Constant-speed noise subjective and objective evaluation association method

Abstract

The invention belongs to the technical field of vehicle noise testing, and particularly relates to a constant-speed noise subjective and objective evaluation association method, which comprises the following steps of: on the basis of steady-state noise sample data under historical constant-speed and high-speed working conditions, obtaining a subjective evaluation scoring model based on a power exponent and a sound pressure level; in the testing process, after subjective evaluation is carried out on the standard sound source, a selected standard sound source sample plate is obtained; and according to real vehicle test data, obtaining a real vehicle sound pressure level and a power exponent parameter, and inputting the real vehicle sound pressure level and the power exponent parameter into a conversion formula associated with subjective and objective evaluation to obtain a subjective level score. According to the technical scheme, the subjective evaluation scoring model based on the power exponent and the sound pressure level is analyzed and summarized by starting from the steady-state noise sample under the constant-speed and high-speed working conditions and combining the self-power spectrum distribution characteristics and the subjective evaluation grade evaluation score, the accuracy and feasibility of the subjective evaluation scoring model are verified, and the technical problem of subjective objective results is solved.

Description

Technical Field

[0001] This invention belongs to the field of vehicle noise testing technology, specifically relating to a method for the correlation between subjective and objective evaluation of uniform noise. Background Technology

[0002] The essence of vehicle steady-state noise quality evaluation is to comprehensively consider the objective evaluation of the differences in human hearing caused by noise, and to reflect the subjective perception of noise by the human ear as completely as possible in a simple index. However, with the increase of objective indices as mentioned above, the difficulty of analysis and statistical complexity have increased significantly. For example, noise spectrum sound pressure level cannot clearly reflect the high-frequency characteristics of noise because the energy is mainly concentrated in the mid-low frequency band; speech intelligibility cannot reflect the low-frequency characteristics of noise because the calculation frequency band is 200~6300Hz; loudness mainly assesses the strength of noise; sharpness mainly assesses the proportion of high-frequency components in the sound; and roughness mainly assesses the modulation phenomenon of 20~300Hz in noise. Therefore, in order to make the objective indices of steady-state noise reflect subjective perception in a simple way, reasonable dimensionality reduction and statistical methods are needed.

[0003] Existing technical solutions lack a clear and efficient evaluation method, and cannot closely integrate objective test results with subjective evaluation scores. They mainly evaluate based on segments such as sound pressure level, speech intelligibility, and sound quality parameters, and cannot provide an overall subjective rating.

[0004] The existing technical solutions have the following drawbacks: there is a contradiction between the objective result and the subjective result due to a single indicator, and the subjective score may be affected by the individual's state and experience. Summary of the Invention

[0005] The purpose of this invention is to provide a method for correlating subjective and objective evaluations of uniform noise, so as to solve the problem that objective test data cannot provide subjective evaluation levels.

[0006] To achieve the above objectives, this application employs the following technical solution:

[0007] A method for subjective and objective evaluation correlation of uniform noise includes the following steps:

[0008] S1. Based on the steady-state noise sample data under historical uniform high-speed conditions, and combined with the self-power spectrum distribution characteristics and subjective evaluation score of each noise sample data, a subjective evaluation scoring model based on power exponent and sound pressure level is derived.

[0009] S2. During the testing process, after subjective evaluation of the standard sound source, the selected standard sound source template is obtained;

[0010] S3. Input the subjective evaluation score corresponding to the standard sound source template, the step size of the subjective score of the standard sound source, the change in sound pressure level corresponding to the step size of the subjective score of the standard sound source, the change in power exponent corresponding to the step size of the subjective score of the standard sound source, and the A-weighted sound pressure level corresponding to the objective sound source into the subjective evaluation scoring model to obtain the conversion formula for the relationship between subjective and objective evaluation.

[0011] S4. Based on the actual vehicle test data, obtain the actual vehicle sound pressure level and power exponent parameters. Input the actual vehicle sound pressure level and power exponent parameters into the conversion formula of subjective and objective evaluation in step S3 to obtain the subjective level score.

