Sound quality detection limit confirmation method and a sound quality detection fault confirmation method

Abstract

The present invention first provides a method for confirming the limit of sound quality detection, and further provides a method for confirming the fault of sound quality detection based on the obtained data. By using the on-site 6δ or 3δ control principle, the limit is determined, so that noise detection is no longer simply measuring the sound level, the detection result is more real and accurate, the after-sales feedback is good, and the fault information can be obtained to guide the maintenance and improve the maintenance efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of noise detection, and in particular to a method for confirming a sound quality detection limit and a sound quality detection failure. Background Art

[0002] At present, the noise problem of range hoods has become an important problem complained by users. There is a lack of noise testing and screening process from the entry of components into the factory to the overall delivery of range hoods, which leads to the failure to detect noise failures in advance, resulting in noise hazards in range hoods entering the market, which is easy to cause customer complaints and damage the corporate image. Among existing companies, most only conduct overall testing of noise, and it is sufficient to measure the sound quality as long as it meets the standard requirements. There is no limit requirement for sound quality in the standard. In actual production, there is no detailed testing of the sound quality of noise. In the actual feedback of after-sales data, there are after-sales repairs and user complaints due to a certain value of sound quality exceeding the standard. Therefore, it is urgent to confirm the limit of sound quality. Summary of the invention

[0003] The main purpose of the present invention is to solve the above problems and shortcomings, and provide a method for confirming a sound quality detection limit and a method for confirming a sound quality detection fault.

[0004] To achieve the above object, the present invention first provides a method for confirming a sound quality detection limit, comprising the following steps:

[0005] S1, the user conducts a subjective evaluation of the noise of N test samples according to the evaluation criteria and records the score of each sample;

[0006] S2, calculating the average score of the subjective evaluation of noise of each sample, sending the samples with the average score above the predetermined value n to the noise detection room for sound quality detection using the noise sound quality detection device, and recording the data of various indicators of the sound quality of each sample respectively;

[0007] S3, statistically analyzing the recorded data of each indicator, counting the maximum and minimum values ​​of each sound quality detection data indicator and calculating the average value μ and standard deviation δ of each data;

[0008] S4, determining the evaluation principles for each sound quality, and determining the target limit values ​​for each sound quality data based on the data calculated in S3.

[0009] Furthermore, in step S1, the test sample size N is ≥ 1000.

[0010] Furthermore, in step S2, a ten-point system is adopted, and the predetermined value of subjective evaluation qualification n≥5.

[0011] Furthermore, the sound quality test data includes speech clarity, loudness, sound pressure level, sharpness and 1 / 3 octave.

[0012] Furthermore, in step S4, the evaluation principle of speech intelligibility is that the higher the value, the better, and its target limit value is higher than the minimum value in the speech intelligibility detection data, and higher than the calculated μ-6δ value of speech intelligibility, or its target limit value is higher than the minimum value in the speech intelligibility detection data, and higher than the calculated μ-3δ value of speech intelligibility.

[0013] Furthermore, in step S4, the evaluation principle of loudness, sound pressure level and sharpness is that the lower the value, the better, each target limit value is lower than the maximum value of the corresponding detection data, and lower than the μ+6δ value calculated by statistics of each detection data, or each target limit value is lower than the maximum value of the corresponding detection data, and lower than the μ+3δ value calculated by statistics of each detection data.

[0014] Furthermore, the standard deviation δ is given by the formula It is calculated that, where N is the sample size, i is any sample from 1 to N samples, and μ is the arithmetic mean of the N test data.

[0015] The present invention further provides a method for confirming a sound quality detection fault, which has the following technical features:

[0016] A sound quality detection fault confirmation method is provided, which performs sound quality noise detection on a noise fault machine, converts the noise signal into a corresponding octave spectrum signal, filters and replays the formed spectrum signal, determines the frequency band where the noise fault is located, thereby determining the sound quality fault category of the noise fault, repairs the noise fault machine, obtains the product fault category, detects and repairs a large number of different types of noise fault machines, records and counts the sound quality fault information and product fault categories accordingly, forms a database of noise and fault categories, and draws out the envelope limit spectrum of the noise according to the test data as a detection standard, performs noise sound quality detection on products produced on site, compares the detection data with the limit value obtained by the sound quality detection limit determination method described above, determines whether the product sound quality is qualified, and for unqualified products, converts the sound quality data into a spectrum signal, and obtains the fault category after comparing it with the database information.

[0017] Furthermore, the envelope limit spectrum has two values, including a failure line and a warning line.

