transformation on a time t and a position coordinate x of a microphone with
respect to a two-dimensional acoustic signal en (t) shown in FIG. is there. In FIG. 7, the horizontal
axis represents time frequency f1, and the vertical axis represents spatial frequency f2.

04-05-2019
15
Here, as described in Non-Patent Document 1, the spectrum of the acoustic signal is a straight
line 30 of inclination (sin φ) / c passing through the origin of the two-dimensional frequency
plane, and 2 in the xt plane shown in FIG. It has a relation orthogonal to the inclination −c / sin
φ of the two-dimensional acoustic signal en (t). Therefore, as described in Non-Patent Document
1, the following equation 1 holds for the relationship between the time frequency f1 and the
space frequency f2.


In Equation 1, f2 = f1 / c is satisfied when φ = 90 °, and therefore, it is represented as a straight
line 31 having a slope of 1 / c in FIG.
At this time, in FIG. 5, the sound wave 14 comes from the + x direction, and the arrangement
direction of the microphones coincides with the traveling direction of the sound wave. Further, in
the formula 1, f2 = -f1 / c is established when φ = -90 °, and therefore, it is represented as a
straight line 32 with a slope of -1 / c in FIG. At this time, in FIG. 5, the sound wave 14 comes
from the −x direction, which is the opposite direction to the case of φ = 90 °.

In FIG. 7, the region where the absolute value of the ratio f2 / f1 of f2 to f1 exceeds the
reciprocal of the speed of sound c is a non-physical region where the spectrum of the acoustic
signal can not physically exist. Therefore, the spectrum of the acoustic signal is present in the
region where the absolute value of f2 / f1 is equal to or less than the speed of sound c. In
addition, it is necessary to observe | f2 | ≦ 1 / (2d) for the absolute value of the spatial
frequency f2 so that spatial aliasing does not occur, and assuming that the sampling period T
similarly for the absolute value of the temporal frequency f1, f1 | ≦ 1 / (2T) must be observed.

Here, the time frequency f1 is normalized by the reciprocal 1 / T of the sampling period T, the
04-05-2019
16
space frequency f2 is normalized by the reciprocal 1 / d of the distance d between the
microphones, and the normalized T × f1 is simply represented as f1. If the normalized d × f 2 is
simply expressed as f 2, equation 1 can be expressed as the following equations 2 and 3.


Hereinafter, in the present specification, unless otherwise specified, the temporal frequency f1
and the spatial frequency f2 are used to mean standardized values.
Therefore, the conditions for avoiding aliasing of time and space are | f1 | ≦ 0.5 and | f2 | ≦
0.5.
Further, from Equation 2, the non-physical region is a region in which | f2 / f1 |> ρ holds for the
absolute value of the ratio of f2 and f1.

As described above, the frequency spectrum of the two-dimensional acoustic signal en (t) when
the sound wave 14 arriving at the incident angle φ is received by the microphone array 16 is a
straight line of inclination ・ · sin φ passing through the origin in the two-dimensional frequency
plane. As represented, in the case of a two-dimensional acoustic signal en (t) over a range of
incident angles φ, the frequency spectrum in the two-dimensional frequency plane is
represented by a straight line of a range of slopes passing through the origin Become. Therefore,
by using a fan filter in which the shape of the pass band in the two-dimensional frequency plane
is a fan, it is possible to make the microphone array 16 have a directional characteristic such that
only incoming waves in a specific direction are received. Here, the sound wave in the specific
direction extracted by the fan filter is called an acoustic beam. An acoustic beam may be
described simply as a beam.

04-05-2019
17
FIG. 8 is a diagram for explaining the relationship between the spectrum of the two-dimensional
acoustic signal and the characteristics of the fan filter in the two-dimensional frequency plane. In
FIG. 8, the horizontal axis represents time frequency f1, and the vertical axis represents spatial
frequency f2.

As described above, in the two-dimensional frequency plane shown in FIG. 8, the non-physical
region 33 has an absolute value | f2 of the value of the ratio of f2 to f1 with the line of φ = 90 °
and the line of φ = −90 ° as a boundary. Is a side larger than 1, which is a region where the
spectrum of the acoustic signal can not exist. Further, a region excluding the non-physical region
33 in the two-dimensional frequency plane is called a physical region. On the other hand, the
passband 34 of the fan filter in the present embodiment is a cross section parallel to the spatial
frequency f2, and is set as a unimodal passband from the physical region to the non-physical
region. In FIG. 8, the blocking area 35 is set in the area between the straight line of φ = φs and
the straight line in the physical area, and the straight line of φ = −90 °, and the passing area
34 is set in the area excluding the blocking area 35. However, in FIG. 8, only the passing area 34
and the blocking area 35 in the physical area are displayed. Further, the passband 34 is set such
that its amplitude is larger in the non-physical area 33 than in the physical area. Therefore, in the
physical region where the frequency spectrum of the two-dimensional acoustic signal is present,
the line of φ = 90 ° has the maximum amplitude, so the central direction of the acoustic beam
is the direction of φ = 90 ° which is the arrangement direction of the microphones.
Unidirectional characteristics can be realized. Furthermore, since the change of the amplitude
can be made steeper in the transient region up to the border line φ = φs with the stop zone 35
in the passband 34, the width of the passband in the cross section parallel to the spatial
frequency axis is Even in the narrow bass range, it is possible to realize a linear array
microphone having the same directivity as the high tone range.

Next, referring to FIG. 1, FIG. 3 and FIG. 4, directivity characteristics in the case where linear
array microphones 15a and 15b are cascade-connected as in the directional array microphone 1
shown in FIG. 1 will be described. .