[0012] Furthermore, the subjective evaluation scoring model is as follows:

[0013] ,in, Subjective scoring based on standard audio sources. The step size for the subjective scoring of the standard audio source. To obtain the objective A-weighted sound pressure level corresponding to the actual vehicle noise to be converted, To objectively correspond the sound pressure level of the standard sound source to the A-weighted sound pressure level, This represents the change in sound pressure level corresponding to the step size of the subjective scoring scale for the standard sound source. The power exponent is the objective corresponding to the actual vehicle noise to be converted. The standard sound source objectively corresponds to the power index. This represents the power-law change in the step size of the standard audio source's subjective scoring scale.

[0014] Furthermore, in step S2, a set number of standard sound sources are subjectively evaluated, and the standard sound source that best meets the test requirements is selected as the standard sound source template.

[0015] Furthermore, standard sound sources are obtained through filtering, and standard exponential sound is obtained through filtering.

[0016] Furthermore, in step S4, the steps for obtaining the actual vehicle sound pressure level and power exponent parameters based on the actual vehicle test data are as follows:

[0017] S41. Obtain the noise test signal of the vehicle under steady-state conditions at high speed and constant speed.

[0018] S42. Convert the obtained noise test signal into frequency domain analysis results to obtain 1 / n octave bands, so as to obtain the energy distribution curve and RMS value of the vehicle model.

[0019] S43. Perform univariate linear regression on the obtained 1 / n octave band curve and solve the regression to obtain the power exponent α and the goodness of fit R. 2 ;

[0020] S44. Solve using the formula and convert the objective test results into subjective evaluation scores.

[0021] Furthermore, the subjective evaluation method is a rating scale, with rating levels including unacceptable, terrible, tolerable, acceptable, reasonable, good, excellent, and outstanding.

[0022] The beneficial effects of this invention are:

[0023] This technical solution starts with steady-state noise samples under uniform high-speed operating conditions, combines their self-power spectrum distribution characteristics and subjective evaluation scores, analyzes and summarizes a subjective evaluation scoring model based on power exponent and sound pressure level, and verifies its accuracy and feasibility, thus solving the technical problem of subjectifying objective results. Attached Figure Description

[0024] Figure 1 This is a framework diagram of the method for the correlation between subjective and objective evaluation of uniform noise in this invention.

[0025] Figure 2 This is a spectral feature diagram of two sets of data processed by univariate linear regression in this invention.

[0026] Figure 3 This is a schematic diagram of the 1 / 3 octave band spectrum of a portion of the sample vehicles after univariate regression in this invention.

[0027] Figure 4 This is a comparison chart of the subjective evaluation scores and sound pressure levels of the sample vehicles under the same power exponent conditions.

[0028] Figure 5 This is a comparison chart of the subjective evaluation scores and power exponents of sample vehicles under the same sound pressure level conditions.

[0029] Figure 6 This is a schematic diagram of the 1 / 3 octave band spectrum of a 67 dBA standard power exponent sound. Detailed Implementation

[0030] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings. The following embodiments are merely exemplary and can only be used to explain and illustrate the technical solution of the present invention, and should not be construed as limiting the technical solution of the present invention.

[0031] like Figures 1 to 6 As shown, this application provides a method for the subjective and objective evaluation correlation of uniform noise, including the following steps:

[0032] S1. Based on the steady-state noise sample data under historical uniform high-speed operating conditions, and combined with the self-power spectrum distribution characteristics and subjective evaluation score corresponding to each noise sample data, a subjective evaluation scoring model based on power exponent and sound pressure level is derived.

[0033] This invention is a subjective evaluation scoring model that converts objective test data into subjective evaluation levels. It is calculated by combining standard sound sources and objective results of the actual vehicle noise to be converted. The core formula is as follows:

[0034] ,in, Subjective scoring based on standard audio sources. The step size for the subjective scoring of the standard audio source. To obtain the objective A-weighted sound pressure level corresponding to the actual vehicle noise to be converted, To objectively correspond the sound pressure level of the standard sound source to the A-weighted sound pressure level, This represents the change in sound pressure level corresponding to the step size of the subjective scoring scale for the standard sound source. The power exponent is the objective corresponding to the actual vehicle noise to be converted. The standard sound source objectively corresponds to the power index. This represents the power-law change in the step size of the standard audio source's subjective scoring scale.

