A method for improving the accuracy of DPOAE test using spectral estimation method

Abstract

The present invention provides a method for improving the accuracy of DPOAE test by using spectrum estimation method, and belongs to the technical field of otoacoustic emission detection. The method for improving the accuracy of DPOAE test by using spectrum estimation method comprises the following steps: S1: collecting the original signal of preset length; S2: performing spectrum estimation on the original signal, obtaining the response frequency of the original signal and the frequency extreme value of the original signal within the preset value range of the response frequency; S3: obtaining the frequency difference according to the frequency extreme value and the response frequency, if the frequency difference is greater than the preset frequency threshold, then performing coherent algorithm calculation on the signal to obtain the DPOAE response, if the frequency difference is not greater than the preset frequency threshold, then performing coherent algorithm calculation on the signal to obtain the DPOAE response. The present invention weakens the influence of transient noise or other noise signals on the coherent averaging algorithm, improves the signal-to-noise ratio in the presence of noise, and thus improves the test efficiency and accuracy of otoacoustic emissions.

Description

Technical Field

[0001] The present invention relates to the technical field of otoacoustic emission detection, and in particular to a method for improving the accuracy of DPOAE testing by using a spectrum estimation method. Background Art

[0002] Since the DPOAEs signal is very weak and basically submerged in the noise, the current common method to extract the effective DPOAEs signal is the coherent averaging method. When the noise occurs randomly, the coherent averaging method can effectively improve the signal-to-noise ratio of the signal. However, during the measurement process, other transient noise signals will occur. This situation only occurs in one or two tests by chance. If this kind of data is also added to the coherent averaging algorithm, the signal-to-noise ratio of the DPOAEs signal will be greatly reduced. Summary of the invention

[0003] The purpose of the present invention is to propose a method for improving the accuracy of DPOAE testing by using spectral estimation method in view of the above-mentioned deficiencies in the prior art.

[0004] The present invention proposes a method for improving the accuracy of DPOAE testing by using a spectral estimation method, comprising the following steps:

[0005] S1: Collect the original signal of preset length;

[0006] S2: performing frequency spectrum estimation on the original signal to obtain the response frequency of the original signal and the frequency extreme value of the original signal within a preset value range of the response frequency;

[0007] S3: Obtain a frequency difference according to the frequency extreme value and the response frequency. If the frequency difference is greater than a preset frequency threshold, the signal is attenuated and then a coherent algorithm is performed to obtain a DPOAE response. If the frequency difference is not greater than the preset frequency threshold, the signal is performed a coherent algorithm to obtain a DPOAE response.

[0008] Furthermore, step S2 includes: obtaining the frequency value of F1 in the original signal according to the selected frequency type of the original signal, and calculating the response frequency according to the ratio of F2 / F1 and the formula 2F1-F2.

[0009] Furthermore, step S2 specifically includes:

[0010] S21: Perform Hanning window processing on the original signal to obtain a first processed signal;

[0011] S22: Performing FFT transformation on the first processed signal to calculate and obtain each frequency amplitude;

[0012] S23: searching for the maximum value of the frequency amplitude corresponding to each frequency point in the range of a first preset number of frequency points on the left to a second preset number of frequency points on the right of the first processed signal in the response frequency segment, and obtaining the frequency K of the maximum frequency point corresponding to the maximum value of the frequency amplitude;

[0013] S24: Obtain the frequency extreme value according to the frequency K.

[0014] Furthermore, step S3 specifically includes:

[0015] S31: Perform standard deviation calculation on the frequency extreme value and the response frequency to obtain a frequency difference Δf;

[0016] S32: setting a corresponding preset frequency threshold when the original signal is in the response frequency segment, and the difference between the frequency amplitude corresponding to the preset frequency threshold and the frequency amplitude corresponding to the response frequency is within a preset amplitude;

[0017] S33: performing a comparison operation on the frequency difference Δf and a preset frequency threshold;

[0018] S34: If the frequency difference Δf is greater than the preset frequency threshold, the signal is attenuated and then a coherent algorithm is calculated to obtain a DPOAE response; if the frequency difference Δf is not greater than the preset frequency threshold, the signal is calculated by a coherent algorithm to obtain a DPOAE response.

[0019] Further, in step S23, the first preset number and the second preset number are both 3.

[0020] Further, step S21 includes: obtaining the Hanning window Where n = 0, 1, ..., N-1, a = 0.5, the spectrum function is obtained based on the Hanning window in is the Dirichlet kernel, and the main lobe function is obtained based on the spectrum function The original signal is processed based on the main lobe function to obtain a first processed signal.

