Acoustic signal processing system based on data acquisition card

Abstract

The invention relates to the technical field of signal processing, in particular to an acoustic signal processing system based on a data acquisition card, which comprises a sound signal acquisition module, a signal decomposition and extraction module, a spectrum analysis and adjustment module and a spectrum output optimization module. According to the invention, by applying the wavelet transform and the convolutional neural network, the automation and accuracy of acoustic signal processing are significantly improved, and the introduction of the wavelet transform allows more detailed time-frequency decomposition of the sound signal, so that the ability of extracting key signal features from a complex environment is improved; the application of the convolutional neural network in spectrum analysis and adjustment enables the system to automatically identify and adapt to the difference of spectrum modes, improves the operation efficiency and the signal quality, reduces the dependence on manual intervention processing, and improves the accuracy of the system. And the working efficiency and reliability of the system in different application environments are enhanced.

Description

Technical Field

[0001] The invention relates to the technical field of signal processing, and in particular to an acoustic signal processing system based on a data acquisition card. Background Art

[0002] Signal processing is a technical field that focuses on signal representation, transformation and information extraction. Signals can be any physical quantity that transmits information, such as sound, electrical signals, images or more. Technically, signal processing covers the acquisition, analysis, modification and resolution of signals, including two major categories: digital signal processing (DSP) and analog signal processing. Digital signal processing uses computers or dedicated hardware (such as DSP chips) to process signals, involving filtering, Fourier transform, convolution and machine learning algorithms, which can improve, optimize or extract important features from signals.

[0003] Among them, the acoustic signal processing system of the data acquisition card uses the data acquisition card as the main tool to process acoustic signals. The data acquisition card is a hardware that can convert analog sound signals into digital signals so that the computer can further process and analyze the signals. It is usually used for noise assessment, environmental sound monitoring, audio equipment testing, and sound quality control. By using the data acquisition card, accurate time and frequency analysis can be achieved, thereby improving the analysis quality and application efficiency of acoustic data.

[0004] When processing non-standard environmental noise or high dynamic range audio signals, existing technologies cannot effectively distinguish useful signals from background noise, and require complex parameter adjustments and operations. This dependence limits the flexibility and application scope of the technology, especially under rapidly changing environmental conditions, where the processing effect and efficiency are poor. In the applications of noise assessment and sound quality control, existing technologies cannot achieve accurate spectrum adaptation, resulting in limited effects of sound quality improvement and noise control measures. The deficiencies of existing technologies in the automatic adjustment and optimization process result in the inability to achieve the expected accuracy and response speed in applications such as high-end audio equipment testing, affecting user experience and equipment performance. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an acoustic signal processing system based on a data acquisition card.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an acoustic signal processing system based on a data acquisition card, the system comprising:

[0007] The sound signal acquisition module tests the signal strength and frequency distribution based on the sound signal received by the data acquisition card, identifies the environmental noise level, adjusts the signal acquisition sensitivity, extracts key noise features, and generates noise-adjusted data;

[0008] The signal decomposition and extraction module uses wavelet transform to perform time-frequency decomposition on the sound signal based on the noise-adjusted data, processes the changes in multi-band energy, identifies key frequency bands and extracts related features to obtain key morphological features;

[0009] The spectrum analysis and adjustment module uses a convolutional neural network based on the key morphological features to compare the current sound spectrum mode with the standard mode, identify spectrum differences, calculate the required spectrum adjustment parameters, and automatically adjust the processing configuration in the data acquisition card to obtain the spectrum adjustment result;

[0010] The spectrum output optimization module adjusts the spectrum distribution of the sound signal in real time based on the spectrum adjustment result, evaluates the output sound quality of the sound signal, adjusts the output parameters according to the sound quality evaluation result, optimizes the sound quality output quality, and obtains the acoustic signal output result.

