Audio signal processing method and system

Abstract

The present invention relates to the field of data processing technology, and in particular to an audio signal processing method and system. The method comprises the following steps: acoustically simulating site plan data to generate acoustic topology data; collecting sound signals according to the acoustic topology data to generate multi-directional audio data; quantizing the multi-directional audio data to generate spectrum gradient matrix data; selecting high entropy data from the spectrum gradient matrix data to generate key feature data; optimizing the model based on the key feature data to generate a spatial audio reconstruction model; iteratively optimizing the spatial audio reconstruction model according to the site dynamic change parameters to generate a dynamic audio reconstruction model; collecting feedback data using sensor equipment to generate real-time feedback data; performing feedback control according to the real-time feedback data to generate sound propagation optimization data. The present invention improves the real-time transmission quality of audio signals by analyzing the sound field characteristics and environmental changes.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an audio signal processing method and system. Background Art

[0002] In large-scale occasions, making the sound propagate more clearly while improving the perceptibility and coverage of the sound will help enhance the audience's understanding and absorption of the content and improve the user's auditory experience. For the leader, it means being able to communicate better with the audience and ensure that the information is conveyed more accurately and vividly. However, the existing audio signal processing methods and systems are limited by the propagation range or have weak adaptability to changes, which results in them being unable to provide users with a good auditory experience in actual use. Summary of the invention

[0003] Based on this, it is necessary to provide an audio signal processing method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, an audio signal processing method comprises the following steps:

[0005] Step S1: Acquire site plan data; digitize the site plan data to generate site grid data; perform acoustic simulation based on the site grid data to generate propagation simulation result data; create a topology map based on the propagation simulation result data to generate acoustic topology map data;

[0006] Step S2: Arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data;

[0007] Step S3: performing information entropy evaluation on the spectrum gradient matrix data to generate information entropy evaluation data; selecting high entropy data based on the information entropy evaluation data to generate key feature data; optimizing the preset audio reconstruction model based on the key feature data to generate a spatial audio reconstruction model;

[0008] Step S4: Acquire the dynamic change parameters of the site; perform influence assessment according to the dynamic change parameters of the site to generate factor influence degree data; perform dynamic change simulation according to the factor influence degree data to generate dynamic simulation result data; perform reflection network construction according to the dynamic simulation result data to generate dynamic reflection network data;

[0009] Step S5: setting an optimization target based on the dynamic reflection network data to generate optimization target information data; iteratively optimizing the spatial audio reconstruction model using the optimization target information data to generate a dynamic audio reconstruction model;

[0010] Step S6: Use sensor equipment to collect feedback data and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use a dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

[0011] The present invention creates an acoustic topology map of the venue through digital processing and acoustic simulation, which helps to understand the propagation characteristics and paths of sound at different locations. The generated acoustic topology map data is the basis for optimizing the sound propagation system. Through the optimized arrangement and signal acquisition of sensors, as well as the generation of spectrum gradient matrix data, more comprehensive and multi-angle sound information can be provided, which is very critical for understanding the overall picture, characteristics and complexity of sound propagation. Information entropy evaluation and extraction of key feature data help to determine which features are the most representative and important. These data play a key role in adjusting the audio reconstruction model, helping the system to reconstruct the sound more accurately to adapt to different scenes and environments. Dynamically changing parameter acquisition and generation of dynamic simulation result data help to understand the impact of real-time changes in the environment on sound propagation. By evaluating the degree of impact and constructing a dynamic reflection network, the system can more accurately simulate the propagation of sound in different environments, providing more comprehensive data support for sound propagation optimization. Optimizing the generation of target information data and iterative optimization of the dynamic audio reconstruction model can adjust the sound reconstruction model according to real-time changes, ensuring the adaptability and quality of sound propagation in different environments. It helps to continuously improve and adjust the system to adapt to the changing environment. The collection of real-time feedback data and the application of feedback control can achieve dynamic adjustment and optimization of sound propagation. Through the real-time collection of audio feedback data and environmental impact data, timely adjustments can be made to keep sound propagation stable and clear in a changing environment, thereby improving the auditory experience. Therefore, the audio signal processing method and system of the present invention obtains the audio propagation characteristics corresponding to different environments by analyzing the sound field characteristics to obtain a spatial audio reconstruction model, and performs a change propagation simulation by considering environmental changes, thereby comprehensively considering the site and environmental factors, and dynamically reconstructing the audio signal, thereby improving the transmission quality of the audio signal to improve the user's auditory experience.

[0012] Preferably, step S1 comprises the following steps:

[0013] Step S11: Acquire site plan data; digitize the site plan data to generate site grid data;

[0014] Step S12: Presetting the location of the sound amplification equipment for the site grid data based on the acoustic propagation principle to generate equipment orientation preset data;

[0015] Step S13: marking the site grid data with equipment information according to the equipment position preset data to generate the position preset grid data;

[0016] Step S14: formulating facility material parameters based on the site plan data to generate material parameter data;

[0017] Step S15: performing acoustic simulation on the azimuth preset grid data through the material parameter data to generate propagation simulation result data;

[0018] Step S16: selecting the best result according to the propagation simulation result data to generate the best simulation result data;

[0019] Step S17: performing multi-dimensional acoustic characteristic analysis according to the optimal simulation result data to generate propagation acoustic characteristic data;

[0020] Step S18: Create a topology map based on the propagation acoustic characteristic data to generate acoustic topology map data.

[0021] The present invention provides an accurate description of the site through digital processing of the site plan and generation of raster data, which helps to understand the layout and structure of the site. The device position is preset and the device information is marked, and the reasonable position and orientation of the sound amplification device are determined through the acoustic propagation principle and the site raster data. It helps to ensure the rationality and uniformity of the sound coverage and improve the effect of sound propagation. The formulation of facility material parameters is crucial to the influence of sound propagation. By formulating the site material parameters, the reflection, absorption and propagation of sound on the surfaces of different materials can be more accurately simulated, which helps to optimize the design and effect of the sound propagation system. The formulation of facility material parameters provides important data on the acoustic properties of the site materials such as sound absorption, reflection, and conduction. It has an important impact on the propagation mode of sound in a specific environment. Acoustic simulation can simulate the propagation of sound in the site based on the facility material parameters and the preset raster data of the orientation. Through the simulation result data, the propagation of sound in different areas can be understood. By optimally selecting the propagation simulation results, the best sound propagation plan is determined. The optimal result data can help the system find the most suitable sound propagation strategy for a specific scene and improve the sound propagation effect. Multidimensional acoustic characteristics analysis is an in-depth study of propagation results, providing more detailed information about sound propagation characteristics, such as sound pressure level, sound wave propagation direction, etc., which helps to understand the details of sound propagation more comprehensively. The creation of acoustic topology data visualizes the propagation path and characteristics of sound in the venue. It helps to intuitively understand the patterns and laws of sound propagation.

[0022] Preferably, step S2 comprises the following steps:

[0023] Step S21: performing a propagation key area analysis based on the acoustic topology data to generate propagation key area data; performing an actual layout analysis based on the propagation key area data to generate sensor location data;

[0024] Step S22: Calculate the frequency weight of the sensor position data using the sampling frequency evaluation formula to generate sampling weight frequency data;

[0025] Step S23: Correcting the position and sampling frequency of the sensor position data according to the sampling weight frequency data to generate position information correction data; arranging sensors based on the position information correction data, and collecting sound signals to generate multi-directional audio data;

[0026] Step S24: performing spectrum time domain analysis on the multi-directional audio data to generate collected acoustic feature data;

[0027] Step S25: using the propagation acoustic characteristic data to perform characteristic error evaluation on the collected acoustic characteristic data to generate actual sampling error data; performing error cause analysis based on the actual sampling error data to generate error adjustment data;

[0028] Step S26: Optimizing the sensor arrangement according to the error adjustment data, and performing re-collection analysis and evaluation, thereby generating optimized collection feature data;

[0029] Step S27: quantizing the optimized acquisition feature data based on the acoustic topology data to generate spectrum gradient matrix data.

[0030] The present invention helps to determine the key positions and areas in sound propagation through the analysis of key area data, so as to determine which positions need special attention and optimize sound propagation. Frequency weight calculation is to determine the importance of sensor position. By evaluating the sampling frequency of the sensor position, the sensor layout can be optimized to ensure that more comprehensive and representative sound data are collected in the key area. Position information correction and sensor layout are crucial for collecting multi-directional audio data. By correcting the position information and rearranging the sensors, the comprehensiveness and accuracy of data collection can be ensured. Spectral time domain analysis of multi-directional audio data can extract and quantify the characteristics of sound, which is very important for understanding the propagation characteristics and acoustic characteristics of sound in different areas, and is helpful for subsequent sound processing and optimization. Feature error evaluation can identify and quantify the errors of acoustic feature data, which helps to understand the accuracy and reliability of data during the collection process. Analyzing the actual sampling error helps to find the cause of the error. It can reveal possible problems or defects in the collection process. Error adjustment data is used to correct the optimal layout of the sensor. By adjusting the error, the sensor can be rearranged to improve the accuracy and reliability of data collection to obtain higher quality collection feature data. The generation of spectral gradient matrix data is based on the quantitative processing of optimized collection feature data. These data are abstract representations of sound characteristics and can more fully describe the characteristics of sound propagation.

[0031] Preferably, the sampling frequency evaluation formula in step S22 is as follows:

[0032] ;

[0033] In the formula, is the sampling frequency weight factor, For sensor The sensitivity of For sensor The sensitivity of For sensor and The maximum distance between is the maximum sensitivity of all sensors. is the maximum distance between all sensors, is the average distance between all sensors, For sensor The horizontal axis, For sensor The horizontal axis, For sensor and The Euclidean distance between .

