Echo suppression method, device and system of spherical screen display screen system

Abstract

The invention discloses an echo suppression method, device and system of a spherical screen display screen system, and the method comprises the steps: obtaining the structural data and acoustic parameters of the spherical screen display screen system, and constructing an initial acoustic model; acquiring real-time acoustic data in the dome screen, and dynamically updating the initial acoustic model to obtain a real-time acoustic model; calculating a sound wave propagation path based on the real-time acoustic model and the current audio output characteristics, and generating a dynamic echo risk map identifying potential echo positions, intensities, time and propagation trends; according to the dynamic echo risk map, calculating real-time offset sound wave parameters for offsetting potential echoes; based on the real-time counteracting sound wave parameters, accurately generating time-space directional counteracting sound waves in the spherical screen; and monitoring an offset effect in real time, and correcting the real-time acoustic model according to a monitoring result. According to the technical scheme, echoes can be prevented from being generated in the spherical screen display screen.

Description

Technical Field

[0001] The present invention relates to the technical field of echo suppression of a spherical screen display system, and in particular to an echo suppression method, device and system of a spherical screen display system. Background Art

[0002] The dome display is a new large-scale immersive display system (see the attached Figure 1 ), usually composed of multiple curved LED display unit panels, which are installed on a curved frame to form a hemispherical or spherical display surface. Dome screens are widely used in planetariums, science and technology exhibitions, virtual reality experience centers and other places, providing audiences with an immersive visual experience. Its unique curved surface structure can not only expand the audience's field of vision, but also create a more realistic and three-dimensional image effect, making the audience feel as if they are in the displayed environment.

[0003] However, the existing dome display system mainly focuses on optimizing visual effects, but the control of the acoustic environment is relatively insufficient. Due to the special structure of the dome, complex acoustic reflections and echoes are easily formed inside it, which not only interferes with the clarity of the audio content, but also may cause the audience to feel inaccurate sound positioning, thus affecting the overall immersive experience. Summary of the invention

[0004] The main purpose of the present invention is to solve the technical problem that complex acoustic reflection and echo phenomena are easily formed inside the existing spherical screen display screen.

[0005] A first aspect of the present invention provides an echo suppression method for a spherical screen display system, the echo suppression method for the spherical screen display system comprising: Obtain the structural data and acoustic parameters of the dome display system and build an initial acoustic model; Collecting real-time acoustic data inside the dome, dynamically updating the initial acoustic model, and obtaining a real-time acoustic model; Based on the real-time acoustic model and current audio output characteristics, the sound wave propagation path is calculated to generate a dynamic echo risk map that identifies the potential echo location, intensity, time and propagation trend; Calculating real-time cancellation acoustic wave parameters for canceling potential echoes according to the dynamic echo risk map; Based on the real-time cancellation sound wave parameters, accurately generate time-space directional cancellation sound waves in the spherical screen; The cancellation effect is monitored in real time, and the real-time acoustic model is corrected according to the monitoring result.

[0006] Optionally, the calculating the sound wave propagation path based on the real-time acoustic model and the current audio output characteristics, and generating a dynamic echo risk map identifying the potential echo position, intensity, time and propagation trend, includes: Discretize the internal space of the dome into multiple acoustic units; For each of the acoustic units, calculating the propagation characteristics of the sound wave in the acoustic unit based on the real-time acoustic model; In combination with the time-frequency characteristics of the current audio output, an echo risk assessment is performed on each of the acoustic units; Based on the spatial relationship between the acoustic units, establishing an acoustic unit association network; Using a graph neural network algorithm, the acoustic unit association network is dynamically analyzed to predict the spatiotemporal evolution of the echo; Based on the prediction results, a dynamic echo risk map including echo location, intensity, time and propagation trend is generated.

