JP2002271885A - Microphone system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a microphone system that can optionally control the principal axis of the directivity, enhance the sharpness of the directivity, separate left and right sound sources around microphones and record or recognize sound in real time. SOLUTION: The microphone system is provided with a reference microphone MIC0, a 1st couple of microphones MIC1, 2 placed around the reference microphone, a 2nd couple of microphones MIC3, 4 placed orthogonally to the 1st couple of the microphones around the reference microphone, and a 3rd couple of microphones MIC 5,6 that are placed at an angle of 45 degrees with respect to the 1st couple of the microphones and the 2nd couple of the microphones around the reference microphone on the same plane, and variably controls the principal axis of the directivity and/or the sharpness of the directivity on the basis of the 1st, 2nd and 3rd couples of the microphones and the reference microphone.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of easily changing the direction of directivity in an environment such as a living room in a home or a conference room in an office where the position of a speaker as a target sound source constantly changes. The present invention relates to a microphone device for voice recording and voice recognition that can perform voice recording.

[0002]

2. Description of the Related Art A reference [1] described below expands a three-microphone microphone system described in [2] and uses a three-directional microphone capsule to provide a 300-microphone system.
A wide-band narrow-angle directional microphone with a frequency of 5 Hz to 5 kHz and a beam width of about 120 degrees has been successfully produced. FIG. 15 shows a block diagram of a conventional microphone system. In FIG. 15, microphones MIC0, MIC2, MIC3,
The MICs 1A and 1B are accommodated in a plane area of 4 cm × 7 cm. Then, adders 150 and 151 add 2
Microphones MIC2 and MIC3, MIC1A
And MIC1B. Further, the integrator 15 outputs the difference output of the subtractors 150 and 151 described above.
6, 152 are used to eliminate the phase component and widen the band. In addition, the directivity function to be obtained is obtained first, and the necessary components 152 and 1 are calculated using the Fourier series.
Seeking 53. Further, a low-pass filter (LPF) 155 having a resonance point is used to perform higher-frequency correction.

References [1] Kono, Nakamura, Yamato, Takashima,
<< Broadband Narrow Angle Directional Microphone System >> IEICE Technical Report EA99-85 December 1999 Reference [2] Nakamura, Kouno, Yama
ato, Sakiyama, @Realization
of Wide-Band Directivity
with Three Microphones ",
IEICE Trans, Fundamentals,
vol. E82-A, no. 4, April 1999

[0004]

However, in the above-described conventional microphone system, the main axis of directivity is fixed because hardware such as a transistor and an operational amplifier is used. In particular, the main axis of the beam cannot be arbitrarily controlled. There was an inconvenience.

In addition, there has been a disadvantage that errors in constants such as resistors and capacitors affect the control of directivity.

Accordingly, the present invention has been made in view of the above-mentioned circumstances, and it is possible to arbitrarily control the main axis of directivity and improve the sharpness of directivity.
It is an object of the present invention to provide a microphone device capable of recording and recognizing a voice in real time by using only one microphone device and separating, for example, left and right sound sources centering on the microphone.

[0007]

A microphone device according to the present invention controls a directional characteristic of a microphone by using a microphone to which a sound wave from a sound source is input. The microphone device is arranged around the reference microphone and the reference microphone. A first pair of microphones, a second pair of microphones arranged orthogonal to the first pair of microphones around the reference microphone, and a first pair of microphones around the reference microphone. And the second one
The third is arranged at an angle of 45 degrees with respect to the pair of microphones.
And a reference microphone, a first, a second, and a third pair of microphones are arranged on the same plane, and a reference microphone, a first, a second, and a third pair of microphones are arranged. The main axis and / or the sharpness of the directivity of the directivity characteristic based on the microphone is variably controlled.

Therefore, according to the present invention, the following operations are performed. After each digital signal is obtained by the A / D converter with respect to the outputs of the reference microphone, the first, second, and third pairs of microphones, mixing, phase conversion, and the like of each digital signal are performed by the digital signal processor. The amplitude characteristics are corrected.

The output of each digital signal is subjected to a Fourier transform (DFT) (however,
Hereinafter, fast Fourier transform (FFT) will be taken as an example. ) Is used to perform phase conversion, and the output of each digital signal is corrected for amplitude characteristics using fast Fourier transform (FFT). This facilitates signal processing using the corrected output of the phase conversion.

Further, the characteristics of a desired frequency band are made constant by mixing the digital signals and correcting the amplitude characteristics by the digital signal processing section. This allows
A characteristic without frequency dependence on the incident angle is obtained.