[0034] The present invention provides a sampling frequency evaluation formula for calculating the frequency weight of sensor position data and generating sampling weight frequency data. Sensitivity ,sensor Sensitivity ,sensor and The maximum distance between , the highest sensitivity among all sensors , the maximum distance between all sensors , the average distance between all sensors ,sensor The horizontal axis ,sensor The horizontal axis ,sensor and The Euclidean distance between As well as the interaction effects between variables, the following functional relationship is formed:

[0035] ;

[0036] pass and The product of the sensor and The higher the sensitivity, the stronger the ability to capture sound signals, so the weight factor will increase. The maximum sensitivity of all sensors Used to normalize the relative sensitivity. and The maximum distance between The larger the value, the corresponding sampling frequency will also increase, which means that when the distance between sensors is large, the sensors and The maximum distance between all sensors is The larger the value, the corresponding sampling frequency decreases, and the sensor with the largest distance difference in the sensor arrangement contributes less to the sampling frequency. The smaller it is, the corresponding sampling frequency increases, which means that the average distance in the sensor arrangement is smaller, which is conducive to capturing the sound signal more comprehensively. and The Euclidean distance between , indicating the sensor and The distance in three-dimensional space. The farther the sensor is, and the more different the direction of the sensor is, the smaller its weight is, reflecting the weaker the association between them. This part represents the sensor and This part measures the impact of the angle between sensors on the calculation. A smaller angle means that the directions between sensors are more similar, which may introduce unnecessary redundant information. A larger angle means that the directionality between sensors is weaker, which may lead to inaccurate data. By considering the impact of the angle, the spatial relationship between sensors can be better measured, thereby improving the reliability and accuracy of the data. The purpose of this part is to take into account the ratio of the distance between sensors to the maximum distance. A larger distance ratio means that the distribution between sensors is sparse, which may result in missing data. A smaller distance ratio means that the distribution between sensors is denser, which can provide more comprehensive and accurate data. By considering the impact of the distance ratio, the distance between sensors can be better balanced, thereby improving the comprehensiveness and accuracy of the data. This functional relationship can reflect the importance or priority of sensors at different locations in sound sampling. A higher weight indicates that the sensor For adjacent sensors It is even more important in sound sampling.

[0037] Preferably, step S3 comprises the following steps:

[0038] Step S31: feature labeling the spectrum gradient matrix data to generate labeling matrix data; perform normalization processing based on the labeling matrix data to generate normalized matrix data;

[0039] Step S32: performing data association analysis on the normalized matrix data to generate association analysis result data; performing information entropy evaluation based on the association analysis result data to generate information entropy evaluation data;

[0040] Step S33: selecting high entropy data from the normalized matrix data according to the information entropy evaluation data to generate key feature data; performing coding mapping conversion on the key feature data to generate feature coding data;

[0041] Step S34: optimizing the preset audio reconstruction model based on the feature coding data to generate a spatial audio reconstruction model.

[0042] The present invention helps to extract and identify key features in the data by feature labeling the spectrum gradient matrix data. It helps to understand the meaning and importance of the data. Normalization processing can standardize the data so that different data are comparable. This processing method helps to eliminate the dimension and size differences between different data and provide a consistent benchmark for data analysis. Data association analysis can help identify the relationship and correlation between data. It is very important to understand the connection between data and how they affect each other, and provide support for further analysis and interpretation. High entropy data selection is based on information entropy evaluation data, which helps to determine the most important and representative features in the data. It helps to simplify the data set, reduce redundant information, and make the data more compact and effective. Through encoding mapping conversion, the selected key feature data can be converted and encoded. This conversion may include recombination or encoding of features, so that the data is more suitable for subsequent model processing and interpretation. Model tuning is based on feature encoding data, which helps to optimize the preset audio reconstruction model. By adjusting and optimizing the model, the performance and accuracy of the model can be improved, so that it can better adapt to the characteristics of the data corresponding to the sound field, thereby generating a more accurate and effective spatial audio reconstruction model.

[0043] Preferably, step S4 comprises the following steps:

[0044] Step S41: Acquire the dynamic change parameters of the venue;

[0045] Step S42: Perform influence assessment based on the site dynamic change parameters to generate factor influence degree data;

[0046] Step S43: formulating environmental simulation parameters according to the factor influence degree data to generate multi-environment parameter data;

[0047] Step S44: using the multi-environment parameter data and based on the azimuth preset grid data to perform dynamic change simulation and generate dynamic simulation result data;

[0048] Step S45: constructing a reflection network according to the dynamic simulation result data to generate dynamic reflection network data.

[0049] The present invention obtains the dynamic change parameters of the site to collect data on the changes of the site environment over time. These data may include parameters such as temperature, humidity, and flow of people, which are helpful to understand the real-time state of the environment. The generation of factor influence degree data involves the evaluation and analysis of the dynamic change parameters of the site, which can determine the degree of influence of different factors on sound propagation. It helps to quantify the impact of different factors on sound propagation and provide an important basis for subsequent simulations. The formulation of environmental simulation parameters is based on the factor influence degree data, which can convert the influence of factors into simulatable parameters. These parameters can be used to simulate different states of the site environment and simulate sound propagation more accurately. The generation of dynamic simulation result data is based on the simulation of multi-environment parameter data. These data provide simulation results of sound propagation under different environmental conditions, which helps to understand the behavior and characteristics of sound in different environments.

[0050] Preferably, step S5 comprises the following steps:

[0051] Step S51: Analyze the sound propagation characteristics based on the dynamic reflection network data to generate reflection characteristic data; set optimization targets based on the reflection characteristic data to generate optimization target information data;

[0052] Step S52: Calculate the sound pressure change at the receiving point of the multiple environmental parameter data using the sound pressure change evaluation formula to generate multiple sound pressure change data;

[0053] Step S53: using the multiple sound pressure change data to perform associated target correction on the optimization target information data to generate optimization target correction data;

[0054] Step S54: adjusting the model parameters of the spatial audio reconstruction model according to the optimization target correction data to generate a parameter adjustment model;

[0055] Step S55: Utilize the multi-environment parameter data to iteratively optimize and evaluate the parameter adjustment model to generate a dynamic audio reconstruction model.

[0056] The present invention can generate reflection characteristic data based on dynamic reflection network data through sound propagation characteristic analysis, so as to gain an in-depth understanding of the reflection of sound in different environments. These data are helpful in identifying the special characteristics of sound propagation and provide a basis for optimization. Using the sound pressure change evaluation formula to calculate the sound pressure change can generate multiple sound pressure change data, which reflects the change of sound pressure at the receiving point under different environmental conditions and is a key indicator for understanding the sound propagation effect. The associated target correction corrects the optimization target information data through multiple sound pressure change data, which helps to set and adjust the optimization target more accurately and ensure that the optimization target meets the actual situation. By adjusting the model parameters of the spatial audio reconstruction model through the optimization target correction data, the performance and accuracy of the model can be improved, so that it can better adapt to the reconstruction requirements of sound in different environments. The parameter adjustment model is iteratively optimized and evaluated using multi-environment parameter data to generate a dynamic audio reconstruction model. This model can better adapt to the sound propagation characteristics in different environments and provide more accurate model support for sound reconstruction.

[0057] Preferably, the sound pressure change evaluation formula in step S52 is as follows:

[0058] ;

[0059] In the formula, is the sound pressure level at the receiving point, is the number of sound sources, is the index value of the sound source, For sound source The initial sound pressure level, is the attenuation coefficient of the sound propagation medium, For sound source The distance to the receiving point, For sound source The average distance is the sound pressure attenuation coefficient, For sound source The length of the sound propagation path to the receiving point, For sound source The average path length, is the environmental factor constant, For sound source The surface area of is the effective cross-sectional area of ​​the sound propagation path, is the deviation correction value of the sound pressure level at the receiving point.

[0060] The present invention provides a sound pressure change evaluation formula, which is used to calculate the sound pressure change of the receiving point for multiple environmental parameter data and generate multiple sound pressure change data. The formula fully considers the number of sound sources. , the index value of the sound source , sound source Initial sound pressure level , the attenuation coefficient of the sound propagation medium , sound source Distance to receiving point , sound source The average distance , sound pressure attenuation coefficient , sound source The length of the sound propagation path to the receiving point , sound source The average path length , environmental factor constant , sound source Surface area , the effective cross-sectional area of ​​the sound propagation path , the deviation correction value of the sound pressure level at the receiving point As well as the interaction effects between variables, the following functional relationship is formed:

[0061] ;

[0062] The number of sound sources Indicates how many sound sources are present. Initial sound pressure level The starting sound pressure level used to calculate the sound source is the initial point of the sound pressure level change. The logarithm of the ratio of the actual distance to the average distance is calculated to represent the attenuation of sound due to distance. The constant n represents the attenuation coefficient of sound in the propagation medium, which represents the various attenuations encountered by sound during propagation, such as air absorption and the density of the propagation medium. The magnitude of the attenuation coefficient n determines the attenuation rate of sound during propagation. This part of the calculation represents the effect of distance attenuation, which serves to amplify the attenuation to more prominently reflect the effect of sound attenuation as the distance increases. The logarithm of the ratio of the actual path length to the average path length is calculated for the length of the sound propagation path. The effect of the propagation path length is expressed by the sound pressure attenuation coefficient The effect of the path is significantly amplified to more strongly reflect the effect of the sound propagation path, the sound pressure attenuation coefficient in a free field environment The theoretical value is 20, because the sound pressure level drops by 20 decibels for each distance multiple. In actual situations, it may vary depending on the environment, so it is represented by an unknown number in the formula to adapt to the calculation of sound pressure changes in different environmental conditions. Environmental factors and the ratio of the surface area of ​​the sound source to the effective cross-sectional area of ​​the sound propagation path are taken into account. This part represents the impact of the environment on sound propagation. For example, if the environment is complex or the surface area of ​​the sound source is small, it will affect the propagation and attenuation of the sound. This functional relationship can simulate and predict the attenuation and propagation of sound under different environmental conditions, providing an important basis for optimizing sound propagation. And use the deviation correction value of the sound pressure level at the receiving point Adjust and correct the functional relationship to reduce the error caused by this data or error term, so as to more accurately generate the sound pressure level of the receiving point , which improves the accuracy and reliability of the sound pressure receiving value evaluation of the sensor equipment. At the same time, the attenuation coefficient of the sound propagation medium and the sound pressure attenuation coefficient in the formula can be adjusted according to the actual situation, and applied to different multi-environment parameter data to calculate the sound pressure change of the receiving point, which improves the flexibility and applicability of the algorithm.