In the one-stage linear array microphone 15b shown in FIG. 4, the frequency characteristics HmA
(f1, f2) of each microphone are set to constant sensitivity and omnidirectionality, that is, HmA
04-05-2019
18
(f1, f2) = 1, and the second two-dimensional Assuming that the characteristic of the FIR filter 13
is HA (f1, f2), the frequency characteristic HtA (f1, f2) of the linear array microphone 15b shown
in FIG.


Similarly to the case of the linear array microphone 15b shown in FIG. 4, the frequency
characteristics HmB (f1, f2) of each microphone of the frequency array HtB (f1, f2) of the linear
array microphone 15a of the one-stage configuration shown in FIG. Is given by HmB (f1, f2) = 1,
and the frequency characteristic of the first two-dimensional FIR filter 12 is HB (f1, f2).


As described above, the directional array microphone 1 of all two stages shown in FIG. 1 is a
straight line of one-stage configuration shown in FIG. 3 instead of each microphone 10 of the
linear array microphone 15b of one-stage configuration shown in FIG. The frequency
characteristics Ht (f1, f2) of all two stages of the directional array microphone 1 shown in FIG. 1
are linear array microphones 15a and 15b each having a single-stage configuration. Is expressed
by the product of HtB (f1, f2) and HtA (f1, f2), which are frequency characteristics of


As a result, the frequency characteristic Ht (f1, f2) of the directional array microphone 1 shown
in FIG. 1 is the frequency characteristic HB (f1, f2) of the first two-dimensional FIR filter 12 and
the second two-dimensional FIR filter 13 It is represented by the product of the characteristic HA
(f1, f2) of
Therefore, on the two-dimensional frequency plane, the passband of the two-stage directional
array microphone 1 is the common part of the passband of the first two-dimensional FIR filter 12
04-05-2019
19
and the passband of the second two-dimensional FIR filter 13 It is possible to realize a directivity
with a narrower beam width than that of the linear array microphones 15a and 15b configured
in one stage.
The amplitude characteristics of the directional array microphone 1 on the two-dimensional
frequency plane will be specifically described below with reference to FIGS. 1, 3, 4 and 9 to 12.

FIG. 9 is a view representing the amplitude characteristics of the frequency characteristic HB (f1,
f3) of the first two-dimensional FIR filter 12 in the linear array microphone 15a shown in FIG. 3
on a two-dimensional frequency plane.
In FIG. 9, the horizontal axis represents time frequency f1, and the vertical axis represents spatial
frequency f3 with respect to y axis.
Here, in FIG. 3, since the arrangement direction of the microphones 10 is the y-axis direction, the
spatial frequency is the spatial frequency f3 with respect to the y-axis.

As described in FIG. 8, in the present embodiment, the passband of the fan filter is set to an area
ranging from a part of the physical area to the non-physical area 33, and the amplitude of the
passband in the non-physical area 33 is physically It is set to be larger than the amplitude of the
passband in the region.
Therefore, the direction of φ = 0 °, which is the arrangement direction of the microphones 10
for obtaining the maximum amplitude in the physical region, is the beam center direction.
In FIG. 9, the passband 36 b in the physical region is set to a region sandwiched by a straight line
of φ = 0 ° and a straight line of φ = 90 ° −φs.
04-05-2019
20
For ease of illustration, the passband of the fan filter in the non-physical area is not shown. As
described later, since the beam center direction in the final two-stage directional array
microphone 1 is the direction of φ = 45 °, the amplitude of the passband 36b on the line of φ
= 45 ° in FIG. Set to 1.

FIG. 10 is a view representing the amplitude characteristics of the frequency characteristic HA
(f1, f2) of the second two-dimensional FIR filter 13 in the linear array microphone 15b shown in
FIG. 4 on a two-dimensional frequency plane. In FIG. 10, the horizontal axis represents time
frequency f1, and the vertical axis represents spatial frequency f2 with respect to the x axis. In
FIG. 4, since the arrangement direction of the microphones 10 is the x-axis direction, the spatial
frequency is the spatial frequency f2 with respect to the x-axis.

As in the case of FIG. 9, the direction of φ = 90 °, which is the arrangement direction of the
microphones 10 for obtaining the maximum amplitude in the physical region, is the beam center
direction. Therefore, the passband 36a in the physical region is set to a region sandwiched by the
line of φ = 90 ° and the line of φ = φs. For ease of illustration, the passband of the fan filter in
the non-physical area is not shown. As described later, since the beam center direction in the
final two-stage directional array microphone 1 is the direction of φ = 45 °, the amplitude of the
passband 36a on the line of φ = 45 ° in FIG. Set to 1.

FIG. 11 converts the frequency characteristic HB (f1, f3) of the first two-dimensional FIR filter 12
shown in FIG. 9 into the frequency characteristic HB (f1, f2) for the spatial frequency f2 with
respect to the x axis, and the amplitude thereof It is a figure showing the characteristic. In FIG.
11, the horizontal axis represents time frequency f1, and the vertical axis represents spatial
frequency f2 with respect to the x axis.

04-05-2019
21
As described above, the frequency characteristics Ht (f1, f2) of the directional array microphone
1 are the frequency characteristics HB (f1, f2) of the first two-dimensional FIR filter 12 and the
frequency characteristics of the second two-dimensional FIR filter 13 It is represented by a
composite property which is the product of HA (f1, f2). However, the frequency characteristic HB
(f1, f3) of the first two-dimensional FIR filter 12 shown in FIG. 9 is represented by the spatial
frequency f3 with respect to the y-axis and the time frequency f1 (f3 = .rho. (Cos .phi.) F1). In
order to obtain the combined characteristic, the expression by the spatial frequency f2 with
respect to the x-axis and the time frequency f1 similar to the frequency characteristic HA (f1, f2)
of the second two-dimensional FIR filter 13 shown in FIG. Convert to = ρ (sin φ) f1).


