Method for identifying abnormal noises, and device for identifying abnormal noises
The method and device modify sound pressure levels to generate comparison sound data for accurate identification of abnormal noise, addressing the challenge of pinpointing abnormal noise frequencies in vehicles.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- DENSO CORP
- Filing Date
- 2021-12-27
- Publication Date
- 2026-05-19
AI Technical Summary
Existing methods for identifying abnormal noise in vehicles, such as air conditioning devices, struggle to accurately pinpoint the frequency of abnormal noise perceived by users due to the complexity of multiple sound frequencies generated.
A method and device that generate comparison sound data by modifying sound pressure levels for each frequency based on sound data and outputting both sound data and comparison sound data for accurate identification of abnormal noise by comparing these sounds.
Enables highly accurate identification of abnormal noise by modifying sound pressure levels to highlight deviations from normal sounds, allowing precise communication of abnormal sounds to inspectors.
Smart Images
Figure 0007861398000001 
Figure 0007861398000002 
Figure 0007861398000003
Abstract
Description
Technical Field
[0001] The present invention relates to a method for identifying abnormal noise and an apparatus for identifying abnormal noise.
Background Art
[0002] Conventionally, air conditioning devices, power window devices, etc. provided in vehicles may generate abnormal noise when driven. For example, an air conditioning device includes a blower motor, and the blower motor itself may generate abnormal noise. Also, for example, an air conditioning device may generate so-called wind noise, which is abnormal noise generated when air flows through a flow path.Such cases, the user may convey the abnormal noise felt by himself / herself to an inspector such as a vehicle dealer and request repair. In such cases, it is common for the user to convey the abnormal noise felt by himself / herself to the inspector using onomatopoeia. Also, as a method for an inspector or the like to identify abnormal noise, for example, there is a method of making it easier to identify abnormal noise by deleting sound data recorded at a far position from sound data recorded at a position close to the object as noise (see, for example, Patent Document 1). When this method is used, since the sound that can be heard even at a position far from the object is removed, it becomes easier for the inspector to identify abnormal noise.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] However, even when using the above-described method for identifying abnormal noise, it has been difficult for an inspector to accurately identify the abnormal noise felt by the user. That is, when a plurality of sounds having different frequencies are generated from the object, it has been difficult for the inspector to accurately identify which frequency of sound is the abnormal noise felt by the user.
[0005] The present invention was made to solve the above-mentioned problems, and its objective is to provide a method for identifying abnormal noises and a device for identifying abnormal noises that enable highly accurate identification of abnormal noises. [Means for solving the problem]
[0006] A method for identifying abnormal noises that solves the above problems comprises a generation step of generating comparison sound data by changing the sound pressure level that is prominent among the sound pressure levels for each frequency based on sound data, and an identification step of identifying abnormal noises in the sound data by outputting the sound data and the comparison sound data as sounds and comparing the sounds.
[0007] According to this method, in the generation process, the sound pressure levels that are prominent among the sound pressure levels for each frequency based on the sound data are modified to generate comparison sound data. Then, in a specific process, the sound data and the comparison sound data are output as sounds, and by comparing these sounds, abnormal sounds in the sound data are identified. Therefore, for example, when a user communicates an abnormal sound they perceive to an inspector, it is possible to communicate it with higher accuracy compared to methods such as using onomatopoeia. Thus, abnormal sounds can be identified with high accuracy.
[0008] The abnormal sound identification device that solves the above problems includes a generation unit (12) that calculates the sound pressure level for each frequency based on sound data, compares the calculated sound pressure level for each frequency with a preset reference sound pressure level, and generates comparison sound data by changing the sound pressure level that is the largest relative to the reference sound pressure level.