[0011] The present invention is improved in that the step of identifying the ambient noise level is specifically as follows:

[0012] Based on the sound signal received by the data acquisition card, the time series of the signal is recorded to obtain the time series data of the sound signal;

[0013] Based on the time series data of the sound signal, a fast Fourier transform is applied to convert the time series data into a frequency domain representation to obtain frequency distribution data;

[0014] Based on the frequency distribution data, the formula is used:

[0015]

[0016] Calculate the energy at each frequency point, identify the key noise source by comparing the energy density of the difference frequency, and calculate its total energy to obtain the ambient noise level, where F(f) represents the Fourier transform result at frequency f, and E(f) is the energy density of frequency f.

[0017] The present invention is improved in that the step of acquiring the noise-adjusted data is specifically as follows:

[0018] Evaluate the current signal acquisition sensitivity, adjust the sensitivity setting according to the environmental noise level, filter non-critical frequency noise, and obtain optimized acquisition parameters;

[0019] Based on the optimized acquisition parameters, the signal is collected again, and the signal in the key frequency range is highlighted by setting the frequency range of the filter, using the formula:

[0020]

[0021] Optimizing the signal strength of the key frequency to obtain processed signal data S′(f), where S(f) is the original signal spectrum, f0 is the center frequency, B is the bandwidth, and f is the frequency;

[0022] Based on the processed signal data, noise suppression processing is performed again, and spectral subtraction is used to remove background noise to obtain noise-adjusted data.

[0023] The present invention is improved in that the step of performing time-frequency decomposition on the sound signal is specifically as follows:

[0024] Based on the noise-adjusted data, wavelet transform is applied to perform time-frequency decomposition, the number of decomposition layers is determined, the instantaneous changes of the signal are captured, and preliminary time-frequency analysis data is obtained;

[0025] Analyze the preliminary time-frequency analysis data using the formula:

[0026] ES(f,t)=∑|WS(f,t)| 2

[0027] Highlight the key features in the signal and identify the key energy areas to obtain time-frequency decomposition data, where ES(f,t) represents the energy value at time t and frequency f, WS(f,t) represents the wavelet coefficient at time t and frequency f, and |WS(f,t)| represents the modulus of the wavelet coefficient, which is used to calculate the energy size of the time-frequency point.

[0028] The present invention is improved in that the step of acquiring the key morphological features is specifically as follows:

[0029] Using a data acquisition card to process the energy changes of the multi-bands, calculating the energy mean and standard deviation of each band, screening the key bands of energy changes, and obtaining preliminary band dynamic data;

[0030] The preliminary frequency band dynamics data are analyzed using the formula:

[0031]

[0032] Identify the key frequency bands and obtain the frequency band identification results, where CY(f) represents the energy concentration at frequency f, ES(f,t) represents the energy value at time t and frequency f, is the average energy value of frequency f, N is the total number of time points;

[0033] Based on the frequency band identification result, key statistical parameters of the signal morphology are extracted, including energy peak value, peak occurrence time and energy change speed data, to obtain key morphological features.

[0034] The present invention is improved in that the step of identifying the spectrum difference is specifically as follows:

[0035] Based on the key morphological features, a convolutional neural network is used to process the current sound spectrum pattern to obtain a spectrum feature representation;

[0036] Based on the spectral feature representation, the extracted sound spectral features are compared with the standard mode, using the formula:

[0037]

[0038] Calculate the key difference score ΔSG to obtain the spectrum difference data, where FG c (i) and FG s (i) represents the characteristic values ​​of the current spectrum and the standard spectrum at point i, respectively, n G is the number of spectral feature points.

[0039] The present invention is improved in that the step of obtaining the spectrum adjustment result is specifically:

[0040] Based on the frequency difference, the required spectrum adjustment parameters are calculated, and the parameter adjustment values ​​are determined according to the difference size and direction to obtain a preliminary adjustment configuration;

[0041] Based on the preliminary adjustment configuration, the processing configuration of the data acquisition card is automatically adjusted using the formula:

[0042] PH new =PH old +α H ΔPH

[0043] Get the spectrum adjustment result, where PH new is the adjusted parameter setting, PH old is the original parameter setting, ΔPH is the adjustment value, α H is the learning rate of the adjustment, which is used to control the adjustment amplitude.