As shown in FIG. 11, the passband 36c of the amplitude characteristic of the first twodimensional FIR filter 12 is a region sandwiched by the line of φ = φs-90 ° and the line of φ =
90 ° -φs, and the beam Center coincides with the direction of φ = 0 °.
The amplitude = 1 is on the line of φ = ± 45 °.

12 shows the frequency characteristic HA (f1, f2) of the second two-dimensional FIR filter 13
shown in FIG. 10 and the frequency characteristic HB (f1, f2) of the first two-dimensional FIR
filter 12 shown in FIG. It is a figure showing the amplitude characteristic about frequency
characteristic Ht (f1, f2) of directional array microphone 1 which is a product.

As shown in FIG. 12, the passband 36d of the directional array microphone 1 is the passband
36a of the second two-dimensional FIR filter 13 shown in FIG. 10 and the pass of the first twodimensional FIR filter 12 shown in FIG. It is a common part with the area 36c, and coincides with
the area sandwiched by the line of φ = φs and the line of φ = 90 ° −φs.
04-05-2019
22
Further, the central direction of the acoustic beam by the directional array microphone 1 is φ =
90 ° of the central direction of the beam by the second two-dimensional FIR filter 13 shown in
FIG. 10, and the first two-dimensional FIR filter 12 shown in FIG. The central value of φ = 0 °,
which is the direction of the center of the beam, is φ = 45 °.

Next, a procedure for specifically designing the first and second two-dimensional FIR filters 12
and 13 having the directivity characteristics as described above will be described.

As described in Non-Patent Documents 3 and 5, first, the amplitude characteristics as shown in
FIG. 8 and FIG. 9 are given as the target amplitude characteristics Hd (f1, f2).
Here, the target amplitude characteristic Hd (f1, f2) is a periodic characteristic having a basic
period of f1 = [− 0.5, 0.5] and f2 = [− 0.5, 0.5]. Given the target amplitude characteristic Hd (f1,
f2), the two-dimensional discrete Fourier inverse transform generates a two-dimensional FIR
filter of the order N1 of time and the order N2 of space (where N1 and N2 are positive even
numbers). The filter coefficient h (n1, n2) can be obtained according to the following equation 8.
Here, n1 and n2 represent discrete time and space variables, n1 is an integer of 0 to N1, and n2
is an integer of -N2 / 2 to N2 / 2. Further, in Equation 8, M1 and M2 are selected to be integers
of powers of 2 by about 10 times N1 and N2 in order to perform two-dimensional discrete
Fourier inverse transform by fast Fourier transform. Further, k1 and k2 respectively represent
discrete time frequency and space frequency variables, k1 is an integer of -M1 / 2 to M1 / 2-1,
and k2 is -M2 / 2 to M2 / 2-1. Is an integer of


As described in Non-Patent Document 5, the target amplitude characteristic Hd (f1, f2) (where f1
= k1 / M1, f2 = k2 / M2) is designed using a Chebyshev polynomial.
04-05-2019
23
Since the fan filter characteristics are symmetrical with respect to the origin, the target amplitude
characteristic Hd (f1, f2) for negative f1 is designed to be a point (f1, f1 with respect to the origin
symmetry). The value of the target amplitude characteristic Hd (f1, f2) in f2) is used.

FIG. 13 is a diagram for explaining the procedure for designing the target amplitude
characteristics Hd (f1, f2) of the first and second two-dimensional FIR filters 12 and 13 of the
directional array microphone 1 shown in FIG. FIG. 13 (b) is a diagram giving a target amplitude
characteristic on a two-dimensional frequency plane consisting of a temporal frequency f1 and a
spatial frequency f2. FIG. 13 (a) is a spatial frequency at a temporal frequency f1 at a twodimensional frequency plane It is the figure which gave the amplitude characteristic in the cross
section 40 parallel to an axis | shaft by the Chebyshev polynomial. Here, the Chebyshev
polynomial is a regular expression for the variable x, and the Nth-order Chebyshev polynomial
TN (x) can be expressed by the following recursion equation.


As shown in FIG. 13A, in the present embodiment, the amplitude characteristic at the cross
section 40 parallel to the spatial frequency axis is set to have the maximum amplitude in the nonphysical region.
In FIG. 13A, the maximum amplitude in the non-physical region is A0, the spatial frequency f2 at
φ = 90 ° of the boundary between the physical region and the non-physical region is F0, and
the amplitude is 1. Further, spatial frequencies f2 at both ends of the stop band are Fs1 and Fs2,
amplitudes at that time are δs, values of φ corresponding to the spatial frequencies Fs1 and Fs2
are φ = φs, and φ = −90 °. The amplitude characteristic of the directional array microphone 1
of the two-stage configuration is finally K times the whole so that the amplitude becomes 1 at the
beam center φ = φb (φb = 45 ° in the present embodiment). It is obtained by FIG. 13A shows
the final amplitude obtained by multiplying the whole by K. The spatial frequency f2 at φ = φb
is Fb. From the above relationship, the following equations 10 to 13 hold.
04-05-2019
24


Here, the amplitude characteristic A of the filter whose order is N and whose stopband is equal
ripple can be given by an N-order Chebyshev polynomial using a frequency parameter f as shown
by the following equation 14.
However, in equation 14, x1 is the value of the variable x of the Chebyshev polynomial (equation
9) when giving the maximum amplitude, and x = x1 · cos (πf) holds.


In Equation 14, δ is the magnitude of the stop band ripple and fst is the stop band end
frequency. When x = x1, f = 0, the maximum amplitude A0 is obtained, and when x = x0, the
amplitude is 1 when f = f0. Assuming that the amplitude becomes 1 / K at f = fb when x = xb, the
following equations 15 to 20 hold.


