JPH07240993A - Sound field controller - Google Patents

Abstract

PURPOSE:To prevent howling and coloration in a system provided with an acoustic feedback system. CONSTITUTION:A microphone 12 and a speaker 14 for PA are disposed inside an acoustic space 10 such as a hall or the like. The microphone 12 gathers the sound of a human and a musical instrument, etc. The gathered sound is amplified in a microamiplifier 16, filter characteristics are imparted in a filtering means 18, amplification is performed in a power amplifier 20 and the sound is emitted from the speaker 14. The sound emitted from the speaker is fed back to the microamplifier 16 and the acoustic feedback system 22 is constituted. The filtering means 18 is set to the inverse characteristics of the minimum phase element of the open loop transfer function of the acoustic feedback system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sound field control device for suppressing howling and coloration in an acoustic feedback system in which a microphone and speaker means are arranged in an acoustic space to acoustically form a feedback loop. With respect to the above, it is possible to easily suppress howling and coloration.

[0002]

2. Description of the Related Art When a microphone and a speaker are in the same acoustic space as in a hall or a lecture hall, the sound emitted from the speaker is acoustically fed back to the microphone to form an acoustic feedback loop. However, there is a problem that a sharp peak appears on the frequency axis to cause audible coloration and howling at the peak position. Further, in order to adjust (extend) the reverberation time of a hall or the like, a so-called electroacoustic reverberation support device in which a microphone and a speaker are arranged in an acoustic space to constitute an acoustic feedback system similarly causes coloration and howling. There's a problem.

As a method of suppressing coloration and howling in an acoustic feedback system, conventionally, a closed loop transfer function G (ω) of the acoustic feedback system is measured and its inverse transfer function G is measured.
It has been practiced to find ^-1 (ω) and insert a filter having the characteristic into the loop. However, in general, the response of the closed loop transfer function G (ω) is a very long response, and it has been difficult to obtain the inverse transfer function thereof. Therefore, in practice, the following method (a) or (b) has been adopted.

(A) The peak and dip are smoothed manually while monitoring the amplitude characteristic in a closed loop (normal equalizing).

(B) The peak and dip of the minimum phase element are manually smoothed while simultaneously monitoring the amplitude characteristic and the phase characteristic in a closed loop (parametric equalization).

[0006]

The above methods (a) and (b) require experience and have a problem that adjustment takes a very long time.

The present invention intends to solve the above problems in the prior art and to provide a sound field control device capable of easily suppressing howling and coloration.

[0008]

The invention according to claim 1 is
In an acoustic feedback system in which a microphone and speaker means are disposed in an acoustic space to acoustically form a feedback loop, a filter having substantially the inverse characteristic of the minimum phase element of the open loop transfer function of the feedback loop in the feedback loop. Means are arranged.

According to a second aspect of the present invention, in an acoustic feedback system in which a microphone and speaker means are arranged in an acoustic space to acoustically form a feedback loop, an open loop transfer function of the loop is provided in the feedback loop. Let H _min ⁻¹ (ω) be the inverse characteristic of the minimum phase element, H _ap (ω) be the all-pass element, and A be the constant part of the open-loop transfer function, and be approximately −H _min (ω) ⁻¹ / (1 The filter means has a characteristic of −A · H _ap (ω)).

According to a third aspect of the invention, in an acoustic feedback system in which a microphone and a speaker means are arranged in an acoustic space to acoustically form a feedback loop, the open loop transmission characteristic of the loop is measured. From the characteristic measuring means and the measured open-loop transfer characteristic, the inverse characteristic of the minimum phase element of the open-loop transfer function of the loop is H.
_min ⁻¹ (ω), the all-pass element is H _ap (ω), and the constant part of the open-loop transfer function is A, approximately H _min ⁻¹ (ω) or −H _min (ω) ⁻¹ / (1- A ・ H
calculating means for obtaining an FIR filter coefficient substantially corresponding to the characteristic of _ap (ω)), and FIR filter means arranged in the feedback loop for convolving the obtained FIR filter coefficient with the signal in the loop. And is provided.