[0044] The present invention is improved in that the step of obtaining the acoustic signal output result is specifically:

[0045] Based on the spectrum adjustment result, adjusting the sound signal processing configuration of the data acquisition card in real time to obtain an adjusted sound signal;

[0046] Based on the adjusted sound signal, the formula is adopted:

[0047]

[0048] Analyze the output sound quality of the sound signal and generate a sound quality evaluation result, where QT represents the sound quality score and SQ k is the frequency component of the adjusted sound signal, RQ kis the frequency content of the reference or standard signal, N Q is the number of frequency components;

[0049] Based on the sound quality evaluation result, output parameter adjustment is performed to optimize the sound quality output quality and obtain an acoustic signal output result.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are:

[0051] In the present invention, by applying wavelet transform and convolutional neural network, the automation and accuracy of acoustic signal processing are significantly improved. The introduction of wavelet transform allows a more detailed time-frequency decomposition of sound signals, thereby improving the ability to extract key signal features from complex environments, thereby achieving higher standards in sound quality control and noise assessment applications. The application of convolutional neural network in spectrum analysis and adjustment enables the system to automatically identify and adapt to differences in spectrum patterns, improves operational efficiency and signal quality, reduces reliance on manual intervention processing, and enhances the system's work efficiency and reliability in different application environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a system flow chart of the present invention;

[0053] Figure 2 is a flow chart for identifying the environmental noise level in the present invention;

[0054] Figure 3 is a flow chart for obtaining noise-adjusted data in the present invention;

[0055] Figure 4 This is a flow chart of performing time-frequency decomposition of a sound signal in the present invention;

[0056] Figure 5 It is a flow chart for obtaining key morphological features in the present invention;

[0057] Figure 6 This is a flow chart of identifying spectrum differences in the present invention;

[0058] Figure 7 A flowchart of obtaining spectrum adjustment results in the present invention;

[0059] Figure 8 This is a flow chart for obtaining the acoustic signal output result in the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0062] Example

[0063] See also Figure 1 The present invention provides a technical solution: an acoustic signal processing system based on a data acquisition card comprises:

[0064] The sound signal acquisition module tests the signal strength and frequency distribution based on the sound signal received by the data acquisition card, identifies the environmental noise level, adjusts the signal acquisition sensitivity, extracts key noise features, and generates noise-adjusted data;

[0065] The signal decomposition and extraction module uses wavelet transform to perform time-frequency decomposition of the sound signal based on the noise-adjusted data, uses the data acquisition card to process the changes in multi-band energy, identifies key frequency bands and extracts related features to obtain key morphological features;

[0066] The spectrum analysis and adjustment module uses a convolutional neural network based on key morphological features to compare the current sound spectrum pattern with the standard pattern, identify spectrum differences, calculate the required spectrum adjustment parameters, and automatically adjust the processing configuration in the data acquisition card to obtain the spectrum adjustment result;

[0067] The spectrum output optimization module adjusts the spectrum distribution of the sound signal in real time based on the spectrum adjustment result, evaluates the output sound quality of the sound signal through the data acquisition card, adjusts the output parameters according to the sound quality evaluation result, optimizes the sound quality output quality, and obtains the acoustic signal output result.

[0068] The data after noise adjustment includes frequency range, adjustment amplitude, and noise characteristic value. The key morphological characteristics include key frequency points, energy peak, and spectral line width. The spectrum adjustment results include adjustment coefficient, spectrum matching index, and comparison before and after adjustment. The acoustic signal output results include signal output quality, frequency response balance, and sound quality level.

