Sound quality assessment device, sound quality assessment method, and sound quality assessment program

Abstract

To provide a sound quality determination device, a sound quality determination method, and a sound quality determination program that can identify the sound quality of a sound generated from a test object and classify the sound quality according to abnormality in the test object.SOLUTION: A sound quality determination device, a sound quality determination method, and a sound quality determination program generate frequency data indicating the intensity of each of frequency components of a first evaluation sound based on sound data obtained by recording the first evaluation sound generated from a test object, extract a first feature quantity of the first evaluation sound from the frequency data, and based on the first feature quantity and a second feature quantity of a second evaluation sound with determined sound quality, determine that the sound quality of the first evaluation sound is the same as the sound quality of the second evaluation sound associated with the first evaluation sound.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a sound quality assessment device, a sound quality assessment method, and a sound quality assessment program. [Background technology]

[0002] Patent Document 1 discloses a technology for performing order analysis on a sound (evaluation sound) generated from a vehicle under inspection as an evaluation target, and quantitatively evaluating the consonance of the evaluation sound based on the slope of a straight line relative to the sound pressure levels of the order components that are the fundamental tone of the order-analyzed evaluation sound and its harmonic order components. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2008-298474 A Summary of the Invention [Problem to be solved by the invention]

[0004] The technology described in Patent Document 1 evaluates an evaluation sound based on the slope of a straight line relative to the sound pressure level of the order components that are the fundamental tone of the evaluation sound and its harmonic order components. Therefore, depending on the evaluation sound, it may be difficult to identify the sound quality of the evaluation sound or to classify the sound quality according to the abnormality of the test subject.

[0005] The present invention has been made in consideration of the above problems, and an object of the present invention is to provide a sound quality assessment device, a sound quality assessment method, and a sound quality assessment program that are capable of identifying the sound quality of a sound generated from an object to be tested and classifying the sound quality according to an abnormality in the object to be tested. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, a sound quality judgment device, a sound quality judgment method, and a sound quality judgment program according to an embodiment of the present invention generate frequency data indicating the intensity of each frequency component of a first evaluation sound based on sound data recording a first evaluation sound generated from a test object. Then, a first feature amount of the first evaluation sound is extracted from the frequency data, and based on the first feature amount and the second feature amount of a second evaluation sound whose sound quality has been determined, the sound quality of the first evaluation sound is judged to be the same as the sound quality of a second evaluation sound associated with the first evaluation sound. Effect of the Invention

[0007] According to the present invention, the quality of the sound generated from the test subject can be identified and the quality of the sound can be classified according to an abnormality in the test subject. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram showing a configuration of a sound quality assessment device according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a flowchart showing the process of the sound quality assessment device according to one embodiment of the present invention. [Diagram 3] FIG. 3 is a diagram illustrating an example of frequency data. [Figure 4] FIG. 4 is a diagram showing an example of fundamental frequencies and harmonic frequencies in frequency data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Next, an embodiment of the present invention will be described in detail with reference to the drawings. In the description, the same parts are designated by the same reference numerals and duplicated explanations will be omitted.

[0010] [Configuration of sound quality assessment device] An example of the configuration of a sound quality assessment device 1 according to this embodiment will be described with reference to Fig. 1. As shown in Fig. 1, the sound quality assessment device 1 includes a sound data receiving unit 21, a storage unit 25, and a controller 100. In addition, the sound quality assessment device 1 may include a measurement unit 11 and a presentation unit 400. The measurement unit 11 is connected to the sound data receiving unit 21, and the sound data receiving unit 21, the storage unit 25, and the presentation unit 400 are connected to the controller 100.

[0011] The measurement unit 11 acquires a sound (first evaluation sound) generated from the test object, and generates sound data that records the sound generated from the test object. For example, the measurement unit 11 is a microphone. The directivity of the microphone and the frequency characteristics of the acquired sound are selected according to the test object.

[0012] The inspection object may be, for example, a horn, an engine, a motor, etc. mounted on an automobile, etc. In addition, the inspection object may be any part that produces sound when in operation, and is not limited to the examples given here.