[0021] Furthermore, step S24 specifically includes:

[0022] S241: By formula Get the frequency correction value ΔK, where Y K is the frequency amplitude of the maximum frequency point, Y K-1 is the frequency amplitude on the left side of the maximum frequency point, Y K+1 is the frequency amplitude on the right side of the maximum frequency point;

[0023] S242: According to formula f m =K+ΔK calculates the frequency extreme value f m .

[0024] The method of improving the accuracy of DPOAE testing by using the spectrum estimation method of the present invention has the following beneficial effects:

[0025] The spectrum estimation method is used to identify whether there is transient noise or other noise in this coherent calculation. If there is interference, the influence of the collected data on the coherent averaging is weakened; if not, normal coherent averaging calculation is performed to weaken the influence of transient noise or other noise signals on the coherent averaging algorithm and improve the signal-to-noise ratio in the presence of noise, thereby improving the test efficiency and accuracy of otoacoustic emissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings, which are incorporated into the specification and constitute a part of the specification, illustrate embodiments of the present invention and are used together with the description to explain the principles of the present invention. In these drawings, similar reference numerals are used to represent similar elements. The drawings described below are some embodiments of the present invention, but not all embodiments. For those of ordinary skill in the art, other drawings can be obtained from these drawings without paying creative work.

[0027] Figure 1 The present invention is a flowchart of a method for improving the accuracy of DPOAE testing by using a spectral estimation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily without conflict.

[0029] See also Figure 1 A method for improving the accuracy of DPOAE testing by using a spectral estimation method according to an embodiment of the present invention comprises the following steps:

[0030] S1: Collect the original signal of preset length;

[0031] S2: performing frequency spectrum estimation on the original signal to obtain the response frequency of the original signal and the frequency extreme value of the original signal within a preset value range of the response frequency;

[0032] S3: Obtain a frequency difference according to the frequency extreme value and the response frequency. If the frequency difference is greater than a preset frequency threshold, the signal is attenuated and then a coherent algorithm is performed to obtain a DPOAE response. If the frequency difference is not greater than the preset frequency threshold, the signal is performed a coherent algorithm to obtain a DPOAE response.

[0033] Specifically, the present application belongs to a method for detecting otoacoustic emissions, which is applied to instruments with otoacoustic emission detection functions. The present application is mainly used in the detection of distortion product otoacoustic emissions (DPOAE), mainly to remove transient noise during the DPOAE test. The recorded evoked response signal x(n, i) is composed of the actual signal s(n, i) and the noise signal u(n, i) added together, that is: x(n, i) = s(n, i) + u(n, i); i = 1, 2, ..., M, where M represents the total number of stimuli and i is the sequence number of the stimulus. For each stimulus, a response x(n, i) will be generated, in which the noise is often stronger than the response, and the noise is basically random, so the x(n, i) obtained each time is different, so it is difficult to judge the shape of the signal s(n) from a single record. However, because OAE is repeatable, when the stimulus conditions remain unchanged, s(n, i) remains basically unchanged each time. Therefore, it can be considered that under the same stimulus conditions, s(n) can be approximately considered as a certain signal. Assume u(n) is a stationary random signal with zero mean and variance w^2, and for each stimulus, they are independent of each other, that is: E{u(n,i)u(n,j)}=0. When n!=j, add the corresponding M records of x(n) and take the average: The operation of this formula is called "coherent averaging", and the degree of smoothing of the signal depends on the size of the stimulus M at this time. If the power of s(n) is denoted as P, then for each stimulus, the signal-to-noise ratio of x(n,i) is P / w^2. After averaging M samples, the power of the signal is still P, the mean of the noise is still zero, but the variance becomes w^2 / M. In this way, the signal-to-noise ratio of the signal becomes: SNR = P / (w^2 / M) = M*P / w^2, which is M times higher than the signal-to-noise ratio without coherent averaging. In the formula The noise signal u(n, i) is required to be random, and the coherent average will cancel the noise signal.

[0034] Specifically, in the actual measurement process, other transient noise signals may occur, such as friction between the test line and clothing, swallowing of saliva, environmental noise, etc. This signal does not always exist, but only occurs in one or two tests by chance. If this data is also added to In the process of DPOAEs, the signal-to-noise ratio will be greatly reduced, so that the signal-to-noise ratio must be improved again by increasing the number of coherent averaging. The measurement time becomes longer.