Further, a Fourier series is used for synthesizing an intermediate output obtained by applying a phase characteristic and an amplitude characteristic to the output of the reference microphone, the first, second and third pairs of microphones by the digital signal processing unit. The main axis of directivity and the width of directivity are made variable by weighted addition with the Fourier series coefficient.

Further, the first,
The main axis of directivity is variably set in a plurality of directions using an intermediate output obtained by performing phase correction and amplitude correction on the outputs of the second and third pairs of microphones. As a result, the intermediate generation outputs are weighted and added, and voices from a plurality of spindles are separated and obtained in real time.

With these digital signal processors, the main axis of directivity can be easily controlled arbitrarily, and the sharpness of directivity is further improved.

[0014]

Embodiments of the present invention will be described below. Three to seven microphone devices according to the present embodiment
By combining two microphones and performing digital signal processing on these microphones, the main axis can be easily changed, realizing a wide band and narrow directivity suitable for speech recognition, and separating sounds from multiple main axis directions. Because it can be acquired, it is optimal for voice recognition systems and video conference recording systems.

FIG. 1 is a layout diagram of a microphone capsule to which the present embodiment is applied. In FIG. 1, a reference microphone MIC0 and a first pair of microphones MIC1 and MIC2 arranged around the reference microphone MIC0.
A second pair of microphones MIC3 and MIC4 arranged orthogonally to the first pair of microphones MIC1 and MIC2 around the reference microphone MIC0, and a first pair of microphones around the reference microphone MIC0 MIC1, MIC2 and a second pair of microphones M
Third that is arranged at 45 degrees to IC3 and MIC4
And a pair of microphones MIC5 and MIC6. These microphones MIC0 to MIC6
Are arranged in a plane space. MIC1, MI
The incident angle of the sound wave with respect to the reference axis based on C0 and MIC2 is s degrees.

Here, with respect to the position P of the reference microphone MIC0, a first pair of microphones MIC1, MI
C2 are respectively arranged at equal intervals d1, the second pair of microphones MIC3 and MIC4 are respectively arranged at equal intervals d2, and the third pair of microphones MIC5 and MIC are respectively arranged.
6 are arranged at equal intervals d3.

FIG. 2 is a diagram showing a difference in distance from a sound source according to the arrangement of each microphone. FIG. 2A shows a time difference from the sound source according to the arrangement of the first pair of microphones MIC1 and MIC2. In FIG. 2A, the sound from the sound source incident at an incident angle of s degrees is located at a distance (+ d1) between the position P of the reference microphone MIC0 and the microphone MIC1.
(cos (s)) in a short time,
Distance (+ d1cos (s)) for microphone MIC2
It will arrive in a long time corresponding to the time.

FIG. 2B shows a second pair of microphones MIC.
3, a distance difference from the sound source according to the arrangement of the MIC 4 is shown.
In FIG. 2B, the sound from the sound source incident at an incident angle of s degrees reaches the microphone MIC3 with respect to the position P of the reference microphone MIC0 in a short time corresponding to the distance (+ d2 sin (s)). MIC
No. 3 arrives in a long time corresponding to the distance (+ d2 sin (s)).

FIG. 2C shows a third pair of microphones MIC.
5, a distance difference from the sound source according to the arrangement of the MIC 6 is shown.
In FIG. 2C, the sound from the sound source incident at an incident angle of s degrees is located at a distance (+ d3 sin (45−) to the microphone MIC6 with respect to the position P of the reference microphone MIC0.
s)), and arrives at the microphone MIC5 in a short time corresponding to the distance (+ d3 sin (45−s)).

FIG. 3 is a hardware configuration diagram of a microphone device using the above-described microphone. In FIG. 3, the microphone devices are MIC0 (1), MIC1
(2), MIC2 (3), MIC3 (4), MIC4
(5), MIC5 (6), MIC6 (7) and each MIC
0 (1), MIC1 (2), MIC2 (3), MIC3
(4), MIC4 (5), MIC5 (6), MIC6
An amplifier 8 for amplifying the signal from (7) so that the signal can be processed;
Amplifier 9, Amplifier 10, Amplifier 11, Amplifier 12, Amplifier 13, and Amplifier 14, and Amplifier 8, Amplifier 9, Amplifier 10, Amplifier 11, Amplifier 12, Amplifier 13, and Amplifier 1
A / D for converting the signal amplified in step 4 into a digital signal
Converter 15, A / D converter 16, A / D converter 17, A
/ D converter 18, A / D converter 19, A / D converter 2
0, A / D converter 21, each A / D converter 15, A / D
Converter 16, A / D converter 17, A / D converter 18, A
/ D converter 19, A / D converter 20, A / D converter 21
An arithmetic processing unit 22 that performs signal processing on the digital signal converted by the above, and a recording device or a voice recognition device 23 that performs a recording process or a voice recognition process on the result of the signal processing performed by the arithmetic processing device 22. Is done.