[0013] Alternatively, the measurement unit 11 may generate sound data for sounds having frequencies within the audible range of the human ear based on the characteristics of human hearing. For example, the measurement unit 11 may generate sound data for only sounds within a predetermined frequency range using a high-pass filter, a low-pass filter, or the like.

[0014] The measurement unit 11 may also perform intensity correction to calculate the intensity of a sound perceived by a human being based on the human hearing characteristics and the intensity of a sound actually generated from a test object, and generate sound data for the sound after the intensity correction.

[0015] The sound data receiving unit 21 acquires sound data from the measurement unit 11. The sound data receiving unit 21 may also read a recording medium on which the sound data generated by the measurement unit 11 is recorded, and acquire the recorded sound data.

[0016] The storage unit 25 is connected to the controller 100, and stores the sound data received by the sound data receiving unit 21. The storage unit 25 may store the received sound data for a predetermined time. In other words, the stored sound data may be held by the storage unit 25 for a predetermined time after it is stored, and may be deleted after the predetermined time has elapsed since it was stored.

[0017] In addition, the storage unit 25 may store a feature calculated by the controller 100 (described later) in association with sound data. More specifically, the storage unit 25 may store the feature of a sound whose sound quality has been specified as reference data for determining the sound quality of a sound represented by newly received sound data.

[0018] The storage unit 25 may store sound data and features related to the sound data for each "sound quality" set in advance. The storage unit 25 may store sound data in which a sound (second evaluation sound) whose "sound quality" has been determined and features related to the sound data (second evaluation sound) in association with the sound (second evaluation sound).

[0019] Here, a note should be made regarding the "sound quality" of the sound produced by the test object. When a test object in which an abnormality has occurred operates, the test object may produce abnormal sounds such as humming, rasping, rattling, and cracking. Note that the types of abnormal sounds are not limited to the examples given here. Also, a test object that is operating normally produces operating sounds that are not classified as abnormal sounds.

[0020] "Sound quality" is thus a categorization of sounds generated by the test object, such as abnormal sounds and other operating sounds. Note that "sound quality" can be said to be a typological expression of sounds that can be distinguished from other sounds based on the characteristics of the sounds. The categorization of sounds may be performed by humans based on the characteristics of the sounds, or may be performed automatically based on the characteristics of the sounds using machine learning or other statistical clustering methods.

[0021] Alternatively, the storage unit 25 may store abnormal sound type data indicating the type of abnormal sound contained in the second evaluation sound in association with a feature amount (second feature amount) related to the sound whose sound quality has been specified (second evaluation sound). According to the above example, the storage unit 25 may also store abnormal sound type data indicating the type of abnormal sound, such as a humming sound, a rasping sound, a rattling sound, or a cracking sound.

[0022] In addition, the storage unit 25 may store cause data indicating the cause of an abnormal sound included in the second evaluation sound, in association with a feature amount (second feature amount) related to the sound with a specified sound quality (second evaluation sound). For example, the cause data may be data indicating causes such as "looseness of fastening parts such as bolts / nuts" and "distortion or deformation of mounting members" in the inspection target.

[0023] For example, a situation in which the test subject may produce an abnormal sound may be assumed in advance, and a second evaluation sound generated by intentionally creating the situation and a second feature amount related to the second evaluation sound may be stored in the storage unit 25. When the second feature amount is stored in the storage unit 25, abnormal sound type data and cause data may be stored in association with the second feature amount.

[0024] The presentation unit 400 presents frequency data, feature quantities, information indicating sound quality, and the like, calculated by the controller 100 described later, to a user operating the sound quality assessment device 1. For example, the presentation unit 400 may display the frequency data and feature quantities as an image (frequency graph, spectrogram). In addition, the presentation unit 400 may display the cause of an abnormal sound based on the output from the controller 100 described later when an abnormal sound is generated in the inspection target. The presentation unit 400 may be a liquid crystal display, an organic EL display, or the like. In addition, when only normal / abnormal is displayed, a rotating light (a device status indicator in a factory) or the like may be used.

