Noise data processing method and apparatus, and electronic device and vehicle

Abstract

A noise data processing method and apparatus, and an electronic device and a vehicle, which relate to the technical field of noise data processing. The method comprises: acquiring first noise data, corresponding to a spectrum component to be determined, in a first time-domain window (101), wherein the length of the first time-domain window is an integer multiple of a step size; determining second noise data corresponding to the spectrum component within the first step size in the first time-domain window, and determining third noise data corresponding to the spectrum component within the last step size in a second time-domain window (102), wherein the second time-domain window is a time-domain window that is obtained after the first time-domain window is moved by the step size; and subtracting the second noise data from the first noise data, and adding the third noise data, so as to obtain fourth noise data corresponding to the spectrum component in the second time-domain window (103). By means of applying the technical solution of the present application, with regard to overlapping data between time-domain windows, the computation of the overlapping data can be reduced, thus saving on computing power.

Description

Cross-reference to related applications

[0001] The application claims the priority of Chinese patent application No. 202310724699.5, filed on June 19, 2023, titled "Noise Data Processing Method and Apparatus, and Electronic Device and Vehicle", the entire content of which is incorporated herein by reference.TECHNICAL FIELD

[0002] The disclosure relates to a field of noise data processing technologies, and relates to, but is not limited to, a noise data processing method, an apparatus, an electronic device and a vehicle.BACKGROUND

[0003] A noise audio generated by in-vehicle devices such as a motor and an engine may be used to analyze whether the devices are running normally. In order to ensure the continuity of data analysis, time-domain windows are overlapped. For example, the noise audio between 0 and 1 second is collected in a first time-domain window, while the noise audio between 0.1 second and 1.1 second is collected in a second time-domain window. The overlapped part of these two time-domain windows is the noise audio between 0.1 second and 1 second.

[0004] Currently, when analyzing and calculating the noise audio corresponding to each time-domain window, although there is data overlap between time-domain windows, an independent and complete calculation is performed for the noise audio corresponding to each time-domain window, which leads to repeated analysis and calculation of the overlapping data, resulting in calculation redundancy and waste of computing power.SUMMARY

[0005] In view of this, the disclosure provides a noise data processing method, an apparatus, an electronic device and a vehicle, which reduces the calculation of overlapping and saves computing power.

[0006] In a first aspect, the disclosure provides a noise data processing method. The noise data processing method includes: obtaining first noise data corresponding to a spectrum component to be determined in a first time-domain window, in which a length of the first time-domain window is an integer multiple of a step size; determining second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determining third noise data corresponding to the spectrum component within a last step size of a second time-domain window, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; and obtaining fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and then adding the third noise data.

[0007] In a second aspect, the disclosure provides a noise data processing apparatus. The noise data processing apparatus includes: an obtaining unit, configured to obtain first noise data corresponding to a spectrum component to be determined in a first time-domain window, in which a length of the first time-domain window is an integer multiple of a step size; a determining unit, configured to determine second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determine third noise data corresponding to the spectrum component within a last step size of a second time-domain window, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; and a processing unit, configured to obtain fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and then adding the third noise data.

[0008] In a third aspect, the disclosure provides a computer-readable storage medium. The computer-readable storage medium stores computer programs, and when the computer programs are executed by a processor, the noise data processing method of the first aspect is implemented.

[0009] In a fourth aspect, the disclosure provides an electronic device. The electronic device includes: a memory, configured to store computer programs; and a processor, configured to, when executing the computer programs stored in the memory, implement the noise data processing method of the first aspect.

[0010] In a fifth aspect, the disclosure provides a vehicle. The vehicle includes the noise data processing apparatus of the second aspect or the electronic device of the fourth aspect.

[0011] In a sixth aspect, the disclosure provides a computer program product. The computer program product includes computer-executable instructions or computer programs. When the computer-executable instructions or computer programs are executed by a processor, the noise data processing method of the first aspect is implemented.

[0012] According to the noise data processing method, an apparatus, an electronic device and a vehicle provided in the disclosure with the above technical solutions, firstly, the first noise data corresponding to the spectrum component to be determined in the first time-domain window is obtained, in which the length of the first time-domain window is the integer multiple of the step size; the second noise data corresponding to the spectrum component within the first step size of the first time-domain window is determined, and the third noise data corresponding to the spectrum component within the last step size of the second time-domain window is determined, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; then, the fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by processing the first noise data, the second noise data from and the third noise data. Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to a latter time-domain window can be accurately derived by using the noise data of a previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power.

[0013] The above description is only an overview of the technical solution of the disclosure. In order to understand the technical means of the disclosure more clearly and to implement these means in accordance with the contents of the specification, and to make the above and additional objects, features and advantages of the disclosure more obvious and understandable, specific implementations of the disclosure are illustrated below.BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings, which are incorporated in and constitute part of the specification, illustrate embodiments consistent with the disclosure and, together with the specification, serve to explain the principles of the disclosure.

[0015] To more clearly illustrate the technical solutions of the embodiments of the disclosure, the accompanying drawings used in the descriptions of the embodiments are briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without inventive work. FIG. 1 is a flowchart of a noise data processing method according to an embodiment of the disclosure. FIG. 2 is a flowchart of a noise data processing method according to another embodiment of the disclosure. FIG. 3 is a schematic diagram of an example according to an embodiment of the disclosure. FIG. 4 is a schematic diagram of an example according to an embodiment of the disclosure. FIG. 5 is a schematic diagram of an example according to an embodiment of the disclosure. FIG. 6 is a schematic flowchart of an example according to an embodiment of the disclosure. FIG. 7 is a schematic diagram of a noise data processing apparatus according to an embodiment of the disclosure. DETAILED DESCRIPTION

[0016] To better understand the above-mentioned objectives, features and advantages of the disclosure, the solutions of the disclosure are further described below. It should be noted that, without contradiction, the embodiments of the disclosure and features of the embodiments can be combined with each other.

[0017] To address the technical problem in current related arts where, when analyzing and calculating noise data corresponding to each time-domain window, although there is data overlap between time-domain windows, an independent and complete calculation is performed for the noise audio corresponding to each time-domain window, leading to repeated analysis and calculation of overlapping data, resulting in calculation redundancy and waste of computing power, the present embodiment provides a noise data processing method. As illustrated in FIG. 1, the noise data processing method includes the following steps.

[0018] At step 101, first noise data corresponding to a spectrum component to be determined in a first time-domain window is obtained.

[0019] A length of the first time-domain window is an integer multiple of a step size.

[0020] The time domain is the real world, and the only domain that actually exists. When evaluating the performance of digital products, analysis is usually performed in the time domain because the performance of the product is ultimately measured in the time domain.

[0021] The frequency domain is used wildly, especially in radio frequency and communication systems, as well as high-speed digital applications. The most important property of the frequency domain is that it is not real, but a mathematical construct. The time domain is the only domain that exists objectively, while the frequency domain is a mathematical category that follows specific rules, and is also referred to as "God's perspective" by some scholars.

[0022] The time domain represents a relationship of dynamic signals by using coordinates in a time axis, while the frequency domain transforms signals to be represented by using coordinates in a frequency axis. Generally, representations in the time-domain are more visual and intuitive, while representations in the frequency-domain are more concise, making it deeper and more convenient way to analyze problems. Currently, signal analysis in the time domain and the frequency domain are interrelated, indispensable, and complementary.

[0023] In the embodiment of the disclosure, a time-domain window refers to a time window corresponding to time-domain data. In some embodiments, the time window may refer to a specific time period, and first time-domain data may refer to time-domain data within a time window starting from 0.

[0024] In some embodiments, the step size includes a time step size, which is obtained by a difference between two adjacent time points. In process simulation, a model discretizes the entire process into multiple smaller processes, and a time required for each step is the time step size. When simulating a time response of a system, it is often necessary to set a time step size. The magnitude of the time step size generally depends on system attributes and model objectives. The larger the time step size's absolute value, the less a calculation time; the smaller the absolute value, the longer the calculation time, but simulation is more refined and process is more complex.