[0007] Optionally, the using of a graph neural network algorithm to dynamically analyze the acoustic unit association network to predict the spatiotemporal evolution of the echo includes: Construct a graph neural network model, represent each acoustic unit as a node, and represent the sound wave propagation relationship between acoustic units as edges; Assign a feature vector to each node, which contains the acoustic characteristics, current sound pressure level and historical echo data of the acoustic unit, as the input of the graph neural network; Based on the node feature vector, a graph convolution layer is designed and applied to capture the acoustic wave interaction between adjacent acoustic units to obtain updated node features; Using the updated node features, construct and apply a temporal convolution layer to process the temporal characteristics of sound wave propagation and generate a temporally related node representation; Applying an attention mechanism to the time-series related node representations to calculate the importance weight of each acoustic unit in the echo formation process; Based on the importance weights and the time-series related node representations, the echo intensity and phase of each acoustic unit at a future time point are predicted through iterative calculation of a multi-layer graph neural network; The prediction results of each acoustic unit are integrated to generate a prediction of the spatiotemporal evolution of the echo in the entire dome space.

[0008] Optionally, the accurately generating the time-space directional cancellation sound waves in the spherical screen based on the real-time cancellation sound wave parameters includes: Determine the distribution structure of the sound-transmitting holes on the spherical screen display screen, each sound-transmitting hole corresponds to at least one acoustic unit; Based on the acoustic unit association network, a mapping relationship between the sound-permeable holes and the acoustic units is established; Calculating the sound wave characteristics that each sound-permeable hole needs to output according to the real-time sound wave cancellation parameters and the mapping relationship; The acoustic impedance characteristics of the sound-permeable holes are used to modulate the sound waves passing through each sound-permeable hole to achieve directional propagation of the sound waves. By controlling the phase difference of the sound waves between adjacent sound-permeable holes, a controllable sound wave interference structure is formed; Based on the acoustic wave interference structure, a local cancellation sound field is formed at the position of the target acoustic unit; By integrating multiple local cancellation sound fields, the overall space-time directional cancellation sound waves in the spherical screen are constructed.

[0009] Optionally, utilizing the acoustic impedance characteristics of the sound-transmitting holes to modulate the sound waves passing through each sound-transmitting hole to achieve directional propagation of the sound waves includes: Classify the sound-transmitting holes on the spherical screen and group similar sound-transmitting holes into one group according to geometric parameters and material properties; Establish a representative acoustic impedance model for each group of sound-permeable holes; Based on each group of representative acoustic impedance models, the overall acoustic characteristic mapping of the dome display screen is constructed; Designing an adaptive acoustic wave modulation algorithm using the overall acoustic property mapping; Inputting the to-be-outputted cancelling sound wave signal into the adaptive sound wave modulation algorithm; According to the adaptively modulated signal, the propagation characteristics of the sound waves through each group of sound-permeable holes are controlled; By adjusting the phase and amplitude relationship of the sound waves between different groups of sound-permeable holes, the directional propagation of the sound waves can be achieved.

[0010] Optionally, the core formula of the adaptive sound wave modulation algorithm is: ; in, represents the output directional sound wave field, is the angular frequency, and Represent the azimuth and elevation angles respectively, N is the total number of sound-permeable hole groups, is the input cancellation sound wave signal, is the weight function of the i-th group of sound-transmitting holes, which is used to adjust the contribution of this group of sound-transmitting holes. is the transfer function of the i-th group of sound-transmitting holes, which describes the propagation characteristics of the sound wave after passing through this group of sound-transmitting holes, and Σ represents the sum of the contributions of all groups.

[0011] Optionally, the sound-transmitting holes on the spherical screen are classified, and similar sound-transmitting holes are grouped together according to geometric parameters and material properties, including: Obtaining geometric parameters of each sound-transmitting hole, including aperture, depth, shape factor and spatial position coordinates; Measure the acoustic properties of the material in the area where each sound-permeable hole is located, including acoustic impedance, sound absorption coefficient and sound wave propagation speed; Construct a multidimensional feature space, where each dimension corresponds to a geometric parameter or material property; Mapping each sound-permeable hole to a point in the multidimensional feature space; Apply clustering algorithms to group points in feature space; Calculate the consistency of acoustic responses of the sound-permeable holes in each group to evaluate the classification effect; Based on the consistency evaluation results, clustering parameters are dynamically adjusted and classification results are iteratively optimized; According to the final clustering results, the sound-permeable holes with similar geometric parameters and material properties are grouped together to complete the classification.