[0025] The presentation unit 400 is not limited to presenting various types of information by visual information. The presentation unit 400 may present information to the user by audio information such as sound, or may generate vibrations and present information to the user by vibration stimulation. For example, the presentation unit 400 may be a sound source, an amplifier, a speaker, etc. for generating and emitting sound.

[0026] Information is input to the controller 100 from the sound data receiving unit 21 and the storage unit 25 , and information is output from the controller 100 to the storage unit 25 and the presentation unit 400 .

[0027] The controller 100 is a general-purpose microcomputer equipped with a CPU (Central Processing Unit), a memory, and an input / output unit. A computer program (sound quality judgment program) for functioning as a part of the sound quality judgment device 1 is installed in the controller 100. By executing the computer program, the controller 100 functions as a plurality of information processing circuits (110, 120, 130).

[0028] Here, an example is shown in which the multiple information processing circuits (110, 120, 130) included in the controller 100 are realized by software. However, it is also possible to configure the information processing circuits (110, 120, 130) by preparing dedicated hardware for executing each information process described below. Also, the multiple information processing circuits (110, 120, 130) may be configured by individual hardware.

[0029] The controller 100 includes a frequency data generating unit 110, a feature extracting unit 120, and a determining unit 130 as a plurality of information processing circuits (110, 120, 130).

[0030] The frequency data generating unit 110 generates frequency data indicating the intensity of each frequency component of the sound (first evaluation sound) represented by the sound data, based on the sound data. More specifically, the frequency data generating unit 110 generates the frequency data using a frequency analysis method such as FFT (Fast Fourier Transform) or wavelet analysis. For example, as shown in Fig. 3, the frequency data is represented as a two-dimensional waveform with frequency on the horizontal axis and intensity on the vertical axis. The frequency analysis method used by the frequency data generating unit 110 is not limited to the examples given here.

[0031] Alternatively, the frequency data generating unit 110 may generate frequency data for frequencies in the audible range of the human ear based on the human hearing characteristics. Also, the frequency data generating unit 110 may perform intensity correction based on the human hearing characteristics for the sound intensity for each frequency obtained by the frequency analysis.

[0032] The feature extraction unit 120 extracts a feature (first feature) of a sound (first evaluation sound) represented by the sound data from the frequency data. More specifically, the feature extraction unit 120 focuses on a frequency (peak frequency) having a prominent intensity compared to the intensity of other surrounding frequencies in the frequency data, and extracts the position of the peak frequency, the half-width of the frequency waveform in the vicinity of the peak frequency, etc. as the feature.

[0033] For example, the feature extraction unit 120 sets a frequency window indicating a predetermined frequency range for the frequency data, and for each frequency window, identifies a frequency having the maximum intensity in the frequency window as a peak frequency. In Fig. 3 and Fig. 4, a window W1, a window W2, and a window W3 are shown as frequency windows. In Fig. 4, a peak frequency F1 having the maximum intensity in the window W1, a peak frequency F2 having the maximum intensity in the window W2, and a peak frequency F3 having the maximum intensity in the window W3 are shown. The frequency data generation unit 110 identifies peak frequencies F1, F2, and F3 for the window W1, the window W2, and the window W3, respectively.

[0034] 3 and 4 show an example in which three frequency windows are set, but the number of frequency windows set by the feature extraction unit 120 is not limited to this. The feature extraction unit 120 may set any number of frequency windows.

[0035] The feature extraction unit 120 identifies the fundamental frequency of the sound represented by the sound data based on the frequency data. Here, the "fundamental frequency" means the smallest frequency among the peak frequencies. In the following description, the sound of the fundamental frequency is referred to as the "fundamental tone." For example, in FIG. 4, the feature extraction unit 120 identifies the peak frequency F1 as the fundamental frequency.