[0035] Specifically, the present application belongs to a method for detecting otoacoustic emissions, which is applied to instruments with otoacoustic emission detection functions. The present invention is mainly used in the detection of distortion product otoacoustic emissions (DPOAE), and proposes an improved coherent averaging algorithm to remove noise in a specific frequency band during the DPOAE test, improve the accuracy of the DPOAE test, and avoid false positives and false negatives. The spectrum estimation method is used to identify whether there is transient noise or other noise in this coherent calculation. If there is interference, the influence of the collected data on the coherent averaging is weakened; if not, normal coherent averaging is performed. The influence of transient noise or other noise signals on the coherent averaging algorithm is weakened, and the signal-to-noise ratio in the presence of noise is improved, thereby improving the test efficiency and accuracy of otoacoustic emissions.

[0036] Specifically, a normal DPOAE test generally consists of four frequency tests, namely 2k, 3k, 4k, and 5k.

[0037] The following examples all take a 2k frequency test as an example, which makes the explanation method clearer.

[0038] In principle, F2 / F1 should be between 1.2 and 1.25. For the convenience of calculation, we choose the ratio of 1.22, then F1 = 1640Hz, and the response frequency of DPOAE is 2F1-F2 = 2*1640-2000 = 1280Hz.

[0039] The idea of ​​the present invention is to perform spectrum estimation on the collected signal and make judgments near the response frequency. If the frequency extreme value near the response frequency changes significantly, it means that it is interfered by the outside world and will affect the subsequent data calculation. Then the coherent algorithm calculation is performed after the measurement data is attenuated. Continue the subsequent test calculation. By judging the frequency extreme value, the interfering signal in the coherent average calculation is eliminated, and the response signal is accurately identified. Prevent false negative and false positive situations caused by noise signals.

[0040] Specifically, in the process of audio analysis, the most commonly used method is the Fourier transform method. Fast Fourier transform (FFT) is a commonly used algorithm in audio analysis. Fast Fourier transform has fence effect and spectrum leakage phenomenon, which makes the calculated signal parameters, namely frequency, amplitude and phase, inaccurate. For the traditional method of using FFT for audio analysis, the magnitude of the frequency error depends on whether the actual frequency value just falls on the corresponding frequency channel of the N-point fast Fourier transform. If the actual frequency falls between two adjacent frequency channels, the corresponding error will be relatively large, and the error amount is uncontrollable. At the same time, the magnitude of the frequency error also depends on the number of FFT points. The smaller the number of FFT points, the further the directly calculated frequency error will deteriorate, and it will not be able to meet the requirements of accurate frequency measurement.

[0041] FFT is a commonly used algorithm in audio analysis and a fast algorithm for discrete Fourier transform (DFT). It decomposes the DFT of a long sequence into the DFT of a short sequence for calculation, thereby greatly reducing the amount of calculation. The discrete spectrum of a finite-length signal can be calculated using the fast Fourier transform.

[0042] Step S2 includes: obtaining the frequency value of F1 in the original signal according to the selected frequency type of the original signal, and calculating the response frequency according to the ratio of F2 / F1 and the formula 2F1-F2.

[0043] Step S2 specifically includes:

[0044] S21: Perform Hanning window processing on the original signal to obtain a first processed signal;

[0045] S22: Performing FFT transformation on the first processed signal to calculate and obtain each frequency amplitude;

[0046] S23: searching for the maximum value of the frequency amplitude corresponding to each frequency point in the range of a first preset number of frequency points on the left to a second preset number of frequency points on the right of the first processed signal in the response frequency segment, and obtaining the frequency K of the maximum frequency point corresponding to the maximum value of the frequency amplitude;

[0047] S24: Obtain the frequency extreme value according to the frequency K.

[0048] Step S3 specifically includes:

[0049] S31: Perform standard deviation calculation on the frequency extreme value and the response frequency to obtain a frequency difference Δf;

[0050] S32: setting a corresponding preset frequency threshold when the original signal is in the response frequency segment, and the difference between the frequency amplitude corresponding to the preset frequency threshold and the frequency amplitude corresponding to the response frequency is within a preset amplitude;

[0051] S33: performing a comparison operation on the frequency difference Δf and a preset frequency threshold;

[0052] S34: If the frequency difference Δf is greater than the preset frequency threshold, the signal is attenuated and then a coherent algorithm is calculated to obtain a DPOAE response; if the frequency difference Δf is not greater than the preset frequency threshold, the signal is calculated by a coherent algorithm to obtain a DPOAE response.