[0011]

According to the first aspect of the present invention, by inserting a filter having substantially the inverse characteristic of the minimum phase element of the open loop transfer function into the loop, the unstable element caused by the open loop transfer function can be eliminated. It can be removed from the closed loop.
As a result, the time response characteristic can be shortened, the peak peak factor of the closed loop transfer function can be smoothed (the unevenness of the peak point level can be reduced), the howling margin can be expanded, and howling can be effectively suppressed. can do.

According to the second aspect of the invention, a filter having a characteristic of approximately -H _min (ω) ^-1 / (1-A · H _ap (ω)) is inserted in the loop. The time response can be further shortened, and the peak itself can be eliminated or reduced to a very low level. Therefore, the howling margin is further expanded and the coloration is reduced.

Since the filter characteristics of the filter inserted in the loop in the inventions of claims 1 and 2 can be obtained from the open loop transfer characteristics measured with the acoustic feedback system in the open loop state, It can be calculated more easily than when it is calculated from the closed-loop transfer characteristic.

According to the present invention, the open loop transfer characteristic is measured, the FIR filter coefficient is calculated based on the measurement result, and the actual sound field is generated by the FIR filter means using the calculated FIR filter coefficient. A series of control processes can be performed, and the inventions of claims 1 and 2 can be easily realized.

[0015]

【Example】

(Embodiment 1) An embodiment of the invention described in claim 1 is shown in FIG. A microphone 12 is provided in the acoustic space 10 such as a hall.
And a PA (Public Address) speaker 14 are provided. The microphone 12 picks up sounds of people, musical instruments, and the like. The collected sound is amplified by the microphone amplifier 16, is given a filter characteristic by the filter means 18 (for example, an impulse response corresponding to this filter characteristic is convoluted with the input signal), is amplified by the power amplifier 20, and is a speaker. It is uttered from 14. The sound uttered from the speaker 14 is returned to the microphone amplifier 16 to form an acoustic feedback system 22. The filter means 18 is set to have substantially the inverse characteristic of the minimum phase element of the open loop transfer function of the acoustic feedback system.

Here, the closed-loop transfer characteristics will be compared with and without the filter means 18 in the feedback loop. FIG. 2A shows the filter means 1
2 schematically shows the acoustic feedback system of FIG. 1 when 8 is not included. The same (b) is a block diagram thereof. Here, A: constant part of open-loop transfer function H (ω): frequency function part of open-loop transfer function (in the present specification, this may be referred to as an open-loop transfer function) | A · H ( ω) | <1. According to FIG. 2B, the closed loop transfer function G
(Ω) is expressed as follows.

G (ω) = A · H (ω) + (A · H
(Ω)) ² + (A · H (ω)) ³ + ··· = A · H (ω) / (1-A · H (ω))

FIG. 3 shows simulation results of impulse response characteristics and frequency characteristics of the closed loop transfer function G (ω). Here, a case is shown in which the amplification factor A is changed every 3 dB. According to this, the impulse response does not converge to zero at -11.5 dB or more, a number of peak points occur in the frequency characteristic, and the levels of these peak points are not uniform. Due to the instability of these responses and the occurrence of frequency peaks, the howling margin of this system is narrow, and the utterance level from the speaker 14 cannot be raised so much.

On the other hand, FIG. 4A schematically shows the acoustic feedback system 22 of FIG. 1 in which the filter means 18 is inserted. The same (b) is the block diagram. Here, H _min (ω): minimum phase component H _min of the open-loop transfer function (omega) ^-1: inverse characteristic of _{H min (ω) (H min} (ω) -1 = 1 / H min (ω)) | A · H (ω) | <1. According to FIG. 4B, the closed loop transfer function G
(Ω) is expressed as follows.