[0036] Furthermore, the feature extraction unit 120 calculates the amount of deviation between the peak frequency other than the fundamental frequency and the frequency of the harmonic corresponding to the fundamental tone. For example, in Fig. 4, when the peak frequency F1 is the fundamental frequency, the feature extraction unit 120 calculates the difference between the value twice the peak frequency F1 (harmonic frequency) and the peak frequency F2 as the amount of deviation. Furthermore, the feature extraction unit 120 calculates the difference between the value three times the peak frequency F1 (harmonic frequency) and the peak frequency F3 as the amount of deviation.

[0037] In this way, the feature extraction unit 120 sets the difference between the harmonic frequency (a frequency obtained by multiplying the fundamental frequency by an integer value) included in the set frequency window and the peak frequency in the frequency window as the deviation amount. In other words, the feature extraction unit 120 extracts the deviation amount of the peak frequency from the harmonic frequency as the feature amount.

[0038] In order to calculate the amount of deviation, it is necessary to narrow down the harmonic frequencies included in each frequency window to one. Therefore, the width of the set frequency window may be set to a value approximately equal to the magnitude of the fundamental frequency. Furthermore, since the frequency resolution (the width of discriminable frequencies) changes depending on the number of data when calculating the frequency, sound data collected at an appropriate sampling frequency (e.g., 44.1 kHz or higher) with a predetermined frequency resolution (e.g., 1 Hz) may be used.

[0039] Alternatively, the feature extraction unit 120 may extract the number of peak frequencies included in a predetermined frequency range in the frequency data as the feature. The feature extraction unit 120 may also extract the number of peak frequencies having an intensity exceeding a predetermined threshold as the feature. Here, the predetermined frequency range and the predetermined threshold are set in advance.

[0040] The feature extraction unit 120 may also extract the half-width of a peak as a feature, targeting a peak in a frequency waveform indicated by a peak frequency. For example, the feature extraction unit 120 may generate an intensity histogram for each target frequency window, and calculate the median and maximum intensities in the target frequency window from the histogram. Then, the width of the frequency range having an intensity equal to the average value of the median and maximum values ​​may be calculated as the half-width.

[0041] The method of obtaining the half-width is not limited to the method of generating a histogram. For example, the half-width may be estimated based on the obtained approximation function by approximating the waveform using a Gaussian function or the like for the peaks of the frequency waveform.

[0042] The feature extraction unit 120 may also extract the ratio of the intensity at the peak frequency to the intensity at the fundamental frequency as the feature. Furthermore, the feature extraction unit 120 may extract the sum of the intensities of the peak frequencies as the feature. Alternatively, the feature extraction unit 120 may calculate the feature by combining the amount of deviation that the peak frequency has from the harmonic frequency, the number of peak frequencies, the half-width, the ratio, the intensity, and the like.

[0043] The feature amount may be calculated using, for example, a part of the peak frequencies identified using a threshold value of sound pressure intensity, or all of the identified peak frequencies. In other words, the feature amount extraction unit 120 may select a peak to be subjected to the process of extracting the feature amount.

[0044] Alternatively, the feature extraction unit 120 may output the frequency data and the feature extracted from the frequency data to the presentation unit 400. In this case, the feature extraction unit 120 may convert the frequency data and the feature into an image (frequency graph, spectrogram) and output the converted image.

[0045] The determination unit 130 determines whether the first evaluation sound related to the first feature amount and the second evaluation sound related to the second feature amount are classified as having the same sound quality, based on the feature amount (first feature amount) calculated by the feature amount extraction unit 120 and the feature amount (second feature amount of the sound whose sound quality has been determined) stored in the storage unit 25. That is, the determination unit 130 determines the sound quality of the first evaluation sound related to the first feature amount by determining whether the first evaluation sound related to the first feature amount and the second evaluation sound related to the second feature amount can be associated with each other.

[0046] For example, the determination unit 130 determines whether or not each of the second features stored in the storage unit 25 is similar to the first feature calculated by the feature extraction unit 120. Here, the determination unit 130 may calculate the Euclidean distance between the first feature and the second feature (the square of the sum of the difference between the components constituting the features and all the components) and determine that they are similar when the calculated Euclidean distance is equal to or smaller than a predetermined threshold value.