[0025] In the embodiment of the disclosure, the time window contains any number of time step sizes, and the length of the time window is an integer multiple of the time step size. For example, when the length of the time window is 1, the time step size may be 0.1, 0.2, etc., which is not completely listed herein.

[0026] Optionally, the first noise data refers to spectrum component data corresponding to data in the first time-domain window.

[0027] At step 102, second noise data corresponding to the spectrum component within a first step size of the first time-domain window is determined, and third noise data corresponding to the spectrum component within a last step size of a second time-domain window is determined.

[0028] The second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size.

[0029] A spectrum refers to a representation of a time-domain signal in the frequency domain, which is obtained by performing the Fourier Transform on the signal. As a result, two graphs are obtained, one with frequency as a horizontal axis and amplitude as a vertical axis, and the other with frequency as a horizontal axis and phase as a vertical axis. However, sometimes phase information is omitted, leaving only data of amplitudes corresponding to different frequencies. Sometimes, an "amplitude spectrum" is used to represent changes of amplitude with frequency, and a "phase spectrum" is used to describe changes of phase with frequency. Simply put, the spectrum may indicates which frequency sine waves a signal is composed of, and may also show information such as magnitude and phase of each frequency sine wave.

[0030] Sound is a mechanical vibration that can penetrate substances in various states of matter. These substances capable of transmitting sound are referred to as media. Sound cannot travel in a vacuum. Sounds we hear are also sound waves with specific frequencies. A frequency range of human auditory perception is approximately 20-20,000 Hz, and a sound wave with a frequency outside this range is imperceptible to the human ear. Waves below 20Hz are called infrasonic waves, and frequencies above 20kHz are called ultrasonic waves. The higher the frequency of a sound, the higher the sound's pitch; the lower the frequency, the lower the pitch.

[0031] In the embodiment of the disclosure, taking the first time-domain window ranging from 0 to 1 with the step size of 0.1 as an example, the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size, i.e., ranging from 0.1 to 1.1. A range of the first step size of the first time-domain window is 0 to 0.1, and a range of the last step size of the second time-domain window is 1 to 1.1. Correspondingly, an overlapping part of the first time-domain window and the second time-domain window is 0.1 to 1.

[0032] For example, the second noise data is noise spectrum data corresponding to the first step size of the first time-domain window.

[0033] At step 103, fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by subtracting the second noise data from the first noise data and then adding the third noise data.

[0034] In the embodiment of the disclosure, the fourth noise data is spectrum component data corresponding to the second time-domain window obtained by moving the first time-domain window by one step size.

[0035] Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, by using the noise data of a previous time-domain window (e.g., the first time-domain window), the noise data within the first step size of the previous time-domain window (e.g., the first time-domain window), and the noise data within the last step size of a latter time-domain window (e.g., the second time-domain window), the noise data corresponding to the latter time-domain window (e.g., the second time-domain window) can be accurately derived. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window (e.g., the second time-domain window), which reduces the calculation of overlapping data, thereby saving the computing power.

[0036] Further, as a refinement and extension of the above embodiment, the embodiment provides a noise data processing method as shown in FIG. 2. The noise data processing method includes the following steps.

[0037] At step 201, first time-domain data corresponding to a spectrum component in a first time-domain window is obtained.

[0038] At step 202, first noise data is obtained by performing a butterfly operation of Fast Fourier Transform (FFT) on the first time-domain data.

[0039] In the embodiment of the disclosure, the first noise data is noise spectrum data corresponding to the first time-domain data, and the noise spectrum data satisfies the matrix shown in the following Equation 1: Y 0 t Y 1 t ⋯ Y k t Y k + 1 t ⋯ ⋯ Y K t = W N 0 0 W N 0 1 ⋯ ⋯ ⋯ ⋯ ⋯ W N 0 N − 1 W N 1 0 ⋱ ⋯ ⋱ W N k 0 ⋱ W N k + 1 0 ⋱ ⋯ ⋱ ⋯ ⋱ W N K − 1 0 ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ W N K − 1 N − 1 x 0 x 1 ⋯ x k x k + 1 ⋯ ⋯ x n

[0040] Optionally, a root of unit in Equation 1 is an operator of fast Discrete Fourier Transform (DFT), which is defined as shown in Equation 2 below: W N k = e 2 π N k

[0041] For example, a single frequency domain component Y k t can be obtained by corresponding multiplication of time-domain data and the root of unit, and a corresponding multiplication relationship is shown in FIG. 3.

[0042] Optionally, step 202 may include: obtaining a target frequency corresponding to the spectrum component, and determining a Nyquist frequency corresponding to the target frequency; obtaining, according to the Nyquist frequency, a number of time-domain sampling points Z required for the target frequency, where Z is an integer not less than 1; and obtaining the first noise data by performing a butterfly recursive calculation of the FFT on a target sequence corresponding to first Z sampling points in a permutation sequence corresponding to the first time-domain data, in which the permutation sequence is obtained by performing a butterfly odd-even permutation of the FFT according to the first time-domain data.

[0043] FFT, also known as the fast Fourier transform, is a fast algorithm of DFT. FFT is obtained by improving DFT algorithms according to properties of DFT such as oddness, evenness, imaginary part and real part.

[0044] The butterfly recursive calculation of FFT refers to accelerating algorithms by using the point-value polynomial representation and performing recursive calculation of polynomial odd-even decomposition using properties of roots of unity. A premise of recursion is to perform the butterfly odd-even permutation, that is, rearranging an original sequence (such as the first time-domain data) according to a butterfly odd-even distribution.

[0045] In the embodiment of the disclosure, an example is given where the permutation sequence contains 16 numbers, i.e., 0-15, the number of time-domain sampling points Z required for the target frequency is 4, and the minimum number of samples in a sample set required for down-sampling at the current target frequency is 8. After performing the butterfly odd-even permutation, the first K numbers after the permutation are the permutation sequence after down-sampling the original sequence of N numbers to a sequence of K numbers. For example, if the permutation sequence [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15] is down-sampled to 8 numbers, a new permutation sequence is [0 2 4 6 8 10 12 14].

[0046] For example, in the permutation sequence [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15], the target sequences corresponding to the first 4 sampling points are [0 4 8 12] [2 6 10 14] [1 5 9 13] [3 7 11 15]. Other situations can be analogized, and specific examples are be given here.

[0047] New sequence permutation means that after permutation of the original sequence, the butterfly recursive calculation of FFT is performed on the first Z numbers to obtain a sequence with the same number of samples as the minimum number of samples in the sample set, so as to obtain the noise spectrum data corresponding to the down-sampled sequence.

[0048] Optionally, the above mode is a first mode to obtain the first noise data corresponding to the first time-domain window, which includes: obtaining the first noise data corresponding to the first time-domain window by performing the butterfly calculation of FFT on the first time-domain data corresponding to the spectrum component to be determined in the first time-domain window.

[0049] Optionally, step 202 further includes: obtaining second time-domain data corresponding to the spectrum component within each step size in the first time-domain window; and obtaining the first noise data by performing the butterfly operation of the FFT on the second time-domain data respectively and combining calculation results in series.

[0050] Optionally, step 202 further includes: obtaining a polynomial corresponding to the second time-domain data by performing a butterfly operation of FFT on the second time-domain data respectively, and combining the polynomial corresponding to the second time-domain data with a root of unity respectively to synthesize a target polynomial of the first noise data, in which the target polynomial is used for determining the first noise data.

[0051] Optionally, step 202 further includes: calculating the first noise data by using the following equation as shown below: Y k 1 = ∑ r = 0 R − 1 Y r ⋅ W N k sr where k represents the spectrum component, Y k 1 represents the target polynomial of the first noise data, R represents a number of step sizes in the first time-domain window, Y r represents the polynomial of the second time-domain data corresponding to the r th< step size, s represents a number of data points in the second time-domain data and W N k represents the root of unity.