[0012] Optionally, the collecting of real-time acoustic data inside the dome and dynamically updating the initial acoustic model to obtain a real-time acoustic model includes: Collect real-time acoustic data through the acoustic sensor network pre-installed in the dome display system; Preprocessing the real-time acoustic data to eliminate noise and extract effective information; Compare the preprocessed real-time acoustic data with the initial acoustic model to identify changes in the acoustic environment; Based on the identified changes, determining acoustic model parameters that need to be updated; Using real-time acoustic data to update and calculate the determined model parameters; Integrate the updated parameters into the initial acoustic model to form a temporary updated model; Verifying the temporarily updated model and evaluating its adaptability to the current acoustic environment; According to the verification results, if the performance of the temporary updated model is better than the current model, the temporary updated model is adopted as the new real-time acoustic model; otherwise, the current model is retained and the reasons for the update failure are recorded for subsequent optimization reference.

[0013] A second aspect of the present invention provides an echo suppression device for a spherical screen display system, comprising: Model building module, used to obtain the structural data and acoustic parameters of the dome display system and build an initial acoustic model; A model updating module, used for collecting real-time acoustic data inside the dome, dynamically updating the initial acoustic model, and obtaining a real-time acoustic model; A risk mapping module, for calculating the sound wave propagation path based on the real-time acoustic model and current audio output characteristics, and generating a dynamic echo risk map identifying the potential echo location, intensity, time and propagation trend; A parameter calculation module, used for calculating real-time cancellation sound wave parameters for canceling potential echoes according to the dynamic echo risk map; A sound wave generation module, used for accurately generating a time-space directional cancellation sound wave in the spherical screen based on the real-time cancellation sound wave parameters; The effect monitoring module is used to monitor the cancellation effect in real time and calibrate the real-time acoustic model according to the monitoring result.

[0014] A third aspect of the present invention provides a spherical screen display system, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via lines; the at least one processor calls the instructions in the memory so that the spherical screen display system executes the steps of the above-mentioned echo suppression method for the spherical screen display system.

[0015] The echo suppression method for the spherical screen display system proposed in the present invention effectively solves the echo problem in a complex spherical screen environment through a series of innovative technical features. First, the method constructs an initial acoustic model by acquiring system structure data and acoustic parameters, and dynamically updates the model using real-time acoustic data, ensuring that the acoustic model can always accurately reflect the acoustic environment inside the current spherical screen. This real-time adaptive model lays a solid foundation for subsequent echo location and suppression. Secondly, based on the updated real-time acoustic model and the current audio output characteristics, the method calculates the sound wave propagation path and generates a dynamic echo risk map. This mapping technology not only accurately locates the position, intensity and time of potential echoes, but also predicts their propagation trend, enabling the system to respond to possible echo problems in advance. In this way, the present method achieves accurate echo location in a complex spherical screen environment.

[0016] In terms of achieving dynamic echo suppression, the innovation of this method is reflected in the following aspects: First, according to the dynamic echo risk map, the system can calculate the targeted compensation sound wave parameters in real time. These parameters take into account the spatiotemporal characteristics of the echo and ensure the accuracy of the compensation sound waves. Secondly, based on these parameters, the system accurately generates spatiotemporal directional compensation sound waves in the dome. This directional sound wave can effectively offset the predicted echo at a specific location and time without causing unnecessary impact on other areas. Finally, the method also includes the steps of real-time monitoring of the compensation effect and correcting the real-time acoustic model accordingly, forming a closed-loop self-optimization system. This continuous monitoring and correction mechanism ensures that the system can adapt to the ever-changing acoustic environment inside the dome and always maintain the best echo suppression effect. Through this dynamic, precise and adaptive echo suppression method, the present invention effectively improves the acoustic environment of the dome display system, significantly improves the user's auditory experience, and provides technical support for creating a more immersive viewing experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0018] Figure 1 It is a structural schematic diagram of an existing ball screen display; Figure 2 Schematic diagram of the flow of the echo suppression method of the spherical screen display system in an embodiment of the present invention; Figure 3 A schematic diagram of an embodiment of an echo suppression device for a spherical screen display system according to an embodiment of the present invention; Figure 4 A schematic diagram of an embodiment of a spherical screen display system in an embodiment of the present invention; Figure 5 A schematic diagram of an embodiment of a curved LED display unit board of a spherical screen display system of the present invention; Figure 6 for Figure 5 A partial enlarged view of point A in the middle.

