JPH0543979B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a method for measuring a noise level, and particularly to a method for reducing the influence of noise at a sound receiving point.

[Summary of the invention]

In the noise measurement method of the present invention, the components of the noise acoustic signal can be removed by converting the acoustic signal into a Fourier transform signal and performing arithmetic processing, thereby making it possible to measure noise with even higher accuracy.

[Conventional technology]

Conventionally, as countermeasures against noise generated from noise generation sources, that is, noise acoustic sources, the first method is to reduce noise generation by implementing soundproofing measures on the noise generation source, and the second is to reduce the noise generation by taking soundproof measures to the noise generation source. There is a method to reduce incoming noise by taking soundproofing measures at a sound receiving point, but in either method, it is important to consider how much influence the noise generated from multiple noise sources has on the sound receiving point. It is important to know whether noise is being
The following methods can be considered to achieve this, but each method is still inadequate in practice.

[Problem that the invention seeks to solve]

That is, as a method for measuring the noise generated from multiple noise sources at a sound receiving point, firstly, one possibility is to generate noise one by one from each noise generating source and measure it at the sound receiving point. However, in reality, there are only a very limited number of objects that can be measured using this method, and it is still insufficient in terms of practicality.

Secondly, a unidirectional microphone (so-called gun microphone) is used as a device that can separate and measure the noise coming from multiple noise sources. ), a microphone system having an acoustic reflector, an array microphone, a sound intensity method, a method using a synchronous addition method, etc. can be considered.

However, when a unidirectional microphone or a microphone system having an acoustic reflector is used, the frequency characteristics are included in the measured value as is, resulting in insufficient accuracy of the measured result.

In addition, the method using an array microphone has bidirectional directivity, so
There is a drawback that it cannot be used depending on the background conditions,
Moreover, there is a problem that the influence of frequency characteristics cannot be removed.

In addition, when using the sound intensity method, it has the advantage of having a flat frequency response, but has the disadvantage that the directivity is figure-directional, which makes it unusable depending on the background conditions. There is a problem where you can't sharpen your sexuality too much.

Further, in the method using the synchronous addition method, it is necessary to obtain a signal synchronized with the generation of the noise from the noise source, but there is a problem in that it is practically difficult to obtain the synchronized signal.

The present invention has been made in consideration of the above points, and
By making it possible to effectively remove noise components arriving at the sound receiving point with a relatively simple configuration,
This paper attempts to propose a noise measurement method that can improve noise measurement accuracy.

[Means for solving problems]

In order to solve this problem, in the present invention, the noise acoustic signal generated at the noise acoustic source to be measured by the unidirectional microphone 15 is converted into the first measurement signal S3, and the nondirectional microphone 12 The acoustic signal including the noise acoustic signal is converted into a second measurement signal S1, and the first and second measurement signals S3 and S1 are Fourier-transformed to obtain first and second measurement Fourier-transformed signals S4 and S2 is obtained, the first Fourier transformed signal S4 is conjugated and multiplied by the second measured Fourier transformed signal S2, and the first Fourier transformed signal S4 is divided by the square operation result, and the division operation result is The noise measurement signal S10 is obtained by averaging S8.

[Effect]

The measurement is performed by Fourier transforming the first and second measurement signals S3 and S1 obtained by the unidirectional microphone 15 and the omnidirectional microphone 12, obtaining a cross spectrum, and then obtaining an ensemble average. A noise acoustic signal S10 can be obtained by removing noise acoustic signals other than the noise acoustic signal generated in the intended noise acoustic source.

In this way, a noise measuring device capable of measuring noise with high accuracy can be easily realized with a simple configuration.

〔Example〕

An embodiment of the present invention will be described in detail below with reference to the drawings.

[1] Noise measurement principle In Fig. 1, the noise acoustic Fourier transform signal S(f), which is expressed by Fourier transforming the acoustic signal generated by the noise generation source 1 constituting the noise acoustic source, that is, the noise acoustic signal, is An output Fourier transformed signal Y which arrives at the omnidirectional microphone 3 through the first transmission path 2 and is expressed by Fourier transforming the output signal obtained from the omnidirectional microphone 3.
Obtain (f).

Further, the noise acoustic Fourier transform signal S(f) generated in the noise source 1 is received by the unidirectional microphone 5 through the second transmission path 4, and the output signal of the unidirectional microphone 5 is received. An output Fourier transform signal Z(f) is obtained which is expressed by Fourier transform.

Here, the omnidirectional microphone 3 receives a noise acoustic signal N(f) (represented as a Fourier transform signal) arriving from around the sound receiving point based on its directivity.
In contrast, the unidirectional microphone 5 does not perform acoustoelectric conversion on the noise acoustic signal N(f) due to its directivity, and therefore does not perform acoustoelectric conversion on the noise acoustic signal N(f) as a part of the output signal. The output Fourier transform signal Z(f) of the directional microphone 5 is expressed as , a complex transfer characteristic that combines the transfer characteristic of the transfer path 4 and the acousto-electrical conversion characteristic of the unidirectional microphone 5 is expressed by a formula multiplied by a transfer characteristic Fourier transform coefficient B(f). I can do it.

On the other hand, the output Fourier transform signal Y(f) of the omnidirectional microphone 3 is transmitted from the noise source 1 as shown in the following equation: Y(f)=X(f)+N(f)...(2) A noise acoustic Fourier transform signal X(f) expressed by Fourier transforming the noise acoustic signal that reached the omnidirectional microphone 3 through the path 2, and a noise acoustic Fourier transform signal X(f) expressed by Fourier transforming the noise acoustic signal. It can be expressed by the sum with signal N(f).

Here, the noise acoustic Fourier transformed signal X(f) is transmitted through the transmission path 4 to the noise acoustic Fourier transformed signal S(f) as shown in Transfer characteristic Fourier transform coefficient A expressed by Fourier transforming the overall characteristic that combines the characteristic and the acoustoelectric conversion characteristic of the omnidirectional microphone 3
It can be expressed by the expression multiplied by (f).

Among the above relationships, from equation (1), S(f)=Z(f)/B(f)... (4) As shown in (4), the noise acoustic Fourier transform signal S(f) is Output Fourier transform signal Z
(f) and the transfer characteristic of the second transfer path 4, and by substituting this into equation (3), X(f)=A(f)/B(f)・Z(f)...As shown in (5), the noise acoustic Fourier transformed signal X(f) of the output Fourier transformed signal Y(f) obtained from the omnidirectional microphone 3 is transferred to the first and second transmission paths 2. and 4, and the conversion characteristics of the omnidirectional microphone 3 and the unidirectional microphone 5.

In addition to this, by substituting equation (5) into equation (2), we get the following equation: Y(f)=A(f)/B(f)・Z(f)+N(f)...(6) Then, the output Fourier transform signal Y(f) of the omnidirectional microphone 3 can be obtained.

The two observation signals observed in the omnidirectional microphone 3 and the unidirectional microphone 5 in this way, that is, the output Fourier transform signal Y(f)
and Z(f), but in this noise measurement method, the noise acoustic Fourier transform signal N(f) can be obtained by processing the two observed signals.
Obtain measurement results that do not include.

That is, by finding the cross spectrum of the output Fourier transform signals Y(f) and Z(f) and dividing it by the square of the absolute value of the output Fourier transform signal Z(f), the following formula Z ^* (f)・Y( f)／｜Z(f)｜ ² =Z＊(f)／｜Z(f)｜ ² {A(f
)/B(f)・Z(f) +N(f)}=A(f)/B(f)・Z ^* (f)・Z(f)/｜Z(f)｜
² +N(f)・Z ^* (f)/｜Z(f)｜ ² =A(f)/(f)B+N(f)
・Z ^* (f)/｜Z(f)｜ ² ...(7), the first term expressed by the transfer characteristics of the first and second transfer paths 2 and 4, and the noise acoustic An expression for the sum with the second term represented by the Fourier transform signal N(f) can be obtained.

Therefore, if we calculate the ensemble average for both sides of equation (7), we get E[Z ^* (f)・Y(f)/|Z(f)| ² ]=E[A(f)/B(f)
] +E[N(f)・Z ^* (f)/｜Z(f)｜ ² ] ...(8), but the second term in equation (8) has a correlation. Since it is a signal component expressed by a noisy acoustic Fourier transform signal N(f), E[N(f)・Z ^* (f)/|Z(f)| ² ]=0...(9) As in, the ensemble average value is 0.

Here, the ensemble average calculation calculates the average of signal values over a predetermined period of time.

Therefore, the ensemble average value of both sides of equation (7) is as follows: E[Z ^* (f)・Y(f)/|Z(f)| ² ]=A(f)/B(f)...
As shown in (10), only the terms related to the transfer characteristics of the first and second transfer paths 2 and 4 remain, which means that the coefficient A(f)/B(f) in equation (5) can be obtained. means.

Therefore, by substituting the result of equation (10) into equation (5), the noise acoustic Fourier transform signal X included in the output Fourier transform signal Y(f) obtained from the omnidirectional microphone 3
(f) is calculated as follows: X(f)=K・Z(f) ...(11) As shown in =E[Z ^* (f)・Y(f)/|Z(f)| ² ]...(12) It can be seen that it can be expressed by the multiplication formula.

Here, the coefficient K in equation (12) is equivalently calculated by inputting the output Fourier transform signal Z(f) obtained from the unidirectional microphone 5 and using the omnidirectional microphone 3 as an input.
When considering a transmission system whose output is the output Fourier transformed signal Y(f), it can be said that it represents the transfer function of the transmission system. 2 channel input type FFT
(fast Fourier trasform) means that it can be calculated by an analyzer.

This means that if noise is measured based on the noise measurement principle shown in Figure 1, the noise acoustic Fourier transform signal X(f) expressed by equation (11) can be easily calculated without using special calculation means. Therefore, according to this method, noise can be easily and reliably measured without being affected by noise.

[2] Noise Measuring Device In FIG. 2, reference numeral 11 indicates a noise measuring device as a whole, which has an arithmetic circuit that realizes the noise measurement principle described above with respect to FIG.

That is, the measurement signal S1 obtained from the omnidirectional microphone 12 corresponding to the omnidirectional microphone 3 in FIG. 1 is amplified in the amplifier circuit 13.
By giving it to the FFT calculation circuit 14, the measured Fourier transform signal S2 consisting of the output Fourier transform signal Y(f) expressed by equation (2) from the FFT calculation circuit 14
get.

Furthermore, a measurement signal S3 obtained from a unidirectional microphone 15 corresponding to the unidirectional microphone 5 in FIG. 1 is amplified in an amplifier circuit 16.
The measured Fourier transform signal S4 is supplied to the FFT calculation circuit 17, and thereby the FFT calculation circuit 17 outputs the measured Fourier transform signal S4, which is the output Fourier transform signal S4 (Z(f)) expressed by equation (1).

The measured Fourier transformed signal S4 is sent to the conjugate circuit 18
, thereby representing the conjugate function Z ^* (f) of the output Fourier transform signal Z(f).
5 and feeds it to the multiplier circuit 19 together with the measured Fourier transformed signal S2.

Thus, the output terminal of the multiplication circuit 19 supplies the multiplication output S6, which is represented by the numerator arithmetic expression Z ^* (f)·Y(f) on the left side of equation (7), to the division circuit 20 as the dividend.

On the other hand, the measured Fourier transformed signal S4 is squared in the squaring circuit 21, and thus 2
At the output end of the multiplication circuit 21, the denominator equation on the left side of equation (7) |Z
(f) A square output S7 representing ² is obtained and supplied to the division circuit 20 as a divisor.

As a result, a division output S8 representing the right side of equation (7) is obtained at the output end of the division circuit 20, and this is the output of the averaging circuit 20.
given to 2.

The averaging circuit 22 performs the operation of equation (8) to transmit the signal to its output terminal by removing the noise signal component of equation (7), as represented by the right side of equation (8). The average output S9 having information A(f)/B(f) representing only the characteristics of the route is sent out and is applied to the multiplication circuit 23.
as a multiplication input.

The measured Fourier transform signal S4 is given to the multiplier circuit 23 as a multiplier input, so that the multiplier circuit 2
3, a multiplication output S10 representing the noise acoustic Fourier transform signal X(f) expressed by equation (11) (therefore, equation (3)) is obtained, and this is the measurement output of the noise measuring device 11. Sent as .

In the above configuration, as shown in FIG. 3, there are a plurality of noise generation sources, for example two, as the noise generation sources 1, and noise acoustic Fourier transform signals S ₁ (f) and S ₂ (f) are generated from each noise generation source. A noise acoustic signal represented by is generated, and these noise acoustic Fourier transform signals S ₁ (f) and
The omnidirectional microphone 12 is set at the sound receiving point so that S ₂ (f) directly reaches the omnidirectional microphone 3 and also reaches the omnidirectional microphone 12 by being reflected on the wall 30. 1 to be measured by the unidirectional microphone 15
It is designed to receive a noise acoustic signal, for example, a noise acoustic Fourier transform signal S ₂ (f) generated in two noise generation sources.

Under such measurement conditions, for the noise acoustic Fourier transform signal S ₂ (f) to be measured,
Noise acoustic Fourier transform signals of other noise sources
Arithmetic processing is performed by considering that S ₁ (f) has arrived at the omnidirectional microphone 3 as a noise acoustic Fourier transformed signal N(f).

That is, as described above for equations (2) and (6), the omnidirectional microphone 12 receives the noise acoustic Fourier transform signal X(f) obtained based on the noise acoustic Fourier transform signal S ₂ (f), Noise acoustic Fourier transform signal N comprising noise acoustic Fourier transform signal S ₂ (f)
A measurement signal S1 represented by the sum of (f) can be obtained.

Along with this, from unidirectional microphone 15 (1)
As described above with respect to the equation, a measurement signal S3 consisting of a noise-acoustic Fourier transform signal S ₂ (f) can be obtained.

These two measurement signals S1 and S3 are converted into Fourier transform signals in FFT calculation circuits 14 and 17 (FIG. 2), and then calculation to obtain a cross spectrum is performed in a conjugation circuit 18 and a multiplication circuit 19, and the calculation result is is squared and divided in the squaring circuit 21 and the division circuit 20, thereby performing the above-mentioned calculation regarding equation (7).

This calculation result includes a component of the noisy acoustic Fourier transform signal N(f), but when the averaging circuit 22 calculates the ensemble average, the noisy acoustic Fourier transform signal N(f) is Based on the fact that the component has no correlation with the noise acoustic Fourier transform signal S ₂ (f) to be measured, ensemble averaging is performed so that the noise acoustic Fourier transform signal component is not included in the ensemble averaging result. .

In this way, among the acoustic signals arriving at the omnidirectional microphone 3, the component of the noise acoustic signal (i.e.
A noise measurement output obtained by removing the noise acoustic signal components other than the noise acoustic signal components to be measured can be obtained as the multiplication output S10 of the multiplication circuit 23.

According to the above configuration, when measuring the noise acoustic signal arriving from the noise source at the sound receiving point,
Noisy acoustic signals arriving at the microphone constituting the acoustoelectric conversion means can be reliably removed with a relatively simple configuration, thereby making it possible to obtain noise measurement results with even higher accuracy.

〔Effect of the invention〕

As described above, according to the present invention, the noise acoustic signal component arriving at the microphone at the sound receiving point can be reliably separated from the noise acoustic signal component and measured by arithmetic processing as a Fourier transform signal. With this configuration, highly accurate noise measurement results can be obtained.

[Brief explanation of the drawing]

Fig. 1 is a schematic diagram for explaining the noise measurement principle of the noise measurement method according to the present invention, Fig. 2 is a block diagram showing a noise measurement device constructed by the noise measurement method according to the invention, and Fig. 3 is its operation. It is a schematic diagram used for explanation. 1... Noise source, 2, 4... Transmission path, 3,
12... Omnidirectional microphone, 5, 15... Unidirectional microphone, S(f), S ₁ (f), S ₂ (f)... Noise acoustic Fourier transform signal, Y(f), Z( f)...Output Fourier transform signal, X(f)...Noise acoustic Fourier transform signal, N(f)...Noise acoustic Fourier transform signal, A
(f), B(f)...Transfer characteristic Fourier transform coefficient.

Claims (1)

[Claims] 1. Converting a noise acoustic signal generated at a noise acoustic source to be measured with a unidirectional microphone into a first measurement signal, and converting the noise acoustic signal with an omnidirectional microphone. converting the acoustic signal containing the signal into a second measurement signal, Fourier transforming the first and second measurement signals to obtain first and second measurement Fourier transform signals, and conjugating the first Fourier transform signal. and multiplying the second measured Fourier transform signal by
A noise measurement method comprising: dividing the first Fourier transform signal by a squared calculation result, and averaging the division calculation results to obtain a noise measurement signal.