[0052] For example, there are 4096 data points in the first time-domain window and 1024 data points in one step size, so R=4096 / 1024=4. Then, one window = step size 1+ step size 2+ step size 3+ step size 4 in series, and r represents 1, 2, 3 and 4.

[0053] Optionally, the above mode is a second mode to obtain the first noise data corresponding to the first time-domain window, and includes: obtaining the noise spectrum data corresponding to each step size of the first time-domain window by performing the butterfly recursive calculation of FFT on the time-domain data corresponding to each step size of the first time-domain window, and obtaining the first noise data corresponding to the first time-domain window by combining the noise spectrum data corresponding to each step size of the first time-domain window in series. In the embodiment of the disclosure, the processing of the first time-domain data includes the above two modes. The first mode is to directly perform the butterfly recursive calculation of FFT to complete spectrum calculation of the first window. An advantage of the first mode includes simple logic, which a disadvantage of the first mode is that when updating the first step size, it is necessary to recalculate the time-domain data within the first step size that has been updated in the first window through the butterfly recursive calculation of FFT again. The second mode is to perform local butterfly recursive calculation of FFT on data within each step size, and then obtain the spectrum of the first window by combining calculation results in series.

[0054] At step 203, second noise data corresponding to the spectrum component within a first step size of the first time-domain window is determined, and third noise data corresponding to the spectrum component within a last step size of a second time-domain window is determined.

[0055] The second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size.

[0056] For example, based on step 202, during acquisition of time-domain window data, the time-domain window moves by the step size. When the time-domain window moves forward by one step size, a corresponding relationship of single time-domain data is shown in FIG. 4. The left block and the middle block are noise spectrum data corresponding to the first time-domain window, the middle block and the right block are noise spectrum data corresponding to the second time-domain window. The time-domain data of the left block moves out of the window, the time-domain data of the right block moves into the window, and the time-domain data in the middle block is the overlapping data of the first time-domain window and the second time-domain window. To calculate a new frequency component Y k t + 1 , it is necessary to subtract the left block part and add the right block part on the basis of Y k t . Because a corresponding relationship between the time-domain data in the middle block and the roots of unity has changed, it is necessary to correct the middle block. For example, a corrected corresponding relationship can be shown in the following Equation 3: Y k t + 1 = Y k t ⋅ W N k − s − x 0 ⋅ W N k 0 + x 1 ⋅ W N k 1 … + x s − 1 ⋅ W N k s − 1 ⋅ W N k − s + x N ⋅ W N k N − s + x N + 1 ⋅ W N k N − s + 1 … + x N + s − 1 ⋅ W N k N − 1

[0057] Optionally, the first item in Equation 3 is entirely multiplied by W N k − s , so when subtracting the left block from the second item, it is also necessary to multiply by W N k − s . The content in the right block of the third part is DFT of time-domain data entering the window within the new step size. Since W N kN = 1, the third item in Equation 3 can extract W N kN , W N − sk in sequence, such that Equation 3 can be re-written as Equation 4 as shown below: Y k t + 1 = Y k t ⋅ W N k − s − x 0 ⋅ W N k 0 + x 1 ⋅ W N k 1 … + x s − 1 ⋅ W N k s − 1 ⋅ W N k − s + x N ⋅ W N k 0 + x 1 ⋅ W N k 1 … + x s − 1 ⋅ W N k s − 1 ⋅ W N k − s

[0058] For example, when the third noise data corresponding to the third time-domain data in the last step size of the second time-domain window is calculated independently, the number of times of recursion is log2(S). The butterfly recursion calculation of local FFT is performed according to the number of times of recursion. Since the extracted root of unity is also different from the root of unity in the above step, this calculation process is called local FFT. For example, based on third time-domain data corresponding to a spectrum component a in the last step size of the second time-domain window, when it is necessary to calculate a sound pressure level corresponding to the spectrum component a, a target frequency corresponding to the spectrum component a is first obtained. Then, a Nyquist frequency corresponding to the target frequency is determined, and a number of time-domain samples X required for the target frequency is obtained according to the Nyquist frequency. A target sequence corresponding to the first X sampling points is selected from a permutation sequence firstly, and then the butterfly recursive calculation of the FFT is performed on the target sequence, with the resulting outcome being used to determine the sound pressure level corresponding to the spectrum component a. The permutation sequence is obtained by performing the butterfly odd-even permutation of the FFT according to the third time-domain data. As illustrated in FIG. 5, a first polynomial of a third layer is selected from the permutation sequence as the target sequence, and then then according to the calculated number of times of recursions = 1, the butterfly recursive calculation is performed on the first polynomial of the third layer to obtain a first polynomial of a second layer. Finally, the sound pressure level corresponding to the spectrum component a can be determined according to the calculation result of the first polynomial of the second layer.

[0059] Optionally, except for the number of times of recursion, the root of unity extracted from the odd-even polynomial of each layer is shown in the following Equation 5: W c = W k N 2 log 2 S − c

[0060] At step 204, fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by subtracting the second noise data from the first noise data and adding the third noise data.

[0061] Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to the latter time-domain window can be accurately derived by using the noise data of the previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power.

[0062] As illustrated in FIG. 6, for example, firstly, the spectrum calculation of the first window is first completed as a window spectrum Y t0 corresponding to a moment t0. When calculating the next window spectrum Y t1 (the window spectrum corresponding to a moment t1, where the window size is an integer multiple R of the step size), data corresponding to the first step size in the window spectrum Y t0 corresponding to the moment t0, namely S1, is obtained, and data corresponding to the last step size in the window spectrum Y t1 corresponding to the moment t1, namely S R+1 , is also obtained. According to Y t1 = Y t0 -S1+S R+1 , the window spectrum Y t1 corresponding to the moment t1 is then calculated.

[0063] By applying the technical solution of the embodiment, repeated calculation can be avoided, thereby reducing the computing power. The number of times of multiplication in the butterfly recursive calculation process of original FFT is Nlog2(N), while the number of times of multiplication in the butterfly recursive calculation of rolling FFT is Nlog2(S), so the amount of computation saved per unit time is as shown in in the following Equation 6: R N log 2 N − N log 2 S = RN log 2 N S

[0064] Therefore, significant computational savings can be achieved in scenarios with a low fundamental frequency and a small step size. For example, for a common audio sampling frequency of 16384HZ and the step size of 2048 sampling points, the amount of computation saved per second can reach 393216 multiplication calculations, saving more than 20% of the computation amount.

[0065] Compared with the current related arts, by applying the technical solution of the embodiment, when analyzing and calculating noise data corresponding to each time-domain window, for the overlapping data between time-domain windows, the calculation of overlapping data can be reduced, thereby saving the computing power.

[0066] The embodiment provides a noise data processing apparatus. As illustrated in FIG. 7, the noise data processing apparatus includes: an obtaining section 31, a determining section 32 and a processing section 33.

[0067] The obtaining section 31 is configured to obtain first noise data corresponding to a spectrum component to be determined in a first time-domain window, in which a length of the first time-domain window is an integer multiple of a step size.

[0068] The determining section 32 is configured to determine second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determine third noise data corresponding to the spectrum component within a last step size of a second time-domain window, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size.

[0069] The processing section 33 is configured to obtain fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and adding the third noise data.

[0070] In some embodiments, the obtaining section 31 is further configured to: obtain first time-domain data corresponding to the spectrum component in the first time-domain window; and obtain the first noise data by performing a butterfly operation of FFT on the first time-domain data.

[0071] In some embodiments, the obtaining section 31 is further configured to: obtain a target frequency corresponding to the spectrum component, and determine a Nyquist frequency corresponding to the target frequency; obtain, according to the Nyquist frequency, a number of time-domain samples Z required for the target frequency; and obtain the first noise data by performing a butterfly recursive calculation of the FFT on a target sequence corresponding to first Z sampling points in a permutation sequence corresponding to the first time-domain data, in which the permutation sequence is obtained by performing a butterfly odd-even permutation of the FFT according to the first time-domain data.