Using the equation 14, the spatial frequency f2 is converted to the frequency parameter f of the
equation 14 to give an amplitude characteristic at the cross section 40 parallel to the spatial
frequency axis on the two-dimensional frequency plane shown in FIG. I do.
FIG. 14 is a view showing the correspondence between the variable x of the Chebyshev
polynomial TN (x) used in the equation 9, the frequency parameter f used in the equation 14, and
the spatial frequency f2.
04-05-2019
25
The horizontal axis represents the correspondence between the variables x, cos (πf), the
frequency parameter f, and the spatial frequency f2, and the vertical axis represents the value of
the Chebyshev polynomial TN (x) for the variable x.
Further, FIG. 15 is a diagram showing the Chebyshev polynomial shown in FIG. 14 with the
horizontal axis as a frequency parameter f. The horizontal axis represents frequency parameter f,
and the vertical axis is Chebyshev polynomial TN (x) for frequency parameter f multiplied by K ·
δ to obtain φ = φb (in this embodiment, φb = 45 °) and the amplitude is 1 It is supposed to
be

If it is assumed that F0, Fs1, Fs2, and Fb on the f2 axis correspond to f0, fst, 1-fst, and fb on the f
axis, respectively, as shown in the correspondence between the frequency parameter f and the
spatial frequency f2 in FIG. The following equations 21 to 25 hold.


When the time frequency f1 (where f1 = [0, 0.5]) is given by using the relational expressions of
the above equations 10 to 25, the amplitude characteristic A (f) in a cross section parallel to the
spatial frequency axis Can be determined.
The specific setting method differs depending on the value of f1 for each of five areas from the
first area to the fifth area.

In the first region, the upper limit of f1 is 0.5, and as the lower limit, the amplitude characteristic
in the cross section is given for f1 up to the following fh.

04-05-2019
26

In this region, since the non-physical region is narrower than the passband of the physical region,
the amplitude characteristic is determined by setting f0 = 0 and A0 = 1 without using the
relational expression of Equation 12.
In this case, when f1 is determined, fst is determined by Equations 10, 11, and 21. When fst is
determined, x1 and δ are determined by Equations 16, 18, 19. As a result, the amplitude
characteristic A (f) is determined by Equations 24 and 14. It is decided.
Further, fb is determined by Equations 13 and 25, and when fb is determined, K is determined by
Equations 17 and 20.
Thus, the final amplitude characteristic is given by K · A (f).

In the second region, the upper limit of f1 is fh, and the lower limit is selected as f1 = fg when the
stopband ripple δ is equal to δs determined in advance. In the second region, the method of
setting the amplitude characteristic A (f) differs depending on whether the stop zone ripple δh
at the upper limit f1 = fh is δh> δs or δh <δs.

In the case of δh> δs, in order to make the ripple δ smaller, a design is made such that the
maximum amplitude A0> 1 by using the relational expression of Equation 12. Specifically, when
fst and f0 are determined from the equations 10-12, 22 and 23, and fst and f0 are determined,
x1 and δ are determined from the equations 16, 18 and 19. As a result, the amplitude
characteristics are determined by the equations 24 and 14. A (f) is determined. Furthermore,
since K is determined in the same manner as in the first region, the final amplitude characteristic
is given by K · A (f).
04-05-2019
27

On the other hand, in the case of δh <δs, the amplitude is determined by the same setting
method as in the first region in order to make the ripple δ larger.

In the third region, the upper limit of f1 is fg, and the lower limit is selected as f1 = fl when the
maximum amplitude A0 in the non-physical region is equal to the predetermined A0max.
In order to set A0> 1 for the maximum amplitude A0, the stop band ripple δ is fixed to a
predetermined δs without using the relational expression of Expression 12. In this case, when f1
is determined, fst-f0 is determined from equations 10, 11 and 21. When fst-f0 is determined, fst,
f0 and x1 are determined by equations 16, 18 and 19. As a result, Characteristic A (f) is
determined. Furthermore, since K is determined in the same manner as in the first region, the
final amplitude characteristic is given by K · A (f).

In the fourth region, the stop band ripple δ is fixed to a predetermined δs without using the
relational expressions of Equations 11 and 12, and the maximum amplitude A0 in the nonphysical region is fixed to a predetermined A0max. In this case, since x1 is determined by
Equation 15 and f0 is determined by Equations 16 and 19, the amplitude characteristic A (f) is
determined by Equations 24 and 14. Furthermore, since K is determined in the same manner as
in the first region, the final amplitude characteristic is given by K · A (f). The upper limit of f1 in
this region is fl, and the lower limit is selected as f1 = fL when the amplitude characteristic to be
set is 0.5 on a straight line of φ = φs. Assuming that x = xc, f = fc, and F = Fc when the amplitude
is 0.5, since the following relational expression 27 holds, FL is determined.


04-05-2019
28
In the fifth region, an amplitude characteristic from the upper limit fL of f1 to the origin is given.
In the present embodiment, since the amplitude at the origin is set to 1, the maximum amplitude
A0 of the non-physical region in this region is given by the following equation 28.


Further, with the stop band ripple δ fixed at a predetermined δs, the final amplitude
characteristic K · A (f) is determined by the same procedure as the fourth region.
The target amplitude characteristic of the two-dimensional FIR filter of the present embodiment
can be determined according to the above-described design procedure of the first to fifth regions.

FIG. 16 is a plan view showing an example of the configuration of a microphone attachment
device 50 for attaching the microphones constituting the directional array microphone 1 of the
present embodiment. 17 is a cross-sectional view showing the configuration of the microphone
attachment 50 seen from the section line XVII-XVII in FIG. 16, and FIG. 18 is a side view showing
the configuration of the microphone attachment 50 shown in FIG. .