[0063] Preferably, step S6 comprises the following steps:

[0064] Step S61: using sensor equipment to collect real-time audio signals and generate real-time audio feedback data;

[0065] Step S62: using sensor equipment to collect environmental change data and generate environmental change data; filtering the environmental change data according to the factor impact degree data and generating real-time environmental impact data;

[0066] Step S63: formulating reconstruction parameters based on the real-time audio feature data and the real-time environmental impact data to generate device reconstruction parameter data;

[0067] Step S64: Utilize the dynamic audio reconstruction model, perform feedback control based on the device reconstruction parameter data, and transmit the audio signal to generate sound propagation optimization data.

[0068] The present invention can provide instant audio feedback data through real-time audio signal collection, which helps to understand the performance of sound in the current environment in real time. The collection of environmental change data and the screening of factor influence data can generate real-time environmental impact data. These data record the real-time impact of environmental changes on sound propagation, helping to better understand the degree of impact of environmental changes on sound. Equipment reconstruction parameter data is formulated based on real-time audio feature data and real-time environmental impact data. These parameters are important factors for optimizing sound propagation according to real-time conditions, and help to improve the adaptability and effect of sound propagation. A dynamic audio reconstruction model is used and feedback control is performed based on the equipment reconstruction parameter data, and audio signals are transmitted to generate sound propagation optimization data. These data show the sound propagation after dynamic adjustment and feedback control, which helps to evaluate the propagation optimization effect and the effect of real-time adjustment of sound propagation.

[0069] The present invention also provides an audio signal processing system, which is used to execute the audio signal processing method as described above, comprising:

[0070] The acoustic field topology construction module is used to obtain the site plan data; digitize the site plan data to generate site grid data; perform acoustic simulation based on the site grid data to generate propagation simulation result data; create a topology map based on the propagation simulation result data to generate acoustic topology map data.

[0071] The sound field feature analysis module is used to arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data;

[0072] The data coding optimization module is used to perform information entropy evaluation on the spectrum gradient matrix data to generate information entropy evaluation data; select high entropy data based on the information entropy evaluation data to generate key feature data; and optimize the preset audio reconstruction model based on the key feature data to generate a spatial audio reconstruction model.

[0073] The environmental change simulation module is used to obtain the dynamic change parameters of the site; conduct influence assessment based on the dynamic change parameters of the site to generate factor influence degree data; conduct dynamic change simulation based on the factor influence degree data to generate dynamic simulation result data; construct a reflection network based on the dynamic simulation result data to generate dynamic reflection network data;

[0074] The model reconstruction enhancement module is used to set optimization targets based on dynamic reflection network data and generate optimization target information data; the optimization target information data is used to iteratively optimize the spatial audio reconstruction model to generate a dynamic audio reconstruction model;

[0075] The sound dynamic adjustment module is used to collect feedback data using sensor equipment and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use the dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

[0076] The beneficial effect of the present application is that the present invention generates site grid data by acquiring and digitally processing site plan data. The propagation simulation result data is generated according to the grid data using acoustic simulation technology, and then the acoustic topology map data is created. It is helpful to have a deep understanding of the propagation mode and acoustic characteristics of sound in the site. The sensor is optimally arranged through multi-directional audio data, and the signal is re-collected to generate optimized acquisition feature data. It is helpful to collect more comprehensive and optimized sound data. The information entropy of the spectrum gradient matrix data is evaluated to generate information entropy evaluation data. It is helpful to understand the complexity and information volume of the data, determine the key features of the data, and then tune the audio reconstruction model. Finally, a spatial audio reconstruction model is generated to improve the accuracy and effectiveness of audio processing. By acquiring the dynamic change parameters of the site and evaluating the influence according to these parameters, factor influence degree data is generated. Dynamic change simulation is performed based on these data to generate dynamic simulation result data. The reflection network is constructed through these data to generate dynamic reflection network data, which helps to more accurately simulate and understand the propagation and reflection of sound in different environments. Optimization targets are set based on dynamic reflection network data to generate optimization target information data. It helps to clarify the direction and goal of optimization, and then use the optimization target data to iteratively optimize the spatial audio reconstruction model to generate a dynamic audio reconstruction model. The model can better adapt to dynamically changing scenes and improve the accuracy and real-time performance of sound reconstruction. Sensor equipment is used to collect feedback data to generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data. Real-time feedback data is used for feedback control to achieve real-time monitoring and adjustment of the sound propagation process, which helps to optimize the sound propagation effect and improve the quality and adaptability of the sound. Therefore, the audio signal processing method and system of the present invention obtains the spatial audio reconstruction model by analyzing the sound field characteristics, and performs change propagation simulation by considering environmental changes to obtain audio propagation characteristics corresponding to different environments, thereby comprehensively considering the site and environmental factors, and dynamically reconstructing the audio signal, thereby improving the transmission quality of the audio signal to improve the user's auditory experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 A schematic diagram of a step flow of an audio signal processing method;

[0078] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0079] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0080] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0081] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0082] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0083] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0084] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0085] To achieve this, please refer to Figures 1 to 4 , an audio signal processing method, comprising the following steps:

[0086] Step S1: Acquire site plan data; digitize the site plan data to generate site grid data; perform acoustic simulation based on the site grid data to generate propagation simulation result data; create a topology map based on the propagation simulation result data to generate acoustic topology map data;

[0087] Step S2: Arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data;

[0088] Step S3: performing information entropy evaluation on the spectrum gradient matrix data to generate information entropy evaluation data; selecting high entropy data based on the information entropy evaluation data to generate key feature data; optimizing the preset audio reconstruction model based on the key feature data to generate a spatial audio reconstruction model;

[0089] Step S4: Acquire the dynamic change parameters of the site; perform influence assessment according to the dynamic change parameters of the site to generate factor influence degree data; perform dynamic change simulation according to the factor influence degree data to generate dynamic simulation result data; perform reflection network construction according to the dynamic simulation result data to generate dynamic reflection network data;

[0090] Step S5: setting an optimization target based on the dynamic reflection network data to generate optimization target information data; iteratively optimizing the spatial audio reconstruction model using the optimization target information data to generate a dynamic audio reconstruction model;

[0091] Step S6: Use sensor equipment to collect feedback data and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use a dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

[0092] The present invention creates an acoustic topology map of the venue through digital processing and acoustic simulation, which helps to understand the propagation characteristics and paths of sound at different locations. The generated acoustic topology map data is the basis for optimizing the sound propagation system. Through the optimized arrangement and signal acquisition of sensors, as well as the generation of spectrum gradient matrix data, more comprehensive and multi-angle sound information can be provided, which is very critical for understanding the overall picture, characteristics and complexity of sound propagation. Information entropy evaluation and extraction of key feature data help to determine which features are the most representative and important. These data play a key role in adjusting the audio reconstruction model, helping the system to reconstruct the sound more accurately to adapt to different scenes and environments. Dynamically changing parameter acquisition and generation of dynamic simulation result data help to understand the impact of real-time changes in the environment on sound propagation. By evaluating the degree of impact and constructing a dynamic reflection network, the system can more accurately simulate the propagation of sound in different environments, providing more comprehensive data support for sound propagation optimization. Optimizing the generation of target information data and iterative optimization of the dynamic audio reconstruction model can adjust the sound reconstruction model according to real-time changes, ensuring the adaptability and quality of sound propagation in different environments. It helps to continuously improve and adjust the system to adapt to the changing environment. The collection of real-time feedback data and the application of feedback control can achieve dynamic adjustment and optimization of sound propagation. Through the real-time collection of audio feedback data and environmental impact data, timely adjustments can be made to keep sound propagation stable and clear in a changing environment, thereby improving the auditory experience. Therefore, the audio signal processing method and system of the present invention obtains the audio propagation characteristics corresponding to different environments by analyzing the sound field characteristics to obtain a spatial audio reconstruction model, and performs a change propagation simulation by considering environmental changes, thereby comprehensively considering the site and environmental factors, and dynamically reconstructing the audio signal, thereby improving the transmission quality of the audio signal to improve the user's auditory experience.

[0093] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of an audio signal processing method of the present invention. In this example, the audio signal processing method includes the following steps:

[0094] Step S1: Acquire site plan data; digitize the site plan data to generate site grid data; perform acoustic simulation based on the site grid data to generate propagation simulation result data; create a topology map based on the propagation simulation result data to generate acoustic topology map data;

[0095] In the embodiment of the present invention, the floor plan data of the required site is obtained, which can be data obtained by field measurement or digitalization. The obtained site floor plan data is digitized and converted into raster data for acoustic simulation. The site raster data is used for acoustic simulation, and the propagation and reflection of sound in the site are simulated by applying acoustic principles and numerical simulation methods to obtain propagation simulation result data. Based on the propagation simulation result data, acoustic topology data is created. This step includes determining the nodes and edges of the acoustic topology and describing the path and relationship of sound propagation.

[0096] Step S2: Arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data;

[0097] In the embodiment of the present invention, by arranging sensors based on acoustic topology data, placing sensors at appropriate positions to collect sound signals, and generating multi-directional audio data. The collected multi-directional audio data is analyzed and evaluated, the position arrangement of the sensors is optimized, and the signal is collected again to obtain optimized collected feature data. The optimized collected feature data is quantified, and a suitable algorithm or method is applied to generate spectrum gradient matrix data.

[0098] Step S3: performing information entropy evaluation on the spectrum gradient matrix data to generate information entropy evaluation data; selecting high entropy data based on the information entropy evaluation data to generate key feature data; optimizing the preset audio reconstruction model based on the key feature data to generate a spatial audio reconstruction model;

[0099] In an embodiment of the present invention, information entropy evaluation is performed on the spectrum gradient matrix data, and information entropy is calculated using an appropriate method or algorithm to generate information entropy evaluation data. Based on the information entropy evaluation data, data analysis and processing are performed, representative data are selected, and key feature data are generated. These data are important or representative in the spectrum gradient matrix. The preset audio reconstruction model is optimized and adjusted using the key feature data to ensure that the model reconstructs the spatial audio more accurately. This may include parameter adjustment, algorithm improvement, or model structure optimization.