[0072] In some embodiments, the obtaining section 31 is further configured to: obtain second time-domain data corresponding to the spectrum component within each step size in the first time-domain window; and obtain the first noise data by performing a butterfly operation of the FFT on the second time-domain data respectively and combining calculation results in series.

[0073] In some embodiments, the obtaining section 31 is further configured to: obtain a polynomial corresponding to the second time-domain data by performing a butterfly operation of FFT on the second time-domain data respectively, and combine the polynomial corresponding to the second time-domain data with a root of unity respectively to synthesize a target polynomial of the first noise data, in which the target polynomial is used for determining the first noise data.

[0074] In some embodiments, the obtaining section 31 is further configured to: calculate the first noise data by using the following equation: Y k 1 = ∑ r = 0 R − 1 Y r ⋅ W N k sr , where k represents the spectrum component, Y k 1 represents the target polynomial of the first noise data, R represents a number of step sizes in the first time-domain window, Y, represents the polynomial of the second time-domain data corresponding to the r th< step size, s represents a number of data points in the second time-domain data, and W N k< represents the root of unity.

[0075] It should be noted that for other corresponding descriptions of functional units involved in the noise data processing apparatus provided in the embodiment, reference may be made to the corresponding descriptions of the noise data processing method provided in the embodiments of the disclosure, such as descriptions in FIGS. 1 and 2, which will not be repeated herein.

[0076] Based on the noise data processing method provided in the disclosure, correspondingly, the embodiment also provides a computer-readable storage medium. The computer-readable storage medium stores computer programs, and when the computer programs are executed by a processor, the noise data processing method provided in the disclosure is implemented.

[0077] The disclosure provides a computer program product. The computer program product includes computer executable instructions or computer programs. When the computer executable instructions or computer programs are executed by a processor, the noise data processing method provided in the disclosure is implemented.

[0078] Based on this understanding, the technical solution of the disclosure may be embodied in the form of a software product, which is stored in a non-transitory storage medium (e.g., a CD-ROM, a USB flash drive, or a mobile hard drive) and includes instructions to cause a computer device (e.g., a personal computer, a server, or a network device) to execute the methods of various implementation scenarios of the disclosure.

[0079] Based on the embodiments of the noise data processing method and the embodiments of the noise data processing apparatus provided in the disclosure, in order to achieve the above objectives, the disclosure also provides an electronic device that can be deployed on a vehicle (such as a new energy vehicle), a server, and the like. The electronic device includes a memory and a processor. The memory is configured to store computer programs, and the processor is configured to implement the noise data processing method provided in the disclosure when executing the computer programs stored in the memory.

[0080] Optionally, the above physical devices may also include user interfaces, network interfaces, cameras, radio frequency (RF) circuits, sensors, audio circuits, wireless fidelity (Wi-Fi) modules, etc. The user interfaces may include displays and input units such as keyboards. Optionally, the user interfaces may also include USB interfaces, card reader interfaces, etc. The network interfaces may include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0081] Those skilled in the art understand that the structure of the physical device provided in the embodiment does not constitute a limitation on the physical device, and the physical device may include more or fewer components, combine with certain components, or have different arrangements of components.

[0082] The memory may also include an operating system and a network communication section. The operating system is a program that manages hardware and software resources of the physical device, and supports information processing programs and other software and / or programs to operate. The network communication section is used to enable communications between various components within the memory, as well as communications with hardware and software in an information processing device.

[0083] Based on the embodiments of the noise data processing method provided in the disclosure, the embodiments of the noise data processing apparatus provided in the disclosure, and the embodiments of the electronic device, the embodiment also provides a vehicle provided with the above electronic device that can implement the noise data processing method provided in the disclosure.

[0084] Through the above descriptions of the embodiments, those skilled in the art can clearly understand that the disclosure can be implemented by means of software in combination with necessary general hardware platforms or by hardware. By applying the technical solution of the embodiment, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to the latter time-domain window can be accurately derived by using the noise data of the previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power.

[0085] It should be noted that in this article, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprise" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or apparatus. Without further restrictions, for a process, method, article or apparatus "comprising one element", the presence of other identical elements in the process, method, article or apparatus that includes the element is not excluded.

[0086] The foregoing descriptions are merely specific implementations of the disclosure, enabling those skilled in the art to understand or implement the disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the disclosure will not be limited to the embodiments described herein, but shall conform to the broadest scope consistent with the principles and novel features disclosed herein.INDUSTRIAL APPLICABILITY

[0087] The disclosure provides a noise data processing method, an apparatus, an electronic device and a vehicle. Firstly, the first noise data corresponding to the spectrum component to be determined in the first time-domain window is obtained, in which the length of the first time-domain window is the integer multiple of the step size; the second noise data corresponding to the spectrum component within the first step size of the first time-domain window is determined, and the third noise data corresponding to the spectrum component within the last step size of the second time-domain window is determined, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; then, the fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by processing the first noise data, the second noise data from and the third noise data. Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to a latter time-domain window can be accurately derived by using the noise data of a previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power

Examples

Embodiment Construction

[0016]To better understand the above-mentioned objectives, features and advantages of the disclosure, the solutions of the disclosure are further described below. It should be noted that, without contradiction, the embodiments of the disclosure and features of the embodiments can be combined with each other.

[0017]To address the technical problem in current related arts where, when analyzing and calculating noise data corresponding to each time-domain window, although there is data overlap between time-domain windows, an independent and complete calculation is performed for the noise audio corresponding to each time-domain window, leading to repeated analysis and calculation of overlapping data, resulting in calculation redundancy and waste of computing power, the present embodiment provides a noise data processing method. As illustrated in FIG. 1, the noise data processing method includes the following steps.

[0018]At step 101, first noise data corresponding to a spectrum component t...

Cross-reference to related applications

[0001] The application claims the priority of Chinese patent application No. 202310724699.5, filed on June 19, 2023, titled "Noise Data Processing Method and Apparatus, and Electronic Device and Vehicle", the entire content of which is incorporated herein by reference.TECHNICAL FIELD

[0002] The disclosure relates to a field of noise data processing technologies, and relates to, but is not limited to, a noise data processing method, an apparatus, an electronic device and a vehicle.BACKGROUND

[0003] A noise audio generated by in-vehicle devices such as a motor and an engine may be used to analyze whether the devices are running normally. In order to ensure the continuity of data analysis, time-domain windows are overlapped. For example, the noise audio between 0 and 1 second is collected in a first time-domain window, while the noise audio between 0.1 second and 1.1 second is collected in a second time-domain window. The overlapped part of these two time-domain windows is the noise audio between 0.1 second and 1 second.

[0004] Currently, when analyzing and calculating the noise audio corresponding to each time-domain window, although there is data overlap between time-domain windows, an independent and complete calculation is performed for the noise audio corresponding to each time-domain window, which leads to repeated analysis and calculation of the overlapping data, resulting in calculation redundancy and waste of computing power.SUMMARY

[0005] In view of this, the disclosure provides a noise data processing method, an apparatus, an electronic device and a vehicle, which reduces the calculation of overlapping and saves computing power.

[0006] In a first aspect, the disclosure provides a noise data processing method. The noise data processing method includes: obtaining first noise data corresponding to a spectrum component to be determined in a first time-domain window, in which a length of the first time-domain window is an integer multiple of a step size; determining second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determining third noise data corresponding to the spectrum component within a last step size of a second time-domain window, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; and obtaining fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and then adding the third noise data.

[0007] In a second aspect, the disclosure provides a noise data processing apparatus. The noise data processing apparatus includes: an obtaining unit, configured to obtain first noise data corresponding to a spectrum component to be determined in a first time-domain window, in which a length of the first time-domain window is an integer multiple of a step size; a determining unit, configured to determine second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determine third noise data corresponding to the spectrum component within a last step size of a second time-domain window, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; and a processing unit, configured to obtain fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and then adding the third noise data.