The microphone attachment shown in FIG. 16 includes a movable plate 51 for attaching and
fixing a microphone, a case 53 supporting the movable plate 51, and a bolt 52 for fixing the
movable plate 51 to the case 53. . In order to fix the movable plate 51 to the housing 53 with the
bolts 52, projecting portions 51b projecting in one thickness direction are provided at both ends
of the movable plate 51, and screw holes are formed in the projecting portions 51b. On the other
04-05-2019
29
hand, the housing 53 is composed of a bottom surface portion 53b and side surface portions 53a
erected at both ends of the bottom surface portion 53b. Through holes are provided in both side
surface portions 53a of the housing 53, and the movable plate 51 is formed on both side
surfaces of the housing 53 by inserting the bolts 52 into the through holes and screwing the
screw holes of the projecting portion 51b. It is sandwiched between 53a and fixed.

The microphone 10 attached to the movable plate 51 is attached such that the diaphragm faces
one side in the thickness direction of the movable plate 51. In FIG. 16, nine microphones 10 are
provided in a 3 × 3 square lattice, the direction of one diagonal line of which coincides with the
axial direction of the screw hole, and the direction of the other diagonal line is the axial direction
of the screw hole It is arranged to be orthogonal to By loosening the tightening of the bolt 52,
the movable plate 51 can be angularly displaced around the axis of the screw hole, so the
direction of the diaphragm of the microphone 10 can be changed.

As described above, the directional characteristics of the directional array microphone 1 of the
present embodiment are the directional characteristics of the microphone array arranged in the
x-axis direction of the microphone array and the y-axis direction when each microphone 10 is
omnidirectional. The directional characteristics of the microphone array arrayed in FIG.
Therefore, in the microphone array using the microphone attachment device 50 shown in FIG.
16, the central direction of the beam can be made orthogonal to the axial direction of the screw
hole among the diagonal directions of the square grid.

In practice, it may be considered that the directivity characteristic is affected by the influence of
the microphone attachment 50, and that the microphone has directivity. In this case, the
direction of the diaphragm of the microphone can be inclined to the direction of the sound
source to adjust to the optimum directivity characteristic as a directional array microphone. For
example, as shown in FIG. 18, when the surface direction of the bottom surface 53b of the
housing 53 is horizontal, when the omnidirectional microphone 10 is used, the movable plate 51
as shown in FIG. By fixing the surface direction of H horizontally, theoretically, it is possible to
04-05-2019
30
receive the sound most sensitive to the sound waves coming from the horizontal direction. On
the other hand, when the directivity of the microphone 10 needs to be adjusted, the directivity of
the microphone 10 is adjusted by inclining the surface direction of the movable plate 51 from
the horizontal direction as shown in FIG. The sensitivity of the microphone can be improved.

FIG. 19 is a block diagram showing the configuration of a directional array speaker 101
according to an embodiment of the present invention. FIG. 20 is a block diagram showing a
configuration of a two-dimensional FIR filter 120 used as the first and second two-dimensional
FIR filters 112 and 113 in the directional array speaker 101 shown in FIG.

A directional array speaker 101 shown in FIG. 19 includes a speaker array 102 including a
plurality of speakers 110 arranged along both the y-axis direction as the first direction and the xaxis direction as the second direction; It includes a first two-dimensional FIR filter 112 and a
second two-dimensional FIR filter 113 provided for each loudspeaker array 111 along the axial
direction. In the speaker array 102 shown in FIG. 19, along the x-axis direction, NA + 1 (NA is an
even number of 2 or more) loudspeakers 110 are arranged at a predetermined interval DA
mutually, and along the y-axis direction And NB + 1 (NB is an even number of 2 or more)
speakers 110 are arranged at a predetermined interval DB. In FIG. 19, the case of the minimum
configuration of NA = NB = 2 is illustrated. Here, a group of NB + 1 (NB is an even number of 2 or
more) loudspeakers 110 along the y-axis direction is referred to as a loudspeaker array 111.

The difference between the directional array speaker 101 of the present embodiment and the
directional array microphone 1 shown in FIG. 1 described above is that the speaker 110 is
disposed instead of the microphone 10. For this reason, in the present embodiment, the direction
of signal flow is opposite to that of the directional array microphone 1 shown in FIG. For
example, in the case of the directional array speaker shown in FIG. 19, the input signal is filtered
by the second two-dimensional FIR filter 113, and the output signal is input to the first twodimensional FIR filter 112 as many as the speaker string 111. Be done. Subsequently, the signals
filtered by the first two-dimensional FIR filter 112 are subjected to D / A (digital / analog) to the
04-05-2019
31
speakers 110 constituting the speaker string 111 corresponding to the first two-dimensional FIR
filter 112. Converted and input. Each speaker 110 outputs a sound corresponding to the input
sound signal. Also in the two-dimensional FIR filter 120 shown in FIG. 20, the two-dimensional
FIR filter 120 has a parallel configuration of a plurality of one-dimensional FIR filters 122 as in
the case shown in FIG. It is. In the two-dimensional FIR filter 120 shown in FIG. 20, the acoustic
signal input from the input terminal 124 is branched and input to each one-dimensional FIR filter
122, and filtering according to the arrangement position of the speaker 110 to which the
acoustic signal is supplied. Is done.