FIG. 4 shows a signal processing performed by the arithmetic processing unit 22 on a signal input to each microphone.
Is shown in the flowchart of FIG. In FIG. 4, step S1
Thus, the initialization of the process for setting i = 0 has already been performed.

Here, the sounds from the sound source separated by R from MIC0 are summarized by the following equations (1), (2), (3).
Equations (4), (5), (6), (7) and (8) are obtained. In the above, Xs (t) in Equation 1 represents a sound source signal, xMIC0 (t) in Equation 2 is a signal observed at time t at the position of MIC0, and xMIC1 in Equation 3
(T) is a signal observed at time t at the position of MIC1, xMIC2 (t) in Expression 4 is a signal observed at time t at the position of MIC2, and xMIC3 in Expression 5
(T) is a signal observed at time t at the position of MIC3, xMIC4 (t) in Expression 6 is a signal observed at time t at the position of MIC4, and xMIC5 in Expression 7
(T) is a signal observed at time t at the position of MIC5, and xMIC6 (t) in Expression 8 is a signal observed at time t at the position of MIC6. Here, k = ω / c, ω represents the angular frequency of the signal, c represents the speed of sound, and θ represents the angle of incidence of the sound source with respect to the reference axis of the microphone.

[0023]

(Equation 1)

[0024]

(Equation 2)

[0025]

(Equation 3)

[0026]

(Equation 4)

[0027]

(Equation 5)

[0028]

(Equation 6)

[0029]

(Equation 7)

[0030]

(Equation 8)

These audio signals are output from the respective MIC0 (1),
MIC1 (2), MIC2 (3), MIC3 (4), M
After being converted into electric signals in the ICs 4 (5), MIC 5 (6), and MIC 6 (7) and amplified by the amplifiers 8, 9, 9, 11, 12, 13, and 14, A / D converter 15, A / D converter 16, A / D converter 17, A / D converter 18, A / D converter
The digital signal is converted by a D converter 19, an A / D converter 20, and an A / D converter 21. However, it is assumed that the sensitivity of each microphone and the gain of each amplifier described above are constant.

This digital signal is supplied to the arithmetic processing unit 22
The following processing is performed in the processing. In FIG. 4, sampling is performed in step S2. Specifically, sampling of a digital signal is performed for each frame period. In step S3, the microphone output is mixed. Specifically, the digital signal obtained from each microphone is expressed by the following equation 9 by the arithmetic processing unit 22:
By performing the mixing processing as shown in Expression 10, Expression 11, and Expression 12, xA (t), xB
(T), xC (t), and xD (t).

[0033]

(Equation 9)

[0034]

(Equation 10)

[0035]

[Equation 11]

[0036]

(Equation 12)

XA (t), xB (t), xC shown in the above-mentioned equations (9), (10), (11) and (12)
The signals of (t) and xD (t) are respectively expressed by the following equations (13), (14), and (15) with respect to the signal at MIC0.
This is a signal as shown in Expression 16 and Expression 16.

[0038]

(Equation 13)

[0039]

[Equation 14]

[0040]

(Equation 15)

[0041]

(Equation 16)

That is, Equations (13), (14) and (15)
XA (t), xB (t), xC shown in Expression 16
The signals (t) and xD (t) are respectively jsin (kd1cosθ) and jsin (kd2s) with respect to the xMIC0 (t) signal of Expression 2 observed at MIC0.
in θ), cos (kd2 sin θ), cos (kd3
It can be seen that the characteristic of sin (π / 4−θ) is added.

That is, Expression 13, Expression 14, Expression 15
XA (t), xB (t), xC shown in Expression 16
The characteristics of the signals (t) and xD (t) vary depending on the angular frequency ω (where k = ω / c) and the incident angle θ of the input signal. In addition, Equation 1 including an imaginary component j
XA (t) shown in Equation 3 and xB shown in Equation 14
(T) is xMIC0 of Equation 2 observed at MIC0.
(T) It can be seen that the phase is advanced by 90 degrees with respect to the signal.

In step S4, the mixed signals are stored in a buffer. Specifically, Equation 13
XA shown in the formula, the formula 14, the formula 15, and the formula 16
Each signal of (t), xB (t), xC (t) and xD (t) is stored in a frame buffer having a buffer number N corresponding to the sample number N used in the frame processing.

In step S5, i indicating the number of times of processing is incremented. In step S6, it is determined whether or not i = N. If i = N is not satisfied in step S6, the process returns to step S2, and the processes and determinations in steps S2 to S6 are repeated.