[0008] In a third aspect, the disclosure provides a computer-readable storage medium. The computer-readable storage medium stores computer programs, and when the computer programs are executed by a processor, the noise data processing method of the first aspect is implemented.

[0009] In a fourth aspect, the disclosure provides an electronic device. The electronic device includes: a memory, configured to store computer programs; and a processor, configured to, when executing the computer programs stored in the memory, implement the noise data processing method of the first aspect.

[0010] In a fifth aspect, the disclosure provides a vehicle. The vehicle includes the noise data processing apparatus of the second aspect or the electronic device of the fourth aspect.

[0011] In a sixth aspect, the disclosure provides a computer program product. The computer program product includes computer-executable instructions or computer programs. When the computer-executable instructions or computer programs are executed by a processor, the noise data processing method of the first aspect is implemented.

[0012] According to the noise data processing method, an apparatus, an electronic device and a vehicle provided in the disclosure with the above technical solutions, firstly, the first noise data corresponding to the spectrum component to be determined in the first time-domain window is obtained, in which the length of the first time-domain window is the integer multiple of the step size; the second noise data corresponding to the spectrum component within the first step size of the first time-domain window is determined, and the third noise data corresponding to the spectrum component within the last step size of the second time-domain window is determined, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; then, the fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by processing the first noise data, the second noise data from and the third noise data. Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to a latter time-domain window can be accurately derived by using the noise data of a previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power.

[0013] The above description is only an overview of the technical solution of the disclosure. In order to understand the technical means of the disclosure more clearly and to implement these means in accordance with the contents of the specification, and to make the above and additional objects, features and advantages of the disclosure more obvious and understandable, specific implementations of the disclosure are illustrated below.BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings, which are incorporated in and constitute part of the specification, illustrate embodiments consistent with the disclosure and, together with the specification, serve to explain the principles of the disclosure.

[0015] To more clearly illustrate the technical solutions of the embodiments of the disclosure, the accompanying drawings used in the descriptions of the embodiments are briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without inventive work. FIG. 1 is a flowchart of a noise data processing method according to an embodiment of the disclosure. FIG. 2 is a flowchart of a noise data processing method according to another embodiment of the disclosure. FIG. 3 is a schematic diagram of an example according to an embodiment of the disclosure. FIG. 4 is a schematic diagram of an example according to an embodiment of the disclosure. FIG. 5 is a schematic diagram of an example according to an embodiment of the disclosure. FIG. 6 is a schematic flowchart of an example according to an embodiment of the disclosure. FIG. 7 is a schematic diagram of a noise data processing apparatus according to an embodiment of the disclosure. DETAILED DESCRIPTION

[0016] To better understand the above-mentioned objectives, features and advantages of the disclosure, the solutions of the disclosure are further described below. It should be noted that, without contradiction, the embodiments of the disclosure and features of the embodiments can be combined with each other.

[0017] To address the technical problem in current related arts where, when analyzing and calculating noise data corresponding to each time-domain window, although there is data overlap between time-domain windows, an independent and complete calculation is performed for the noise audio corresponding to each time-domain window, leading to repeated analysis and calculation of overlapping data, resulting in calculation redundancy and waste of computing power, the present embodiment provides a noise data processing method. As illustrated in FIG. 1, the noise data processing method includes the following steps.

[0018] At step 101, first noise data corresponding to a spectrum component to be determined in a first time-domain window is obtained.

[0019] A length of the first time-domain window is an integer multiple of a step size.

[0020] The time domain is the real world, and the only domain that actually exists. When evaluating the performance of digital products, analysis is usually performed in the time domain because the performance of the product is ultimately measured in the time domain.

[0021] The frequency domain is used wildly, especially in radio frequency and communication systems, as well as high-speed digital applications. The most important property of the frequency domain is that it is not real, but a mathematical construct. The time domain is the only domain that exists objectively, while the frequency domain is a mathematical category that follows specific rules, and is also referred to as "God's perspective" by some scholars.

[0022] The time domain represents a relationship of dynamic signals by using coordinates in a time axis, while the frequency domain transforms signals to be represented by using coordinates in a frequency axis. Generally, representations in the time-domain are more visual and intuitive, while representations in the frequency-domain are more concise, making it deeper and more convenient way to analyze problems. Currently, signal analysis in the time domain and the frequency domain are interrelated, indispensable, and complementary.

[0023] In the embodiment of the disclosure, a time-domain window refers to a time window corresponding to time-domain data. In some embodiments, the time window may refer to a specific time period, and first time-domain data may refer to time-domain data within a time window starting from 0.

[0024] In some embodiments, the step size includes a time step size, which is obtained by a difference between two adjacent time points. In process simulation, a model discretizes the entire process into multiple smaller processes, and a time required for each step is the time step size. When simulating a time response of a system, it is often necessary to set a time step size. The magnitude of the time step size generally depends on system attributes and model objectives. The larger the time step size's absolute value, the less a calculation time; the smaller the absolute value, the longer the calculation time, but simulation is more refined and process is more complex.

[0025] In the embodiment of the disclosure, the time window contains any number of time step sizes, and the length of the time window is an integer multiple of the time step size. For example, when the length of the time window is 1, the time step size may be 0.1, 0.2, etc., which is not completely listed herein.

[0026] Optionally, the first noise data refers to spectrum component data corresponding to data in the first time-domain window.

[0027] At step 102, second noise data corresponding to the spectrum component within a first step size of the first time-domain window is determined, and third noise data corresponding to the spectrum component within a last step size of a second time-domain window is determined.

[0028] The second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size.

[0029] A spectrum refers to a representation of a time-domain signal in the frequency domain, which is obtained by performing the Fourier Transform on the signal. As a result, two graphs are obtained, one with frequency as a horizontal axis and amplitude as a vertical axis, and the other with frequency as a horizontal axis and phase as a vertical axis. However, sometimes phase information is omitted, leaving only data of amplitudes corresponding to different frequencies. Sometimes, an "amplitude spectrum" is used to represent changes of amplitude with frequency, and a "phase spectrum" is used to describe changes of phase with frequency. Simply put, the spectrum may indicates which frequency sine waves a signal is composed of, and may also show information such as magnitude and phase of each frequency sine wave.

[0030] Sound is a mechanical vibration that can penetrate substances in various states of matter. These substances capable of transmitting sound are referred to as media. Sound cannot travel in a vacuum. Sounds we hear are also sound waves with specific frequencies. A frequency range of human auditory perception is approximately 20-20,000 Hz, and a sound wave with a frequency outside this range is imperceptible to the human ear. Waves below 20Hz are called infrasonic waves, and frequencies above 20kHz are called ultrasonic waves. The higher the frequency of a sound, the higher the sound's pitch; the lower the frequency, the lower the pitch.

[0031] In the embodiment of the disclosure, taking the first time-domain window ranging from 0 to 1 with the step size of 0.1 as an example, the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size, i.e., ranging from 0.1 to 1.1. A range of the first step size of the first time-domain window is 0 to 0.1, and a range of the last step size of the second time-domain window is 1 to 1.1. Correspondingly, an overlapping part of the first time-domain window and the second time-domain window is 0.1 to 1.

[0032] For example, the second noise data is noise spectrum data corresponding to the first step size of the first time-domain window.

[0033] At step 103, fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by subtracting the second noise data from the first noise data and then adding the third noise data.

[0034] In the embodiment of the disclosure, the fourth noise data is spectrum component data corresponding to the second time-domain window obtained by moving the first time-domain window by one step size.

[0035] Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, by using the noise data of a previous time-domain window (e.g., the first time-domain window), the noise data within the first step size of the previous time-domain window (e.g., the first time-domain window), and the noise data within the last step size of a latter time-domain window (e.g., the second time-domain window), the noise data corresponding to the latter time-domain window (e.g., the second time-domain window) can be accurately derived. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window (e.g., the second time-domain window), which reduces the calculation of overlapping data, thereby saving the computing power.