[0100] Step S4: Acquire the dynamic change parameters of the site; perform influence assessment according to the dynamic change parameters of the site to generate factor influence degree data; perform dynamic change simulation according to the factor influence degree data to generate dynamic simulation result data; perform reflection network construction according to the dynamic simulation result data to generate dynamic reflection network data;

[0101] In an embodiment of the present invention, the relevant parameters of the dynamic changes of the venue are obtained through sensors, equipment or other data collection methods. Based on these dynamic change parameters, the impact degree is evaluated. It may be necessary to use a specific model or algorithm to evaluate the impact degree of these factors on sound propagation and generate factor impact degree data. The obtained factor impact degree data is used to simulate the dynamic changes of the venue through simulation tools or models to generate dynamic simulation result data. Based on the dynamic simulation result data, a reflection network is established to describe the reflection and propagation of sound inside the venue, and dynamic reflection network data is generated.

[0102] Step S5: setting an optimization target based on the dynamic reflection network data to generate optimization target information data; iteratively optimizing the spatial audio reconstruction model using the optimization target information data to generate a dynamic audio reconstruction model;

[0103] In an embodiment of the present invention, the optimization target of sound propagation is determined by using dynamic reflection network data. This may include improving certain specific aspects of sound propagation, such as reducing echoes, optimizing sound coverage, or improving sound quality. According to the set optimization target, optimization target information data is generated, which may include specific parameters or configurations required for sound propagation. The existing spatial audio reconstruction model is iteratively optimized using the optimization target information data. This may involve adjusting the parameters, structure, or other relevant acoustic properties of the model to meet the set optimization target.

[0104] Step S6: Use sensor equipment to collect feedback data and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use a dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

[0105] In an embodiment of the present invention, real-time feedback data is collected by using sensor devices, including real-time audio feedback data and real-time environmental impact data. For example, sound intensity, frequency distribution, ambient temperature, humidity, etc. A previously established dynamic audio reconstruction model is used. The real-time collected audio feedback data and environmental impact data are input into the dynamic audio reconstruction model. Feedback control is performed based on these data, which may include dynamically adjusting the parameters or configuration of sound propagation to optimize the sound propagation effect. In combination with the real-time data and the feedback control of the model, sound propagation optimization data is generated, that is, audio signal data adjusted and optimized according to the actual feedback data.

[0106] Preferably, step S1 comprises the following steps:

[0107] Step S11: Acquire site plan data; digitize the site plan data to generate site grid data;

[0108] Step S12: Presetting the location of the sound amplification equipment for the site grid data based on the acoustic propagation principle to generate equipment orientation preset data;

[0109] Step S13: marking the site grid data with equipment information according to the equipment position preset data to generate the position preset grid data;

[0110] Step S14: formulating facility material parameters based on the site plan data to generate material parameter data;

[0111] Step S15: performing acoustic simulation on the azimuth preset grid data through the material parameter data to generate propagation simulation result data;

[0112] Step S16: selecting the best result according to the propagation simulation result data to generate the best simulation result data;

[0113] Step S17: performing multi-dimensional acoustic characteristic analysis according to the optimal simulation result data to generate propagation acoustic characteristic data;

[0114] Step S18: Create a topology map based on the propagation acoustic characteristic data to generate acoustic topology map data.

[0115] The present invention provides an accurate description of the site through digital processing of the site plan and generation of raster data, which helps to understand the layout and structure of the site. The device position is preset and the device information is marked, and the reasonable position and orientation of the sound amplification device are determined through the acoustic propagation principle and the site raster data. It helps to ensure the rationality and uniformity of the sound coverage and improve the effect of sound propagation. The formulation of facility material parameters is crucial to the influence of sound propagation. By formulating the site material parameters, the reflection, absorption and propagation of sound on the surfaces of different materials can be more accurately simulated, which helps to optimize the design and effect of the sound propagation system. The formulation of facility material parameters provides important data on the acoustic properties of the site materials such as sound absorption, reflection, and conduction. It has an important impact on the propagation mode of sound in a specific environment. Acoustic simulation can simulate the propagation of sound in the site based on the facility material parameters and the preset raster data of the orientation. Through the simulation result data, the propagation of sound in different areas can be understood. By optimally selecting the propagation simulation results, the best sound propagation plan is determined. The optimal result data can help the system find the most suitable sound propagation strategy for a specific scene and improve the sound propagation effect. Multidimensional acoustic characteristics analysis is an in-depth study of propagation results, providing more detailed information about sound propagation characteristics, such as sound pressure level, sound wave propagation direction, etc., which helps to understand the details of sound propagation more comprehensively. The creation of acoustic topology data visualizes the propagation path and characteristics of sound in the venue. It helps to intuitively understand the patterns and laws of sound propagation.

[0116] In an embodiment of the present invention, a site plan is obtained from a source such as a surveying and mapping agency, satellite imagery, or a professional map service. An image or paper map is digitized, for example, by converting it into an electronic format using a scanner or professional software. The image is cleaned and repaired to ensure its accuracy and consistency. The image or other format is converted into a common geographic information system (GIS) format, such as Shapefile, GeoJSON, or KML. Using acoustic simulation software or algorithms, combined with terrain, material, and sound propagation characteristics, the optimal device position is preset according to the acoustic principle and the geographic information of the site grid data. The preset device position is determined based on the simulation results, taking into account acoustic factors such as propagation path, reflection, and absorption. The selected device position is marked in the site grid data, which may be represented by a number or symbol to indicate the device position and orientation. The device orientation, type, and other relevant information are combined into the site grid data. Ensure that the marked device position is accurate and corresponds to the acoustic prediction. Investigate and analyze materials such as buildings, ground, obstacles, etc. on the site. Determine corresponding acoustic parameters such as reflectivity, absorptivity, etc. based on the acoustic properties, sound absorption, and other properties of the material. Arrange the collected material parameters into a data format suitable for processing. Use professional acoustic simulation software to import site grid data and determined material parameter data. Set the simulated sound source, receiver location and acoustic characteristics, taking into account the sound parameters and material characteristics of the equipment. Run the acoustic simulation model to simulate the propagation process of sound in the site. Collect the simulation result data, including sound pressure level, sound field map, sound propagation path, etc. Perform a detailed analysis of the acoustic simulation result data, including sound propagation path, sound field distribution, etc. Select the best sound propagation result based on the analysis results, considering factors such as sound pressure level and sound coverage. Select the best simulation result and record the corresponding position, sound pressure level or other key parameters. Extract multi-dimensional acoustic characteristic data of the best simulation result, which may include sound pressure level, spectrum distribution, sound field map, etc. Analyze the extracted acoustic characteristic data to explore the various characteristics and changes of sound propagation in the site. Integrate and explain the relationship between multiple acoustic characteristics, including sound directionality, propagation path, sound intensity, etc. Based on the acoustic characteristic data, use graphics software or specific tools to draw the acoustic topology of the site. Mark the sound propagation path, sound pressure level distribution and important acoustic characteristics in the topology map. Create acoustic topology data to ensure that the map represents the sound propagation and acoustic characteristics of the venue.

[0117] Preferably, step S2 comprises the following steps:

[0118] Step S21: performing a propagation key area analysis based on the acoustic topology data to generate propagation key area data; performing an actual layout analysis based on the propagation key area data to generate sensor location data;

[0119] Step S22: Calculate the frequency weight of the sensor position data using the sampling frequency evaluation formula to generate sampling weight frequency data;

[0120] Step S23: Correcting the position and sampling frequency of the sensor position data according to the sampling weight frequency data to generate position information correction data; arranging sensors based on the position information correction data, and collecting sound signals to generate multi-directional audio data;

[0121] Step S24: performing spectrum time domain analysis on the multi-directional audio data to generate collected acoustic feature data;

[0122] Step S25: using the propagation acoustic characteristic data to perform characteristic error evaluation on the collected acoustic characteristic data to generate actual sampling error data; performing error cause analysis based on the actual sampling error data to generate error adjustment data;

[0123] Step S26: Optimizing the sensor arrangement according to the error adjustment data, and performing re-collection analysis and evaluation, thereby generating optimized collection feature data;

[0124] Step S27: quantizing the optimized acquisition feature data based on the acoustic topology data to generate spectrum gradient matrix data.

[0125] The present invention helps to determine the key positions and areas in sound propagation through the analysis of key area data, so as to determine which positions need special attention and optimize sound propagation. Frequency weight calculation is to determine the importance of sensor position. By evaluating the sampling frequency of the sensor position, the sensor layout can be optimized to ensure that more comprehensive and representative sound data are collected in the key area. Position information correction and sensor layout are crucial for collecting multi-directional audio data. By correcting the position information and rearranging the sensors, the comprehensiveness and accuracy of data collection can be ensured. Spectral time domain analysis of multi-directional audio data can extract and quantify the characteristics of sound, which is very important for understanding the propagation characteristics and acoustic characteristics of sound in different areas, and is helpful for subsequent sound processing and optimization. Feature error evaluation can identify and quantify the errors of acoustic feature data, which helps to understand the accuracy and reliability of data during the collection process. Analyzing the actual sampling error helps to find the cause of the error. It can reveal possible problems or defects in the collection process. Error adjustment data is used to correct the optimal layout of the sensor. By adjusting the error, the sensor can be rearranged to improve the accuracy and reliability of data collection to obtain higher quality collection feature data. The generation of spectral gradient matrix data is based on the quantitative processing of optimized collection feature data. These data are abstract representations of sound characteristics and can more fully describe the characteristics of sound propagation.