[0069] See also Figure 2 , the specific steps for identifying the environmental noise level are:

[0070] Based on the sound signal received by the data acquisition card, the time series of the signal is recorded to obtain the time series data of the sound signal;

[0071] When using a data acquisition card to capture sound signals from the environment, the first thing to do is to set up and debug the hardware equipment to ensure that the sensitivity of the data acquisition card and the signal receiving range meet the test standards. Next, the system uses the data acquisition card to record the voltage fluctuations of the ambient sound in real time. The voltage fluctuation data is converted by the system into time series data of the sound signal, thereby forming raw data that can be further analyzed. The process involves signal acquisition, digital conversion, and data storage. Each step needs to be controlled to avoid data loss or interference, and finally the time series data is stored.

[0072] Based on the time series data of the sound signal, the fast Fourier transform is applied to convert the time series data into frequency domain representation to obtain frequency distribution data;

[0073] When converting the sound signal time series data into a frequency domain representation using the Fast Fourier Transform (FFT), it is necessary to select a suitable window function to reduce edge effects. Then, the discrete Fourier transform of the time series is quickly calculated using the FFT algorithm to convert the time series into a series of frequency components and their corresponding amplitude information. By analyzing the frequency components, the distribution of each frequency component in the sound signal can be clearly seen, providing accurate data support for subsequent environmental noise level analysis.

[0074] Based on the frequency distribution data, the formula is used:

[0075]

[0076] Calculate the energy at each frequency point, identify the key noise source by comparing the energy density of the difference frequency, and calculate its total energy to obtain the environmental noise level. Normalize the energy to a total energy of 1 to facilitate the comparison of noise levels under different environments or conditions, where F(f) represents the Fourier transform result at frequency f, and E(f) is the energy density of frequency f;

[0077] If at frequencies f1, f2, ..., f n The Fourier transform results measured at are:

[0078] F(f1)=20,F(f2)=30,...,F(f n )=10

[0079] Calculate the energy of each frequency point as:

[0080] |F(f1)| 2 =400,|F(f2)| 2 =900,...,|F(f n )| 2 =100

[0081] The total energy is:

[0082] ∑|F(f)| 2 =1400

[0083] then:

[0084]

[0085] The proportion of each frequency component in the total energy is obtained. By calculating E(f), the dominant noise source frequency can be identified and its contribution to the total noise can be accurately quantified, thereby helping to evaluate environmental noise and formulate control measures.

[0086] See also Figure 3 , the specific steps for obtaining the noise-adjusted data are:

[0087] Evaluate the current signal acquisition sensitivity, adjust the sensitivity setting according to the ambient noise level, filter non-critical frequency noise, and obtain optimized acquisition parameters;

[0088] In the process of evaluating and recording the current signal acquisition sensitivity, it is necessary to evaluate the initial sensitivity setting through environmental noise data, which involves analyzing the noise data collected in the early stage to determine the existence of non-critical frequencies and adjusting the acquisition parameters accordingly to filter out noise. The optimized acquisition parameters will be used to improve the quality and effect of signal acquisition, thereby ensuring that the device can more accurately capture sound signals within the critical frequency range in subsequent operations.

[0089] Based on the optimized acquisition parameters, the signal is collected again, and the signal in the key frequency range is highlighted by setting the frequency range of the filter. The formula is used:

[0090]

[0091] Optimize the signal strength of the key frequency to obtain the processed signal data S′(f), where S(f) is the original signal spectrum, rect is the rectangular filtering operation, f0 is the center frequency, B is the bandwidth, and f is the frequency;

[0092] If the spectrum of the original signal has significant energy at the key frequency f0 = 1000 Hz, the bandwidth B = 200 Hz, the spectrum value S(1000) = 50, the rectangular filter operation is selected That is, the signal between 900 and 1100 Hz, so The output of this frequency is:

[0093] S′(1000)=50×1=50

[0094] The result shows that the signal strengthens the signal at 1000 Hz, making this frequency component more prominent, which is helpful for further signal analysis and noise identification.