[0019] Description of Figure Numbers: 1. Curved LED display unit board; 2. Display unit; 3. Sound-permeable hole.

[0020] 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

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0023] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, which must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0024] An embodiment of the present application provides an echo suppression method for a spherical screen display system. Figure 2 A flow chart of an echo suppression method for a spherical screen display system provided in an embodiment of the present application. In this embodiment, the method includes: See also Figure 2 , obtain the structural data and acoustic parameters of the ball screen display system and build an initial acoustic model; Specifically, the ball screen display system is usually composed of an arc frame, a plurality of arc LED display unit boards 1 and an audio system. The arc frame is the supporting structure of the entire system, a plurality of arc LED display unit boards 1 are mounted on the frame to form a hemispherical or spherical display surface, and the audio system is responsible for the playback and transmission of sound.

[0025] Acquiring the structural data of the dome display system first involves measuring the geometric parameters of the arc skeleton. This includes the radius of curvature, height, diameter of the skeleton, and the number and arrangement of the skeleton structural units. These data can be measured by a 3D laser scanner. For example, using the FARO Focus 3D laser scanner, accurate 3D point cloud data of the entire skeleton structure can be obtained in a short time.

[0026] For the curved LED display unit board 1, it is necessary to measure the size, thickness, weight of each board, as well as the diameter, depth, number and spatial distribution of the sound-transmitting holes 3 on the board. These data can be obtained using a high-precision digital caliper and a weight sensor. For example, the diameter of the sound-transmitting hole 3 can be accurately measured using a digital caliper.

[0027] Acoustic parameters are key indicators that describe the propagation characteristics of sound waves in a dome system. They mainly include the following aspects: 1. Sound absorption coefficient: It indicates the ability of a material to absorb sound energy. It can be measured using the impedance tube method. For example, the Brüel&Kjær Type 4206-T impedance tube is used.

[0028] 2. Scattering coefficient: describes the degree to which sound waves are scattered by a surface. It can be measured in an anechoic chamber using a rotating microphone array, such as the NTi Audio FX100 audio analyzer with a rotating mechanism.

[0029] 3. Sound transmission loss: It indicates the energy loss when sound waves pass through the structure. It can be measured using the double chamber method. For example, a Norsonic Nor150 sound level meter can be used.

[0030] 4. Reverberation time: It indicates the speed at which sound decays in space. It can be measured using the impulse response method, such as using EASE software with an omnidirectional sound source and a test microphone.

[0031] The process of obtaining these acoustic parameters is as follows: First, an omnidirectional sound source (such as Brüel & Kjær Type 4295) and a test microphone array (such as PCB 378B02) are arranged inside the dome. Then, a standard test signal, such as a logarithmic sweep signal or a maximum length sequence signal, is played. By analyzing the received signal, various acoustic parameters can be calculated.

[0032] For example, in a hemispherical dome with a diameter of 30 meters, at least 16 test points can be arranged at different heights and azimuths. A logarithmic sweep signal from 20Hz to 20kHz is played at each point for 10 seconds. By analyzing the signal received at each test point, the frequency response, reverberation time and other parameters of the point can be obtained.

[0033] These structural data and acoustic parameters form the basis of the initial acoustic model. By importing the geometric data into acoustic simulation software (such as ODEON or EASE) and inputting the measured acoustic parameters, a digital model that accurately reflects the acoustic characteristics of the dome screen system can be constructed. This model can simulate the propagation, reflection and absorption of sound waves in the system, providing an accurate theoretical basis for subsequent echo prediction and suppression.