[0126] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0127] Step S21: performing a propagation key area analysis based on the acoustic topology data to generate propagation key area data; performing an actual layout analysis based on the propagation key area data to generate sensor location data;

[0128] In the embodiment of the present invention, by analyzing the optimal simulation result data, attention is paid to the propagation path, sound pressure level change, diffraction, reflection and other phenomena of sound in the venue. Find key nodes or areas in the propagation path, such as bottlenecks of sound propagation or areas that have a significant impact on sound propagation. Or places where sound waves focus or areas that hinder sound wave propagation. Determine which areas are critical for sound propagation. Consider obstacles, materials, structures or propagation characteristics on the sound propagation path, and identify areas that have a significant impact on sound propagation. Mark the identified key areas on the acoustic topology map to highlight their importance. Reports, charts or graphical displays can be generated to clearly show the role of these key areas in sound propagation. Verify and further analyze the identified key areas to ensure their accuracy and practicality. If necessary, the acoustic topology map can be adjusted to better reflect the real scene or modify the simulation parameters to improve accuracy, and generate propagation key area data. According to the analysis results, mark or determine the location, characteristics or attributes of these key areas on the acoustic topology map. Record the data of the key propagation area, such as location, sound pressure level, propagation path, etc., to generate sensor position data.

[0129] Step S22: Calculate the frequency weight of the sensor position data using the sampling frequency evaluation formula to generate sampling weight frequency data;

[0130] In an embodiment of the present invention, by preparing sensor position data and related frequency sensitivity data, a preset sampling frequency evaluation formula is used, which fully considers parameters such as the sensitivity of the sensor, the maximum distance between sensors, and the maximum sensitivity to accurately calculate the frequency weights between the sensors. The calculated frequency weights are recorded in the form of data for subsequent position correction and sensor layout. Or the frequency weights are evaluated by field testing and simulation, simulation and other methods to obtain sampling weight frequency data.

[0131] Step S23: Correcting the position and sampling frequency of the sensor position data according to the sampling weight frequency data to generate position information correction data; arranging sensors based on the position information correction data, and collecting sound signals to generate multi-directional audio data;

[0132] In the embodiment of the present invention, a series of iterations are performed by using the calculated frequency weights to continuously adjust the sensor positions and recalculate the frequency weights until the required sampling quality and acoustic requirements are achieved. Position information correction data is generated, including a fine-tuning correction value for the sensor position. Sensors are rearranged according to the corrected position data, and sound signals are collected through these sensors.

[0133] Step S24: performing spectrum time domain analysis on the multi-directional audio data to generate collected acoustic feature data;

[0134] In an embodiment of the present invention, the multi-directional audio data acquired by the sensors arranged in the above steps are converted into spectrum data by using algorithms such as Fourier transform to obtain the distribution and energy distribution of sound at different frequencies. Features in terms of frequency, amplitude, and sound duration are extracted. Common feature extraction methods include short-time Fourier transform, power spectral density analysis, zero crossing rate, volume, spectrum centroid, spectrum slope, etc. The features of the sound spectrum are extracted, such as frequency range, frequency intensity distribution, dominant frequency, etc. Including time domain features such as the duration, start time, and end time of the audio signal. Identify the intensity and distribution characteristics of sound within different frequency ranges. The extracted features are screened and selected to remove redundant or unimportant features. Machine learning models (such as support vector machines, neural networks, etc.) or statistical models can be used to classify, cluster, or identify acoustic features to generate collected acoustic feature data.

[0135] Step S25: using the propagation acoustic characteristic data to perform characteristic error evaluation on the collected acoustic characteristic data to generate actual sampling error data; performing error cause analysis based on the actual sampling error data to generate error adjustment data;

[0136] In an embodiment of the present invention, acoustic features such as spectrum, amplitude, frequency distribution, etc. in the propagation acoustic feature data and the collection acoustic feature data are compared. Differences and errors are calculated, and indicators such as mean square error (MSE), absolute error, and correlation coefficient are used for quantification. For each acoustic feature, the difference between the propagation data and the collection data is calculated to generate actual sampling error data. Check the impact of changes in environmental conditions, such as noise level, temperature, humidity, etc. on acoustic features. Consider factors such as recording equipment, sensor performance, or calibration, which may affect data accuracy. If there is a signal processing process, evaluate its impact on acoustic features, such as filtering, gain, and other operations. Display the error data in a visual manner in the form of a bar graph, a scatter plot, an error curve, etc., highlighting the differences and trends. Record error statistics, such as average error, maximum / minimum error, standard deviation, etc., for quantitative description of the error level. Draw a conclusion based on the evaluation results, explain the degree of difference and the source of error between the propagation acoustic feature data and the collection acoustic feature data, and generate error adjustment data.

[0137] Step S26: Optimizing the sensor arrangement according to the error adjustment data, and performing re-collection analysis and evaluation, thereby generating optimized collection feature data;

[0138] In the embodiment of the present invention, the position or parameters of the sensor are adjusted according to the error adjustment data to minimize the acquisition error. After the optimized arrangement, data is collected again and compared with the previous data to evaluate whether the error reduction has been achieved. Based on the results of the re-collection analysis, data after the optimized arrangement is generated, that is, optimized acquisition feature data, which should show better accuracy and consistency.

[0139] Step S27: quantizing the optimized acquisition feature data based on the acoustic topology data to generate spectrum gradient matrix data.

[0140] In an embodiment of the present invention, the time domain data is converted into frequency domain data by performing Fourier transform or spectrum analysis on the optimized collected feature data. Specific acoustic features, such as frequency range, amplitude, sound intensity, etc., are extracted from the spectrum data. The extracted acoustic feature data is organized into a matrix form. The feature data is spatially analyzed using acoustic topology data, and the acoustic feature data is associated with position information, taking into account propagation paths, sound wave diffusion, etc. The acoustic feature data after spatial analysis is quantized and represented in a matrix form, where the rows or columns of the matrix represent different spatial positions or sound sources, and the values ​​in the matrix represent the acoustic feature data. The spatial distribution and changes of the acoustic features in the matrix are described, and necessary normalization is performed to eliminate unit or scale differences to generate spectrum gradient matrix data.

[0141] Preferably, the sampling frequency evaluation formula in step S22 is as follows:

[0142] ;

[0143] In the formula, is the sampling frequency weight factor, For sensor The sensitivity of For sensor The sensitivity of For sensor and The maximum distance between is the maximum sensitivity of all sensors. is the maximum distance between all sensors, is the average distance between all sensors, For sensor The horizontal axis, For sensor The horizontal axis, For sensor and The Euclidean distance between .

[0144] The present invention provides a sampling frequency evaluation formula for calculating the frequency weight of sensor position data and generating sampling weight frequency data. Sensitivity ,sensor Sensitivity ,sensor and The maximum distance between , the highest sensitivity among all sensors , the maximum distance between all sensors , the average distance between all sensors ,sensor The horizontal axis ,sensor The horizontal axis ,sensor and The Euclidean distance between As well as the interaction effects between variables, the following functional relationship is formed:

[0145] ;

[0146] pass and The product of the sensor and The higher the sensitivity, the stronger the ability to capture sound signals, so the weight factor will increase. The maximum sensitivity of all sensors Used to normalize the relative sensitivity. and The maximum distance between The larger the value, the corresponding sampling frequency will also increase, which means that when the distance between sensors is large, the sensors and The maximum distance between all sensors is The larger the value, the corresponding sampling frequency decreases, and the sensor with the largest distance difference in the sensor arrangement contributes less to the sampling frequency. The smaller it is, the corresponding sampling frequency increases, which means that the average distance in the sensor arrangement is smaller, which is conducive to capturing the sound signal more comprehensively. and The Euclidean distance between , indicating the sensor and The distance in three-dimensional space. The farther the sensor is, and the more different the direction of the sensor is, the smaller its weight is, reflecting the weaker the association between them. This part represents the sensor and This part measures the impact of the angle between sensors on the calculation. A smaller angle means that the directions between sensors are more similar, which may introduce unnecessary redundant information. A larger angle means that the directionality between sensors is weaker, which may lead to inaccurate data. By considering the impact of the angle, the spatial relationship between sensors can be better measured, thereby improving the reliability and accuracy of the data. The purpose of this part is to take into account the ratio of the distance between sensors to the maximum distance. A larger distance ratio means that the distribution between sensors is sparse, which may result in missing data. A smaller distance ratio means that the distribution between sensors is denser, which can provide more comprehensive and accurate data. By considering the impact of the distance ratio, the distance between sensors can be better balanced, thereby improving the comprehensiveness and accuracy of the data. This functional relationship can reflect the importance or priority of sensors at different locations in sound sampling. A higher weight indicates that the sensor For adjacent sensors It is even more important in sound sampling.

[0147] Preferably, step S3 comprises the following steps:

[0148] Step S31: feature labeling the spectrum gradient matrix data to generate labeling matrix data; perform normalization processing based on the labeling matrix data to generate normalized matrix data;

[0149] Step S32: performing data association analysis on the normalized matrix data to generate association analysis result data; performing information entropy evaluation based on the association analysis result data to generate information entropy evaluation data;

[0150] Step S33: selecting high entropy data from the normalized matrix data according to the information entropy evaluation data to generate key feature data; performing coding mapping conversion on the key feature data to generate feature coding data;

[0151] Step S34: optimizing the preset audio reconstruction model based on the feature coding data to generate a spatial audio reconstruction model.

[0152] The present invention helps to extract and identify key features in the data by feature labeling the spectrum gradient matrix data. It helps to understand the meaning and importance of the data. Normalization processing can standardize the data so that different data are comparable. This processing method helps to eliminate the dimension and size differences between different data and provide a consistent benchmark for data analysis. Data association analysis can help identify the relationship and correlation between data. It is very important to understand the connection between data and how they affect each other, and provide support for further analysis and interpretation. High entropy data selection is based on information entropy evaluation data, which helps to determine the most important and representative features in the data. It helps to simplify the data set, reduce redundant information, and make the data more compact and effective. Through encoding mapping conversion, the selected key feature data can be converted and encoded. This conversion may include recombination or encoding of features, so that the data is more suitable for subsequent model processing and interpretation. Model tuning is based on feature encoding data, which helps to optimize the preset audio reconstruction model. By adjusting and optimizing the model, the performance and accuracy of the model can be improved, so that it can better adapt to the characteristics of the data corresponding to the sound field, thereby generating a more accurate and effective spatial audio reconstruction model.