As described above, although the directional array speaker 101 of the present embodiment is
different from the directional array microphone 1 shown in FIG. 1, the frequency spectrum of the
acoustic signal is represented on a straight line passing through the origin in a two-dimensional
frequency plane. Common points. Therefore, since the design method of the two-dimensional FIR
filter is completely common to that of the directional array microphone 1 shown in FIG. 1, the
directional array speaker 101 of this embodiment is the directional array microphone 1 shown in
FIG. You can get the same effect as

FIG. 21 is a perspective view schematically showing the structure of a microphone attaching
device 60 which is a holding means for attaching the directional array microphone 1 shown in
FIG. 1 as another embodiment of the present invention. The microphone attachment device 60
shown in FIG. 21 has three first arms 61 arranged in the longitudinal direction along the x-axis
direction and three second arms 62 arranged in the longitudinal direction along the y-axis
direction. And the like, and has a configuration in which they are arranged in a lattice. A
connecting member 63 for connecting the first arm and the second arm is provided at a point
where the first arm 61 and the second arm 62 intersect, and the crossing angle of the two arms
at the point where the two arms intersect is It is connected angularly freely. In the xy plane, the
operating lever 64 is attached to the longitudinal end of the first arm 61 disposed at the end
along the longitudinal direction, and the longitudinal end of the second arm adjacent to the first
arm 61 The control lever 64 is attached to the portion along the longitudinal direction thereof.
When the operation lever 64 is opened and closed, the angle formed by the first arm 61 and the
second arm 62 can be changed by the same connection mechanism as that of the magic hand.
The first arm 61 and the second arm connected are supported below a central intersection of the
arms by a support post 65 which projects vertically from a support 66. The microphones 10
04-05-2019
32
used in the directional array microphone 1 shown in FIG. 1 are attached and held at the position
of the connecting member 63 where the first arm 61 and the second arm 62 intersect. Therefore,
the longitudinal direction of the first arm 61 and the second arm 62 coincides with the
arrangement direction of the microphones 10.

FIG. 22 is a plan view for explaining the operation of the microphone attachment device 60
shown in FIG. In the microphone attachment device 60, the angle between the longitudinal
direction (x-axis direction) of the first arm 61 and the longitudinal direction (y-axis direction) of
the second arm 62 can be angularly displaced by the same principle as the magic hand There is.
The angle α1 between the + x direction and the + y direction can be an obtuse angle as shown in
FIG. 22 (a), and the angle α2 between the + x direction and the + y direction is acute as shown in
FIG. 22 (b). You can also As a result, the angle between the + x direction which is the first
direction which is the arrangement direction of the microphones 10 and the + y direction which
is the second direction also changes.

FIG. 23 is an explanatory view showing the relationship between the operation of the
microphone attachment device 60 shown in FIG. 22 and the directivity characteristic of the
directional array microphone 1. FIG. 23 (a) corresponds to the case where the angle formed by
the + x direction and the + y direction in FIG. 22 (a) is an obtuse angle α1, and FIG. 23 (b)
corresponds to the + x direction and the + y direction in FIG. This corresponds to the case where
the angle formed by is an acute angle α2. FIG. 23C shows the case where the angle between the
+ x direction and the + y direction is a right angle α3.

As described above, the characteristic of the directional array microphone 1 is a characteristic in
which the directivity characteristic of the microphones 10 arrayed along the x-axis direction and
the directivity characteristic of the microphones 10 arrayed along the y-axis direction are
superimposed. And has a directional characteristic of the common part of both characteristics. As
shown in FIG. 23C, the directivity characteristic by the arrangement of the microphones 10 along
the x-axis direction assumes that the beam width is approximately 180 ° and the center
04-05-2019
33
direction of the beam is the + x direction, and the microphones along the y-axis direction
Assuming that the directivity characteristic by the arrangement of 10 has a beam width of
approximately 180 ° and the central direction of the beam is the + y direction, the angle
between the + x direction and the + y direction is a right angle α3 as shown in FIG. In some
cases, the common part 72 of the bi-directional characteristic has a beam width of a right angle
α3. On the other hand, when the angle formed by the + x direction and the + y direction is an
obtuse angle α1 as shown in FIG. 23A, a common line between the straight line 70
perpendicular to the x axis and the straight line 71 perpendicular to the y axis Since the portion
72 has the directivity characteristics of the entire directional array microphone 1, the beam
width is 180 ° -α1, and an acute beam width can be obtained. When an angle formed by the + x
direction and the + y direction is an acute angle α2 as shown in FIG. 23B, the common portion
72 sandwiched by the straight line 70 perpendicular to the x axis and the straight line 71
perpendicular to the y axis is Since the directivity characteristic of the directional array
microphone 1 is obtained, the beam width is 180 ° -α2, and an obtuse beam width can be
obtained. As described above, the beam width of the directional array microphone 1 can be
changed by changing the angle formed by the + x direction and the + y direction which are the
arrangement directions of the microphones 10 constituting the directional array microphone 1.

EXAMPLES The present invention will be specifically described below by giving design examples
of the directional array microphone of the present invention.

FIGS. 24 to 33 show a design example of a directional array microphone in which 3 × 3 nine
microphones are specifically arranged according to the design procedure of the directional array
microphone of the present invention described above. .

The design is as follows: beam center φ b = 45 °, side lobe level (side lobe at each microphone
row) δs = -28 dB, spacing between microphone arrays DA = DB = 2.5 cm, number of microphone
arrays NA + 1 = NB + 1 It carried out as = 3, the stop zone end phi S = -10 degrees, rho = 0.6634
(sampling frequency fs = 9023 Hz) of Formula (3), A0max = 79.8, and the time order N1 of the
filter = 100.
Assuming that the size of the microphone is 10 mm in diameter, the array size is about 6 cm per
04-05-2019
34
side, and about 8 cm diagonally of the microphone array.

As a result of design, the beam width defined by 0.5 times the amplitude at the center of the
beam (6 dB reduction) is about 70 °, the side lobe level in the whole directional array
microphone is -20 dB, and the back surface opposite to the beam direction The side lobe level at
(φ = −90 ° to −180 °) was −40 dB, and the bandwidth fL to fH of the directional array
microphone was 220 Hz to 4300 Hz (f1 = 0.024 to 0.475).