When i = N in step S6, preprocessing is performed in step S7. Specifically, all the frame buffers with the number of buffers N are represented by Expressions 13 and 14,
XA (t), xB shown in Equations 15 and 16
(T), xC (t), and xD (t) are stored, but when this frame buffer is completely filled, the effect of framing of continuous sound is reduced as preprocessing of frame processing. Window processing such as a humming window or a Hanning window is performed.

In step S8, frame processing is performed. Specifically, each processing of the phase conversion and the correction of the amplitude characteristic is performed using the fast Fourier transform (FFT).

First, the output XA (ω) of the FFT for xA (t) shown in Expression 13 will be described. x A
FIG. 5 shows the incident angle dependence of sin (kd1 cos θ) (here, d1 = 0.008 m), which is the amplitude component of (t). In FIG. 5, sin (kd1cos
θ) is dependent on the angular frequency of the signal (where k = ω / c) (1000 Hz, 2000 Hz, 30 Hz).
00Hz, 4000Hz, 5000Hz, 6000H
It can be seen that it changes according to z).

Therefore, sin (kd1cosθ) / si
FIG. 6 shows the incident angle dependence of n (kd). Now, X
Consider A (ω) / sin (kd1). In FIG. 6, sin (kd1cosθ) / sin (kd
The incident angle dependence characteristic of 1) is that the angular frequency ω of the signal (where k = ω / c) (1000 Hz, 2000 Hz, 30
00Hz, 4000Hz, 5000Hz, 6000H
It can be seen that the variation due to z) almost disappears.

Further, as described above, xA (t) shown in Equation 13 including the imaginary component j has a phase difference of 90 from the xMIC0 (t) signal of Equation 2 observed by MIC0.
X'R as shown in equations (17) and (18)
Assuming that A (ω) and X′IA (ω), there is no phase advance. Here, φ _A (ω) in Expressions 17 and 18
Represents the phase of X _A (ω).

[0051]

[Equation 17]

[0052]

(Equation 18)

Here, X'A (ω) = X'RA (ω) +
jX′IA (ω), and equations (17) and (18) represent the spectrum after phase conversion. Further, if kd1 << 1, sin (kd1cos (θ)) can be approximated to kd1cosθ, so that the following equation (19) is obtained.
Output X A of FFT for x A (t) shown in the equation
From (ω), for the signal input to MIC0, co
A component whose amplitude changes with sθ can be obtained.

[0054]

[Equation 19]

Similarly, considering the signal input to MIC0, the following is obtained. X shown in Equation 14
The output XB (ω) of the FFT for B (t) will be described. sin (kd2) which is an amplitude component of xB (t)
FIG. 7 shows the incident angle dependent characteristic of sin (θ) (here, d1 = 0.008 m). In FIG. 7, the incident angle dependence of sin (kd2 sin (θ)) is represented by the angular frequency ω of the signal (where k = ω / c) (100
0Hz, 2000Hz, 3000Hz, 4000Hz,
5000 Hz and 6000 Hz).

Therefore, sin (kd2sin (θ)) /
FIG. 8 shows the incident angle dependence of sin (kd2). Now, let us consider XB (ω) / sin (kd2). In FIG. 8, sin (kd2sin θ) / sin
The incident angle dependence of (kd2) is represented by the angular frequency ω of the signal.
(However, k = ω / c) (1000 Hz, 2000H
z, 3000Hz, 4000Hz, 5000Hz, 60
00 Hz).

As described above, xB (t) shown in Equation 14 including the imaginary component j has a phase of 90 times that of the xMIC0 (t) signal of Equation 2 observed by MIC0.
X'R as shown in Equation 20 and Equation 21
Assuming that B (ω) and X′IB (ω), there is no phase advance. Here, φ _B (ω) in Expressions 20 and 21
Represents the phase of X _B (ω).

[0058]

(Equation 20)

[0059]

(Equation 21)

Here, X'B (ω) = X'RB (ω) +
jX′IB (ω), and Equations 20 and 21 represent the spectrum after the phase conversion. Therefore, assuming that kd2 << 1, sin (kd2sinθ) can be approximated to kd2sinθ, so that the following equation 22 holds, and the output XB (ω) of the FFT with respect to xB (t) shown in equation 14 is obtained.
Accordingly, a component whose amplitude changes by sin θ with respect to the signal input to MIC0 can be obtained.

[0061]

(Equation 22)

Next, the output XC (ω) of the FFT for xC (t) shown in Expression 15 will be described. x C
Cos (kd2sin (θ)) which is the amplitude component of (t)
Is expressed using the Taylor expansion as in the following Expression 23. Here, λ indicates an approximation error.