[0153] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0154] Step S31: feature labeling the spectrum gradient matrix data to generate labeling matrix data; perform normalization processing based on the labeling matrix data to generate normalized matrix data;

[0155] In an embodiment of the present invention, the labeling standard of each feature in the spectrum gradient matrix is ​​defined. For example, features such as frequency range and sound intensity are labeled. The features in the spectrum gradient matrix are labeled and marked as corresponding categories or feature labels. The value range of the spectrum gradient matrix data is analyzed to determine the maximum and minimum values. Use a suitable normalization method (such as linear normalization, Z-score standardization, etc.) to scale the spectrum gradient matrix data to a specific numerical range to ensure that the data is processed on the same scale.

[0156] Step S32: performing data association analysis on the normalized matrix data to generate association analysis result data; performing information entropy evaluation based on the association analysis result data to generate information entropy evaluation data;

[0157] In an embodiment of the present invention, the correlation or correlation law between the features is checked by calculating the degree of correlation between different features using a statistical method or a correlation coefficient. The correlation analysis result data is generated based on the correlation calculation, and the relationship between the features is recorded, which may include a correlation coefficient matrix or a correlation network diagram. Based on the correlation analysis result data, the information entropy between the features is calculated. Related methods of information theory, such as Shannon entropy, etc., can be used. The information entropy calculation results between the features are recorded as information entropy evaluation data, which is used to represent the uncertainty or correlation between the features.

[0158] Step S33: selecting high entropy data from the normalized matrix data according to the information entropy evaluation data to generate key feature data; performing coding mapping conversion on the key feature data to generate feature coding data;

[0159] In an embodiment of the present invention, the important features in the normalized matrix data are determined by using information entropy evaluation data. These include features with higher information entropy or correlation. Feature screening is performed based on the information entropy evaluation results, and the features that have the greatest impact on the audio reconstruction model are selected. Appropriate encoding methods are designed, such as unique hot encoding, label encoding, etc., to encode and convert the selected key features. The key feature data is converted according to the designed encoding method to generate feature encoding data to ensure that the data format is suitable for model input.

[0160] Step S34: optimizing the preset audio reconstruction model based on the feature coding data to generate a spatial audio reconstruction model.

[0161] In the embodiment of the present invention, by determining the specific type of audio reconstruction, it is to enhance a specific frequency band, remove noise, improve sound clarity, etc. The expected effect and quality standard of audio reconstruction are clarified. According to the task requirements, select an appropriate model type, such as deep neural network (CNN, RNN, Transformer), autoencoder, GAN, etc. Determine the number of layers of the model, the number of neurons in each layer, and the connection method between layers. According to the characteristics of the audio data, select an appropriate feature extraction method, such as time domain features, frequency domain features, Mel spectrum features, etc. Ensure that the data format meets the model input requirements and perform preprocessing such as normalization and encoding. Define hyperparameters such as learning rate, number of iterations, and regularization parameters. Use an appropriate initialization method to initialize model parameters, such as random initialization, pre-trained weights, etc. Select an appropriate loss function to measure the difference between the reconstructed model output and the expected output, such as MSE, MAE, cross entropy, etc. Select an appropriate optimization algorithm, such as Adam, SGD, etc., to minimize the loss function. Use training data to train the model, monitor the loss function and adjust the model parameters. Use a validation set or cross-validation to evaluate model performance, avoid overfitting, and adjust the model structure and parameters. Deploy the trained model to the actual environment, evaluate the model performance in the real environment, adjust and optimize the model to meet the actual needs, and thus generate a preset audio reconstruction model. Select appropriate optimization algorithms, such as gradient descent, genetic algorithm, etc., to tune the audio reconstruction model. Adjust and optimize the model parameters based on feature encoding data to ensure that the model can better fit the data. Use cross-validation and other techniques to evaluate model performance to prevent the model from overfitting or underfitting.

[0162] Preferably, step S4 comprises the following steps:

[0163] Step S41: Acquire the dynamic change parameters of the venue;

[0164] Step S42: Perform influence assessment based on the site dynamic change parameters to generate factor influence degree data;

[0165] Step S43: formulating environmental simulation parameters according to the factor influence degree data to generate multi-environment parameter data;

[0166] Step S44: using the multi-environment parameter data and based on the azimuth preset grid data to perform dynamic change simulation and generate dynamic simulation result data;

[0167] Step S45: constructing a reflection network according to the dynamic simulation result data to generate dynamic reflection network data.

[0168] The present invention obtains the dynamic change parameters of the site to collect data on the changes of the site environment over time. These data may include parameters such as temperature, humidity, and flow of people, which are helpful to understand the real-time state of the environment. The generation of factor influence degree data involves the evaluation and analysis of the dynamic change parameters of the site, which can determine the degree of influence of different factors on sound propagation. It helps to quantify the impact of different factors on sound propagation and provide an important basis for subsequent simulations. The formulation of environmental simulation parameters is based on the factor influence degree data, which can convert the influence of factors into simulatable parameters. These parameters can be used to simulate different states of the site environment and simulate sound propagation more accurately. The generation of dynamic simulation result data is based on the simulation of multi-environment parameter data. These data provide simulation results of sound propagation under different environmental conditions, which helps to understand the behavior and characteristics of sound in different environments.

[0169] As an example of the present invention, refer to Figure 4 As shown, in this example, step S4 includes:

[0170] Step S41: Acquire the dynamic change parameters of the venue;

[0171] In the embodiment of the present invention, by collecting the dynamically changing parameters of the venue, such as temperature, humidity, wind speed, material property changes, number of people, seating arrangement, etc., a sensor network or device is used to monitor relevant parameters in the venue and ensure the accuracy and real-time nature of the data.

[0172] Step S42: Perform influence assessment based on the site dynamic change parameters to generate factor influence degree data;

[0173] In an embodiment of the present invention, an evaluation model is created based on an acoustic propagation model to quantitatively measure the impact of different parameters on sound propagation. This model may include weights of various parameters to reflect the degree of their impact on sound propagation. Using the established evaluation model, analyze the impact of each dynamically changing parameter on sound propagation one by one. Compare how changes in each parameter affect the propagation path, reflection, refraction, absorption, etc. of the sound. Formulate quantitative evaluation criteria, which may include defining the degree range or numerical range of the impact of different parameter changes on sound propagation. For example, use a scale (1-10 levels) or a percentage to indicate the degree of impact. Based on the evaluation results, prioritize the parameters with greater impact and generate factor impact data.

[0174] Step S43: formulating environmental simulation parameters according to the factor influence degree data to generate multi-environment parameter data;

[0175] In the embodiment of the present invention, the importance of each factor to sound propagation is understood by analyzing the data of the degree of influence of the factors. These factors may include acoustic environment parameters (such as temperature, humidity, air pressure), material parameters (such as wall material, floor material), etc. According to the degree of influence data and the expected acoustic effect, the parameters to be simulated are selected. For example, the numerical range or change trend of the acoustic environment parameters and material parameters to be adjusted is determined according to the degree of influence of the factors.

[0176] Step S44: using the multi-environment parameter data and based on the azimuth preset grid data to perform dynamic change simulation and generate dynamic simulation result data;

[0177] In an embodiment of the present invention, multiple environmental parameter data is integrated with azimuth grid data. Using acoustic simulation software or model, the preset azimuth grid data is dynamically simulated based on multiple environmental parameters. Environmental changes, material properties and sound propagation modes are considered. According to the environmental parameter data, the parameters of the simulation model are set, such as the location of the sound source, the material properties of the site, the acoustic properties, etc. The acoustic simulation model is run to simulate the dynamic changes of the sound on the azimuth grid under different environmental conditions. Model verification is performed to check whether the configured parameters can accurately simulate the expected acoustic properties. This may require simulating the model under different environmental conditions and comparing them with the actual observation results. If the model verification finds that there are differences with the actual situation, the parameters need to be adjusted. The parameter values ​​are gradually adjusted to observe their effects on sound propagation until the simulation results are consistent with the actual situation, and dynamic simulation result data is generated.

[0178] Step S45: constructing a reflection network according to the dynamic simulation result data to generate dynamic reflection network data.

[0179] In the embodiment of the present invention, the dynamic simulation result data is analyzed, including the sound propagation path, reflection, refraction and absorption. According to the analysis results, a reflection network model is established. The paths and characteristics of sound reflection and propagation in the venue are identified and marked, such as the reflectivity, propagation path, sound absorption characteristics of different surfaces, etc. The accuracy and effectiveness of the reflection network model are verified through simulation and observation results.

[0180] Preferably, step S5 comprises the following steps:

[0181] Step S51: Analyze the sound propagation characteristics based on the dynamic reflection network data to generate reflection characteristic data; set optimization targets based on the reflection characteristic data to generate optimization target information data;

[0182] Step S52: Calculate the sound pressure change at the receiving point of the multiple environmental parameter data using the sound pressure change evaluation formula to generate multiple sound pressure change data;

[0183] Step S53: using the multiple sound pressure change data to perform associated target correction on the optimization target information data to generate optimization target correction data;

[0184] Step S54: adjusting the model parameters of the spatial audio reconstruction model according to the optimization target correction data to generate a parameter adjustment model;

[0185] Step S55: Utilize the multi-environment parameter data to iteratively optimize and evaluate the parameter adjustment model to generate a dynamic audio reconstruction model.

[0186] The present invention can generate reflection characteristic data based on dynamic reflection network data through sound propagation characteristic analysis, so as to gain an in-depth understanding of the reflection of sound in different environments. These data are helpful in identifying the special characteristics of sound propagation and provide a basis for optimization. Using the sound pressure change evaluation formula to calculate the sound pressure change can generate multiple sound pressure change data, which reflects the change of sound pressure at the receiving point under different environmental conditions and is a key indicator for understanding the sound propagation effect. The associated target correction corrects the optimization target information data through multiple sound pressure change data, which helps to set and adjust the optimization target more accurately and ensure that the optimization target meets the actual situation. By adjusting the model parameters of the spatial audio reconstruction model through the optimization target correction data, the performance and accuracy of the model can be improved, so that it can better adapt to the reconstruction requirements of sound in different environments. The parameter adjustment model is iteratively optimized and evaluated using multi-environment parameter data to generate a dynamic audio reconstruction model. This model can better adapt to the sound propagation characteristics in different environments and provide more accurate model support for sound reconstruction.