[0018] Furthermore, the gap between the unqualified line and the warning line is 2db.

[0019] In summary, the sound quality detection limit confirmation method and the sound quality detection fault confirmation method provided by the present invention have the following advantages compared with the prior art:

[0020] 1. The sample size is large enough and the data is sufficient to obtain accurate data;

[0021] 2. Use score limits to effectively remove unqualified data and ensure the validity of the data;

[0022] 3. The confirmation method is simple and effective limit values ​​can be obtained;

[0023] 4. The fault confirmation method is simple. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 : A fault frequency band diagram of a faulty machine noise in a sound quality detection fault confirmation method of the present invention;

[0025] Figure 2 : A schematic diagram of a limit envelope spectrum line in a sound quality detection fault confirmation method of the present invention; DETAILED DESCRIPTION

[0026] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0027] The present invention first provides a method for confirming a sound quality detection limit, comprising the following steps:

[0028] S1, subjective evaluation:

[0029] A grading method is adopted to formulate clear evaluation standards according to the requirements of the testing standards and after-sales feedback information. In this embodiment, the evaluation is divided into "ten levels" based on the principle of a ten-point system. In practice, the levels can be set as needed. The user conducts a subjective evaluation of the noise and the sound quality of the noise of N test samples according to the evaluation standards, and subjectively scores the sound quality indicators, and records the scores of each sample.

[0030] In this embodiment, ordinary users and technical personnel are selected respectively. The number of people in each group can be determined according to needs. Each group selects samples of no less than 1,000 products. Simple noise analysis training is conducted for ordinary users to make them understand the noise categories and standard noise conditions, and then conduct subjective evaluations. The subjective evaluations of each sample are scored and recorded. If conditions permit, the two groups of evaluators can cross-evaluate the products to obtain accurate data. The sample size is large enough to obtain sufficient reference data. The scoring standards refer to the following table:

[0031] Subjective evaluation criteria for sound quality

[0032]

[0033] S2, the subjective evaluation results are scored and the average score of each sample is calculated. The samples with an average score above the predetermined value n are sent to the noise detection room for sound quality detection using a professional noise sound quality detection device, and the sound quality index data of each sample are recorded separately.

[0034] In this embodiment, the evaluation adopts a 10-point system, and the predetermined value of subjective evaluation qualification n≥5 can be set. The best solution is n≥6, which means that the noise performance of the sample is good, the noise value meets the standard requirements, the noise is low, the noise perception is good, and the user acceptance is high. Products with sample scores above 6 points will be subjected to professional sound quality testing to achieve the screening of normally distributed data. It is simple and intuitive, and unqualified data and products with unqualified noise can be directly eliminated.

[0035] The samples enter the noise detection room one by one for noise sound quality detection. In this embodiment, the sound quality includes but is not limited to speech clarity, loudness, sound pressure level, sharpness and 1 / 3 octave. The detection results of each detection index are output in the form of data, and the detection data of each sound quality index of the sample are recorded separately. After all samples are tested, the test results of all samples are summarized.

[0036] S3, the test results obtained in the above steps are summarized, and the recorded index data are statistically analyzed respectively. In the summary table, the maximum value (MAX value) and minimum value (MIN value) of each sound quality index test data can be obtained through statistical data, and the average value μ and standard deviation δ of each index data can be calculated.

[0037] Among them, the average value μ of each indicator data is the arithmetic mean of each data, and the standard deviation δ is calculated using the formula:

[0038]

[0039] Among them, δ is the standard deviation to be calculated, N is the sample size, i is any sample from 1 to N samples, xi is the test data of any sample, and μ is the arithmetic mean of the N test data. It should be noted that N in this formula is not necessarily the same as the sample size N in step S1, and is the sample size with a score of 6 or above after step S1, which is less than or equal to the sample size in step S1.

[0040] And further according to the control requirements, the value of 3δ or 6δ is calculated. It should be noted that the control requirements of 3δ or 6δ adopt the management method commonly used in Japanese quality management. The calculation method is simple and clear, easy to use for calculation management, and accurate normal distribution data can be obtained.

[0041] After statistical analysis and calculation, the summary table shown in the following table is obtained:

[0042] Sound quality test data summary table (excerpt)

[0043] In this embodiment, the sample size is relatively large, and the above table is only an excerpt of part of the data.

[0044]

[0045] S4, determining the evaluation principles for each sound quality, and determining the target limit values ​​for each sound quality data based on the data calculated in S3.