G (ω) = A · H (ω) · H _min (ω) ⁻¹
+ (A ・ H (ω) ・ H _min (ω) ^-1 ) ² + (A ・ H
(Ω) · H _min (ω) ⁻¹ ) ³ + …… = A · H (ω) · H _min (ω) ⁻¹ / (1-A · H (ω)
H _min (ω) ⁻¹ ) = A · H _ap (ω) / (1−A · H _ap (ω)) where H _ap (ω) is an all-pass element (all-pass filter). That is, a function whose amplitude is constant regardless of frequency and which changes with frequency by the phase angle

FIG. 5 shows the simulation result of the impulse response characteristic and frequency characteristic of the closed loop transfer function G (ω). Here, a case is shown in which the amplification factor A is changed every 3 dB. According to this, the impulse response is converged to zero even at -2.5 dB, and many peak points are generated, but the levels of the peak points are averaged. That is, the inverse characteristic H _min (ω) ^{−1 of} the minimum phase element
By inserting the filter means 18 having the above, the unstable element due to the open loop transfer function can be removed from the closed loop transfer function, whereby the time response can be shortened and the peak factor of the closed loop transfer function can be smoothed. . Therefore, the howling margin is expanded, and the utterance level from the speaker 14 can be increased.

Next, based on the measurement of the open loop transfer characteristic H (ω), the inverse filter characteristic H of the minimum phase element is obtained.
An example of a procedure for obtaining _min (ω) ^-1 will be described. Open loop transfer function H (ω) and its minimum phase element H
_The following relation holds between _min (ω) and its all-pass element H _ap (ω).

H (ω) = H _min (ω) × H _ap (ω) The arrangement of the poles and zeros of these H (ω), H _min (ω) and H _ap (ω) is shown on the complex plane. It becomes like 6. That is, the minimum phase element H _min (ω) is the open loop transfer function H
The position of the zero point on the real axis i δ in the unstable region of (ω) is moved from the − side to the + side with the imaginary axis ω as the axis of symmetry. In addition, the all-pass element H _ap (ω) has a zero point at the position of the zero point in the unstable region of the open loop transfer function H (ω), and the position of this zero point on the real axis i δ is symmetrical with respect to the imaginary axis ω. The pole is arranged at a position moved from the − side to the + side as an axis.

Further, H (ω), H _min (ω), H
_The following relation holds for _ap (ω).

| H (ω) | = | H _min (ω) | × | H _ap
(Ω) | | H _ap (ω) | = 1 (that is, H _ap (ω) has only phase characteristics) ∴ | H _min (ω) | = | H (ω) | ∴ | H _min (ω) ^{- 1} | = 1 / | H (ω) | Further, the phase characteristic θ (ω) of the open loop transfer function H (ω) is expressed by the following equation.

Θ (ω) = θ _min (ω) + θ _ap (ω) + θ
_lin (ω) where θ _min (ω): Phase characteristic of minimum phase element θ _ap (ω): Phase characteristic of all-pass element θ _lin (ω): Phase characteristic of linear element

FIG. 7 shows an example of a procedure for obtaining the inverse characteristic H _min ⁻¹ (ω) of the minimum phase element of the open loop transfer function H (ω) and setting the filter characteristic of the filter means 18 of FIG. That is, first, (1) open loop transfer characteristic H
(Ω) is measured, and based on this, (2) the phase characteristic θ _min (ω) of the minimum phase element is obtained, and (3) the inverse characteristic H _min (ω) ⁻¹ of the minimum phase element is obtained. Then, (4) the FIR filter coefficient corresponding to the obtained inverse characteristic H _min (ω) ⁻¹ is obtained, and (5) this FIR filter coefficient is set in the filter means 18. The specific contents of each step will be described.

(1) Measurement of open loop transfer characteristic H (ω) With the acoustic feedback system 22 in the open state, an impulse signal is emitted from the speaker 14 and picked up by the microphone 16, and this impulse response is Fourier transformed. The open loop transfer characteristic H (ω) is obtained. This open loop transfer characteristic H
The frequency amplitude characteristic | H (ω) | and the frequency phase characteristic θ (ω) are uniquely determined from (ω).

(2) Phase characteristic θ of the minimum phase element
8, 9 the calculation procedure of the _min (omega) phase characteristic theta _min minimum phase element calculating the (omega). Each step will be described.