[0095] Based on the processed signal data, noise suppression processing is performed again, and the background noise is removed by using spectral subtraction to obtain noise-adjusted data;

[0096] The processed signal data is subjected to noise suppression processing again, which involves using spectral subtraction to further suppress the noise of the signal by subtracting the estimated noise spectrum. This process can clearly distinguish the effective components and noise components in the signal, thereby effectively improving the clarity and usability of the signal, generating noise-adjusted data, and showing a significant improvement in signal quality.

[0097] See also Figure 4 , the steps of time-frequency decomposition of sound signals are as follows:

[0098] Based on the noise-adjusted data, wavelet transform is applied to perform time-frequency decomposition, determine the number of decomposition layers, capture the instantaneous changes of the signal, and obtain preliminary time-frequency analysis data;

[0099] The noise-adjusted data is extracted from the data acquisition card. The data has been preliminarily processed to optimize the signal quality. Next, the wavelet transform is applied to perform time-frequency decomposition of the signal. The appropriate wavelet mother function and decomposition level are selected to ensure that all instantaneous changes of the signal can be captured, determine the quality and availability of the decomposed data, and obtain preliminary time-frequency analysis data, which will provide a basis for further energy feature analysis.

[0100] Analyze the preliminary time-frequency analysis data using the formula:

[0101] ES(f,t)=∑|WS(f,t)| 2

[0102] Highlight the key features in the signal and identify the key energy areas to obtain time-frequency decomposition data, where ES(f,t) represents the energy value at time t and frequency f, WS(f,t) represents the wavelet coefficient at time t and frequency f, and |WS(f,t)| represents the modulus of the wavelet coefficient, which is used to calculate the energy size of the time-frequency point;

[0103] If t = 1 second, the wavelet coefficient WS(1000,1) = 5 + 3i at the frequency f = 1000Hz, then its modulus is and then:

[0104] ES(1000,1)=(5.83) 2 =33.9889

[0105] The calculation process shows that the energy concentration at t = 1 second and f = 1000 Hz is 33.9889. The result can help understand the energy characteristics of the signal at this moment, which is crucial for further analyzing the behavior of the signal and extracting key features.

[0106] See also Figure 5 , the steps to obtain key morphological features are as follows:

[0107] The data acquisition card is used to process the changes in multi-band energy, reflecting the energy changes of each band at different time points, calculating the energy mean and standard deviation of each band, screening the key bands of energy changes, and obtaining preliminary band dynamic data;

[0108] The multi-band energy data processed by the data acquisition card reflects the energy changes of each frequency band at different time points, providing a basis for subsequent data analysis. By calculating the energy mean and standard deviation of each frequency band, not only can the average energy level of the frequency band during the entire observation period be analyzed, but also the degree of its energy fluctuation can be evaluated. Statistical data is the basis for screening those key frequency bands with significant changes. The selected key frequency bands show the areas with the largest energy changes, which plays a key role in identifying potential important events or state changes, thereby generating preliminary frequency band dynamic data.

[0109] The preliminary band dynamics data were analyzed using the formula:

[0110]

[0111] Identify the key frequency bands and obtain the frequency band identification results, where CY(f) represents the energy concentration at frequency f, ES(f,t) represents the energy value at time t and frequency f, is the average energy value of frequency f, N is the total number of time points;

[0112] There is a frequency f whose energy values ​​at three time points t=1, 2, and 3 are ES(f,1)=10, ES(f,2)=20, ES(f,3)=15, and the average energy value is The total number of time points N = 3, then:

[0113]

[0114] The results show that at frequency f, the concentration of energy change is 16.67, reflecting that the energy fluctuation of this frequency band is large and it is a key frequency band.

[0115] Based on the frequency band recognition results, the key statistical parameters of the signal morphology are extracted, including energy peak, peak occurrence time and energy change speed data, to obtain the key morphological features;

[0116] Based on the frequency band identification results, further analysis involves extracting key statistical parameters of the signal morphology, including energy peak, peak occurrence time, and speed of energy change. Statistical characteristics are the key to understanding signal behavior. The energy peak reflects the moment and size of maximum energy release, which is one of the most significant features of the signal. The peak occurrence time indicates the specific moment when the maximum energy release occurs, and the energy change rate reveals the speed of signal energy change, which is especially important for dynamic signal processing. The parameters together constitute the key morphological characteristics of the signal and provide necessary data support for applications such as monitoring, diagnosis or feature recognition.