[0063]

(Equation 23)

From the above, the relationship of Expression 24 is obtained, and Expression 15 is obtained.
Output x C of FFT for x C (t) shown in equation
From (ω), for the signal input to MIC0, co
A component whose amplitude changes by s2θ can be obtained. Note that λ uses the reference document [1].

[0065]

(Equation 24)

Next, the output XD (ω) of the FFT with respect to xD (t) shown in Expression 16 will be described. xD
Cos (kd3sin (π / 4) which is the amplitude component of (t)
−θ)) is expressed by the following Expression 25 using Taylor expansion. Here, γ indicates an approximation error.

[0067]

(Equation 25)

From the above, the relationship of Expression 26 is obtained, and Expression 16 is obtained.
Output X D of FFT for x D (t) shown in equation
From (ω), for the signal input to MIC0,
A component whose amplitude changes by n2θ can be obtained. Note that γ uses the reference document [1].

[0069]

(Equation 26)

FIG. 9 shows a directional characteristic ψ (θ) to be approximated by a Fourier series. When the directivity ψ (θ) shown in FIG. 9 and the output of MIC0 are added, the directivity D (θ) = 1 + ψ
If (θ) is obtained, sensitivity other than the beam can be suppressed. Here, the central angle of the main axis is θc (degree), and the width of the beam is θw (degree). At this time, ψ (θ) is expressed by the following Equation 27 by Fourier series expansion.

[0071]

[Equation 27]

Actually, in the processing of the above-mentioned equations (13) to (26), cos θ, sin θ, cos 2θ, sin
Since only up to 2θ has been obtained, θw = 60 degrees is a value suitable for suppressing the sensitivity outside the beam. The coefficients α0, αi, βi are given by the following equations (28), (29), and (30).
It is obtained by the formula.

[0073]

[Equation 28]

[0074]

(Equation 29)

[0075]

[Equation 30]

FIG. 10 shows an example of ψ (θ) in the Fourier series when θc = 60 degrees and θw = 60 degrees.

In Equation 27 described above, assuming that M = 2, as shown in Equation 31, weighted addition of the above-described intermediate generation output can provide a characteristic having directivity only in the main axis direction.

[0078]

(Equation 31)

However, in equation 31, each intermediate output Ycos (ω), Y Rcos (ω), YIcos
(Ω) is the following equation 32, equation 33, and equation 34, respectively.
It is expressed by an equation. Further, each intermediate generation output Ysin (ω),
Y Rsin (ω) and Y Isin (ω) are expressed by the following equations 35, 36, and 37, respectively. Here, φA (ω) and φB (ω) are X A
(Ω) and the phase of XB (ω).

[0080]

(Equation 32)

[0081]

[Equation 33]

[0082]

(Equation 34)

[0083]

(Equation 35)

[0084]

[Equation 36]

[0085]

(37)

In equation 31, the intermediate output Y cos (2ω) and Y sin (2ω) are expressed by the following equations 38 and 39, respectively.

[0087]

(38)

[0088]

[Equation 39]

Here, d1 = d2 = d3 = 0.008 m
FIGS. 11 and 12 show the simulation results when. FIG. 11 shows a simulation result of the directional characteristics when θc = 0 degrees, and FIG. 12 shows a simulation result of the directional characteristics when θc = 135 degrees. It can be seen that each shows directivity without frequency dependence. Further, since these directivities are finally determined by the coefficients αi and βi of the Fourier series, if a plurality of pairs of αi and βi are prepared in advance for θc, weighted addition of each intermediate generation signal is performed. With only this, it is possible to separate and acquire sounds from a plurality of spindles in real time.

In the above processing, the reference microphone MIC0 is used.
By using IC1 to MIC4, the reference microphone MIC0 can be substituted. That is, Equation 40 shows the output sum of MIC1 to MIC4.

[0091]

(Equation 40)

Here, when d1 = d2 = 0.008 m, it is the amplitude component in the above equation (40).
s (kd1cosθ) + cos (kd2sinθ)) /
The value of 2 is as shown in the incident angle dependence of the output sum of MIC1 to MIC4 shown in FIG. This gives M
The output sum of the IC1 to MIC4 depends on the value of the incident angle θ in the high frequency range, but the angular frequency ω of the signal (where k = ω / c) (1000 Hz, 2000 Hz, 30 Hz).
00Hz, 4000Hz, 5000Hz, 6000H
It can be seen that z) takes a substantially constant value. Since these values take an average value at θ = 22.5 degrees, the following Expression 41 is obtained.
By performing the correction represented by the equation, it is possible to obtain a characteristic that is substantially independent of the angular frequency ω, and it is possible to perform an approximation.