[0046] The products participating in the sound quality test in the noise detection room are noise qualified products. Therefore, all data of sound quality are standard data of qualified products, and all data are acceptable to users and below the standard requirements. In this embodiment, sound quality includes but is not limited to speech clarity, loudness, sound pressure level, sharpness and 1 / 3 octave. Among them, speech clarity can reflect whether noise affects speech clarity. Therefore, the evaluation principle is that the larger the test data, the better, which means that under the test environment, noise has no effect on speech or has little effect. Therefore, the detection limit of language clarity needs to be greater than the minimum value in the language clarity data in the above table. At the same time, according to the control requirements in step S3, when the on-site requirements are based on the 3δ control requirements, the language clarity must be greater than the difference between the average value μ and 3 times the standard deviation δ calculated above. When the on-site requirements are based on the 6δ control requirements, the language clarity must be greater than the difference between the average value and 6 times the standard deviation δ calculated above. That is: when the on-site requirements are based on the 3δ control requirements, the limit of language clarity is greater than or equal to the smaller value of the language clarity MIN value and μ-3δ, or when the on-site requirements are based on the 6 control requirements, the limit of language clarity is greater than or equal to the smaller value of the language clarity MIN value and μ-6δ. The evaluation principles of loudness, sound pressure level, sharpness and 1 / 3 octave are opposite to those of language clarity. The smaller the limit, the better, which means low noise and high user acceptance rate. Therefore, when the site requirements are based on the 3δ control requirements, the respective limits of loudness, sound pressure level, sharpness and 1 / 3 octave shall be less than or equal to the larger value of the MIX value of the corresponding data and μ+3δ; when the site requirements are based on the 6 control requirements, the respective limits of loudness, sound pressure level, sharpness and 1 / 3 octave shall be less than or equal to the larger value of the MIX value in the respective data and μ+6δ.

[0047] Through the above evaluation principles and statistical calculation of the test data, the limit values ​​of each sound quality index are obtained, as shown in the following table:

[0048]

[0049] After the above steps, according to the 3δ or 6δ control principle, the limit values ​​(target values) of various indicators such as speech clarity, loudness, sharpness, sound pressure level, etc. in the noise sound quality are obtained through online testing. In actual production testing, the values ​​of each indicator obtained through testing are compared with the above-mentioned limit values ​​to determine whether the sound quality of the product is qualified.

[0050] The present invention further provides a method for determining whether a product is qualified and determining the fault type according to the limit value determined by the above method.

[0051] Collect products as test samples at the production site, or directly use unqualified noise test products or noise faulty machines returned for after-sales repair (preferably without other faults) as test samples. Before the test, it is impossible to determine which item of the sound quality exceeds the standard and does not meet the standard requirements. Therefore, it is necessary to use the faulty machine as a sample to enter the sound quality test room to test all sound quality indicators, obtain the corresponding test signal of each indicator, convert the noise signal into the corresponding 1 / 3 octave spectrum signal, obtain multiple continuous peaks of each indicator, filter and replay the formed spectrum signal, filter the outer peaks of each indicator one by one, determine the sound quality frequency band where the noise fault is located, and when it passes By filtering a certain peak, the noise problem is solved. It can be seen that the noise problem is related to the frequency band where the filtered peak is located, so as to determine the sound quality fault category of the noise fault, repair the noise fault machine, obtain the product fault category, detect and repair a large number of different types of noise fault machines, record and count the sound quality fault information and product fault category, form a database of noise and fault category, and statistically sample the test data of the product according to the 3δ or 6δ principle, and count the frequency band range where the common sound quality indicators have the largest data. The maximum value of the data in each frequency band range is plotted to draw the envelope limit spectrum of the noise. The envelope limit spectrum can be two, among which, if Figure 2 As shown, including the upper black unqualified line and the lighter warning line, the noise sound quality test is carried out on the products produced on site, and the test data is compared with the limit value obtained by the sound quality test limit determination method as described above to determine whether the product sound quality is qualified. For unqualified products, the sound quality data is converted into a spectrum signal, which is compared with the envelope limit spectrum, and the fault category is obtained after comparison with the database information. Furthermore, the envelope limit spectrum is two, including the unqualified line and the warning line, and the difference between the unqualified line and the warning line is 2db. During the on-site test, if any frequency band detected by each sound quality indicator exceeds the unqualified line, it will be displayed as a red light. If a peak sound is configured, it will exceed the warning line and display a yellow light. If it does not exceed the warning line, it will display a green light.