[0047] In addition, the judgment unit 130 may calculate a cosine similarity between the first feature amount and the second feature amount (the angle formed in a vector space between a vector representing the first feature amount and a vector representing the second feature amount), and judge that they are similar when the calculated cosine similarity is equal to or smaller than a predetermined threshold value.

[0048] When the judgment unit 130 judges that the first feature and the second feature are similar, it judges that the first evaluation sound related to the first feature and the second evaluation sound related to the second feature judged to be similar to the first feature are classified into the same sound quality.

[0049] In addition, the judgment unit 130 may judge whether or not the first evaluation sound related to the first feature amount and the second evaluation sound related to the second feature amount are classified into the same sound quality by using machine learning or other statistical clustering methods based on the first feature amount and the second feature amount.

[0050] Alternatively, the determination unit 130 may determine that an abnormal sound included in a second evaluation sound associated with the first evaluation sound is included in the first evaluation sound, based on the abnormal sound type data stored in the storage unit 25. The determination unit 130 may output data indicating the determined type of abnormal sound to the presentation unit 400.

[0051] The determination unit 130 may estimate that the cause of the abnormal sound included in the second evaluation sound associated with the first evaluation sound occurs in the test subject based on the cause data stored in the storage unit 25. The determination unit 130 may output data indicating the estimated cause of the abnormal sound to the presentation unit 400.

[0052] [Processing procedure of sound quality assessment device] Next, a processing procedure of the sound quality judgment device 1 according to this embodiment will be described with reference to the flowchart of Fig. 2. Fig. 2 is a flowchart showing the processing of the sound quality judgment device 1 according to one embodiment of the present invention.

[0053] In step S101, the sound data receiving unit 21 acquires sound data.

[0054] In step S103, the frequency data generating unit 110 generates frequency data based on the sound data.

[0055] In step S105, the feature extraction unit 120 calculates the fundamental frequency and the harmonic frequencies based on the frequency data.

[0056] In step S107, the feature extraction unit 120 selects a peak that is to be subjected to the process of extracting a feature.

[0057] In step S109, the feature extraction unit 120 extracts features.

[0058] In step S111, the judgment unit 130 judges the sound quality of the first evaluation sound related to the first feature.

[0059] In step S113, the determining unit 130 determines whether or not the first evaluation sound includes an abnormal sound, based on the determined sound quality.

[0060] If the first evaluation sound includes an abnormal sound (YES in step S113), the determination unit 130 presumes in step S115 that a cause corresponding to the abnormal sound occurs in the test object. If the first evaluation sound does not include an abnormal sound (NO in step S113), the process proceeds to step S117.

[0061] In step S117, it is determined whether or not to end the measurement based on an instruction from the user, etc. If the measurement is not to be ended (NO in step S117), the process returns to step S101, and if the measurement is to be ended (YES in step S117), the process of the sound quality judgment device 1 is ended.

[0062] [Effects of the embodiment] As described in detail above, the sound quality assessment device, sound quality assessment method, and sound quality assessment program of this embodiment generate frequency data indicating the intensity of each frequency component of the first evaluation sound based on sound data recording a first evaluation sound generated from a test subject, extract a first feature of the first evaluation sound from the frequency data, and determine that the sound quality of the first evaluation sound is the same as the sound quality of a second evaluation sound associated with the first evaluation sound based on the first feature and the second feature of the second evaluation sound whose sound quality has been determined.

[0063] This allows the quality of the sound generated by the test object to be identified. Furthermore, the quality of the sound can be classified according to the abnormality of the test object. Furthermore, since the quality of the sound can be identified based on various features contained in the sound generated by the test object, erroneous identification of the quality of the sound can be suppressed. By evaluating the abnormal sound in detail, the accuracy of the abnormal sound determination can be improved.

[0064] In the sound quality evaluation device, sound quality evaluation method, and sound quality evaluation program according to the present embodiment, the first feature may include a deviation between the peak frequency in the frequency data and the frequency of the harmonic corresponding to the fundamental tone of the first evaluation tone. Here, the frequency of the fundamental tone may be determined, and the frequency of the harmonic tone may be calculated by multiplying the frequency of the fundamental tone by an integer value set in advance. Installation abnormalities of the test object and abnormalities of the components are likely to appear in the deviation from the harmonic frequency. Therefore, by including the deviation in the first feature, it is possible to detect a difference from the normal state of the test object based on the abnormal sound. This improves the accuracy of abnormal sound determination.