[0053] The coherent algorithm calculation to obtain the DPOAE response in step S3 specifically includes: Correction acquisition According to the formula Calculations are performed to obtain DPOAE responses, where x(n, i) represents the induced response signal, s(n, i) represents the actual signal included in the response x(n) generated after each stimulation, u(n, i) represents the conventional noise signal included in the response x(n) generated after each stimulation, and n(n, i) represents the transient noise included in the response x(n) generated after each stimulation.

[0054] In step S23 , the first preset number and the second preset number are both 3.

[0055] Step S21 includes: obtaining the Hanning window Where n = 0, 1, ..., N-1, a = 0.5, the spectrum function is obtained based on the Hanning window in is the Dirichlet kernel, and the main lobe function is obtained based on the spectrum function The original signal is processed based on the main lobe function to obtain a first processed signal.

[0056] Specifically, in spectrum analysis, the Hanning window is defined as Its spectrum function is: in is the Dirichlet kernel, when a=0.5 it is the Hanning window, and when a=0.54 it is the Hamming window.

[0057] The step S21 of obtaining the transformation function based on the spectrum function specifically includes: transforming the spectrum function into a modulus function while ignoring the phase difference When x→0, Y→a, x→±1, applying Robita's rule, we have Substitute a=0.5 into the mode function to convert the mode function into the main lobe function Specifically, The three items in the brackets have phase differences. When N is large, the phase difference can be ignored and the modulus function can be as follows:

[0058] In this formula, when x→0, Y→a, x→±1, Robita's rule is applied. Substitute a=0.5 into the mode function to convert the mode function into the main lobe function

[0059] Step S24 specifically includes:

[0060] S241: By formula Get the frequency correction value ΔK, where Y K is the frequency amplitude of the maximum frequency point, Y K-1 is the frequency amplitude on the left side of the maximum frequency point, Y K+1 is the frequency amplitude on the right side of the maximum frequency point;

[0061] S242: According to formula fm =K+ΔK calculates the frequency extreme value f m .

[0062] Step S241 also includes: based on the main lobe function It is concluded that xf(x)+(x+1)f(x+1)=f(x)-f(x-1), (x-1)f(x)+(x+2)f(x+1)=0; the two spectral lines Y in the main lobe of the amplitude spectrum are K , Y K+1 Translate and obtain the corresponding center of gravity coordinate x0, The two spectral lines Y in the main lobe of the amplitude spectrum K-1 , Y K Translate and obtain the corresponding center of gravity coordinates The frequency correction value can be obtained based on the two centroid coordinates calculated above. where Y K is the maximum value among the frequency values, Y K-1 is the amplitude on the left side of the maximum frequency point, Y K+1 It is the amplitude on the right side of the maximum frequency point.

[0063] Specifically, according to the formula The frequency extremes in any frequency range can be calculated. Since the frequency resolution of FFT transformation is limited by the sampling rate and FFT line bundle, it cannot be infinitely small. In this case, we can use the formula The result of the derivation process is used to perform frequency correction and find the frequency extreme value of the required frequency band.

[0064] Specifically, the following specific implementation steps can be adopted:

[0065] ① Collect a signal I(n) of a certain length;

[0066] ② Perform Hanning window processing on the collected signal;

[0067] ③Perform FFT transformation to calculate the amplitude of each frequency;

[0068] ④Search between the 3 points on the left and the 3 points on the right in the response frequency band, obtain the maximum frequency amplitude, and record the frequency K corresponding to the maximum value;

[0069] 1) According to the maximum value position, determine the size of the left and right sides of the maximum value

[0070] According to the formula Calculate the frequency extreme value f after correction of the response frequency band m =K+ΔK.

[0071] ⑤The extreme frequency and response frequency f rPerform standard deviation calculation to obtain Δf. When this value is greater than the preset frequency threshold, the collected data is attenuated.

[0072] ⑥ Perform coherent averaging to calculate DPOAE response, It can be seen from the formula that the transient noise is not taken into account. After the formula is corrected Transient noise n(n, i) is unstable and can be large or small. Therefore, by using the above steps to weaken the n(n, i) signal, a relatively accurate s(n, i) can be obtained, where s(n, i) represents the actual signal included in the response x(n) generated after each stimulus, u(n, i) represents the conventional noise signal included in the response x(n) generated after each stimulus, and n(n, i) represents the transient noise included in the response x(n) generated after each stimulus.