FIG. 24 is a diagram showing the amplitude characteristics of the second two-dimensional FIR
filter 13 in the range of φ = −90 ° to 90 ° as a design example of the directional array
microphone 1 shown in FIG.
FIG. 24 represents an amplitude in a two-dimensional frequency plane formed by the time
frequency axis f1 and the spatial frequency f2 in the x-axis direction.
The same display is performed also in the following FIGS. 25-29.

Also, the design is performed in the range of f1 = [− 0.5, 0.5] and f2 = [− 0.5, 0.5], and the
amplitude characteristic in that range is determined. To make the amplitude of the non-physical
region zero. Similarly, in the case of FIGS. 25 to 29 shown below, the amplitude of the nonphysical region is set to 0 for ease of illustration.

FIG. 25 is a diagram showing the amplitude characteristic of the second two-dimensional FIR
filter 13 in the range of φ = 90 ° to 270 ° (−90 °). In FIG. 25, the characteristics at φ = 90
° to 270 ° (−90 °), which are the characteristics of the back surface of the amplitude
04-05-2019
35
characteristics in the range of φ = −90 ° to 90 ° shown in FIG. 24, are displayed. Here, since
the second two-dimensional FIR filter 13 provides directivity corresponding to the arrangement
of the microphones 10 in the x-axis direction, its characteristic is symmetrical with respect to φ
= 90 ° which is the arrangement direction of the microphones. Therefore, the same amplitude
characteristics are obtained in FIGS. 24 and 25.

FIG. 26 is a diagram showing the amplitude characteristics of the first two-dimensional FIR filter
12 in the range of φ = −90 ° to 90 °. FIG. 27 is a diagram showing the amplitude
characteristic of the first two-dimensional FIR filter 12 in the range of φ = 90 ° to 270 ° (−90
°). The first two-dimensional FIR filter 12 provides directivity corresponding to the arrangement
of the microphones 10 in the y-axis direction, and its characteristic is symmetrical with respect to
φ = 0 °, which is the arrangement direction of the microphones 10. Therefore, in FIG. 27
showing the characteristic of φ = 180 ° opposite to the arrangement direction of the
microphones 10 shown in FIG. 26, 220 Hz to 4300 Hz (f1 = 0.024 to 0. 6) which is the band fL
to fH of the directional array microphone. The amplitude at 475) is very small.

FIG. 28 shows the product of the amplitude characteristic shown in FIG. 24 and the amplitude
characteristic shown in FIG. 26. The amplitude characteristic of the entire directional array
microphone 1 is displayed in the range of φ = −90 ° to 90 °. FIG. FIG. 29 shows the product
of the amplitude characteristic shown in FIG. 25 and the amplitude characteristic shown in FIG.
27. The amplitude characteristic of the entire directional array microphone 1 is φ = 90 ° to
270 ° (-90 °). It is the figure displayed in the range of. As apparent from FIGS. 28 and 29, the
entire directional array microphone 1 has an amplitude characteristic having a peak at φ = 45
°, and the amplitude at φ = 225 ° which is the opposite direction becomes very small. There is.

FIG. 30 is a diagram showing the directivity characteristic of the second two-dimensional FIR
filter 13 as a design result of the directional array microphone 1 shown in FIG. 1. FIG. 31 shows
the directivity of the first two-dimensional FIR filter 12. FIG. 32 is a diagram showing
04-05-2019
36
characteristics, and FIG. 32 is a diagram showing directivity characteristics of the directional
array microphone 1 as a whole. In FIG. 30 to FIG. 32, the amplitude (dB) with respect to the angle
φ in the traveling direction of the acoustic beam is displayed polar-coordinated with the time
frequency f1 as a parameter. In the second two-dimensional FIR filter 13, φ = 90 ° is in the
center direction of the beam, in the first two-dimensional FIR filter 12, φ = 0 ° is in the center
direction of the beam, and in the entire directional array microphone 1 φ = 45 ° is toward the
center of the beam. Also, if the beam width is defined at a position 6 dB lower than the maximum
amplitude, the beam width in the second two-dimensional FIR filter 13 is 120 °, and the beam
width in the first two-dimensional FIR filter 12 is 120 ° The beam width of the directional array
microphone 1 as a whole is 70 °.

FIG. 33 is a diagram showing the frequency characteristics of the directional array microphone 1
as a whole. In FIG. 33, with the beam direction φ as a parameter, the amplitude (dB) of the
vertical axis with respect to the time frequency f1 of the horizontal axis is displayed. From FIG.
33, it can be confirmed that the amplitude is −40 dB or less at φ = −90 ° to −180 ° in the
direction opposite to the traveling direction of the acoustic beam.

As a comparative example, when the number of microphones required to obtain the same beam
characteristic and the same time frequency band is calculated using a linear directional array
microphone, for microphone spacing D = 2.5 cm, The number of microphones required was N2 +
1 = 7, and the array size was about 16 cm. Therefore, in the case of the linear type of the prior
art, the array size is more than twice as large as the case of the present invention.

Each embodiment described above is merely an example of the present invention, and the
configuration can be changed within the scope of the present invention. For example, in the
directional array microphone 1 shown in FIG. 1, the number NA of microphones in each row in
the x-axis direction and the distance DA between the microphones, and the number NB of
microphones in each row in the y-axis direction and the distance DB between microphones are
Although all are the same, the effects of the present invention can be obtained even if they are
04-05-2019
37
not all the same.