[0065] Furthermore, in the sound quality assessment device, the sound quality assessment method, and the sound quality assessment program according to the present embodiment, the first feature may include the number of peak frequencies included in a predetermined frequency range in the frequency data. By including the peak frequencies in the first feature, it is possible to detect a difference from the normal state of the test object based on abnormal sounds.

[0066] In the sound quality assessment device, sound quality assessment method, and sound quality assessment program according to the present embodiment, the first feature may include a half-width of a peak in the frequency data. For example, among abnormal sounds, a crack sound can be easily detected by focusing on a difference in the half-width of the peak. Therefore, by including the half-width of the peak in the first feature, a difference from the normal state of the test object can be detected based on the abnormal sound.

[0067] Furthermore, in the sound quality evaluation device, sound quality evaluation method, and sound quality evaluation program according to the present embodiment, the first feature may include a ratio of the intensity at the peak frequency in the frequency data to the intensity of the fundamental tone of the first evaluation sound. Resonance abnormalities caused by installation abnormalities of the test object or abnormalities in components are easily detected as changes in the intensity ratio between peak frequencies. Therefore, by including the intensity ratio between peak frequencies in the first feature, it is possible to detect a difference from the normal state of the test object based on the abnormal sound.

[0068] Furthermore, the sound quality evaluation device, sound quality evaluation method, and sound quality evaluation program according to the present embodiment may store abnormal sound type data indicating the type of abnormal sound included in the second evaluation sound in association with the second feature amount, and determine that the abnormal sound included in the second evaluation sound associated with the first evaluation sound is included in the first evaluation sound based on the abnormal sound type data. This makes it possible to determine whether a known abnormal sound is generated from the test object. By storing multiple abnormal sound type data corresponding to multiple abnormal sounds, the evaluation of abnormal sounds can be performed in detail, and the accuracy of abnormal sound determination can be improved.

[0069] Furthermore, the sound quality evaluation device, sound quality evaluation method, and sound quality evaluation program according to the present embodiment may store cause data indicating the cause of an abnormal sound contained in the second evaluation sound in association with the second feature amount, and estimate that the cause of the abnormal sound contained in the second evaluation sound associated with the first evaluation sound occurs in the test object based on the cause data. This makes it possible to determine whether or not a known abnormality whose cause is identified occurs in the test object. By storing multiple cause data corresponding to multiple causes, the accuracy of estimating the cause of the abnormal sound can be improved.

[0070] In addition, the sound quality assessment device, the sound quality assessment method, and the sound quality assessment program according to the present embodiment may present the extracted first feature amount via a presentation unit. This allows the user to recognize the features of the sound generated from the test object. As a result, this leads to improved convenience for the user.

[0071] Each of the functions described in the above embodiments may be implemented by one or more processing circuits, including a programmed processor, an electrical circuit, or a device such as an application specific integrated circuit (ASIC), or a circuit component arranged to perform the described functions.

[0072] Although the contents of the present invention have been described above in accordance with the embodiments, the present invention is not limited to these descriptions, and various modifications and improvements are possible, which will be obvious to those skilled in the art. The descriptions and drawings forming part of this disclosure should not be understood as limiting the present invention. Various alternative embodiments, examples, and operating techniques will be apparent to those skilled in the art from this disclosure.