[0117] See also Figure 6 , the specific steps of identifying spectrum differences are:

[0118] Based on key morphological features, a convolutional neural network is used to process the current sound spectrum pattern and obtain the spectrum feature representation;

[0119] Using the data of key morphological features, the current sound spectrum pattern is processed through a convolutional neural network. The network effectively extracts high-level features from the sound data through its multi-layer structure. The process not only includes data pre-processing such as noise filtering and standardization, but also involves the use of activation functions to enhance the recognition accuracy of the pattern. Each convolution layer processes the signal within a specific frequency range to generate a set of feature maps, which are then downsampled through a pooling layer to reduce the dimension while retaining important feature information. In addition, batch normalization is applied after each convolution layer to stabilize the learning process and accelerate the convergence speed of the network. Through a series of processing steps, the network outputs the feature representation of the current sound spectrum. The feature data not only accurately captures the essential characteristics of the sound, but also provides input for subsequent analysis and application.

[0120] Based on the spectral feature representation, the extracted sound spectral features are compared with the standard mode using the formula:

[0121]

[0122] Calculate the key difference score ΔSG to obtain the spectrum difference data, where FG c (i) and FG s (i) represents the characteristic values ​​of the current spectrum and the standard spectrum at point i, respectively, n G is the number of spectral feature points;

[0123] If the current sound spectrum (FG c ) and standard sound spectrum (FG s ) at the three feature points are FG c =[1,3,5 and FG s =[1,2,6], and the number of feature points n G=3, the calculation process is as follows:

[0124] ΔSG=|1-1|+|3-2|+|5-6|=0+1+1=2

[0125] The result shows that the total difference score between the current sound spectrum and the standard sound spectrum is 2, indicating that there is a slight difference, which provides a quantitative basis for subsequent adjustments and optimizations.

[0126] See also Figure 7 , the specific steps for obtaining the spectrum adjustment result are:

[0127] Based on the frequency difference, the required spectrum adjustment parameters are calculated, and the parameter adjustment values ​​are determined according to the difference size and direction to obtain a preliminary adjustment configuration;

[0128] Based on the monitored frequency differences, the required spectrum adjustment parameters are calculated through data analysis technology, including frequency analysis of the sound spectrum pattern in the original data acquisition card to identify the main differences from the preset standard pattern. After that, the system calculates the specific values ​​that need to be adjusted at each frequency point according to the size and direction of the difference. The calculation not only relies on the difference data, but also combines the impact of environmental factors and equipment performance to ensure the practicality and effectiveness of the adjustment configuration. The preliminary adjustment configuration obtained provides a theoretical basis and implementation guide for subsequent practical application adjustments. The configuration will then be applied in the processing configuration of the data acquisition card to achieve optimal adjustment of the sound spectrum.

[0129] Based on the preliminary adjustment configuration, the processing configuration of the data acquisition card is automatically adjusted using the formula:

[0130] PH new =PH old +α H ΔPH

[0131] Get the spectrum adjustment result, where PH new is the adjusted parameter setting, PH old is the original parameter setting, ΔPH is the adjustment value, α H is the adjusted learning rate, which is used to control the adjustment amplitude;

[0132] If the original parameter setting PH old is 50, the adjustment value ΔPH is 5, and the learning rate α H Set to 0.1, according to the formula, the new parameter sets PH new It can be calculated as follows:

[0133] PH new =50+0.1·5

[0134] PH new=50+0.5

[0135] PH new =50.5

[0136] The results show that after applying the adjustments, the new parameter setting is 50.5, an increase of 0.5 compared to the original setting, which reduces the difference between the measurement and the target, ensuring that the data acquisition card more accurately meets the expected operating standard or performance target.