[0093]

[Equation 41]

Thus, by omitting the reference microphone MIC0 and using six microphones MIC1 to MIC6, the main axis of the directional characteristic is variably controlled, and the directivity is easily directed to the target sound source. be able to.

The Y obtained by the equation 31 obtained in this manner is
For (ω), output processing is performed in step S9 in FIG. Specifically, since the output Y (ω) has been subjected to the frequency analysis, it can be treated as it is as a speech analysis result in the arithmetic processing unit 22 or used as an input for further speech analysis. The speech recognition device 23 can be used for speech analysis for speech recognition. In addition, by performing an inverse Fourier transform on Y (ω), a signal in the frequency domain is converted back to a waveform signal in the time domain, so that the recording device 23 can use the signal for voice recording or the like.

Thereafter, in step S10, after initialization processing is performed with i = 0, the process returns to step S2, and the processing and determination of steps S2 to S6 are repeated.

FIG. 14 shows the omission of the MIC. Hereinafter, the omission of the MICs 5 and 6 shown in FIG.
The omission of the MICs 3 and 4 shown in C will be described.

Equations (42), (43), and (44) shown below
From Equation (45), Equation (46) is obtained.

[0099]

(Equation 42)

[0100]

[Equation 43]

[0101]

[Equation 44]

[0102]

[Equation 45]

[0103]

[Equation 46]

In this way, according to Equation 46, Equation 14 is obtained.
Output X B of FFT for x B (t) shown in the equation
(Ω) is expressed by using XA (ω), thereby obtaining si
An nθ component can be generated.

From the following equations (47) and (48), XD (ω) shown in equation (26) can be expressed as
By expressing using XA (ω), a sin2θ component which is a double angle component can be generated, and XMIC0 (ω) cos2θ shown in Expression 24 is expressed as XA (ω ) Can be used to generate a cos2θ component that is a double-angle component.

As a result, it is not necessary to calculate the sin2θ component by the equation (26) according to the equation (49).
Since the mixing output of 5 (t) and x MIC6 (t) becomes unnecessary, the MICs 5 and 6 can be omitted as shown in FIG. 14A.

As a result, the MICs 5 and 6 are omitted and M
By using five microphones from IC0 to MIC4, the main axis of directional characteristics can be easily variably controlled,
Directivity can be easily directed to a target sound source.

It should be noted that MIC0 is obtained from the above equation (41).
Is unnecessary, so that the MIC0 shown in FIG. 14B can be omitted.

Thus, by omitting the MICs 0, 5, and 6, and using the four microphones MIC1 to MIC4, the main axis of the directivity characteristic can be more easily variably controlled to direct the light to the target sound source. Sex can be easily turned.

[0110] Further, si by the equation 22 is obtained by the equation 46.
The calculation of the nθ component becomes unnecessary, and the equation 2
Since the calculation of the cos2θ component in Equation 4 becomes unnecessary,
Since the mixing output of xMIC3 (t) and xMIC4 (t) becomes unnecessary, MIC3 and MIC4 shown in FIG. 14C can be omitted.

By using three microphones MIC0 to MIC2, omitting the MICs 5, 6, 3, and 4, the main axis of the directional characteristics can be easily and variably controlled to achieve the desired sound source. The directivity can be easily turned to.

[0112]

[Equation 47]

[0113]

[Equation 48]

[0114]

[Equation 49]

[0115]

[Equation 50]

In FIG. 14C, when MIC3 and MIC4 are omitted, MIC0 is newly provided.
Was obtained from the signals of MIC1 to MIC4.
This is because the ICs 3 and 4 are omitted, so that they become necessary.

In the above-described embodiment, the quadratic Fourier series expansion indicating the double-angle component has been described.
The present invention is not limited to this, and may be applied to a Fourier series expansion of third order or higher.

That is, by using Equations (51) and (52), it is possible to generate a cos3θ component which is a triple angle component by expressing using XA (ω) as in Equation (53). Further, as shown in Expression 54, X A
By using (ω), the triple-size component si
An n3θ component can be generated.

As a result, it is possible to generate a component having a triple angle or more, thereby approximating the Fourier series to a triple angle or more, thereby enabling a higher-order Fourier series expansion.