[0009] In this configuration, the generation unit calculates the sound pressure level for each frequency based on the sound data. The generation unit then compares the calculated sound pressure level for each frequency with a preset reference sound pressure level, and the sound pressure level with the highest value relative to the reference sound pressure level is changed to generate comparison sound data. In this way, for example, by using a reference sound pressure level corresponding to a normal sound, the sound pressure level of frequencies that deviate from normal can be changed. Therefore, compared to methods such as visually inspecting the waveform and changing the sound pressure level of frequencies that appear to be protruding, it is possible to change the sound pressure level of frequencies corresponding to abnormal sounds with a high probability. Furthermore, for example, by outputting the sound data and comparison sound data as separate sounds and comparing them, it is possible to identify abnormal sounds in the sound data. Therefore, when a user communicates an abnormal sound they perceive to an inspector, etc., it is possible to communicate with higher accuracy compared to methods such as using onomatopoeia. Thus, abnormal sounds can be identified with high accuracy. [Brief explanation of the drawing]
[0010] [Figure 1] A schematic diagram illustrating a device for identifying abnormal noises in one embodiment. [Figure 2] A flowchart illustrating a method for identifying abnormal noises in one embodiment. [Figure 3] Waveform diagram of time-sound pressure level in one embodiment. [Figure 4] Waveform diagram of frequency-sound pressure level and reference sound pressure level in one embodiment. [Figure 5] Waveform diagram of frequency-sound pressure level and reference sound pressure level in one embodiment. [Figure 6] Waveform diagram of frequency-sound pressure level and reference sound pressure level in one embodiment. [Modes for carrying out the invention]
[0011] Hereinafter, an embodiment of the present invention will be described with reference to Figures 1 to 6. (Explanation of the abnormal noise detection device) As shown in Figure 1, the abnormal noise detection device of this embodiment includes a personal computer, or PC 11. The PC 11 comprises a PC main unit 12, a display device, a display 13, and an input device, a keyboard 14, for performing various operations. The PC main unit 12 comprises a CPU capable of performing various calculations, memory capable of storing various programs and data, and a speaker capable of outputting electrical signals as sound, and constitutes a generation unit.
[0012] The PC unit 12 performs various processes using pre-saved programs, etc. For example, the PC unit 12 calculates the sound pressure level for each frequency (see waveform X2 in Figure 4) based on the sound data recorded by the IC recorder 15 (see waveform X1 in Figure 3). The PC unit 12 also compares the calculated sound pressure level for each frequency with a pre-set reference sound pressure level (see waveform Z1 in Figure 4), and generates comparison sound data by changing the largest sound pressure level relative to the reference sound pressure level. The reference sound pressure level is data corresponding to normal sound. For example, if the object is an air conditioning unit 17 with a blower motor installed in a vehicle 16, the reference sound pressure level is created to correspond to the normal sound generated when a normal air conditioning unit 17 is driven, and is data corresponding to a sound that can be considered normal and is acceptable. The PC unit 12 also generates comparison sound data by relatively decreasing the largest sound pressure level relative to the reference sound pressure level. In this embodiment, the personal computer unit 12 generates comparison sound data by changing the highest sound pressure level relative to the reference sound pressure level to 0.
[0013] Furthermore, the computer unit 12 is capable of outputting sound data and comparison sound data as separate sounds. For example, based on the operation of the keyboard 14, the computer unit 12 alternately outputs sound data and comparison sound data as separate sounds from the speakers.
[0014] Then, based on the fact that the abnormal sound cannot be identified, the computer main body 12 treats the comparison sound data as the sound data and generates the comparison sound data again. Specifically, for example, when the user cannot identify the abnormal sound by listening and comparing the sounds of the sound data and the comparison sound data, and an operation to that effect is performed, the computer main body 12 generates the comparison sound data again. The computer main body 12 compares the sound pressure level for each frequency in the previous comparison sound data with a preset reference sound pressure level, changes the sound pressure level that is the largest with respect to the reference sound pressure level, and generates the latest comparison sound data.
[0015] Then, for example, after the computer main body 12 generates the comparison sound data again, based on an operation of the keyboard 14, for example, the previous comparison sound data and the latest comparison sound data are alternately output from the speaker as sounds.