[0035] In practical application, after subjectively evaluating the standard sound source, a selected standard sound source template is obtained (based on power-law priority evaluation), and the subjective evaluation score (S) and step size (Step) corresponding to the template are assigned. RSM ), the change in sound pressure level corresponding to the step size (Δ) SPL_step ), the change in the power exponent corresponding to the step size (Δ) α_step After filling in the core formula, the corresponding solution formula for a certain working condition can be obtained:

[0036] In the formula: the relevant information of the corresponding standard sound source (in combination with subjective and objective evaluation) can be described as a sound source with an A-weighted sound pressure level of 67 dB(A) and a power exponent of 1.4. Its subjective evaluation score is 8. Under the condition of keeping the power exponent unchanged, if the sound pressure level increases by 2 dB, the subjective evaluation score decreases by 0.5. Under the condition of keeping the sound pressure level unchanged, if the power exponent changes by 0.1, the subjective evaluation score decreases by 0.5.

[0037] S2. During the testing process, after subjective evaluation of the standard sound source, the selected standard sound source template is obtained;

[0038] S3. Input the subjective evaluation score corresponding to the standard sound source sample, the step size of the standard sound source subjective evaluation score, the change in sound pressure level corresponding to the step size of the standard sound source subjective evaluation score, the change in power exponent corresponding to the step size of the standard sound source subjective evaluation score, and the A-weighted sound pressure level corresponding to the standard sound source objective evaluation into the subjective evaluation scoring model to obtain the conversion formula for the relationship between subjective and objective evaluation, i.e., the core formula:

[0039] .

[0040] S4. Based on the actual vehicle test data, obtain the actual vehicle sound pressure level and power exponent parameters. Input the actual vehicle sound pressure level and power exponent parameters into the conversion formula of subjective and objective evaluation in step S3 to obtain the subjective level score.

[0041] S41. First, the noise test signal of the vehicle under steady-state conditions at high speed and constant speed should be obtained and converted into frequency domain analysis results to obtain its 1 / 3 octave band curve (theoretically, this method can be used for 1 / n octave bands, but considering the accuracy of numerical solution and the amount of data, it is recommended to use the 1 / 3 octave band curve. The larger n is, the larger the amount of data to be analyzed). This step mainly obtains its energy distribution curve and RMS value. No weighting is required here.

[0042] S42. Perform univariate linear regression on the obtained 1 / 3 octave band curve and solve the regression to obtain the power exponent α and the goodness of fit R. 2 (The standard value should be ≥0.95).

[0043] S43, Combining formulas To ensure that the scoring results are in increments of 0.5, adjustments can be made accordingly when changing the increment. The objective test results are converted into subjective evaluation scores by solving the equation, where: S step0.5 The subjective evaluation score (0.5 scoring step size) can be replaced by Round(2S,0) / 2 in Excel, where S is the subjective evaluation score, LRMS is the total sound pressure level in dB(A), and α is the exponent.

[0044] This application has a certain degree of versatility and can be combined with the subjective evaluation systems of different companies. It can be adjusted based on standard sound samples. In this summary case, for intuitive representation, the step size is 0.5 points, the sound pressure level of the standard sound sample is 67 dB(A), the power exponent α is 1.4, and the subjective evaluation score is 8 points.

[0045] To find the standard value, The segmented summation value uses 0.5 as the step size for the subjective rating level difference.

[0046] In this embodiment, for the purpose of intuitive demonstration, a constant speed condition of 120km / h is used as the sample for analysis. Other high-speed conditions can be referred to the formula. The solution process is performed, and the parameters of RSM (the subjective evaluation method is the Rating Scale Method, RSM) are all calculated based on standard samples. Subjective evaluation ratings for standard samples A-weighted sound pressure level of standard samples For the standard sample unweighted power exponent, For the same power exponent Changes in subjective ratings of standard samples under certain conditions The corresponding change in sound pressure level, Changes in subjective scores of standard samples under the same sound pressure level conditions Corresponding power index Change.

[0047] Standard samples can be obtained through filtering, and filtering pink noise is relatively convenient (the filter is set to a linear filter corresponding to different power exponents). Standard power exponent sound (R 2 >0.99) was obtained using a filtering method.