[0187] In an embodiment of the present invention, a sound source is set in a simulated environment and the sound pressure level change at the receiving point is recorded. The path of sound wave propagation, reflection characteristics, and sound pressure level distribution in various regions are analyzed. Reflection characteristics are extracted from the data obtained from the acoustic simulation, including the reflection coefficient of sound waves on different surfaces, the change of sound pressure level, etc. These data can describe the propagation of sound in the environment. According to the collected reflection characteristic data, the optimization target of sound propagation is set. This may include reducing sound reflection, maximizing sound pressure level, reducing sound attenuation, etc. A preset sound pressure change evaluation formula is used, which fully considers the number of sound sources, the initial sound pressure level, the attenuation coefficient of the sound propagation medium, and other parameters to calculate the sound pressure change, and obtain the change of sound pressure received by different sensors under different environmental conditions. These data can represent the change of sound pressure level of sound in different environments. Multiple sound pressure change data are used for comparative analysis with the previously set optimization target data. The difference and correlation between the actual situation of sound propagation and the optimization target are determined. The optimization target is modified and the target parameters are adjusted to be more consistent with the actual sound propagation situation. It may be necessary to modify the target value, weight or other parameters to optimize the performance of the model. According to the optimization target correction data, the parameters of the spatial audio reconstruction model are calibrated and adjusted to make the model more accurately match the actual sound propagation situation. This may include model parameter optimization in terms of sound enhancement, noise reduction, and sound field adjustment. The parameter adjustment model is iteratively evaluated through multi-environment parameter data, and the parameter adjustment model is optimized to try to make it more applicable in different environments to generate a dynamic audio reconstruction model that is more adaptable to multiple environments. This may include evaluating the effect of the model under different environmental conditions and comparing the accuracy and performance of the model in sound reconstruction.

[0188] Preferably, the sound pressure change evaluation formula in step S52 is as follows:

[0189] ;

[0190] In the formula, is the sound pressure level at the receiving point, is the number of sound sources, is the index value of the sound source, For sound source The initial sound pressure level, is the attenuation coefficient of the sound propagation medium, For sound source The distance to the receiving point, For sound source The average distance is the sound pressure attenuation coefficient, For sound source The length of the sound propagation path to the receiving point, For sound source The average path length, is the environmental factor constant, For sound source The surface area of is the effective cross-sectional area of ​​the sound propagation path, is the deviation correction value of the sound pressure level at the receiving point.

[0191] The present invention provides a sound pressure change evaluation formula, which is used to calculate the sound pressure change of the receiving point for multiple environmental parameter data and generate multiple sound pressure change data. The formula fully considers the number of sound sources. , the index value of the sound source , sound source Initial sound pressure level , the attenuation coefficient of the sound propagation medium , sound source Distance to receiving point , sound source The average distance , sound pressure attenuation coefficient , sound source The length of the sound propagation path to the receiving point , sound source The average path length , environmental factor constant , sound source Surface area , the effective cross-sectional area of ​​the sound propagation path , the deviation correction value of the sound pressure level at the receiving point As well as the interaction effects between variables, the following functional relationship is formed:

[0192] ;

[0193] The number of sound sources Indicates how many sound sources are present. Initial sound pressure level The starting sound pressure level used to calculate the sound source is the initial point of the sound pressure level change. The logarithm of the ratio of the actual distance to the average distance is calculated to represent the attenuation of sound due to distance. The constant n represents the attenuation coefficient of sound in the propagation medium, which represents the various attenuations encountered by sound during propagation, such as air absorption and the density of the propagation medium. The magnitude of the attenuation coefficient n determines the attenuation rate of sound during propagation. This part of the calculation represents the effect of distance attenuation, which serves to amplify the attenuation to more prominently reflect the effect of sound attenuation as the distance increases. The logarithm of the ratio of the actual path length to the average path length is calculated for the length of the sound propagation path. The effect of the propagation path length is expressed by the sound pressure attenuation coefficient The effect of the path is significantly amplified to more strongly reflect the effect of the sound propagation path, the sound pressure attenuation coefficient in a free field environment The theoretical value is 20, because the sound pressure level drops by 20 decibels for each distance multiple. In actual situations, it may vary depending on the environment, so it is represented by an unknown number in the formula to adapt to the calculation of sound pressure changes in different environmental conditions. Environmental factors and the ratio of the surface area of ​​the sound source to the effective cross-sectional area of ​​the sound propagation path are taken into account. This part represents the impact of the environment on sound propagation. For example, if the environment is complex or the surface area of ​​the sound source is small, it will affect the propagation and attenuation of the sound. This functional relationship can simulate and predict the attenuation and propagation of sound under different environmental conditions, providing an important basis for optimizing sound propagation. And use the deviation correction value of the sound pressure level at the receiving point Adjust and correct the functional relationship to reduce the error caused by this data or error term, so as to more accurately generate the sound pressure level of the receiving point , which improves the accuracy and reliability of the sound pressure receiving value evaluation of the sensor equipment. At the same time, the attenuation coefficient of the sound propagation medium and the sound pressure attenuation coefficient in the formula can be adjusted according to the actual situation, and applied to different multi-environment parameter data to calculate the sound pressure change of the receiving point, which improves the flexibility and applicability of the algorithm.

[0194] Preferably, step S6 comprises the following steps:

[0195] Step S61: using sensor equipment to collect real-time audio signals and generate real-time audio feedback data;

[0196] Step S62: using sensor equipment to collect environmental change data and generate environmental change data; filtering the environmental change data according to the factor impact degree data and generating real-time environmental impact data;

[0197] Step S63: formulating reconstruction parameters based on the real-time audio feature data and the real-time environmental impact data to generate device reconstruction parameter data;

[0198] Step S64: Utilize the dynamic audio reconstruction model, perform feedback control based on the device reconstruction parameter data, and transmit the audio signal to generate sound propagation optimization data.

[0199] The present invention can provide instant audio feedback data through real-time audio signal collection, which helps to understand the performance of sound in the current environment in real time. The collection of environmental change data and the screening of factor influence data can generate real-time environmental impact data. These data record the real-time impact of environmental changes on sound propagation, helping to better understand the degree of impact of environmental changes on sound. Equipment reconstruction parameter data is formulated based on real-time audio feature data and real-time environmental impact data. These parameters are important factors for optimizing sound propagation according to real-time conditions, and help to improve the adaptability and effect of sound propagation. A dynamic audio reconstruction model is used and feedback control is performed based on the equipment reconstruction parameter data, and audio signals are transmitted to generate sound propagation optimization data. These data show the sound propagation after dynamic adjustment and feedback control, which helps to evaluate the propagation optimization effect and the effect of real-time adjustment of sound propagation.

[0200] In an embodiment of the present invention, real-time audio signal acquisition and environmental change data acquisition are performed by the sensors arranged in step S23 to generate real-time audio feedback data and environmental change data. Real-time audio feedback data needs to be standardized, such as noise reduction, filtering, etc., to ensure that the collected audio data accurately reflects the actual sound. Environmental change data may include temperature, humidity, air pressure, etc., and the collected environmental change data is screened and filtered out in combination with the factor influence degree data. Determine the environmental parameters that have an important impact on sound propagation, extract these parameters, and generate real-time environmental impact data. Use real-time audio feature data (spectrum, amplitude, frequency, etc.) and real-time environmental impact data (temperature, humidity, etc.) as input to formulate the reconstruction parameters of the device. It can be formulated by establishing a rule set based on domain knowledge. For example, rising temperature will increase noise, and the clarity of high-frequency sound may be affected by humidity. Design rules to map audio features and environmental data to optimal reconstruction parameters, which can be a series of if-then rules or other logical expressions. Test the validity and applicability of the rule set and verify it based on actual data. The rules can be adjusted and optimized based on the verification results. Using experimental data or verification in a simulated environment, adjust and verify the developed equipment reconstruction parameters to ensure their robustness and effectiveness under different environmental conditions. Use the previously generated dynamic audio reconstruction model, combined with the equipment reconstruction parameter data, with real-time audio and environmental impact data as input. Apply the dynamic audio reconstruction model to process the audio signal, which may involve removing environmental noise, sound optimization, etc. Perform real-time feedback control based on the equipment reconstruction parameters and the dynamic audio reconstruction model. Generate sound propagation optimization data based on the data generated during the control process, such as sound propagation path, signal quality, etc., to evaluate and optimize the effect of sound propagation.

[0201] The present invention also provides an audio signal processing system, which is used to execute the audio signal processing method as described above, comprising:

[0202] The acoustic field topology construction module is used to obtain the site plan data; digitize the site plan data to generate site grid data; perform acoustic simulation based on the site grid data to generate propagation simulation result data; create a topology map based on the propagation simulation result data to generate acoustic topology map data.

[0203] The sound field feature analysis module is used to arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data;

[0204] The data coding optimization module is used to perform information entropy evaluation on the spectrum gradient matrix data to generate information entropy evaluation data; select high entropy data based on the information entropy evaluation data to generate key feature data; and optimize the preset audio reconstruction model based on the key feature data to generate a spatial audio reconstruction model.