[0016] (Explanation of the abnormal sound identification method) Next, the abnormal sound identification method will be explained. The abnormal sound identification method of this embodiment is performed using the above-described abnormal sound identification device.
[0017] The abnormal sound identification method includes a "generation step" and an "identification step". In the "generation step", the comparison sound data is generated by changing the prominent sound pressure level among the sound pressure levels for each frequency based on the sound data. In the "generation step" of this embodiment, the sound pressure level for each frequency is compared with a preset reference sound pressure level, and the sound pressure level that is the largest with respect to the reference sound pressure level is changed as the prominent sound pressure level to generate the comparison sound data. Also, in the "generation step", the sound pressure level that is the largest with respect to the reference sound pressure level is relatively decreased and changed to 0 in this embodiment to generate the comparison sound data.
[0018] In the "identification step", the sound data and the comparison sound data are output as sounds respectively, and the abnormal sound in the sound data is identified by listening and comparing those sounds. Also, when abnormal sounds cannot be identified in the "specific process", the comparison sound data is treated as sound data, and the "generation process" and "specific process" are performed again.
[0019] Specifically, the abnormal sound identification method is executed according to the flowchart shown in FIG. 2. For example, when the inspector inputs the sound data recorded by the IC recorder 15 (see the waveform X1 in FIG. 3) into the computer main body 12 and performs an operation to indicate the generation process, the computer main body 12 performs the following processing of step S1 and below in the generation process. At this time, the sound data handled by the computer main body 12 is the sound data at the timing when abnormal sounds occur in the sound data recorded by the IC recorder 15.
[0020] In step S1, the computer main body 12 calculates the sound pressure level for each frequency based on the sound data (see the waveform X1 in FIG. 3), and automatically proceeds to step S2. At this time, the computer main body 12 executes a fast Fourier transform to calculate the sound pressure level for each frequency from the sound data.
[0021] Note that FIG. 3 shows a schematic waveform X1 of the sound data. Also, FIG. 4 shows a schematic waveform X2 of the sound pressure level for each frequency. The computer main body 12 of this embodiment displays the calculated sound pressure level for each frequency (see the waveform X2 in FIG. 4) on the display 13.
[0022] In step S2, the computer main body 12 compares the calculated sound pressure level for each frequency with a preset reference sound pressure level (see the waveform Z1 in FIG. 4), and automatically proceeds to step S3.
[0023] In step S3, the PC unit 12 generates comparison sound data by changing the highest sound pressure level relative to the reference sound pressure level to 0 (see waveform X3 in Figure 5). Specifically, the PC unit 12 changes the sound pressure levels to 0 in the frequency band that includes the frequency of the highest sound pressure level relative to the reference sound pressure level, and in the frequency bands of several tens of Hz to several hundred Hz before and after the highest sound pressure level frequency. When generating comparison sound data after changing the sound pressure levels, the PC unit 12 calculates the comparison sound data by performing an inverse fast Fourier transform.
[0024] Figure 5 shows a schematic waveform X3 in which the sound pressure level in the frequency band including approximately 4500 Hz, which is the highest relative to the reference sound pressure level, is changed to 0. In this embodiment, the PC main unit 12 displays the modified sound pressure level (see waveform X3 in Figure 5) on the display 13.
[0025] Next, for example, when an inspector performs an operation to indicate that a specific process should be carried out, the PC unit 12 performs the process of step S4 in the specific process. In step S4, the computer unit 12 alternately outputs the sound data and the comparison sound data as sound from the speakers.
[0026] Next, in step S5, for example, the user compares the sounds output from the speakers. Next, in step S6, it is determined whether the user was able to identify the abnormal sound in the sound data. That is, the user determines whether the abnormal sound heard in the sound data is no longer audible in the comparison sound data. If it is determined in step S6 that the abnormal sound was identified, i.e., the abnormal sound heard in the sound data is no longer audible in the comparison sound data, the abnormal sound identification method is terminated. If it is determined in step S6 that the abnormal sound was not identified, i.e., the abnormal sound heard in the sound data is also audible in the comparison sound data, for example, the inspector performs an operation to that effect and proceeds to step S7.