[0034] Please continue reading Figure 2 , collecting real-time acoustic data inside the dome, dynamically updating the initial acoustic model, and obtaining a real-time acoustic model; In one embodiment of the present invention, the real-time acoustic data inside the dome is collected, and the initial acoustic model is dynamically updated to obtain the real-time acoustic model, including: collecting real-time acoustic data through an acoustic sensor network preset in the dome display system; preprocessing the real-time acoustic data to eliminate noise and extract effective information; comparing the preprocessed real-time acoustic data with the initial acoustic model to identify changes in the acoustic environment; based on the identified changes, determining the acoustic model parameters that need to be updated; using the real-time acoustic data to update and calculate the determined model parameters; integrating the updated parameters into the initial acoustic model to form a temporary update model; verifying the temporary update model to evaluate its adaptability to the current acoustic environment; according to the verification result, if the performance of the temporary update model is better than the current model, the temporary update model is used as the new real-time acoustic model; otherwise, the current model is retained and the reason for the update failure is recorded for subsequent optimization reference.

[0035] Specifically, pre-installing an acoustic sensor network in the dome display system is the key to achieving real-time acoustic data collection. This network consists of multiple microphone arrays distributed at different locations inside the dome. For example, in a hemispherical dome with a diameter of 30 meters, 32 microphone arrays can be evenly arranged, each array containing 4 omnidirectional microphones. This arrangement can fully capture the sound field distribution inside the dome. The real-time acoustic data collected includes information such as sound pressure level, spectrum distribution, and sound field directionality. The sampling frequency is set to 48kHz to ensure that all sound details within the audible range of the human ear are captured.

[0036] Preprocessing the collected real-time acoustic data is an important step to improve the accuracy of subsequent analysis. The preprocessing process first uses a digital filter to eliminate environmental noise, such as low-frequency noise generated by the air conditioning system. Then, the time domain signal is converted into a frequency domain signal using Fourier transform to extract effective acoustic features, such as energy distribution and phase information in each frequency band. This step can significantly improve the signal-to-noise ratio of the acoustic data and provide a clear data basis for subsequent model comparison and update.

[0037] Comparing the preprocessed real-time acoustic data with the initial acoustic model is key to identifying changes in the acoustic environment. This process uses numerical analysis methods to calculate the error between the actual measured data and the model's predictions. For example, for each measurement point, the root mean square error between the measured sound pressure level and the model's predicted sound pressure level is calculated. If the error in a certain area exceeds a preset threshold (such as 3dB), the acoustic environment in that area is considered to have changed significantly.

[0038] Based on the identified changes, determine the acoustic model parameters that need to be updated. These parameters may include the sound absorption coefficient, scattering coefficient, or geometric structure parameters of a specific area. For example, if the actual reverberation time of an area is found to be longer than the model predicts, the sound absorption coefficient of the surface material in that area needs to be adjusted.

[0039] Using real-time acoustic data to update the model parameters is a key step in improving model accuracy. Here, an iterative optimization algorithm, such as the Levenberg-Marquardt algorithm, is used to adjust the parameters by minimizing the error between the model prediction and the measured value. For example, when adjusting the sound absorption coefficient, the algorithm will gradually change the parameter value until the reverberation time predicted by the model is close enough to the measured value.

[0040] Integrate the updated parameters into the initial acoustic model to form a temporary updated model. This process uses acoustic simulation software, such as ODEON, to input the new parameter values ​​into the original geometric model and recalculate the acoustic field distribution.

[0041] Verifying the temporarily updated model is an important step to ensure the validity of the model update. The verification process involves using the updated model to predict the sound field distribution inside the dome, and then comparing it with the actual measurement data. For example, you can select 10 measurement points that are not involved in the model update and compare the measured sound pressure levels at these points with the model prediction values.

[0042] Decide whether to use the temporary update model based on the verification results. If the prediction error of the temporary update model at the verification point is less than that of the current model (such as a 20% decrease in average error), use the temporary update model as the new real-time acoustic model. Otherwise, keep the current model and record the reason for the update failure, such as "the prediction error of a specific frequency band increases", for reference in subsequent optimization.