Phase characteristic θ (ω) The measured phase characteristic θ (ω) is represented by θ (ω) = θ _min (ω) + θ _ap (ω) + θ _lin (ω) as described above. Unwrapping process The measured phase characteristics θ (ω) are connected together to form a series. Removal of linear phase θ _lin (ω) (wave front delay) The linear phase θ _lin (ω) is approximated as θ _lin (ω) = aω from the gradient (= a) of the points (1, 2) obtained in the above process. Therefore, this linear phase θ _lin (ω) is subtracted from the phase characteristic θ (ω) and removed. Linear phase θ
_{When lin} (ω) is removed, the characteristic of θ _min (ω) + θ _ap (ω) is obtained.

Group delay characteristic calculation θ _min (ω) + θ _ap (ω) = φ (ω)
Is differentiated to calculate the group delay characteristic −dψ (ω) / dω. Operation to eliminate the zero point in the unstable region By taking the absolute value | -dψ (ω) / dω | of the group delay characteristic, the element without any delay (that is, the element with no zero point in the unstable region) produce. This corresponds to the operation of producing H _min (ω) from H (ω) in FIG. That is, by this operation, the element H _min (ω) having no delay causing howling is extracted.

Absolute value of group delay characteristic obtained in the integration step | −dψ (ω)
Integrate / dω |

[0033]

[Equation 1] Ask for. Removal of linear phase Linear phase of integrated value obtained in the process θ _lin (ω) ′
Is approximated as θ _lin (ω) ′ = bω from the gradient (= b) of the point (3, 4), so this linear phase θ _{lin is} calculated from the integrated value.
Subtract (ω) ′. Subtracted value

[0034]

[Equation 2] Is obtained as the phase characteristic θ _lin (ω) of the minimum phase element.

(3) Inverse characteristic H of the minimum phase element
_min (omega) that determine the calculation H _min (omega) ^{-1 -1,} the real part H _min
That is, (ω) ⁻¹ Re and the imaginary part H _min (ω) ⁻¹ Im are obtained. Each of these is obtained as follows.

H _min (ω) ^-1 Re = │H _min (ω) ^-1 │
cos (−θ _min (ω)) = (1 / | H _min (ω) |) cos (−θ _min (ω)) H _min (ω) ⁻¹ Im = | H _min (ω) ⁻¹ | sin ( −θ
_min (ω)) = (1 / | H (ω) |) sin (-θ _min (ω))

(4) Calculation of FIR filter coefficient Inverse characteristic H of the minimum phase characteristic obtained as described above
Inverse Fourier transform is performed on _min (ω) ⁻¹ to obtain an impulse response, which is obtained as an FIR filter coefficient. (5) Filter coefficient setting The obtained FIR filter coefficient is set in the FIR filter as the filter means 18. Then, by closing the loop, the closed loop transfer characteristic of FIG. 5 is obtained.

(Embodiment 2) An embodiment of the invention described in claim 2 is shown in FIG. In the acoustic space 10 such as a hall, a microphone 12 and a PA speaker 14 are arranged. The microphone 12 picks up sounds of people, musical instruments, and the like. The collected sound is amplified by the microphone amplifier 16, is given a filter characteristic by the filter means 24 (for example, an impulse response corresponding to this filter characteristic is convoluted with the input signal), is amplified by the power amplifier 20, and is a speaker. It is uttered from 14. The sound uttered from the speaker 14 is returned to the microphone amplifier 16 to form an acoustic feedback system 26. The filter means 24 is approximately -H _min (ω) ^-1.
The characteristic is set to / (1-A · H _ap (ω)).

FIG. 11A shows the acoustic feedback system 26 of FIG.
Is a schematic representation. The same (b) is the block diagram. According to FIG. 11B, the closed loop transfer function G (ω) is expressed as follows.