[0120]

(Equation 51)

[0121]

(Equation 52)

[0122]

(Equation 53)

[0123]

(Equation 54)

[0124]

The microphone device according to the present invention is a microphone device for controlling the directional characteristics of a microphone using a microphone to which a sound wave from a sound source is input. A pair of microphones, a second pair of microphones arranged orthogonal to the first pair of microphones around the reference microphone, a first pair of microphones around the reference microphone, and the second pair of microphones. And a third pair of microphones arranged at an angle of 45 degrees with respect to the pair of microphones, and the reference microphone, the first and second microphones.
And the third pair of microphones are arranged on the same plane, and the principal axis and / or the directivity sharpness of the directional characteristics based on the reference microphone, the first, second, and third pairs of microphones are variable. The main axis of directivity can be arbitrarily controlled, the accuracy of directivity can be improved, and only one microphone device is used. There is an effect that voice recording or voice recognition can be performed in real time by separating the left and right sound sources.

In the microphone device of the present invention, the first, second and third pairs of microphones are arranged at equal intervals around the reference microphone, so that the processing is simplified and common. By using seven microphones, the main axis of the directional characteristics can be variably controlled, and the directivity can be easily directed to a target sound source.

In the microphone device of the present invention, the reference microphone is omitted in the above description, and the first and second microphones are omitted.
And a third pair of microphones,
The microphone device can be miniaturized and easily constructed, and the main axis of the directional characteristic is variably controlled by omitting the reference microphone and using six microphones, thereby facilitating directivity to a target sound source. It has the effect that it can be directed to.

In the microphone device of the present invention, the third pair of microphones is omitted and the reference microphone and the first and second pair of microphones are replaced in the above description, so that the microphone device can be downsized. By using five microphones without using the third pair of microphones, the main axis of the directivity can be easily variably controlled, and the directivity of the target sound source can be improved. Can be easily turned.

Further, in the microphone device of the present invention, the third pair of microphones is omitted and the first and second pairs of microphones are replaced in the above description, so that the reference microphone and the third pair of microphones are replaced. By omitting the microphone and using four microphones, there is an effect that the main axis of the directional characteristic can be more easily variably controlled and the directivity can be easily directed to a target sound source.

Further, in the microphone device of the present invention, the second pair of microphones is omitted and the reference microphone and the first pair of microphones are replaced in the above description, so that the second and third pairs of microphones are replaced. By omitting the microphones and using three microphones, the main axis of the directional characteristics can be more easily variably controlled and the directivity can be easily directed to the target sound source.

Further, in the microphone device of the present invention, after the digital signals are obtained by the A / D converter for the outputs of the first, second and third pairs of microphones, The processing unit performs mixing, phase conversion, and correction of the amplitude characteristics of each digital signal, so that the digital signal processing of mixing, phase conversion, and correction of the amplitude characteristics of each digital signal can be performed flexibly without any error, and the directivity Can be arbitrarily controlled, the accuracy of directivity can be improved, and the signal processing output can be used for voice recording or voice recognition.

Further, in the microphone device of the present invention, since the phase conversion and the amplitude characteristic are corrected by using the Fourier transform for the output of each digital signal in the above description, each of the digital signals is obtained by using a higher-order Fourier series expansion. By performing digital signal processing, it is possible to arbitrarily control the main axis of directivity, and it is possible to improve the sharpness of directivity.

In the microphone device of the present invention, the characteristics of a desired frequency band are fixed by mixing, phase conversion, and correction of the amplitude characteristics of each digital signal in the above description. Accordingly, the signal processing can be facilitated, the main axis of directivity can be arbitrarily controlled, and the sharpness of directivity can be improved.

Further, the microphone device according to the present invention, as described above, comprises the reference microphone, the first, second and third microphones.
The main axis of directivity and the sharpness of directivity are varied by the Fourier series coefficient using the Fourier series to synthesize an intermediate output in which the output of each pair of microphones is corrected for phase characteristics and amplitude characteristics. Therefore, it is possible to obtain a characteristic in which directivity is provided only in the main axis direction by weighting and adding the intermediate generation output using the Fourier series coefficient, and thereby the main axis of directivity can be easily controlled arbitrarily. And the sharpness of directivity can be improved.

Further, the microphone device of the present invention is characterized in that the reference microphone, the first, the second and the third
The main axis of directivity is variably set in a plurality of directions by using the intermediate generation output obtained by correcting the phase characteristics and the amplitude characteristics of the outputs of the pair of microphones. Then, by separating and acquiring sounds from a plurality of main axes in real time, the main axes of directivity can be easily controlled arbitrarily, and the effect of improving the sharpness of directivity can be achieved. .

[Brief description of the drawings]

FIG. 1 is a layout diagram of a microphone capsule to which the present embodiment is applied.