[0052] For example, Figure 1 As shown, the subjective evaluation of range hood No. 1 is that there is obvious buzzing / beeping sound. Through the filtering function, it is found that the statistical frequency band of this feature is 200-500Hz, which has relatively high energy. The maximum peak of this frequency band is about 45dBA. The subjective evaluation of range hood No. 2 is that there is obvious metallic hissing sound. Through the filtering function, it is found that the statistical frequency band of this feature is above 2000Hz, which has relatively high energy. The maximum peak of this frequency band is about 43dBA.

[0053] Through the test statistics of multiple different characteristic noises, we can approximately draw the following Figure 2 The envelope limit spectrum of an entire frequency band shown can be used as a standard for detecting problematic range hoods (3σ normal distribution statistical method).

[0054] If a product fails the inspection, the energy statistics of the 1 / 3 octave band are compared with the system database to identify the fault category and determine the fault type, so as to provide a judgment result and repair suggestion.

[0055] In summary, the sound quality detection limit confirmation method and the sound quality detection fault confirmation method provided by the present invention have the following advantages compared with the prior art:

[0056] 1. The sample size is large enough and the data is sufficient to obtain accurate data;

[0057] 2. Use score limits to effectively remove unqualified data and ensure the validity of the data;

[0058] 3. The confirmation method is simple and effective limit values ​​can be obtained;

[0059] 4. The fault confirmation method is simple.

[0060] As described above, similar technical solutions can be derived from the given solution content. However, any simple modification, equivalent changes and modifications made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A method for confirming a sound quality detection limit, characterized in that: The following steps are included: S1, two groups of users conduct subjective evaluation on the noise crosstalk of N test samples according to the evaluation criteria and record the score of each sample; S2, calculating the average score of the subjective evaluation of noise of each sample, sending qualified samples with an average score above a predetermined value n to the noise detection room for sound quality detection using a noise sound quality detection device, and recording the data of various indicators of sound quality of each sample respectively; S3, statistically analyzing the recorded data of each indicator, counting the maximum and minimum values ​​of each sound quality detection data indicator and calculating the average value μ and standard deviation δ of each data; According to the control requirements, perform numerical calculations of 3δ or 6δ; S4, determine the evaluation principles of various sound qualities, and determine the target limit values ​​of various sound quality data based on the data calculated in S3; The sound quality detection data includes speech clarity, loudness, sound pressure level, sharpness and 1 / 3 octave, wherein, according to the control requirements in step S3, when the site requirements are based on the 3δ control requirements, the limit value of speech clarity is greater than or equal to the smaller value of the minimum speech clarity and μ-3δ, or when the site requirements are based on the 6δ control requirements, the limit value of speech clarity is greater than or equal to the smaller value of the minimum speech clarity and μ-6δ; The evaluation principle for loudness, sound pressure level, sharpness and 1 / 3 octave is that when the site is required to comply with the 3δ control requirements, the respective limits of loudness, sound pressure level, sharpness and 1 / 3 octave shall be less than or equal to the maximum value of the corresponding data.

2. The method for confirming the sound quality detection limit value according to claim 1, characterized in that: In step S1, the test sample size N≥1000.

3. The method for confirming the sound quality detection limit value according to claim 1, characterized in that: In step S2, a ten-point system is adopted, and the predetermined value of subjective evaluation qualification n≥5.

4. The method for confirming the sound quality detection limit value according to claim 1, characterized in that: The standard deviation δ is given by the formula It is calculated that, where N is the sample size, i is any sample from 1 to N samples, and μ is the arithmetic mean of the N test data.

5. A method for confirming a sound quality detection fault, characterized in that: Perform sound quality noise detection on the noise fault machine, convert the noise signal into the corresponding 1 / 3 octave frequency spectrum signal, filter and replay the formed frequency spectrum signal, determine the frequency band where the noise fault is located, thereby determining the sound quality fault category of the noise fault, repair the noise fault machine, obtain the product fault category, detect and repair a large number of different types of noise fault machines, record and count the sound quality fault information and product fault categories accordingly, form a database of noise and fault categories, and draw the envelope limit spectrum of the noise based on the test data as a detection standard, perform noise sound quality detection on the products produced on site, compare the detection data with the limit value obtained by the sound quality detection limit confirmation method as described in any one of claims 1 to 4, determine whether the product sound quality is qualified, and for unqualified products, convert the sound quality data into a frequency spectrum signal, and obtain the fault category after comparing it with the database information.

6. A sound quality detection fault confirmation method as claimed in claim 5, characterized in that: There are two envelope limit spectrums, including a failure line and a warning line.

7. A sound quality detection fault confirmation method as claimed in claim 6, characterized in that: The gap between the unqualified line and the warning line is 2db.

Images