G (ω) = A · H (ω) · [-H
_min (ω) ^-1 / (1-A ・ H _ap (ω))] + {A ・ H
(Ω) ・ [-H _min (ω) ^-1 / (1-A ・ H
_ap (ω))] ² + {A · H (ω) · [−H _min (ω)
^-1 / (1-A ・ H _ap (ω))} ³ + …… = A ・ H (ω) ・ [-H _min (ω) ^-1 / (1-A ・ H _ap
(Ω))] / [1-A ・ H (ω) ・ (-H
_min (ω) ^-1 ) / (1-A · H _ap (ω))] = −A · H _ap (ω)

FIG. 12 shows the simulation result of the impulse response characteristic and the frequency characteristic of the closed loop transfer function G (ω). Here, a case is shown in which the amplification factor A is changed every 3 dB. According to this, a flat frequency characteristic in which the time response waveform is extremely short and the peak disappears is obtained. Therefore, the howling margin is maximized and the coloration is eliminated.

(Embodiment 3) An embodiment of the invention described in claim 3 is shown in FIG. In the acoustic space 10 such as a hall, a microphone 12 and a PA speaker 14 are arranged. The sound picked up by the microphone 12 is amplified by the microphone amplifier 16 and then converted into a digital signal by the A / D converter 28. This digital signal is input to the input terminal 1 of the switch means 32 when the closed loop transfer characteristic is measured. In addition, when the filter characteristic is set, the FI as a filter means is used at the time of actual use (performance, lecture, etc.).
It is input to the R filter 30. The FIR filter 30 is
The filter characteristic H stored in the coefficient memory 34
_min (ω) ^-1 or -H _min (ω) ^-1 / (1-A · H _ap
The input signal is convolved with coefficient information of the impulse response corresponding to (ω)), and a filter characteristic is added to the input signal.

The measuring signal generating means 36 generates, for example, an impulse signal as a measuring signal when measuring the open loop transfer characteristic. This measurement signal is input to the input terminal 3 of the switch means 32. The switch means 32 has two output terminals a and b, and guides the signals input to the input terminals 1, 2 and 3 to either of the outputs a and b. Switching control means 38
Switches the switch means 32 based on the instruction operation of the operator. That is, when the measurement of the open loop transfer characteristic is instructed, the input terminal 3 and the output terminal a are connected, and the input terminal 1 is connected to the output terminal b. When the operator performs a measurement start instruction operation in this state, a measurement signal is output from the output terminal a, converted into an analog signal by the D / A converter 40, amplified by the power amplifier 20, and emitted from the speaker 14. The microphone 12 picks up the sound, converts it into a digital signal by the A / D converter 28, and through the switch means 32, the open loop transfer characteristic measuring means 4
Enter 2.

The open loop transfer characteristic measuring means 42 Fourier transforms the collected impulse response h (t) to obtain the open loop transfer characteristic A · H (ω). The obtained open loop transfer characteristic A · H (ω) is input to the calculating means 44.
Here, since the constant part A is known in advance as the gain of the electric system, the frequency amplitude characteristic | H (ω) | and the frequency phase characteristic θ (ω) can be uniquely understood from the transfer characteristic H (ω). The calculation means 44 is the filter characteristic calculation means 4
At 6, the filter characteristic H _min (ω) ⁻¹ or −H is calculated from the open loop transfer characteristic H (ω) by the above-described method.
_min (ω) ⁻¹ / (1−A · H _ap (ω)) is calculated (which is calculated by an operator's arbitrary selection operation). The convolution coefficient calculation means 48 performs inverse Fourier transform on the calculated filter characteristic to obtain a corresponding impulse response, and stores the coefficient value (delay time and gain) of this impulse response in the coefficient memory 34.

When the setting of the filter characteristic is completed as described above, the switch means 32 switches the connection. That is, the input terminal 2 is connected to the output terminal a and the other input / output terminals are opened. As a result, the acoustic feedback system 50 is configured and enters the standby state. Therefore, if a performance, a lecture, etc. are performed using the microphone 12 in this state,
In the FIR filter 30, the FIR filter coefficient is convolved with the sound pickup signal of the microphone 12,
A filter characteristic is added to suppress howling (when the filter characteristic is set to H _min (ω) ⁻¹ ) or suppress howling and coloration (filter characteristic to −H _min (ω) ⁻¹ / (1- (When A · H _ap (ω)) is set), it is possible to perform the loud sound.