FIG. 1 is a block diagram showing a configuration of a directional array microphone 1 according
to an embodiment of the present invention. FIG. 3 is a block diagram showing a configuration of
a two-dimensional FIR filter 20 used as the first and second two-dimensional FIR filters 12 and
13 in the directional array microphone 1 shown in FIG. 1. FIG. 7 is a block diagram showing a
configuration of a linear array microphone 15a in which the microphones 10 are linearly
arranged in the y-axis direction. FIG. 7 is a block diagram showing a configuration of a linear
array microphone 15b in which the microphones 10 are linearly arranged in the x-axis direction.
It is an explanatory view for showing a situation of sound reception by a microphone array which
consists of a microphone arranged mutually at equal intervals along a straight line. It is a figure
showing the time change of the acoustic signal observed with each microphone 10_n in the
microphone row 16 shown in FIG. FIG. 7 is a diagram in which a spectrum of an acoustic signal
obtained by performing two-dimensional Fourier transform on a time t and a position coordinate
x of a microphone with respect to a two-dimensional acoustic signal en (t) shown in FIG. It is a
figure for demonstrating the relationship between the spectrum of a two-dimensional acoustic
signal, and the characteristic of a fan filter in a two-dimensional frequency plane. FIG. 5 is a
diagram representing, on a two-dimensional frequency plane, an amplitude characteristic of the
frequency characteristic HB (f1, f3) of the first two-dimensional FIR filter 12 in the linear array
microphone 15a shown in FIG. 3; FIG. 5 is a diagram showing, on a two-dimensional frequency
plane, the amplitude characteristics of the frequency characteristics HA (f1, f2) of the second
two-dimensional FIR filter 13 in the linear array microphone 15b shown in FIG. The frequency
characteristic HB (f1, f3) of the first two-dimensional FIR filter 12 shown in FIG. 9 is converted to
the frequency characteristic HB (f1, f2) for the spatial frequency f2 with respect to the x axis, and
the amplitude characteristic is represented FIG. The product of the frequency characteristic HA
(f1, f2) of the second two-dimensional FIR filter 13 shown in FIG. 10 and the frequency
characteristic HB (f1, f2) of the first two-dimensional FIR filter 12 shown in FIG. FIG. 7 is a
diagram showing the amplitude characteristic of the frequency characteristic Ht (f1, f2) of the
directional array microphone 1; It is a figure for demonstrating the procedure which designs the
target amplitude characteristics Hd (f1, f2) of the 1st and 2nd 2nd 20th FIR filters 12 and 13 of
the directional array microphone 1 shown in FIG. FIG. 10 is a diagram showing a correspondence
relationship between a variable x of a Chebyshev polynomial TN (x) used in Expression 9, a
frequency parameter f used in Expression 14, and a spatial frequency f2.
It is a figure which made the horizontal axis the frequency parameter f by the Chebyshev
polynomial shown in FIG. It is a top view which shows an example of a structure of the
04-05-2019
38
microphone attachment fixture 50 for attaching each microphone which comprises the
directional array microphone 1 of this Embodiment. FIG. 17 is a cross-sectional view showing the
configuration of the microphone attachment 50 seen from the section line XVII-XVII in FIG. 16; It
is a side view which shows the structure of the microphone attachment tool 50 shown in FIG.
FIG. 1 is a block diagram showing a configuration of a directional array speaker 101 according to
an embodiment of the present invention. FIG. 20 is a block diagram showing a configuration of a
two-dimensional FIR filter 120 used as the first and second two-dimensional FIR filters 112 and
113 in the directional array speaker 101 shown in FIG. 19.

FIG. 2 is a perspective view schematically showing a structure of a microphone attaching device
60 which is a holding means for attaching the directional array microphone 1 shown in FIG. FIG.
22 is a plan view for illustrating the operation of the microphone attachment device 60 shown in
FIG. 21. FIG. 23 is an explanatory diagram for illustrating the relationship between the operation
of the microphone attachment device 60 shown in FIG. 22 and the directivity characteristic of the
directional array microphone 1. As a design example of the directional array microphone 1
shown in FIG. 1, the amplitude characteristic of the second two-dimensional FIR filter 13 is
displayed in a range of φ = −90 ° to 90 °. It is the figure which displayed the amplitude
characteristic of the 2nd two-dimensional FIR filter 13 in the range of (phi) = 90 degrees-270
degrees (-90 degrees). FIG. 6 is a diagram showing the amplitude characteristic of the first twodimensional FIR filter 12 in a range of φ = −90 ° to 90 °. FIG. 6 is a diagram showing the
amplitude characteristic of the first two-dimensional FIR filter 12 in a range of φ = 90 ° to 270
° (−90 °). FIG. 27 is a diagram displaying the product of the amplitude characteristic shown in
FIG. 24 and the amplitude characteristic shown in FIG. 26. FIG. 28 is a view displaying the
product of the amplitude characteristic shown in FIG. 25 and the amplitude characteristic shown
in FIG. 27. It is a figure which shows the directional characteristic of 2nd two-dimensional FIR
filter 13 which is a design result of the directional array microphone 1 shown in FIG. FIG. 6 is a
diagram showing the directivity of the first two-dimensional FIR filter 12; It is a figure which
shows the directional characteristic in the directional array microphone 1 whole. It is a figure
which shows the frequency characteristic in the directional array microphone 1 whole.
Explanation of sign

DESCRIPTION OF SYMBOLS 1 directional array microphone 2 microphone array 10 microphone
11 microphone row 12 1st two-dimensional FIR filter 13 2nd two-dimensional FIR filter 33 non-
04-05-2019
39
physical area 34 pass band 50, 60 microphone attachment apparatus 101 directional array
speaker 102 speaker array 110 microphone 111 microphone array 112 first two-dimensional
FIR filter 113 second two-dimensional FIR filter
04-05-2019
40
```
###### Документ
Категория
Без категории
Просмотров
0
Размер файла
61 Кб
Теги
jp2008048294
1/--страниц
Пожаловаться на содержимое документа