[0073] Of course, the present invention includes various embodiments not described herein. Therefore, the technical scope of the present invention is defined only by the invention-specifying matters related to the scope of the claims appropriate from the above description. [Explanation of symbols]

[0074] 1 Sound quality determination device 11 Measurement section 21 Sound data reception section 25 Memory section 100 Controller 110 Frequency data generator 120 Feature Extraction Unit 130 Judgment section 400 Presentation section

Claims

1. A sound quality evaluation device that evaluates sound quality of a first evaluation sound generated from a test object, comprising: a sound data receiving unit that receives sound data in which the first evaluation sound is recorded; a frequency data generating unit that generates frequency data indicating an intensity of each frequency component of the first evaluation sound based on the sound data; a feature extraction unit that extracts a first feature of the first evaluation sound from the frequency data; a storage unit that stores a second feature amount of a second evaluation sound whose sound quality has been determined in association with the second evaluation sound; a determination unit that determines, based on the first feature amount and the second feature amount, that a sound quality of the first evaluation sound is the same as a sound quality of the second evaluation sound associated with the first evaluation sound; Equipped with The first characteristic amount includes a deviation amount between a peak frequency in the frequency data and a frequency of a harmonic corresponding to a fundamental tone of the first evaluation tone. A sound quality assessment device comprising:

2. 2. The sound quality assessment device according to claim 1, The feature extraction unit determining a frequency of the fundamental tone; Multiplying the frequency of the fundamental tone by a preset integer value to calculate the frequency of the harmonic tone. A sound quality assessment device comprising:

3. 3. The sound quality assessment device according to claim 1, The first feature amount includes the number of peak frequencies included in a predetermined frequency range in the frequency data. A sound quality assessment device comprising:

4. The sound quality assessment device according to any one of claims 1 to 3, The first feature amount includes a half-width of a peak in the frequency data. A sound quality assessment device comprising:

5. The sound quality assessment device according to any one of claims 1 to 4, The first characteristic amount includes a ratio of an intensity of a peak frequency in the frequency data to an intensity of a fundamental tone of the first evaluation tone. A sound quality assessment device comprising:

6. The sound quality assessment device according to any one of claims 1 to 5, the storage unit stores abnormal sound type data indicating a type of the abnormal sound included in the second evaluation sound in association with the second feature amount, The determination unit determines, based on the abnormal sound type data, that an abnormal sound included in the second evaluation sound associated with the first evaluation sound is included in the first evaluation sound. A sound quality assessment device comprising:

7. 7. A sound quality assessment device according to claim 6, the storage unit stores cause data indicating a cause of the abnormal sound included in the second evaluation sound in association with the second feature amount, The determination unit estimates, based on the cause data, that a cause of an abnormal sound included in the second evaluation sound associated with the first evaluation sound occurs in the test object. A sound quality assessment device comprising:

8. The sound quality assessment device according to any one of claims 1 to 7, further comprising: a presentation unit for presenting the first feature amount A sound quality assessment device comprising:

9. A sound quality evaluation method for evaluating sound quality of a first evaluation sound generated from a test object, comprising the steps of: receiving sound data in which the first evaluation sound is recorded; generating frequency data indicating an intensity of each frequency component of the first evaluation sound based on the sound data; extracting a first feature of the first evaluation sound from the frequency data; a storage unit stores a second feature amount of a second evaluation sound whose sound quality has been determined in association with the second evaluation sound; determining, based on the first feature amount and the second feature amount, that a sound quality of the first evaluation sound is the same as a sound quality of the second evaluation sound associated with the first evaluation sound; The first characteristic amount includes a deviation amount between a peak frequency in the frequency data and a frequency of a harmonic corresponding to a fundamental tone of the first evaluation tone. A sound quality assessment method comprising:

10. A computer having a storage unit, receiving sound data recording a first evaluation sound generated from the test subject; generating frequency data indicating an intensity of each frequency component of the first evaluation sound based on the sound data; extracting a first feature of the first evaluation sound from the frequency data; storing, by the storage unit, a second feature quantity of a second evaluation sound whose sound quality has been determined, in association with the second evaluation sound; determining, based on the first feature amount and the second feature amount, that a sound quality of the first evaluation sound is the same as a sound quality of the second evaluation sound associated with the first evaluation sound; A sound quality judgment program for executing The first characteristic amount includes a deviation amount between a peak frequency in the frequency data and a frequency of a harmonic corresponding to a fundamental tone of the first evaluation tone. A sound quality assessment program characterized by:

Images