The whole or a part of the series of operations from the start of measurement of the open loop transfer characteristic to the setting of the filter characteristic in the coefficient memory 34 can be sequentially programmed and automatically performed.

Further, in each of the above embodiments, the frequency characteristic excluding the peak component is explained as being flattened as a whole. However, in addition to smoothing the peak factor or reducing the peak itself, an arbitrary frequency characteristic is set. It can also be given.

The contents described in the above embodiment are based on the exact solution filter obtained according to the logical expression. Such a filter changes the sound field such as movement of a person in the sound field and temperature change. Very critical to. Therefore, in reality, it is effective to use a filter excluding fine peaks and dips within a range in which the characteristics of the filter are not changed, for stable operation.

For example, it is assumed that the exact amplitude filters H _min (ω) and H _min (ω) ⁻¹ obtained according to the theoretical formula have the frequency amplitude characteristics shown in FIGS. 14 (a) and 14 (b).
This characteristic has fine peaks and dips.
The fine peaks and dips can be removed as follows, for example.

First, H _min (ω) is subjected to inverse Fourier transform to obtain the time response characteristic h of the minimum phase element in FIG. 15 (a).
_{Find min} (t). When the obtained h _min (t) is multiplied by a coefficient that decays with time as shown in FIG. 15 (b) as a time window, the time of the minimum phase element whose time response becomes short as shown in FIG. 15 (c). The response waveform h _min (t) 'is obtained. The frequency amplitude characteristic H _min (ω) ′ of this time response waveform h _min (t) ′ and its inverse characteristic H _min (ω) ′
^-1 is as shown in FIGS. 15 (d) and 15 (e), respectively, and finer peaks and dips are eliminated as compared with FIGS. 14 (a) and 14 (b) (however, FIG. 14 (a), The characteristic feature of (b) remains). Therefore, by using H _min (ω) ′ and H _min (ω) ′ ⁻¹ obtained in this way, stable operation can be achieved against changes in the sound field such as movement of a person in the sound field and temperature changes. You can

Specifically, in the configuration shown in FIG. 13, for example, in the filter characteristic calculating means 46, the minimum phase element H _min (ω) is calculated from the input open loop transfer characteristic H (ω).
Is obtained and the inverse Fourier transform is performed to obtain the time response h _min (t) ′ of the minimum phase element in FIG. 15 (a).
5 (b) is multiplied by a desired time window as shown in FIG.
_Find h _min (t) 'in (c). And h
_{The min} (t) 'is Fourier transformed to obtain H _min (ω)', and the filter characteristic H _min (ω) ' ^-1 is obtained from this.

Further, in the above embodiment, the present invention was used for the purpose of suppressing howling and coloration due to the acoustic feedback system which inevitably occur in the acoustic space. However, the reverberation time of a hall or the like is adjusted (extended). ) For the purpose,
The present invention is similarly applied to a so-called electroacoustic reverberation support device in which a microphone and a speaker are arranged in an acoustic space to actively construct an acoustic feedback system and other systems having various acoustic feedback systems, and coloration and howling are also applied. Can be prevented.

[0053]

As described above, according to the present invention, the open loop transfer function is obtained by inserting the filter having substantially the inverse characteristic of the minimum phase element of the open loop transfer function into the loop. Unstable elements due to can be removed from the closed loop. As a result, the time response waveform of the system can be shortened, unevenness in the level of the peak point of the frequency characteristic can be reduced, the howling margin can be expanded, and howling can be effectively suppressed.

According to the second aspect of the invention, a filter having a characteristic of approximately -H _min (ω) ^-1 / (1-A · H _ap (ω)) is inserted in the loop. , The time response waveform of the system can be shortened, and the peak itself can be eliminated or reduced to a very low level. Therefore, the howling margin can be expanded and
Coloration is also reduced.

Since the filter characteristic of the filter inserted in the loop in the inventions of claims 1 and 2 can be obtained from the open loop transfer characteristic measured with the acoustic feedback system in the open loop state, It can be calculated more easily than when it is calculated from the closed-loop transfer characteristic.