[0205] The environmental change simulation module is used to obtain the dynamic change parameters of the site; conduct influence assessment based on the dynamic change parameters of the site to generate factor influence degree data; conduct dynamic change simulation based on the factor influence degree data to generate dynamic simulation result data; construct a reflection network based on the dynamic simulation result data to generate dynamic reflection network data;

[0206] The model reconstruction enhancement module is used to set optimization targets based on dynamic reflection network data and generate optimization target information data; the optimization target information data is used to iteratively optimize the spatial audio reconstruction model to generate a dynamic audio reconstruction model;

[0207] The sound dynamic adjustment module is used to collect feedback data using sensor equipment and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use the dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

[0208] The beneficial effect of the present application is that the present invention generates site grid data by acquiring and digitally processing site plan data. The propagation simulation result data is generated according to the grid data using acoustic simulation technology, and then the acoustic topology map data is created. It is helpful to have a deep understanding of the propagation mode and acoustic characteristics of sound in the site. The sensor is optimally arranged through multi-directional audio data, and the signal is re-collected to generate optimized acquisition feature data. It is helpful to collect more comprehensive and optimized sound data. The information entropy of the spectrum gradient matrix data is evaluated to generate information entropy evaluation data. It is helpful to understand the complexity and information volume of the data, determine the key features of the data, and then tune the audio reconstruction model. Finally, a spatial audio reconstruction model is generated to improve the accuracy and effectiveness of audio processing. By acquiring the dynamic change parameters of the site and evaluating the influence according to these parameters, factor influence degree data is generated. Dynamic change simulation is performed based on these data to generate dynamic simulation result data. The reflection network is constructed through these data to generate dynamic reflection network data, which helps to more accurately simulate and understand the propagation and reflection of sound in different environments. Optimization targets are set based on dynamic reflection network data to generate optimization target information data. It helps to clarify the direction and goal of optimization, and then use the optimization target data to iteratively optimize the spatial audio reconstruction model to generate a dynamic audio reconstruction model. The model can better adapt to dynamically changing scenes and improve the accuracy and real-time performance of sound reconstruction. Sensor equipment is used to collect feedback data to generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data. Real-time feedback data is used for feedback control to achieve real-time monitoring and adjustment of the sound propagation process, which helps to optimize the sound propagation effect and improve the quality and adaptability of the sound. Therefore, the audio signal processing method and system of the present invention obtains the spatial audio reconstruction model by analyzing the sound field characteristics, and performs change propagation simulation by considering environmental changes to obtain audio propagation characteristics corresponding to different environments, thereby comprehensively considering the site and environmental factors, and dynamically reconstructing the audio signal, thereby improving the transmission quality of the audio signal to improve the user's auditory experience.

[0209] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0210] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. 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 invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An audio signal processing method, characterized in that: The following steps are involved: Step S1: Acquire site plan data; Digitally process the site plan data to generate site raster data; Perform acoustic simulation based on site grid data to generate propagation simulation result data; Create a topology map based on the propagation simulation result data to generate acoustic topology map data; Step S2: Arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data; Step S3: perform information entropy evaluation on the spectrum gradient matrix data to generate information entropy evaluation data; select high entropy data based on the information entropy evaluation data to generate key feature data; perform model tuning on the preset audio reconstruction model based on the key feature data to generate a spatial audio reconstruction model; Step S3 includes the following steps: Step S31: feature labeling the spectrum gradient matrix data to generate labeling matrix data; perform normalization processing based on the labeling matrix data to generate normalized matrix data; Step S32: performing data association analysis on the normalized matrix data to generate association analysis result data; performing information entropy evaluation based on the association analysis result data to generate information entropy evaluation data; Step S33: selecting high entropy data from the normalized matrix data according to the information entropy evaluation data to generate key feature data; performing coding mapping conversion on the key feature data to generate feature coding data; Step S34: optimizing the preset audio reconstruction model based on the feature coding data to generate a spatial audio reconstruction model; Step S4: Acquire the dynamic change parameters of the site; perform influence assessment according to the dynamic change parameters of the site to generate factor influence degree data; perform dynamic change simulation according to the factor influence degree data to generate dynamic simulation result data; perform reflection network construction according to the dynamic simulation result data to generate dynamic reflection network data; Step S5: setting an optimization target based on the dynamic reflection network data to generate optimization target information data; iteratively optimizing the spatial audio reconstruction model using the optimization target information data to generate a dynamic audio reconstruction model; Step S5 includes the following steps: Step S51: Analyze the sound propagation characteristics based on the dynamic reflection network data to generate reflection characteristic data; set optimization targets based on the reflection characteristic data to generate optimization target information data; Step S52: Calculate the sound pressure change at the receiving point of the multiple environmental parameter data using the sound pressure change evaluation formula to generate multiple sound pressure change data; Step S53: using the multiple sound pressure change data to perform associated target correction on the optimization target information data to generate optimization target correction data; Step S54: adjusting the model parameters of the spatial audio reconstruction model according to the optimization target correction data to generate a parameter adjustment model; Step S55: Iteratively optimize and evaluate the parameter adjustment model using the multi-environment parameter data to generate a dynamic audio reconstruction model; the sound pressure change evaluation formula in step S52 is as follows: ； In the formula, is the sound pressure level at the receiving point, is the number of sound sources, is the index value of the sound source, For sound source The initial sound pressure level, is the attenuation coefficient of the sound propagation medium, For sound source The distance to the receiving point, For sound source The average distance is the sound pressure attenuation coefficient, For sound source The length of the sound propagation path to the receiving point, For sound source The average path length, is the environmental factor constant, For sound source The surface area of is the effective cross-sectional area of ​​the sound propagation path, is the deviation correction value of the sound pressure level at the receiving point; Step S6: Use sensor equipment to collect feedback data and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use a dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

2. The audio signal processing method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Acquire site plan data; digitize the site plan data to generate site grid data; Step S12: Presetting the location of the sound amplification equipment for the site grid data based on the acoustic propagation principle to generate equipment orientation preset data; Step S13: marking the site grid data with equipment information according to the equipment position preset data to generate the position preset grid data; Step S14: formulating facility material parameters based on the site plan data to generate material parameter data; Step S15: performing acoustic simulation on the azimuth preset grid data through the material parameter data to generate propagation simulation result data; Step S16: selecting the best result according to the propagation simulation result data to generate the best simulation result data; Step S17: performing multi-dimensional acoustic characteristic analysis according to the optimal simulation result data to generate propagation acoustic characteristic data; Step S18: Create a topology map based on the propagation acoustic characteristic data to generate acoustic topology map data.

3. The audio signal processing method according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing a propagation key area analysis based on the acoustic topology data to generate propagation key area data; performing an actual layout analysis based on the propagation key area data to generate sensor location data; Step S22: Calculate the frequency weight of the sensor position data using the sampling frequency evaluation formula to generate sampling weight frequency data; Step S23: Correcting the position and sampling frequency of the sensor position data according to the sampling weight frequency data to generate position information correction data; arranging sensors based on the position information correction data, and collecting sound signals to generate multi-directional audio data; Step S24: performing spectrum time domain analysis on the multi-directional audio data to generate collected acoustic feature data; Step S25: using the propagation acoustic characteristic data to perform characteristic error evaluation on the collected acoustic characteristic data to generate actual sampling error data; performing error cause analysis based on the actual sampling error data to generate error adjustment data; Step S26: Optimizing the sensor arrangement according to the error adjustment data, and performing re-collection analysis and evaluation, thereby generating optimized collection feature data; Step S27: quantizing the optimized acquisition feature data based on the acoustic topology data to generate spectrum gradient matrix data.

4. The audio signal processing method according to claim 3, characterized in that: The sampling frequency evaluation formula in step S22 is as follows: ； In the formula, is the sampling frequency weight factor, For sensor The sensitivity of For sensor The sensitivity of For sensor and The maximum distance between is the maximum sensitivity of all sensors. is the maximum distance between all sensors, is the average distance between all sensors, For sensor The horizontal axis, For sensor The horizontal axis, For sensor and The Euclidean distance between .

5. The audio signal processing method according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Acquire the dynamic change parameters of the venue; Step S42: Perform influence assessment based on the site dynamic change parameters to generate factor influence degree data; Step S43: formulating environmental simulation parameters according to the factor influence degree data to generate multi-environment parameter data; Step S44: using the multi-environment parameter data and based on the azimuth preset grid data to perform dynamic change simulation and generate dynamic simulation result data; Step S45: constructing a reflection network according to the dynamic simulation result data to generate dynamic reflection network data.

6. The audio signal processing method according to claim 1, characterized in that: Step S6 includes the following steps: Step S61: using sensor equipment to collect real-time audio signals and generate real-time audio feedback data; Step S62: using sensor equipment to collect environmental change data and generate environmental change data; filtering the environmental change data according to the factor impact degree data and generating real-time environmental impact data; Step S63: formulating reconstruction parameters based on the real-time audio feature data and the real-time environmental impact data to generate device reconstruction parameter data; Step S64: Utilize the dynamic audio reconstruction model, perform feedback control based on the device reconstruction parameter data, and transmit the audio signal to generate sound propagation optimization data.

7. An audio signal processing system, characterized in that: For executing the audio signal processing method according to claim 1, the audio signal processing system comprises: The acoustic field topology construction module is used to obtain the site plan data; digitally process the site plan data to generate site grid data; perform acoustic simulation based on the site grid data to generate propagation simulation result data; create a topology map based on the propagation simulation result data to generate acoustic topology map data; The sound field feature analysis module is used to arrange sensors according to the acoustic topology data, collect sound signals, and generate multi-directional audio data; optimize the arrangement of sensors according to the multi-directional audio data, and re-collect signals to generate optimized collection feature data; quantize the optimized collection feature data to generate spectrum gradient matrix data; The data coding optimization module is used to evaluate the information entropy of the spectrum gradient matrix data and generate information entropy evaluation data; select high entropy data based on the information entropy evaluation data and generate key feature data; optimize the preset audio reconstruction model based on the key feature data and generate a spatial audio reconstruction model; The environmental change simulation module is used to obtain the dynamic change parameters of the site; conduct influence assessment based on the dynamic change parameters of the site to generate factor influence degree data; conduct dynamic change simulation based on the factor influence degree data to generate dynamic simulation result data; construct a reflection network based on the dynamic simulation result data to generate dynamic reflection network data; The model reconstruction enhancement module is used to set optimization targets based on dynamic reflection network data and generate optimization target information data; the optimization target information data is used to iteratively optimize the spatial audio reconstruction model to generate a dynamic audio reconstruction model; The sound dynamic adjustment module is used to collect feedback data using sensor equipment and generate real-time feedback data, including real-time audio feedback data and real-time environmental impact data; use the dynamic audio reconstruction model and perform feedback control based on the real-time audio feedback data and real-time environmental impact data to generate sound propagation optimization data.

Images