[0036] Further, as a refinement and extension of the above embodiment, the embodiment provides a noise data processing method as shown in FIG. 2. The noise data processing method includes the following steps.

[0037] At step 201, first time-domain data corresponding to a spectrum component in a first time-domain window is obtained.

[0038] At step 202, first noise data is obtained by performing a butterfly operation of Fast Fourier Transform (FFT) on the first time-domain data.

[0039] In the embodiment of the disclosure, the first noise data is noise spectrum data corresponding to the first time-domain data, and the noise spectrum data satisfies the matrix shown in the following Equation 1: Y 0 t Y 1 t ⋯ Y k t Y k + 1 t ⋯ ⋯ Y K t = W N 0 0 W N 0 1 ⋯ ⋯ ⋯ ⋯ ⋯ W N 0 N − 1 W N 1 0 ⋱ ⋯ ⋱ W N k 0 ⋱ W N k + 1 0 ⋱ ⋯ ⋱ ⋯ ⋱ W N K − 1 0 ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ W N K − 1 N − 1 x 0 x 1 ⋯ x k x k + 1 ⋯ ⋯ x n

[0040] Optionally, a root of unit in Equation 1 is an operator of fast Discrete Fourier Transform (DFT), which is defined as shown in Equation 2 below: W N k = e 2 π N k

[0041] For example, a single frequency domain component Y k t can be obtained by corresponding multiplication of time-domain data and the root of unit, and a corresponding multiplication relationship is shown in FIG. 3.

[0042] Optionally, step 202 may include: obtaining a target frequency corresponding to the spectrum component, and determining a Nyquist frequency corresponding to the target frequency; obtaining, according to the Nyquist frequency, a number of time-domain sampling points Z required for the target frequency, where Z is an integer not less than 1; and obtaining the first noise data by performing a butterfly recursive calculation of the FFT on a target sequence corresponding to first Z sampling points in a permutation sequence corresponding to the first time-domain data, in which the permutation sequence is obtained by performing a butterfly odd-even permutation of the FFT according to the first time-domain data.

[0043] FFT, also known as the fast Fourier transform, is a fast algorithm of DFT. FFT is obtained by improving DFT algorithms according to properties of DFT such as oddness, evenness, imaginary part and real part.

[0044] The butterfly recursive calculation of FFT refers to accelerating algorithms by using the point-value polynomial representation and performing recursive calculation of polynomial odd-even decomposition using properties of roots of unity. A premise of recursion is to perform the butterfly odd-even permutation, that is, rearranging an original sequence (such as the first time-domain data) according to a butterfly odd-even distribution.

[0045] In the embodiment of the disclosure, an example is given where the permutation sequence contains 16 numbers, i.e., 0-15, the number of time-domain sampling points Z required for the target frequency is 4, and the minimum number of samples in a sample set required for down-sampling at the current target frequency is 8. After performing the butterfly odd-even permutation, the first K numbers after the permutation are the permutation sequence after down-sampling the original sequence of N numbers to a sequence of K numbers. For example, if the permutation sequence [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15] is down-sampled to 8 numbers, a new permutation sequence is [0 2 4 6 8 10 12 14].

[0046] For example, in the permutation sequence [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15], the target sequences corresponding to the first 4 sampling points are [0 4 8 12] [2 6 10 14] [1 5 9 13] [3 7 11 15]. Other situations can be analogized, and specific examples are be given here.

[0047] New sequence permutation means that after permutation of the original sequence, the butterfly recursive calculation of FFT is performed on the first Z numbers to obtain a sequence with the same number of samples as the minimum number of samples in the sample set, so as to obtain the noise spectrum data corresponding to the down-sampled sequence.

[0048] Optionally, the above mode is a first mode to obtain the first noise data corresponding to the first time-domain window, which includes: obtaining the first noise data corresponding to the first time-domain window by performing the butterfly calculation of FFT on the first time-domain data corresponding to the spectrum component to be determined in the first time-domain window.

[0049] Optionally, step 202 further includes: obtaining second time-domain data corresponding to the spectrum component within each step size in the first time-domain window; and obtaining the first noise data by performing the butterfly operation of the FFT on the second time-domain data respectively and combining calculation results in series.

[0050] Optionally, step 202 further includes: obtaining a polynomial corresponding to the second time-domain data by performing a butterfly operation of FFT on the second time-domain data respectively, and combining the polynomial corresponding to the second time-domain data with a root of unity respectively to synthesize a target polynomial of the first noise data, in which the target polynomial is used for determining the first noise data.

[0051] Optionally, step 202 further includes: calculating the first noise data by using the following equation as shown below: Y k 1 = ∑ r = 0 R − 1 Y r ⋅ W N k sr where k represents the spectrum component, Y k 1 represents the target polynomial of the first noise data, R represents a number of step sizes in the first time-domain window, Y r represents the polynomial of the second time-domain data corresponding to the r th< step size, s represents a number of data points in the second time-domain data and W N k represents the root of unity.

[0052] For example, there are 4096 data points in the first time-domain window and 1024 data points in one step size, so R=4096 / 1024=4. Then, one window = step size 1+ step size 2+ step size 3+ step size 4 in series, and r represents 1, 2, 3 and 4.

[0053] Optionally, the above mode is a second mode to obtain the first noise data corresponding to the first time-domain window, and includes: obtaining the noise spectrum data corresponding to each step size of the first time-domain window by performing the butterfly recursive calculation of FFT on the time-domain data corresponding to each step size of the first time-domain window, and obtaining the first noise data corresponding to the first time-domain window by combining the noise spectrum data corresponding to each step size of the first time-domain window in series. In the embodiment of the disclosure, the processing of the first time-domain data includes the above two modes. The first mode is to directly perform the butterfly recursive calculation of FFT to complete spectrum calculation of the first window. An advantage of the first mode includes simple logic, which a disadvantage of the first mode is that when updating the first step size, it is necessary to recalculate the time-domain data within the first step size that has been updated in the first window through the butterfly recursive calculation of FFT again. The second mode is to perform local butterfly recursive calculation of FFT on data within each step size, and then obtain the spectrum of the first window by combining calculation results in series.

[0054] At step 203, second noise data corresponding to the spectrum component within a first step size of the first time-domain window is determined, and third noise data corresponding to the spectrum component within a last step size of a second time-domain window is determined.

[0055] The second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size.

[0056] For example, based on step 202, during acquisition of time-domain window data, the time-domain window moves by the step size. When the time-domain window moves forward by one step size, a corresponding relationship of single time-domain data is shown in FIG. 4. The left block and the middle block are noise spectrum data corresponding to the first time-domain window, the middle block and the right block are noise spectrum data corresponding to the second time-domain window. The time-domain data of the left block moves out of the window, the time-domain data of the right block moves into the window, and the time-domain data in the middle block is the overlapping data of the first time-domain window and the second time-domain window. To calculate a new frequency component Y k t + 1 , it is necessary to subtract the left block part and add the right block part on the basis of Y k t . Because a corresponding relationship between the time-domain data in the middle block and the roots of unity has changed, it is necessary to correct the middle block. For example, a corrected corresponding relationship can be shown in the following Equation 3: Y k t + 1 = Y k t ⋅ W N k − s − x 0 ⋅ W N k 0 + x 1 ⋅ W N k 1 … + x s − 1 ⋅ W N k s − 1 ⋅ W N k − s + x N ⋅ W N k N − s + x N + 1 ⋅ W N k N − s + 1 … + x N + s − 1 ⋅ W N k N − 1