[0027] In step S7, the PC unit 12 automatically proceeds to step S2, treating the comparison sound data as sound data. More specifically, the PC unit 12 replaces the data so that the sound pressure levels for each frequency in the comparison sound data (see waveform X3 in Figure 5) are treated as the sound pressure levels for each frequency calculated in step S1, and then proceeds to step S2.
[0028] Therefore, steps S2 to S6 are performed again. At this time, in step S3, the sound pressure level that is highest relative to the reference sound pressure level in the sound pressure level for each frequency in the comparison sound data (see waveform X3 in Figure 5) is changed to 0 (see waveform X4 in Figure 6), and the latest comparison sound data is generated.
[0029] Figure 6 shows a schematic waveform X4 in which the sound pressure level for each frequency in the comparative sound data (see waveform X3 in Figure 5) is changed to 0 in the frequency band including approximately 4900 Hz, which is the highest relative to the reference sound pressure level.
[0030] In other words, if the abnormal noise cannot be identified in step S6, steps S7, S2 through S6 are repeated in sequence to continue the process of identifying the abnormal noise. Next, the operation of the above embodiment will be described.
[0031] In the generation process, the sound pressure levels that are prominent among the sound pressure levels for each frequency based on the sound data are modified to generate comparison sound data. Then, in a specific process, the sound data and the comparison sound data are output as sounds, and by listening to and comparing these sounds, abnormal sounds in the sound data are identified.
[0032] Next, the effects of the above embodiment are described below. (1) For example, when a user communicates an abnormal sound to an inspector, it is possible to communicate the abnormal sound with higher accuracy compared to methods such as using onomatopoeia. Therefore, the abnormal sound can be identified with high accuracy.
[0033] (2) In the generation process by the PC main unit 12, the calculated sound pressure level for each frequency is compared with a preset reference sound pressure level, and the sound pressure level that is highest relative to the reference sound pressure level is changed as the prominent sound pressure level to generate comparison sound data. In this way, for example, by using a reference sound pressure level corresponding to a normal sound, the sound pressure level of frequencies that deviate from normal can be changed. Therefore, compared to, for example, visually inspecting the waveform and changing the sound pressure level of frequencies that seem to be prominent, it is possible to change the sound pressure level of frequencies corresponding to abnormal noise with a high probability. Therefore, abnormal noise can be identified quickly and easily.
[0034] (3) If an abnormal sound cannot be identified in a specific process, the comparison sound data is treated as sound data and the generation process and identification process are performed again. Therefore, by comparing the sounds while excluding sounds of frequencies that were irrelevant in the previous step, abnormal sounds in the sound data can be identified. Thus, abnormal sounds can be identified with high accuracy.
[0035] (4) In the generation process by the PC unit 12, the sound pressure level is changed to be relatively smaller when comparison sound data is generated, so for example, when output as sound, the abnormal noise becomes smaller. Therefore, users and inspectors can identify abnormal noises with high accuracy.
[0036] This embodiment can be implemented with the following modifications. This embodiment and the following modifications can be combined with each other to the extent that they do not contradict each other technically. In the above embodiment, the generation process by the personal computer 12 involves changing the sound pressure level with respect to the reference sound pressure level to the highest sound pressure level as the prominent sound pressure level to generate comparison sound data. However, the invention is not limited to this, and the sound pressure level may be changed without using the reference sound pressure level.
[0037] For example, in the generation process, the sound pressure levels that stand out from the sound pressure levels for each frequency based on the sound data may be visually confirmed from the waveform displayed on the display 13, and the sound pressure levels may be changed to generate comparison sound data. In this case, a personal computer 11 that does not store any specific programs for use as an abnormal sound identification device can be used.