[0137] See also Figure 8 , the steps for obtaining the acoustic signal output result are as follows:

[0138] Based on the spectrum adjustment result, the sound signal processing configuration of the data acquisition card is adjusted in real time to obtain the adjusted sound signal;

[0139] Based on the adjustment results, the data acquisition card is operated in real time to modify the processing configuration of the sound signal. The adjustment operation is based on the existing spectrum adjustment results. This process ensures that the adjustment parameters are applied in real time to respond to the current signal characteristics, including identifying spectrum differences and adjusting the processing method of the sound signal according to the differences, which directly affects the quality and output performance of the sound signal. By adjusting the settings of the data acquisition card, the sound signal is optimized and the adjusted sound signal is obtained.

[0140] Based on the adjusted sound signal, the formula is adopted:

[0141]

[0142] Analyze the output sound quality of the sound signal and generate a sound quality evaluation result, where QT represents the sound quality score and SQ k is the frequency component of the adjusted sound signal, RQ k is the frequency content of the reference or standard signal, N Q is the number of frequency components;

[0143] If N Q =3, the adjusted frequency components SQ1=50, SQ2=60, SQ3=55, the reference frequency components RQ1=40, RQ2=50, RQ3=45, then the calculation is as follows:

[0144]

[0145] QT=(0.25) 2 +(0.20) 2 +(0.22) 2

[0146] QT=0.0625+0.04+0.0484=0.1509≈0.15

[0147] The result shows that the total sound quality score is 0.15, reflecting that there is a significant difference between the adjusted sound signal and the reference signal, indicating that there is room for improvement in the sound quality. If the difference is large, further adjustment is needed.

[0148] Based on the sound quality evaluation results, the output parameters are adjusted to optimize the sound quality output quality. If the evaluation results show that the sound quality does not meet the expected standards, the frequency band balance and dynamic range are further fine-tuned to obtain the acoustic signal output results;

[0149] Based on the sound quality evaluation results, the output parameters are adjusted to optimize the sound output quality. If the evaluation results show that the sound quality does not meet the expected standards, further improvements are made by fine-tuning the frequency band balance and dynamic range. The adjustments are based on the results of systematic sound quality analysis to ensure that the output acoustic signal can meet user needs and application standards. Finally, the optimal acoustic signal output result is obtained through adjustment.

[0150] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An acoustic signal processing system based on a data acquisition card, characterized in that: The system comprises: The sound signal acquisition module tests the signal strength and frequency distribution based on the sound signal received by the data acquisition card, identifies the environmental noise level, adjusts the signal acquisition sensitivity, extracts key noise features, and generates noise-adjusted data; The signal decomposition and extraction module uses wavelet transform to perform time-frequency decomposition on the sound signal based on the noise-adjusted data, processes the changes in multi-band energy, identifies key frequency bands and extracts related features to obtain key morphological features; The spectrum analysis and adjustment module uses a convolutional neural network based on the key morphological features to compare the current sound spectrum mode with the standard mode, identify spectrum differences, calculate the required spectrum adjustment parameters, and automatically adjust the processing configuration in the data acquisition card to obtain the spectrum adjustment result; The spectrum output optimization module adjusts the spectrum distribution of the sound signal in real time based on the spectrum adjustment result, evaluates the output sound quality of the sound signal, adjusts the output parameters according to the sound quality evaluation result, optimizes the sound quality output quality, and obtains the acoustic signal output result.

2. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The step of identifying the environmental noise level is specifically as follows: Based on the sound signal received by the data acquisition card, the time series of the signal is recorded to obtain the time series data of the sound signal; Based on the time series data of the sound signal, a fast Fourier transform is applied to convert the time series data into a frequency domain representation to obtain frequency distribution data; Based on the frequency distribution data, the formula is used: Calculate the energy at each frequency point, identify the key noise source by comparing the energy density of the difference frequency, and calculate its total energy to obtain the ambient noise level, where F(f) represents the Fourier transform result at frequency f, and E(f) is the energy density of frequency f.

3. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The steps for obtaining the noise-adjusted data are specifically as follows: Evaluate the current signal acquisition sensitivity, adjust the sensitivity setting according to the environmental noise level, filter non-critical frequency noise, and obtain optimized acquisition parameters; Based on the optimized acquisition parameters, the signal is collected again, and the signal in the key frequency range is highlighted by setting the frequency range of the filter, using the formula: Optimizing the signal strength of the key frequency to obtain processed signal data S′(f), where S(f) is the original signal spectrum, f0 is the center frequency, B is the bandwidth, and f is the frequency; Based on the processed signal data, noise suppression processing is performed again, and spectral subtraction is used to remove background noise to obtain noise-adjusted data.

4. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The step of performing time-frequency decomposition on the sound signal is specifically as follows: Based on the noise-adjusted data, wavelet transform is applied to perform time-frequency decomposition, the number of decomposition layers is determined, the instantaneous changes of the signal are captured, and preliminary time-frequency analysis data is obtained; Analyze the preliminary time-frequency analysis data using the formula: ES(f,t)=∑|WS(f,t)| 2 Highlight the key features in the signal and identify the key energy areas to obtain time-frequency decomposition data, where ES(f,t) represents the energy value at time t and frequency f, WS(f,t) represents the wavelet coefficient at time t and frequency f, and |WS(f,t)| represents the modulus of the wavelet coefficient, which is used to calculate the energy size of the time-frequency point.

5. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The steps for obtaining the key morphological features are specifically as follows: Using a data acquisition card to process the energy changes of the multi-bands, calculating the energy mean and standard deviation of each band, screening the key bands of energy changes, and obtaining preliminary band dynamic data; The preliminary frequency band dynamics data are analyzed using the formula: Identify the key frequency bands and obtain the frequency band identification results, where CY(f) represents the energy concentration at frequency f, ES(f,t) represents the energy value at time t and frequency f, is the average energy value of frequency f, N is the total number of time points; Based on the frequency band identification result, key statistical parameters of the signal morphology are extracted, including energy peak value, peak occurrence time and energy change speed data, to obtain key morphological features.

6. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The steps of identifying the spectrum difference are specifically as follows: Based on the key morphological features, a convolutional neural network is used to process the current sound spectrum pattern to obtain a spectrum feature representation; Based on the spectral feature representation, the extracted sound spectral features are compared with the standard mode, using the formula: Calculate the key difference score ΔSG to obtain the spectrum difference data, where FG c (i) and FG s (i) represents the characteristic values ​​of the current spectrum and the standard spectrum at point i, respectively, n G is the number of spectral feature points.

7. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The steps of obtaining the spectrum adjustment result are specifically as follows: Based on the frequency difference, the required spectrum adjustment parameters are calculated, and the parameter adjustment values ​​are determined according to the difference size and direction to obtain a preliminary adjustment configuration; Based on the preliminary adjustment configuration, the processing configuration of the data acquisition card is automatically adjusted using the formula: PH new =PH old +α H ·ΔPH Get the spectrum adjustment result, where PH new is the adjusted parameter setting, PH old is the original parameter setting, ΔPH is the adjustment value, α H is the learning rate of the adjustment, which is used to control the adjustment amplitude.

8. The acoustic signal processing system based on the data acquisition card according to claim 1, characterized in that: The steps for obtaining the acoustic signal output result are specifically as follows: Based on the spectrum adjustment result, adjusting the sound signal processing configuration of the data acquisition card in real time to obtain an adjusted sound signal; Based on the adjusted sound signal, the formula is adopted: Analyze the output sound quality of the sound signal and generate a sound quality evaluation result, where QT represents the sound quality score and SQ k is the frequency component of the adjusted sound signal, RQ k is the frequency content of the reference or standard signal, N Q is the number of frequency components; Based on the sound quality evaluation result, the output parameters are adjusted to optimize the sound quality output and obtain the acoustic signal output result.

Images