[0057] Optionally, the first item in Equation 3 is entirely multiplied by W N k − s , so when subtracting the left block from the second item, it is also necessary to multiply by W N k − s . The content in the right block of the third part is DFT of time-domain data entering the window within the new step size. Since W N kN = 1, the third item in Equation 3 can extract W N kN , W N − sk in sequence, such that Equation 3 can be re-written as Equation 4 as shown below: Y k t + 1 = Y k t ⋅ W N k − s − x 0 ⋅ W N k 0 + x 1 ⋅ W N k 1 … + x s − 1 ⋅ W N k s − 1 ⋅ W N k − s + x N ⋅ W N k 0 + x 1 ⋅ W N k 1 … + x s − 1 ⋅ W N k s − 1 ⋅ W N k − s

[0058] For example, when the third noise data corresponding to the third time-domain data in the last step size of the second time-domain window is calculated independently, the number of times of recursion is log2(S). The butterfly recursion calculation of local FFT is performed according to the number of times of recursion. Since the extracted root of unity is also different from the root of unity in the above step, this calculation process is called local FFT. For example, based on third time-domain data corresponding to a spectrum component a in the last step size of the second time-domain window, when it is necessary to calculate a sound pressure level corresponding to the spectrum component a, a target frequency corresponding to the spectrum component a is first obtained. Then, a Nyquist frequency corresponding to the target frequency is determined, and a number of time-domain samples X required for the target frequency is obtained according to the Nyquist frequency. A target sequence corresponding to the first X sampling points is selected from a permutation sequence firstly, and then the butterfly recursive calculation of the FFT is performed on the target sequence, with the resulting outcome being used to determine the sound pressure level corresponding to the spectrum component a. The permutation sequence is obtained by performing the butterfly odd-even permutation of the FFT according to the third time-domain data. As illustrated in FIG. 5, a first polynomial of a third layer is selected from the permutation sequence as the target sequence, and then then according to the calculated number of times of recursions = 1, the butterfly recursive calculation is performed on the first polynomial of the third layer to obtain a first polynomial of a second layer. Finally, the sound pressure level corresponding to the spectrum component a can be determined according to the calculation result of the first polynomial of the second layer.

[0059] Optionally, except for the number of times of recursion, the root of unity extracted from the odd-even polynomial of each layer is shown in the following Equation 5: W c = W k N 2 log 2 S − c

[0060] At step 204, fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by subtracting the second noise data from the first noise data and adding the third noise data.

[0061] Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to the latter time-domain window can be accurately derived by using the noise data of the previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power.

[0062] As illustrated in FIG. 6, for example, firstly, the spectrum calculation of the first window is first completed as a window spectrum Y t0 corresponding to a moment t0. When calculating the next window spectrum Y t1 (the window spectrum corresponding to a moment t1, where the window size is an integer multiple R of the step size), data corresponding to the first step size in the window spectrum Y t0 corresponding to the moment t0, namely S1, is obtained, and data corresponding to the last step size in the window spectrum Y t1 corresponding to the moment t1, namely S R+1 , is also obtained. According to Y t1 = Y t0 -S1+S R+1 , the window spectrum Y t1 corresponding to the moment t1 is then calculated.

[0063] By applying the technical solution of the embodiment, repeated calculation can be avoided, thereby reducing the computing power. The number of times of multiplication in the butterfly recursive calculation process of original FFT is Nlog2(N), while the number of times of multiplication in the butterfly recursive calculation of rolling FFT is Nlog2(S), so the amount of computation saved per unit time is as shown in in the following Equation 6: R N log 2 N − N log 2 S = RN log 2 N S

[0064] Therefore, significant computational savings can be achieved in scenarios with a low fundamental frequency and a small step size. For example, for a common audio sampling frequency of 16384HZ and the step size of 2048 sampling points, the amount of computation saved per second can reach 393216 multiplication calculations, saving more than 20% of the computation amount.

[0065] Compared with the current related arts, by applying the technical solution of the embodiment, when analyzing and calculating noise data corresponding to each time-domain window, for the overlapping data between time-domain windows, the calculation of overlapping data can be reduced, thereby saving the computing power.

[0066] The embodiment provides a noise data processing apparatus. As illustrated in FIG. 7, the noise data processing apparatus includes: an obtaining section 31, a determining section 32 and a processing section 33.

[0067] The obtaining section 31 is configured to obtain first noise data corresponding to a spectrum component to be determined in a first time-domain window, in which a length of the first time-domain window is an integer multiple of a step size.

[0068] The determining section 32 is configured to determine second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determine third noise data corresponding to the spectrum component within a last step size of a second time-domain window, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size.

[0069] The processing section 33 is configured to obtain fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and adding the third noise data.

[0070] In some embodiments, the obtaining section 31 is further configured to: obtain first time-domain data corresponding to the spectrum component in the first time-domain window; and obtain the first noise data by performing a butterfly operation of FFT on the first time-domain data.

[0071] In some embodiments, the obtaining section 31 is further configured to: obtain a target frequency corresponding to the spectrum component, and determine a Nyquist frequency corresponding to the target frequency; obtain, according to the Nyquist frequency, a number of time-domain samples Z required for the target frequency; and obtain the first noise data by performing a butterfly recursive calculation of the FFT on a target sequence corresponding to first Z sampling points in a permutation sequence corresponding to the first time-domain data, in which the permutation sequence is obtained by performing a butterfly odd-even permutation of the FFT according to the first time-domain data.

[0072] In some embodiments, the obtaining section 31 is further configured to: obtain second time-domain data corresponding to the spectrum component within each step size in the first time-domain window; and obtain the first noise data by performing a butterfly operation of the FFT on the second time-domain data respectively and combining calculation results in series.

[0073] In some embodiments, the obtaining section 31 is further configured to: obtain a polynomial corresponding to the second time-domain data by performing a butterfly operation of FFT on the second time-domain data respectively, and combine the polynomial corresponding to the second time-domain data with a root of unity respectively to synthesize a target polynomial of the first noise data, in which the target polynomial is used for determining the first noise data.

[0074] In some embodiments, the obtaining section 31 is further configured to: calculate the first noise data by using the following equation: Y k 1 = ∑ r = 0 R − 1 Y r ⋅ W N k sr , where k represents the spectrum component, Y k 1 represents the target polynomial of the first noise data, R represents a number of step sizes in the first time-domain window, Y, represents the polynomial of the second time-domain data corresponding to the r th< step size, s represents a number of data points in the second time-domain data, and W N k< represents the root of unity.

[0075] It should be noted that for other corresponding descriptions of functional units involved in the noise data processing apparatus provided in the embodiment, reference may be made to the corresponding descriptions of the noise data processing method provided in the embodiments of the disclosure, such as descriptions in FIGS. 1 and 2, which will not be repeated herein.

[0076] Based on the noise data processing method provided in the disclosure, correspondingly, the embodiment also provides a computer-readable storage medium. The computer-readable storage medium stores computer programs, and when the computer programs are executed by a processor, the noise data processing method provided in the disclosure is implemented.

[0077] The disclosure provides a computer program product. The computer program product includes computer executable instructions or computer programs. When the computer executable instructions or computer programs are executed by a processor, the noise data processing method provided in the disclosure is implemented.

[0078] Based on this understanding, the technical solution of the disclosure may be embodied in the form of a software product, which is stored in a non-transitory storage medium (e.g., a CD-ROM, a USB flash drive, or a mobile hard drive) and includes instructions to cause a computer device (e.g., a personal computer, a server, or a network device) to execute the methods of various implementation scenarios of the disclosure.

[0079] Based on the embodiments of the noise data processing method and the embodiments of the noise data processing apparatus provided in the disclosure, in order to achieve the above objectives, the disclosure also provides an electronic device that can be deployed on a vehicle (such as a new energy vehicle), a server, and the like. The electronic device includes a memory and a processor. The memory is configured to store computer programs, and the processor is configured to implement the noise data processing method provided in the disclosure when executing the computer programs stored in the memory.