[0038] In the above embodiment, if an abnormal sound could not be identified in a specific step, the comparison sound data was treated as sound data and the generation and identification steps were repeated, thereby sequentially increasing the sound pressure level to be changed to 0. However, the embodiment is not limited to this. That is, if an abnormal sound could not be identified in a specific step, instead of sequentially increasing the sound pressure level to be changed to 0, the sound pressure level to be changed to 0 could be changed sequentially, for example.
[0039] In the above embodiment, the generation process is described as relatively reducing the prominent sound pressure level to 0, but the invention is not limited to this, and other values may be used. For example, during the generation process, the protruding sound pressure levels may be changed to values greater than 0 while relatively reducing them.
[0040] Furthermore, for example, in the generation process, the prominent sound pressure level may be relatively significantly altered. Specifically, for example, in the generation process, the prominent sound pressure level may be relatively significantly altered by reducing the sound pressure levels other than the prominent one. In this way, the abnormal noise will be louder when output as sound, allowing users and inspectors to identify the abnormal noise with high accuracy.
[0041] In the above embodiment, the object is assumed to be an air conditioning unit 17, but the object may be changed; for example, the object may be a power window device. In this case, the reference sound pressure level must be created to correspond to the normal sound generated when a normal power window device is driven.
[0042] In the above embodiment, the abnormal noise identification device is equipped with a personal computer 11, and the personal computer body 12 constitutes the generation unit. However, the device is not limited to this, and for example, a mobile terminal such as a smartphone may be used as the abnormal noise identification device. [Explanation of symbols]
[0043] 12...The computer unit (generation unit).
Claims
1. A generation process that generates comparative sound data by changing the sound pressure level that is prominent among the sound pressure levels for each frequency based on sound data, A method for identifying abnormal noises, comprising a step of outputting the aforementioned sound data and the aforementioned comparison sound data as sounds, and having the vehicle user compare the sounds to identify abnormal noises in the aforementioned sound data.
2. A generation step of generating comparative sound data by changing the sound pressure level that is prominent among the sound pressure levels for each frequency based on sound data, The system includes a process of identifying an abnormal sound in the sound data by outputting the aforementioned sound data and the aforementioned comparison sound data as sounds and comparing the sounds. In the generation step, the sound pressure level for each frequency is compared with a preset reference sound pressure level, and the sound pressure level that is the largest relative to the reference sound pressure level is changed as the prominent sound pressure level to generate the comparison sound data. Methods for identifying unusual noises.
3. In the generation step, the sound pressure level is changed to be relatively smaller to generate the comparison sound data. A method for identifying abnormal noises according to claim 1 or claim 2.
4. A generation step of generating comparative sound data by changing the sound pressure level that is prominent among the sound pressure levels for each frequency based on sound data, The system includes a process of identifying an abnormal sound in the sound data by outputting the aforementioned sound data and the aforementioned comparison sound data as sounds and comparing the sounds. In the generation step, the sound pressure level is relatively changed to generate the comparison sound data. Methods for identifying unusual noises.
5. If the abnormal sound cannot be identified in the specified step, the comparison sound data is treated as the sound data and the generation step and the identification step are performed again. A method for identifying an abnormal sound according to any one of claims 1 to 4.
6. An abnormal sound identification device comprising a generation unit (12) that calculates the sound pressure level for each frequency based on sound data, compares the calculated sound pressure level for each frequency with a preset reference sound pressure level, and generates comparison sound data by changing the sound pressure level that is the largest relative to the reference sound pressure level.
7. Based on the fact that it could not identify the abnormal sound, the generation unit treats the comparison sound data as sound data and generates the comparison sound data again. The abnormal noise identification device according to claim 6.
8. The generation unit generates the comparison sound data by relatively reducing the sound pressure level. The abnormal noise identification device according to claim 6 or claim 7.
9. The generation unit generates the comparison sound data by relatively changing the sound pressure level significantly. The abnormal noise identification device according to claim 6 or claim 7.