FIG. 2A is a diagram illustrating a distance difference from MIC0 according to the arrangement of each microphone, FIG. 2A is a distance difference from MIC0 according to the arrangement of a first pair of microphones MIC1 and MIC2, and FIG. A pair of microphones MIC3
Distance difference from MIC0 according to arrangement of MIC4, FIG. 2C
Is a distance difference from MIC0 according to the arrangement of the third pair of microphones MIC5 and MIC6.

FIG. 3 is a hardware configuration diagram of a microphone device.

FIG. 4 is a flowchart of signal processing in the arithmetic processing device.

FIG. 5 is a diagram showing the incident angle dependence of sin (kdcos θ).

FIG. 6 is a diagram showing the incident angle dependence of sin (kdcosθ) / sin (kd).

FIG. 7 is a diagram showing the incident angle dependence of sin (kdsin θ).

FIG. 8 is a diagram showing the incident angle dependence of sin (kdsinθ) / sin (kd).

FIG. 9 is a diagram showing directional characteristics to be approximated by a Fourier series.

FIG. 10 is a diagram illustrating an example of directivity characteristics in a Fourier series.

FIG. 11 is a diagram illustrating a simulation result of a directional characteristic when θc = 0 degrees.

FIG. 12 is a diagram showing a simulation result of the directivity characteristics when θc = 135 degrees.

FIG. 13 is a diagram showing the incident angle dependence of the output sum of MIC1 to MIC4.

FIG. 14 is a diagram showing omission of an MIC, and FIG.
5 MICs omitted for ICs 5 and 6, FIG. 14B is MIC
Four MICs omitting 0, 5, and 6; FIG. 14C is the MIC
Three MICs with 3, 4, 5, and 6 omitted are shown.

FIG. 15 is a block diagram of a conventional microphone system.

[Explanation of symbols]

1 MIC0, 2 MIC1, 3 MIC2, 4
MIC3, 5 MIC4, 6 MIC5, 7
MIC6, 8 to 14 Amplifier, 15 to 21 A /
D converter, 22 arithmetic processing unit, 23 recording device or voice recognition device

Claims (12)

[Claims] 1. A microphone device for controlling a directional characteristic of a microphone using a microphone to which a sound wave from a sound source is input, comprising: a reference microphone; a first pair of microphones arranged around the reference microphone; A second pair of microphones arranged orthogonally to the first pair of microphones around the reference microphone, a first pair of microphones and the second pair of microphones around the reference microphone 4 for paired microphones
A third pair of microphones arranged at an angle of 5 degrees, wherein the reference microphone and the first, second and third pairs of microphones are arranged on the same plane;
The reference microphone, the first, second and third ones
A microphone device characterized by variably controlling a main axis and / or sharpness of directivity of a directivity characteristic based on each pair of microphones. 2. The microphone device according to claim 1, wherein the first, second, and third pairs of microphones are arranged at equal intervals around the reference microphone. . 3. The microphone device according to claim 2, wherein the reference microphone is omitted and the first and second microphones are omitted.
A microphone device, wherein the microphone device is replaced with a pair of microphones. 4. The microphone device according to claim 2, wherein the third pair of microphones is omitted and replaced with the first and second pairs of microphones. 5. The microphone device according to claim 3, wherein the third pair of microphones is omitted and replaced with the first and second pairs of microphones. 6. The microphone device according to claim 4, wherein the second pair of microphones is omitted, and replaced with the reference microphone and the first pair of microphones. 7. The microphone device according to claim 1, wherein after the digital signals are obtained by an A / D converter with respect to the outputs of the pair of reference microphones, the first, second, and third microphones. A digital signal processing unit for performing mixing, phase conversion, and correction of amplitude characteristics of the digital signals. 8. The microphone device according to claim 7, wherein the output of each of the digital signals is subjected to a phase conversion by using a Fourier transform. 9. The microphone device according to claim 7, wherein the output of each of the digital signals is subjected to Fourier transform to correct an amplitude characteristic. 10. The microphone device according to claim 7, wherein characteristics of a desired frequency band are fixed by mixing, phase conversion, and correction of amplitude characteristics of each digital signal. 11. The microphone device according to claim 7, wherein said reference microphone, said first, second and third ones.
The Fourier series is used to synthesize the intermediate output in which the output of each pair of microphones has been corrected for phase and amplitude characteristics, and the main axis of directivity and the sharpness of directivity are made variable by the Fourier series coefficient. A microphone device characterized by the above-mentioned. 12. The microphone device according to claim 7, wherein an output of the reference microphone, a first, a second, and a third pair of microphones is subjected to correction of a phase characteristic and an amplitude characteristic, and an intermediate generation output is obtained. A microphone device characterized in that a principal axis of directivity is variably set in a plurality of directions by using a microphone.