[0048] Basis and verification process:

[0049] The subjective evaluation method is the Rating Scale Method (RSM). The rating levels are shown in the table below:

[0050]

[0051] By comparing two sets of data from any standard database and performing univariate linear regression on them, the spectral characteristics, under a double logarithmic coordinate system, appear as a straight line with a negative slope and a power exponent, such as... Figure 2 As shown, Figure 2 The spectrum is a 1 / 3 octave band noise spectrum. The legend line marked with "L" represents the result after univariate linear regression. The horizontal axis format is Decade, and the vertical axis format is Log. This spectrum feature satisfies the power-law distribution and has a general pattern.

[0052] Based on the fact that the noise spectrum satisfies a power-law distribution function and that the power-law distribution is scale-free, the relationship between sound pressure and frequency can be characterized as follows:

[0053] P = cf -α (1)

[0054] In the formula, P is the sound pressure level; c is a constant; f is the frequency; and α is the power exponent.

[0055] Let Y = log 10 P, X = log 10 f, C=log 10 c, then formula (1) can be transformed into:

[0056] Y = C - αX (2)

[0057] Definition 1 (Horizontal Coordinate Parameterization)

[0058] The base-10 octave system as specified in GB / T 3241-2010 Octave and Fractional Octave Filters:

[0059] f cn = 10 3M / 10 (3)

[0060] In the formula, f cn is the center frequency; M is the coefficient.

[0061] When M = x / bb is an odd number (4)

[0062] When M = (2x+1) / bb is even (5)

[0063] In the formula, b is the denominator of the octave bandwidth indication value (a fraction of the octave indicated by the reciprocal of a positive integer, denoted as 1 / b, b=1, 3, 6, 12, ...), and x is any positive integer representing the number of octave bands.

[0064] Definition 2 (Vertical Coordinate Parameterization):

[0065] Human hearing is approximately logarithmic, with a scale of dB values. Sound pressure level (SPL) (dB) is expressed in L. p It indicates that:

[0066] L p = 20 log 10 (P / P ref (6)

[0067] In the formula, P is the effective sound pressure level, P ref The reference sound pressure in the air is 20 μPa.

[0068] Definition 3 (Univariate Linear Regression and Regression Solution):

[0069] From formulas (2) and (3), we can see that:

[0070] 20 log 10 (P oct / P ref ) = 20 log 10 (c / P ref ) – 6αM (7)

[0071] Let the constant D = 20 log 10 (c / P ref Substituting M into formulas (4) and (5), we get:

[0072] L oct = D – 6αx / bb is an odd number (8)

[0073] L oct = D – 3α / b - 6αx / bb is an even number (9)

[0074] In the formula, P oct L is the effective sound pressure level (octave band). oct Sound pressure level (octave band).

[0075] In formulas (8) and (9), D, b, and α are all non-variables. Therefore, the regression fitting model of sound pressure level and octave band number can be simplified as follows:

[0076] L = A + βx (10)

[0077] In the formula, L is the sound pressure level, A is the intercept of the line on the vertical axis, and β is the slope of the line.

[0078] Therefore, when the power exponent is in the octave band on the horizontal axis, its relationship with the slope is as follows (it can be seen that the relationship between the power exponent and the slope remains unchanged regardless of the octave band division):

[0079] α = -bβ / 6 (11)

[0080] Based on the correlation pattern of univariate linear regression, its slope β, intercept A, and goodness of fit R 2 The solution results are as follows:

[0081] (12)

[0082] (13)

[0083] (14)

[0084] Considering the characteristics of human hearing, one-third octave band is often used as a simplified auditory model in evaluation and analysis, which can reflect the spectral characteristics of noise in more detail. That is, if b=3, then the power exponent α = -β / 2.

[0085] The following sample was used for verification, and patterns can be statistically analyzed:

[0086]

[0087] After a comprehensive evaluation of the standard power-law sound samples, the following conclusions can be drawn:

[0088] (1) In the sound sample with a sound pressure level of 67 dB(A), the power exponent α is 1.4, which is the best and its subjective evaluation score is 8.0;

[0089] (2) In the 65dB(A) and 67dB(A) sound samples, the power exponent α is in the range [1.1, 1.4]. When α increases by 0.1, the subjective evaluation score increases by 0.5. This is mainly reflected in the significant reduction of subjective perception of high-frequency noise components as α increases.