[0080] Optionally, the above physical devices may also include user interfaces, network interfaces, cameras, radio frequency (RF) circuits, sensors, audio circuits, wireless fidelity (Wi-Fi) modules, etc. The user interfaces may include displays and input units such as keyboards. Optionally, the user interfaces may also include USB interfaces, card reader interfaces, etc. The network interfaces may include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0081] Those skilled in the art understand that the structure of the physical device provided in the embodiment does not constitute a limitation on the physical device, and the physical device may include more or fewer components, combine with certain components, or have different arrangements of components.

[0082] The memory may also include an operating system and a network communication section. The operating system is a program that manages hardware and software resources of the physical device, and supports information processing programs and other software and / or programs to operate. The network communication section is used to enable communications between various components within the memory, as well as communications with hardware and software in an information processing device.

[0083] Based on the embodiments of the noise data processing method provided in the disclosure, the embodiments of the noise data processing apparatus provided in the disclosure, and the embodiments of the electronic device, the embodiment also provides a vehicle provided with the above electronic device that can implement the noise data processing method provided in the disclosure.

[0084] Through the above descriptions of the embodiments, those skilled in the art can clearly understand that the disclosure can be implemented by means of software in combination with necessary general hardware platforms or by hardware. By applying the technical solution of the embodiment, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to the latter time-domain window can be accurately derived by using the noise data of the previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power.

[0085] It should be noted that in this article, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprise" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or apparatus. Without further restrictions, for a process, method, article or apparatus "comprising one element", the presence of other identical elements in the process, method, article or apparatus that includes the element is not excluded.

[0086] The foregoing descriptions are merely specific implementations of the disclosure, enabling those skilled in the art to understand or implement the disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the disclosure will not be limited to the embodiments described herein, but shall conform to the broadest scope consistent with the principles and novel features disclosed herein.INDUSTRIAL APPLICABILITY

[0087] The disclosure provides a noise data processing method, an apparatus, an electronic device and a vehicle. Firstly, the first noise data corresponding to the spectrum component to be determined in the first time-domain window is obtained, in which the length of the first time-domain window is the integer multiple of the step size; the second noise data corresponding to the spectrum component within the first step size of the first time-domain window is determined, and the third noise data corresponding to the spectrum component within the last step size of the second time-domain window is determined, in which the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; then, the fourth noise data corresponding to the spectrum component in the second time-domain window is obtained by processing the first noise data, the second noise data from and the third noise data. Compared with the current solution where the independent and complete calculation is performed for data in each time-domain window with data overlap, by applying the technical solution of the disclosure, when analyzing and calculating noise data corresponding to each time-domain window, the noise data corresponding to a latter time-domain window can be accurately derived by using the noise data of a previous time-domain window, the noise data within the first step size of the previous time-domain window, and the noise data within the last step size of the latter time-domain window. In this way, for the overlapping data between adjacent time-domain windows, there is no need for repeated calculation in the calculation of the latter time-domain window, which reduces the calculation of overlapping data, thereby saving the computing power

Examples

Embodiment Construction

[0016]To better understand the above-mentioned objectives, features and advantages of the disclosure, the solutions of the disclosure are further described below. It should be noted that, without contradiction, the embodiments of the disclosure and features of the embodiments can be combined with each other.

[0017]To address the technical problem in current related arts where, when analyzing and calculating noise data corresponding to each time-domain window, although there is data overlap between time-domain windows, an independent and complete calculation is performed for the noise audio corresponding to each time-domain window, leading to repeated analysis and calculation of overlapping data, resulting in calculation redundancy and waste of computing power, the present embodiment provides a noise data processing method. As illustrated in FIG. 1, the noise data processing method includes the following steps.

[0018]At step 101, first noise data corresponding to a spectrum component t...

Claims

1. A noise data processing method, comprising: obtaining first noise data corresponding to a spectrum component to be determined in a first time-domain window, wherein a length of the first time-domain window is an integer multiple of a step size; determining second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determining third noise data corresponding to the spectrum component within a last step size of a second time-domain window, wherein the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; and obtaining fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and adding the third noise data.

2. The noise data processing method of claim 1, wherein obtaining the first noise data corresponding to the spectrum component to be determined in the first time-domain window comprises: obtaining first time-domain data corresponding to the spectrum component in the first time-domain window; and obtaining the first noise data by performing a butterfly operation of Fast Fourier Transform (FFT) on the first time-domain data.

3. The noise data processing method of claim 2, wherein obtaining the first noise data by performing the butterfly operation of FFT on the first time-domain data comprises: obtaining a target frequency corresponding to the spectrum component, and determining a Nyquist frequency corresponding to the target frequency; obtaining, according to the Nyquist frequency, a number of time-domain sampling points Z required for the target frequency; and obtaining the first noise data by performing a butterfly recursive calculation of the FFT on a target sequence corresponding to first Z sampling points in a permutation sequence corresponding to the first time-domain data, wherein the permutation sequence is obtained by performing a butterfly odd-even permutation of the FFT according to the first time-domain data.

4. The noise data processing method of claim 1, wherein obtaining the first noise data corresponding to the spectrum component to be determined in the first time-domain window comprises: obtaining second time-domain data corresponding to the spectrum component within each step size in the first time-domain window; and obtaining the first noise data by performing a butterfly operation of the FFT on the second time-domain data respectively and combining calculation results in series.

5. The noise data processing method of claim 4, wherein obtaining the first noise data by performing the butterfly operation of the FFT on the second time-domain data respectively and combining the calculation results in series comprises: obtaining a polynomial corresponding to the second time-domain data by performing a butterfly operation of FFT on the second time-domain data respectively, and combining the polynomial corresponding to the second time-domain data with a root of unity respectively to synthesize a target polynomial of the first noise data, wherein the target polynomial is used for determining the first noise data.

6. The noise data processing method of claim 5, wherein obtaining the polynomial corresponding to the second time-domain data by performing the butterfly operation of FFT on the second time-domain data respectively, and combining the polynomial corresponding to the second time-domain data with the root of unity respectively to synthesize the target polynomial of the first noise data comprise: calculating the first noise data by using the following equation: Y k 1 = ∑ r = 0 R − 1 Y r ⋅ W N k sr where k represents the spectrum component, Yk1 represents the target polynomial of the first noise data, R represents a number of step sizes in the first time-domain window, Yr represents the polynomial of the second time-domain data corresponding to the rth step size, s represents a number of data points in the second time-domain data, and WNk represents the root of unity.

7. The noise data processing method of any one of claims 1-6, wherein determining the third noise data corresponding to the spectrum component within the last step size of the second time-domain window comprises: obtaining third time-domain data corresponding to the spectrum component within the last step size of the second time-domain window; and obtaining the third noise data by performing a butterfly operation of FFT on the third time-domain data.

8. A noise data processing apparatus, comprising: an obtaining section, configured to obtain first noise data corresponding to a spectrum component to be determined in a first time-domain window, wherein a length of the first time-domain window is an integer multiple of a step size; a determining section, configured to determine second noise data corresponding to the spectrum component within a first step size of the first time-domain window, and determine third noise data corresponding to the spectrum component within a last step size of a second time-domain window, wherein the second time-domain window is a time-domain window obtained by moving the first time-domain window by one step size; and a processing section, configured to obtain fourth noise data corresponding to the spectrum component in the second time-domain window by subtracting the second noise data from the first noise data and adding the third noise data.

9. A computer-readable storage medium, wherein the computer-readable storage medium stores computer programs, and when the computer programs are executed by a processor, the noise data processing method of any one of claims 1-7 is implemented.

10. An electronic device, comprises: a memory, configured to store computer programs; and a processor configured to implement the noise data processing method of any one of claims 1-7 when executing the computer programs stored in the memory.

11. A vehicle, comprising the noise data processing apparatus of claim 8 or the electronic device of claim 10.

12. A computer program product, comprising computer executable instructions or computer programs, wherein when the computer executable instructions or computer programs are executed by a processor, the noise data processing method of any one of claims 1-7 is implemented.

Images