According to the third aspect of the invention, the open loop transfer characteristic is measured, the FIR filter coefficient is calculated based on the measurement result, and the sound field control is performed by the FIR filter means using the obtained FIR filter coefficient. A series of processes can be performed, and the inventions according to claims 1 and 2 can be easily realized.

[Brief description of drawings]

FIG. 1 is a system configuration diagram showing an embodiment of the invention described in claim 1.

FIG. 2 is a filter means 1 in the sound field control device of FIG.
8 is a diagram schematically showing a configuration in which 8 is removed and a block diagram showing its system. FIG.

FIG. 3 is an impulse response waveform and frequency characteristic diagram in a closed loop in the sound field control device of FIG.

FIG. 4 is a diagram schematically showing the configuration of FIG. 1 and a block diagram showing its system.

5 is an impulse response waveform and a frequency characteristic diagram in a closed loop in the sound field control device of FIG.

FIG. 6 H of H (ω) = H _min (ω) × H _ap (ω)
_{(Ω), H min (ω} ), is a diagram showing the arrangement of the poles and zeros of H _ap (omega) in the complex plane.

FIG. 7 is a flowchart showing an example of a procedure from measurement of open loop transfer characteristics to setting of filter characteristics.

FIG. 8 is a flowchart showing an example (first half) of a calculation procedure of the phase characteristic θ _min (ω) of the minimum phase element.

FIG. 9 is a flowchart showing an example (second half) of a procedure for calculating the phase characteristic θ _min (ω) of the minimum phase element.

FIG. 10 is a system configuration diagram showing an embodiment of the invention according to claim 2;

FIG. 11 is a diagram schematically showing the configuration of FIG. 10 and a block diagram showing its system.

FIG. 12 is an impulse response waveform and frequency characteristic diagram during closed loop in the sound field control device of FIG.

FIG. 13 is a system configuration diagram showing an embodiment of the invention according to claim 3;

FIG. 14: Exact solution H obtained according to the theoretical formula
_min(Ω), H_min(Ω) ^-1Is the frequency amplitude characteristic of.

FIG. 15 is a diagram showing an example of an operation for removing peaks and dips from the characteristics of FIG.

[Explanation of symbols]

 10 Acoustic Space 12 Microphone 14 Speaker 18, 24 Filter Means 22, 26, 50 Acoustic Feedback System 30 FIR Filter Means 42 Open Loop Transfer Characteristic Means 44 Computation Means

Claims (3)

[Claims] 1. An acoustic feedback system in which a microphone and a speaker means are disposed in an acoustic space to acoustically form a feedback loop, and in the feedback loop, the minimum phase element of the open loop transfer function of the loop is approximately formed. A sound field control device in which filter means having reverse characteristics is arranged. 2. An acoustic feedback system in which a microphone and speaker means are arranged in an acoustic space to acoustically form a feedback loop, wherein an inverse of a minimum phase element of an open loop transfer function of the loop is provided in the feedback loop. Letting H _min ⁻¹ (ω) be the characteristic, H _ap (ω) be the all-pass element, and A be the constant part of the open-loop transfer function, approximately −H _min (ω) ⁻¹ / (1−A · H _ap ( A sound field control device in which filter means having the characteristics of ω)) is arranged. 3. An open loop transfer characteristic measuring means for measuring an open loop transfer characteristic of an acoustic feedback system in which a microphone and a speaker means are arranged in an acoustic space to acoustically form a feedback loop, and From the measured open-loop transfer characteristic, the inverse characteristic of the minimum phase element of the open-loop transfer function of the loop is H
_min ⁻¹ (ω), the all-pass element is H _ap (ω), and the constant part of the open-loop transfer function is A, approximately H _min ⁻¹ (ω) or −H _min (ω) ⁻¹ / (1- A ・ H
_ap (ω)), a calculating means for obtaining an FIR filter coefficient substantially corresponding to the characteristic, and a FIR filter means arranged in the feedback loop for convolving the signal in the loop with the obtained FIR filter coefficient. A sound field control device comprising: