JP7631741B2 - Noise evaluation device, acoustic characteristic measuring device, noise evaluation method, acoustic characteristic measuring method, and noise evaluation program - Google Patents

Description

The present invention relates to a noise evaluation device, an acoustic characteristic measurement device, a noise evaluation method, an acoustic characteristic measurement method, and a noise evaluation program.

Generally, electrical products are made up of multiple modules, such as electronic components, mechanical components, and housings. When manufacturing such electrical products, the modules are assembled in order, and then functional and visual inspections are carried out to confirm that there are no problems before the assembly is complete.

One of the functional tests is a sound test. This is to check, for example, whether the device's operation sounds and warning sounds are sounding correctly and whether there are any problems with the audio quality. In such sound tests, noise from the surrounding environment may be mixed in and cause erroneous judgments. For this reason, it is common to carry out the test with soundproofing measures in place, such as storing the product in a highly soundproof box known as a soundproof box.

However, the method using the soundproof box mentioned above requires that the product that has been inspected be removed from the soundproof box and the next product to be inspected be stored back in the box after each inspection, which reduces productivity. As a result, efforts are being made to collect sound without using soundproofing measures such as soundproof boxes, and then remove noise through signal processing before carrying out inspections. In this case, technology is required that can distinguish whether the collected sound is from the target object or noise from the surrounding environment and remove the noise.

For example, Patent Document 1 discloses an invention for a voice recognition device that performs voice recognition by removing noise from the surrounding environment. In this voice recognition device, first, noise without voice (sound to be measured) is sampled, and the distribution of the time components of the noise is converted into a distribution of frequency components using FFT (Fast Fourier Transform). Next, the user's voice, which includes noise, is sampled, and the distribution of the time components of the user's voice is converted into a distribution of frequency components using FFT. Then, the frequency components of only the voice can be obtained by subtracting the frequency components of the noise from the frequency components of the voice.

JP 2003-005788 A

The above-mentioned Patent Document 1 is based on the assumption that noise does not change over time. In other words, if the noise is the same when sampling the ambient sound and when sampling the voice, it is possible to accurately extract the frequency components of only the voice. However, if the noise changes between the two sampling periods, there is a problem that the frequency components of the noise that are different from the actual ones are subtracted from the frequency components of the voice. The reason for this is that, even though it is not possible to accurately estimate the noise components when the noise is changing, the noise components sampled under the assumption that they do not change are used.

The present invention has been made in consideration of the above problems, and aims to provide a noise evaluation device that determines whether the frequency spectrum of noise other than the sound to be measured that is included in a collected sound signal can be estimated.

In order to solve the above problems, the noise evaluation device of the present invention has a sound collection signal acquisition means, a measurement target sound on/off control means, and a judgment means. The sound collection signal acquisition means acquires a sound collection signal generated by collecting the measurement target sound output by the measurement target and sound from the surrounding environment. The measurement target sound on/off control means performs on/off control to repeatedly turn on/off the output of the measurement target sound generated by the measurement target. The judgment means sets the sound collection signal when the output of the measurement target sound is on as a first sound collection signal, and sets the sound collection signal when the output of the measurement target sound is off as a second sound collection signal. Then, the judgment means calculates the time change in intensity or frequency spectrum of multiple second sound collection signals over a predetermined period, and judges whether or not the frequency spectrum of noise other than the measurement target sound contained in the sound collection signal can be estimated based on this time change.

The acoustic characteristic measurement device of the present invention has the above-mentioned noise evaluation device and an acoustic characteristic calculation unit. The acoustic characteristic calculation unit calculates the acoustic characteristics of the measurement target sound using the first sound collection signal and the second sound collection signal during a period in which the noise evaluation device determines that the frequency spectrum of noise other than the measurement target sound contained in the sound collection signal can be estimated.

The noise evaluation method of the present invention controls the measurement target sound output by the measurement target to be repeatedly turned on and off, and obtains a collected sound signal generated by collecting the measurement target sound. Based on this control, the collected sound signal when the output of the measurement target sound is on is taken as a first collected sound signal, and the collected sound signal when the output of the measurement target sound is off is taken as a second collected sound signal, and the intensity or frequency spectrum of the second collected sound signal is calculated as a change over time. Then, based on this change over time, it is determined whether or not the frequency spectrum of noise other than the measurement target sound contained in the collected sound signal can be estimated.

The noise evaluation program of the present invention causes a computer to execute a process of acquiring a sound collection signal generated by collecting the measurement target sound output by the measurement target, and a process of controlling the output of the measurement target sound of the measurement target to be repeatedly turned on and off. Based on this control, the computer further executes a process of making the sound collection signal when the output of the measurement target sound is on a first sound collection signal, and a process of making the sound collection signal when the output of the measurement target sound is off a second sound collection signal. The computer then executes a process of calculating the change over time in intensity or frequency spectrum of the second sound collection spectrum, and determining whether or not the frequency spectrum of noise other than the measurement target sound contained in the sound collection signal can be estimated based on this change over time.

The effect of the present invention is to provide a noise evaluation device that determines whether the frequency spectrum of noise other than the sound to be measured that is included in a collected sound signal can be estimated.

1 is a block diagram showing a noise evaluation device according to a first embodiment; 4 is a flowchart showing the operation of the noise evaluation device of the first embodiment. FIG. 11 is a block diagram showing an acoustic characteristic measuring apparatus according to a second embodiment. FIG. 13 is a block diagram showing an acoustic characteristic measuring apparatus according to a third embodiment. 13 is a timing chart showing the relationship between a measurement target sound on/off control signal and a collected sound signal in the third embodiment. 13 is a graph showing an example of a difference calculation of a second sound collection signal according to the third embodiment. 13 is a graph showing another example of the difference calculation of the second sound collection signal according to the third embodiment. 10 is a flowchart showing the operation of an acoustic characteristic measuring apparatus according to a third embodiment. 9 is a flowchart showing details of step S104 in the flowchart of FIG. 8.

Below, an embodiment of the present invention will be described in detail with reference to the drawings. However, the embodiment described below has limitations that are technically preferable for implementing the present invention, but does not limit the scope of the invention to the following. Note that similar components in each drawing are given the same numbers, and descriptions may be omitted.

First Embodiment
1 is a block diagram showing a noise evaluation device 10 according to the first embodiment. The noise evaluation device 10 has a collected sound signal acquisition means 1, a measurement target sound on/off control means 2, and a determination means 3.

The sound collection signal acquisition means 1 acquires a sound collection signal generated by collecting the measurement target sound output by the measurement target and the sound from the surrounding environment. Here, the sound from the surrounding environment is noise.

The measurement target sound on/off control means 2 performs on/off control to repeatedly turn on/off the output of the measurement target sound generated by the measurement target.

The determination means 3 determines the collected sound signal when the output of the sound to be measured is on as the first collected sound signal, and determines the collected sound signal when the output of the sound to be measured is off as the second collected sound signal. Then, the determination means 3 calculates the time change in intensity or frequency spectrum of the multiple second collected sound signals over a predetermined period, and determines whether or not the frequency spectrum of the noise contained in the first collected sound signal can be estimated based on this time change. Details of this process will be described later.

2 is a flowchart showing the operation of the noise evaluation device 10. The noise evaluation device 10 controls the measurement target to repeatedly turn on/off the measurement target sound emitted by the measurement target (S1). The determination means 3 sets the collected sound signal acquired by the collected sound signal acquisition means 1 when the measurement target sound is on as the first collected sound signal, and the collected sound signal acquired by the collected sound signal acquisition means 1 when the measurement target sound is off as the second collected sound signal. Then, a number corresponding to the number of times of on/off switching is assigned (S2). For example, the second off time can be the second collected sound signal (2), the fifth off time can be the first collected sound signal (5), and so on. Next, the difference between the second collected sound signal at a certain off time of on/off switching of the measurement target sound output and the second collected sound signal at a certain off time after the on/off time is calculated (S3). The difference can be calculated, for example, using the time change in signal strength within the same time width. This signal strength can be calculated, for example, by the root mean square (effective value). If the difference in the root mean square is equal to or less than a predetermined threshold, it can be determined that the noise level has not changed. Alternatively, in calculating the difference, the time change in the second collected signal strength may be Fourier transformed for each time, converted into a frequency spectrum, and calculated as the difference in sound pressure at each frequency. In this case, for example, first, the frequency spectrum of the second collected signal at a certain time is calculated, and similarly, the frequency spectrum of the second collected signal a predetermined number of times after the certain time is calculated. Next, subtraction is performed for each component to calculate the difference spectrum. Then, for example, the absolute value of each frequency component of the difference spectrum is calculated, and if there is no component equal to or greater than a predetermined threshold in this absolute value, it can be determined that the second collected signal (noise) has not changed, and if there is a component equal to or greater than the threshold, it can be determined that the second collected signal (noise) has changed. In addition, in calculating the difference spectrum, for example, a threshold may be set in the frequency spectrum of the second collected signal, and the difference may be calculated only for frequency components equal to or greater than the threshold. Next, it is determined whether the difference calculated as above, for example, the maximum value of the components of the difference spectrum, is less than a predetermined threshold (S4). If the difference is less than the threshold (S4_Yes), it is determined that the frequency spectrum of the noise during the period in which the difference was calculated can be estimated (S5). This is because it can be determined that the noise has not changed during the period. If the noise has not changed, it is considered that the noise components contained in the first collected signal are the same as those of the second collected signal in the time adjacent to the first collected signal of interest. Therefore, the collected spectrum of the noise contained in the first collected signal can be estimated by calculating the frequency spectrum of the second collected signal in that time. On the other hand, if the difference is equal to or greater than the threshold (S4_No), it is determined that the frequency spectrum of the noise during the period in which the difference was calculated cannot be estimated (S6). This is because it is determined that the noise during that period has changed and the frequency spectrum of the noise contained in the first collected signal cannot be accurately estimated.

By performing the above operations, the noise evaluation device of this embodiment can determine whether or not the frequency spectrum of noise other than the measurement target sound contained in the collected signal can be estimated.

Second Embodiment
3 is a block diagram showing an acoustic characteristic measurement device 20 that uses the noise evaluation device 10 of the first embodiment. The acoustic characteristic measurement device 20 includes an acoustic characteristic calculation means 4 in addition to the noise evaluation device 10. The acoustic characteristic calculation means 4 receives the first and second sound collection signals and a judgment result from the judgment means 3. This judgment result indicates a period during which the frequency spectrum of the noise can be estimated and a period during which the frequency spectrum of the noise cannot be estimated.

In the above configuration, the acoustic characteristic calculation means 4 performs a Fourier transform on the first sound collection signal and the second sound collection signal, respectively, to calculate the first sound collection spectrum and the second sound collection spectrum. Then, during a period in which the determination means 3 determines that the frequency spectrum of the noise can be estimated, the noise is removed by subtracting the second sound collection spectrum of an adjacent time from the first sound collection spectrum of a certain time, and the spectrum of the sound to be measured is extracted. In this way, the acoustic characteristic of the sound to be measured can be obtained as the spectrum of the sound to be measured.

As described above, the acoustic characteristic measurement device of this embodiment calculates acoustic characteristics only during the period in which the frequency spectrum of the noise can be estimated, making it possible to perform highly accurate acoustic characteristic measurements.

Third Embodiment
In this embodiment, a specific configuration example of the acoustic characteristic measurement device of the second embodiment will be described. Fig. 4 is a block diagram showing an acoustic characteristic measurement device 100 of the third embodiment. The acoustic characteristic measurement device 100 has a measurement target sound on/off control unit 110, a collected sound signal acquisition unit 120, a collected sound signal division unit 130, an FFT processing unit 140, a noise spectrum estimation feasibility determination unit 150, and an acoustic characteristic calculation unit 160. Here, FFT is an abbreviation for Fast Fourier Transform. For example, a computer equipped with a processor, memory, and storage can be used as the hardware of the acoustic characteristic measurement device 100.

The measurement target sound on/off control unit 110 outputs a command to the measurement target 300 to output sound (on) or stop sound (off), and controls the output of the measurement target sound of the measurement target 300 to be turned on/off at a predetermined cycle. The measurement target sound on/off control unit 110 also outputs a control signal to the collected sound signal division unit 130 and the FFT processing unit 140.

The sound collection signal acquisition unit 120 acquires the sound collection signal generated by the sound collection microphone 200, which collects the sound generated from the measurement target and the surrounding environmental sound and converts the sound into an electrical signal. The sound collection microphone 200 collects noise coming from the surrounding environment in addition to the measurement target sound output from the measurement target, so the sound collection signal also contains noise.

The sound collection signal division unit 130 divides the received sound collection signal into a first sound collection signal when the output of the sound to be measured is on, and a second sound collection signal when it is off. The sound collection signal division unit 130 also counts the number of times the sound to be measured is switched on/off, and assigns numbers to the first sound collection signal and the second sound collection signal according to the number of times it is switched on/off. The first sound collection signal and the second sound collection signal are then output to the FFT processing unit 140.

The FFT processing unit 140 performs FFT processing on the first collected sound signal and the second collected sound signal for each on period and off period.

The noise spectrum estimation feasibility determination unit 150 calculates the difference between the second sound collection signal at a certain time and the second sound collection signal at a time a predetermined number of times after that time, and determines whether or not the frequency spectrum of the noise can be estimated for the period for which the difference was calculated, based on this difference. Details will be described later.

The acoustic characteristic calculation unit 160 calculates acoustic characteristics using the first sound collection signal and the second sound collection signal for the period in which the noise spectrum estimation feasibility determination unit 150 determined that the noise frequency spectrum can be estimated. The calculation of acoustic characteristics will also be described later.

Next, using a specific example, the estimation of the noise frequency spectrum and the calculation of the acoustic characteristics will be described. First, the collected sound signal will be described. FIG. 5 is a timing chart showing the signal strength of the collected sound signal acquired by the collected sound signal acquisition unit 120 and an example of the output of the on/off control signal output by the measurement target sound on/off control unit 110. When the on/off control signal is sent from the measurement target sound on/off control unit 110 to the measurement target 300, the measurement target 300 turns the measurement target sound output on/off accordingly. Note that the number of on/off switching is counted by the collected sound signal division unit 130, for example, and as shown in FIG. 5, each on period and off period is assigned a number such as the i-th on, the i-th off, the i+1-th on, ....

Next, the estimation of the frequency spectrum of noise will be described. First, FFT processing is performed for each on period and off period, and the time change of the sound collection signal is converted into a frequency spectrum. Here, the sound pressure of the frequency f ₁ in the second sound collection spectrum obtained by FFT processing the second sound collection signal when the sound to be measured is off is P _OFF (sound collection order). Also, the sound pressure of the frequency f ₁ in the first sound collection spectrum obtained by FFT processing the first sound collection signal when the sound to be measured is on is represented by P _ON (sound collection order). For example, in the case of off, if it is the i-1th sound collection, it is P _OFF (i-1), and if it is the ith sound collection, it is P _OFF (i). Whether or not the frequency spectrum of noise can be estimated is determined by whether the absolute value of the difference between P _OFF at a certain time and P _OFF after a predetermined time is less than the threshold value P _th .

FIG. 6 is a graph showing the sound pressure at the frequency f ₁ of the i-1th off and the i-th off. In the case of FIG. 6, the difference ΔP _OFF (i)=P _OFF (i)-P _OFF (i-1)<P _th . Therefore, the change in noise between the i-1th off and the i-th off is small and can be considered constant. Therefore, the component of the frequency f ₁ of the measurement target sound can be obtained by subtracting P _{OFF (i) or P OFF (} i-1), or the average value of these, from the sound pressure P _ON (i) of the frequency f ₁ of the first sound collection spectrum at the i-th on. By performing this component acquisition in a predetermined frequency range, the sound pressure spectrum of the measurement target sound, that is, the acoustic characteristics, can be obtained.

FIG. 7 is a graph showing the sound pressure at frequency f ₁ of the i-1th off and the ith off when the difference in sound pressure during off is large. In this case, ΔP _OFF (i) = P _OFF (i) - P _OFF (i-1) ≧ P _th . In such a case, since the noise component changes significantly between the two off periods, it is difficult to accurately estimate the noise component in the on period between the two. For this reason, for example, if the average value of P _OFF (i) and P _OFF (i-1) is subtracted as the noise component to calculate the measurement target sound component, it is considered that the error will be large. Therefore, in this embodiment, the sound collection spectrum during the period when the absolute value of the difference is equal to or greater than the threshold value is discarded without being used for acoustic characteristic measurement.

As described above, acoustic measurements are performed after removing noise using only the collected sound signal from a period in which the noise components can be accurately estimated. This embodiment therefore makes it possible to avoid erroneous noise removal and perform highly accurate acoustic measurements.

6 and 7, the noise removal possibility judgment was performed using the sound collection signal when the i-th and the previous (i-1)th off-time, but the judgment may be performed by calculating the difference in another predetermined period, such as the second or third previous period. The judgment may be performed at all frequencies, but a certain threshold _C1 may be set and the judgment may be performed only at frequencies where the sound pressure is _C1 or higher. The sound pressure change may be judged not by the difference but by the ratio _POFF (i)/ _POFF (i-1).

Figure 8 is a flowchart showing the operation of the acoustic characteristic measuring device 100 described above. First, the collected sound signal is divided into a first collected sound signal when the output of the sound to be measured is on, and a second collected sound signal when it is off (S101). Next, FFT processing is performed for each on period and off period to convert them into a frequency distribution (S102).

Next, loop processing L from S103 to S106 is performed. In the loop processing, it is determined whether it is possible to estimate the frequency spectrum of the noise from P _ON (i) at the i-th ON (S104). If it is determined that it is possible to estimate the frequency spectrum of the noise, the sound pressure P(i) at the frequency f of the sound to be measured is calculated by P(i)=P _ON (i)-P _OFF (i) (S105). On the other hand, if it is determined that it is impossible to estimate the frequency spectrum of the noise, the collected sound signal for that period is discarded. In this loop processing, the processing of S104 and S105 is performed N times, which is set as a predetermined range, and is repeated until it is completed for a predetermined frequency range (L). In this way, the influence of noise fluctuations is reduced and the acoustic characteristics can be calculated.

Fig. 9 is a flowchart showing the details of the noise removal judgment S104 in the flowchart of Fig. 8. First, the difference ΔP _OFF (i) between the sound pressure at the i-th off time and the off time a predetermined number of times before is calculated by the formula ΔP _OFF (i) = P _OFF (i) - P _OFF (i - a) (S1041). Next, it is determined whether the absolute value of the difference ΔP _OFF (i) is less than a threshold value (S1042). If the absolute value of ΔP _OFF (i) is less than the threshold value (S1042_Yes), it is determined that the frequency spectrum of the noise in the i-th first collected sound signal can be estimated (S1043), and P _ON (i) is held for acoustic characteristic calculation (S1044).

On the other hand, if the absolute value of the difference ΔP _OFF (i) is equal to or greater than the threshold value P _th in S1042 (S1042_No), it is determined that it is impossible to estimate the frequency spectrum of the noise in the i-th first collected signal (S1045), and P _ON (i) is excluded from the calculation of acoustic characteristics (S1046).

As described above, this embodiment makes it possible to avoid erroneous noise removal, thereby enabling acoustic characteristic measurements to be performed that are less susceptible to the effects of noise fluctuations.

The scope of the present invention also includes a program that causes a computer to execute the processes of the first to third embodiments described above, and a recording medium on which the program is stored. As the recording medium, for example, a magnetic disk, a magnetic tape, an optical disk, a magneto-optical disk, a semiconductor memory, etc. can be used.

The present invention has been described above using the above-mentioned embodiment as an exemplary example. However, the present invention is not limited to the above-mentioned embodiment. In other words, the present invention can be applied in various aspects that can be understood by a person skilled in the art within the scope of the present invention.

REFERENCE SIGNS LIST 1 Sound collection signal acquisition means 2 Measurement target sound on/off control means 3 Determination means 4 Acoustic characteristic calculation means 10 Noise evaluation device 20, 100 Acoustic characteristic measurement device 110 Measurement target sound on/off control section 120 Sound collection signal acquisition section 130 Sound collection signal division section 140 FFT processing section 150 Noise spectrum estimation feasibility determination section 160 Acoustic characteristic calculation section 200 Sound collection microphone 300 Measurement target

Claims (9)

A sound collection signal acquisition means for acquiring a sound collection signal generated by collecting a measurement target sound outputted from a measurement target;
a measurement target sound on/off control means for repeatedly turning on/off the output of the measurement target sound of the measurement target;
a determination means for determining, based on the control, a sound collection signal when the output of the sound to be measured is on as a first sound collection signal, a sound collection signal when the output of the sound to be measured is off as a second sound collection signal, calculating a time change in intensity or frequency spectrum of the second sound collection signal during a certain period, determining that the frequency spectrum of the noise contained in the first sound collection signal during that period can be estimated if the time change is less than a preset threshold value, and determining that the frequency spectrum of the noise contained in the first sound collection signal during that period cannot be estimated if the time change is equal to or greater than the threshold value ;
A noise evaluation device comprising: The determination means:
An FFT processing unit that performs FFT processing on the second collected sound signal during each off period,
2. The noise evaluation device according to claim 1, wherein the time change of the second sound collection signal is calculated using a second sound collection spectrum obtained by performing FFT processing on the second sound collection signal. The noise evaluation device according to claim 1 or 2;
an acoustic characteristic calculation unit that calculates acoustic characteristics of the measurement target sound by using the first sound collection signal and the second sound collection signal during a period during which the noise evaluation device has determined that the frequency spectrum of the noise can be estimated;
1. An acoustic characteristic measuring device comprising: The acoustic characteristic calculation unit,
Based on a first sound collection spectrum obtained by FFT processing the first sound collection signal for a certain period and a second sound collection spectrum obtained by FFT processing a second sound collection signal for a period adjacent to the certain period,
4. The acoustic characteristic measuring device according to claim 3, further comprising: a frequency component calculating unit for calculating the acoustic characteristic. Acquire a collected sound signal generated by collecting the measurement target sound output by the measurement target,
performing control to repeatedly turn on/off the output of the measurement target sound of the measurement target;
Based on the control, the sound collection signal when the output of the sound to be measured is on is set as a first sound collection signal, and the sound collection signal when the output of the sound to be measured is off is set as a second sound collection signal,
a time change in the intensity or frequency spectrum of the second sound collection signal during a certain period is calculated, and if the time change is less than a preset threshold value, it is determined that the frequency spectrum of the noise contained in the first sound collection signal during that period can be estimated, and if the time change is equal to or greater than the threshold value, it is determined that the frequency spectrum of the noise contained in the first sound collection signal during that period cannot be estimated;
A noise evaluation method comprising: The second collected sound signal is subjected to FFT processing during each off period;
The noise evaluation method according to claim 5 , further comprising the step of calculating the time change of the second sound collection signal using a second sound collection spectrum obtained by subjecting the second sound collection signal to the FFT processing. The noise is evaluated by the noise evaluation method according to claim 5 or 6,
an acoustic characteristic measuring method comprising: calculating an acoustic characteristic of the measurement target sound using the first sound collection signal and the second sound collection signal during a period in which it is determined that a frequency spectrum of the noise can be estimated. The first sound collection signal for a certain period is subjected to FFT processing to calculate a first sound collection spectrum;
A second sound collection spectrum is calculated by performing FFT processing on a second sound collection signal in a period adjacent to the certain period;
8. The acoustic characteristic measuring method according to claim 7, further comprising the step of calculating a frequency component of the acoustic characteristic based on the first sound collection spectrum and the second sound collection spectrum. A process of acquiring a collected sound signal generated by collecting a measurement target sound output from a measurement target;
A process of periodically turning on/off the output of the measurement target sound of the measurement target;
A process of setting the sound collection signal when the output of the sound to be measured is on as a first sound collection signal and setting the sound collection signal when the output of the sound to be measured is off as a second sound collection signal based on the control;
A process of calculating a time change in the intensity or frequency spectrum of the second sound collection signal during a certain period , determining that the frequency spectrum of the noise contained in the first sound collection signal during the period can be estimated if the time change is less than a preset threshold value, and determining that the frequency spectrum of the noise contained in the first sound collection signal during the period cannot be estimated if the time change is equal to or greater than the threshold value;
A noise evaluation program that causes a computer to execute the above steps.

Images