[0073] The method of the present invention calculates the response frequency f by using the spectrum estimation method. r The frequency extreme value f near m , determine whether this extreme value is at the response frequency f r If it is within a specific range, it is considered to be an accurate signal and participates in the coherent average calculation; if it is not within a specific frequency range, the amplitude of the collected data is first attenuated and then participates in the coherent average calculation. When there is an interference signal, this method is used to identify it and attenuate the collected data. The influence of the interference signal on the amplitude of the measurement of a specific frequency band is greatly reduced, which improves the accuracy of the DPOAE test.

[0074] The contents described above may be implemented individually or in combination in various ways, and these variations are all within the protection scope of the present invention.

[0075] It should be noted that, in the description of the present application, the terms "upper end", "lower end" and "bottom end" indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of the application is usually placed when in use, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the presence of other identical elements in the process, method, article or device comprising the elements.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for improving the accuracy of distortion product otoacoustic emission testing using spectral estimation method, characterized in that: The steps include: S1: Collect the original signal of preset length; S2: performing frequency spectrum estimation on the original signal to obtain the response frequency of the original signal and the frequency extreme value of the original signal within a preset value range of the response frequency; S3: Obtain a frequency difference according to the frequency extreme value and the response frequency. If the frequency difference is greater than a preset frequency threshold, the signal is attenuated and then a coherent algorithm is used to calculate the distortion product otoacoustic emission response. If the frequency difference is not greater than the preset frequency threshold, the signal is used to calculate the distortion product otoacoustic emission response. Step S2 specifically includes: S21: Perform Hanning window processing on the original signal to obtain a first processed signal; S22: Performing FFT transformation on the first processed signal to calculate and obtain each frequency amplitude; S23: searching for the maximum value of the frequency amplitude corresponding to each frequency point in the range of a first preset number of frequency points on the left to a second preset number of frequency points on the right of the first processed signal in the response frequency segment, and obtaining the frequency K of the maximum frequency point corresponding to the maximum value of the frequency amplitude; S24: Obtain the frequency extreme value according to the frequency K.

2. The method for improving the accuracy of distortion product otoacoustic emission testing by using spectral estimation method as claimed in claim 1, characterized in that: Step S2 includes: obtaining the frequency value of F1 in the original signal according to the selected frequency type of the original signal, and calculating the response frequency according to the ratio of F2 / F1 and the formula 2F1-F2.

3. A method for improving the accuracy of distortion product otoacoustic emission testing using spectral estimation method as claimed in claim 1 or 2, characterized in that: Step S3 specifically includes: S31: Perform standard deviation calculation on the frequency extreme value and the response frequency to obtain a frequency difference Δf; S32: setting a corresponding preset frequency threshold when the original signal is in the response frequency segment, and the difference between the frequency amplitude corresponding to the preset frequency threshold and the frequency amplitude corresponding to the response frequency is within a preset amplitude; S33: performing a comparison operation on the frequency difference Δf and a preset frequency threshold; S34: If the frequency difference Δf is greater than the preset frequency threshold, the signal is attenuated and then a coherent algorithm is used to calculate the distortion product otoacoustic emission response; if the frequency difference Δf is not greater than the preset frequency threshold, the signal is used to calculate the distortion product otoacoustic emission response.

4. The method for improving the accuracy of distortion product otoacoustic emission testing by using spectral estimation method as claimed in claim 2, characterized in that: In step S23 , the first preset number and the second preset number are both 3.

5. The method for improving the accuracy of distortion product otoacoustic emission testing by using spectral estimation method as claimed in claim 2, characterized in that: Step S21 includes: obtaining the Hanning window Where n = 0, 1, ..., N-1, a = 0.5, the spectrum function is obtained based on the Hanning window in is the Dirichlet kernel, and the main lobe function is obtained based on the spectrum function The original signal is processed based on the main lobe function to obtain a first processed signal.

6. A method for improving the accuracy of distortion product otoacoustic emission testing using spectral estimation method as claimed in claim 4, characterized in that: Step S24 specifically includes: S241: By formula Get the frequency correction value ΔK, where Y K is the frequency amplitude of the maximum frequency point, Y K-1 is the frequency amplitude on the left side of the maximum frequency point, Y K+1 is the frequency amplitude on the right side of the maximum frequency point; S242: According to formula f m =K+ΔK calculates the frequency extreme value f m .

Images