[0090] (3) In the 65dB(A) and 67dB(A) sound samples, the power exponent α is in the range [1.4, 1.6]. When α increases by 0.1, the subjective evaluation score decreases by 0.5. This is mainly manifested in the significant enhancement of the subjective perception of low-frequency noise components as α increases.

[0091] (4) Under the same power exponent α, the lower the sound pressure level, the higher the subjective evaluation score. A decrease of 2 dB in sound pressure level will increase the subjective evaluation score by 0.5.

[0092] The parameters obtained after solving the sample can be referred to in the table below. The analysis shows that after accurately counting and retaining the values ​​in units of 0.5, the objective evaluation results can be directly correlated with the subjective evaluation results. The evaluation method is highly accurate and feasible.

[0093]

[0094] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for subjective and objective evaluation correlation of uniform noise, characterized in that, Includes the following steps: S1. Based on the steady-state noise sample data under historical uniform high-speed conditions, and combined with the self-power spectrum distribution characteristics and subjective evaluation score of each noise sample data, a subjective evaluation scoring model based on power exponent and sound pressure level is derived. S2. During the testing process, after subjective evaluation of the standard sound source, the selected standard sound source template is obtained; S3. Input the subjective evaluation score corresponding to the standard sound source template, the step size of the subjective score of the standard sound source, the change in sound pressure level corresponding to the step size of the subjective score of the standard sound source, the change in power exponent corresponding to the step size of the subjective score of the standard sound source, and the A-weighted sound pressure level corresponding to the objective sound source into the subjective evaluation scoring model to obtain the conversion formula for the relationship between subjective and objective evaluation. S4. Based on the actual vehicle test data, obtain the actual vehicle sound pressure level and power exponent parameters. Input the actual vehicle sound pressure level and power exponent parameters into the conversion formula of subjective and objective evaluation in step S3 to obtain the subjective level score.

2. The method for subjective and objective evaluation correlation of uniform noise according to claim 1, characterized in that, The subjective evaluation scoring model is as follows: ,in, Subjective scoring based on standard audio sources. The step size for the subjective scoring of the standard audio source. To obtain the objective A-weighted sound pressure level corresponding to the actual vehicle noise to be converted, To objectively correspond the sound pressure level of the standard sound source to the A-weighted sound pressure level, This represents the change in sound pressure level corresponding to the step size of the subjective scoring scale for the standard sound source. The power exponent is the objective corresponding to the actual vehicle noise to be converted. The standard sound source objectively corresponds to the power index. This represents the power-law change in the step size of the standard audio source's subjective scoring scale.

3. The method for subjective and objective evaluation correlation of uniform noise according to claim 1, characterized in that, In step S2, a set number of standard sound sources are subjectively evaluated, and the standard sound source that best meets the test requirements is selected as the standard sound source template.

4. The method for subjective and objective evaluation correlation of uniform noise according to claim 3, characterized in that, Standard sound sources are obtained through filtering, and standard exponential sound is obtained through filtering.

5. The method for subjective and objective evaluation correlation of uniform noise according to claim 1, characterized in that, In step S4, the steps for obtaining the actual vehicle sound pressure level and power exponent parameters based on the actual vehicle test data are as follows: S41. Obtain the noise test signal of the vehicle under steady-state conditions at high speed and constant speed. S42. Convert the obtained noise test signal into frequency domain analysis results to obtain 1 / n octave bands, so as to obtain the energy distribution curve and RMS value of the vehicle model. S43. Perform univariate linear regression on the obtained 1 / n octave band curve and solve the regression to obtain the power exponent α and the goodness of fit R. 2 ; S44. Solve using the formula and convert the objective test results into subjective evaluation scores.

6. The method for subjective and objective evaluation correlation of uniform noise according to claim 1, characterized in that, The subjective evaluation method is a rating scale, with rating levels including unacceptable, terrible, tolerable, acceptable, reasonable, good, excellent, and outstanding.

Images