[0043] It is not difficult to understand that this method of dynamically updating the acoustic model has several advantages. First, it can adapt to changes in the acoustic environment inside the dome in real time, such as changes in the sound absorption characteristics caused by the entry of spectators. Second, by only updating the parameters of the changed area, the amount of calculation is greatly reduced and the system response speed is improved. Third, the verification step ensures the reliability of the model update and avoids model degradation due to measurement errors. Finally, recording the reasons for the update failure provides valuable data support for the long-term optimization of the system. This method enables the dome display system to always maintain an accurate acoustic model, providing a reliable foundation for subsequent echo suppression.

[0044] Please continue reading Figure 2 , based on the real-time acoustic model and current audio output characteristics, calculating the sound wave propagation path, and generating a dynamic echo risk map identifying potential echo location, intensity, time and propagation trend; In one embodiment of the present invention, the method of calculating the sound wave propagation path based on the real-time acoustic model and the current audio output characteristics and generating a dynamic echo risk map that identifies the potential echo position, intensity, time and propagation trend includes: discretizing the internal space of the spherical screen into multiple acoustic units; for each of the acoustic units, calculating the propagation characteristics of the sound wave in the acoustic unit based on the real-time acoustic model; performing echo risk assessment on each of the acoustic units in combination with the time-frequency characteristics of the current audio output; establishing an acoustic unit association network based on the spatial relationship between the acoustic units; using a graph neural network algorithm to dynamically analyze the acoustic unit association network and predict the spatiotemporal evolution of the echo; and generating a dynamic echo risk map including the echo position, intensity, time and propagation trend based on the prediction results.

[0045] Specifically, discretizing the internal space of the dome into multiple acoustic units is a basic step to achieve accurate echo prediction. This process uses three-dimensional meshing technology to divide the dome space into a large number of small cubes or tetrahedron units. Taking a hemispherical dome with a diameter of 30 meters as an example, it can be divided into 100,000 cubic units with a side length of about 30 centimeters. This division method can capture the detailed characteristics of sound wave propagation while ensuring computational efficiency. Each acoustic unit is assigned specific acoustic properties, such as sound absorption coefficient and scattering coefficient, which are derived from the real-time acoustic model constructed previously.

[0046] Calculating the acoustic wave propagation characteristics for each acoustic element is key to predicting the echo. This step uses numerical acoustic simulation methods such as the finite-difference time-domain method (FDTD). The FDTD method simulates the propagation, reflection, and diffraction of sound waves in each element by solving the acoustic wave equation. Specifically, for each time step (usually milliseconds), the sound pressure and particle velocity are calculated for each element. This method can accurately capture the behavior of acoustic waves in complex geometric structures, including edge diffraction and focal effects.

[0047] Performing echo risk assessment on each acoustic unit in combination with the time-frequency characteristics of the current audio output is an important step in identifying potential problem areas. First, a short-time Fourier transform is performed on the current audio output to obtain its time-frequency representation. This time-frequency feature is then combined with the propagation characteristics of each acoustic unit to calculate the echo intensity that may be generated within the unit. For example, if a unit has a high reflection coefficient in a specific frequency band, and the current audio output also has strong energy in that frequency band, then the echo risk of this unit is high. This assessment method takes into account the dynamic changes in sound source characteristics and environmental factors, and can identify echo hotspots in real time.

[0048] The key to global echo analysis is to establish an acoustic unit association network based on the spatial relationship between acoustic units. This network is represented by a graph structure, where each node represents an acoustic unit and the edge represents the sound wave propagation relationship between adjacent units. The weight of the edge can be represented by the time or energy attenuation of the sound wave propagation between units. For example, for any two adjacent acoustic units, the sound wave propagation time can be calculated based on the distance and sound speed between them as the weight of the edge connecting the two units. This network structure can effectively express the sound wave propagation path in complex space.

[0049] The core innovation of this method is to use the graph neural network algorithm to dynamically analyze the acoustic unit association network and predict the spatiotemporal evolution of the echo. Graph neural network is a deep learning model that specializes in processing graph structured data. Here, the graph convolutional network (GCN) is used to process the acoustic unit association network. Each layer of GCN aggregates and updates the node features to simulate the propagation process of sound waves in space. Specifically, for each acoustic unit, GCN considers the characteristics of its neighboring units and performs weighted aggregation according to the weight of the edge to capture the spatial dependence of sound wave propagation. By stacking multiple layers of GCN, the model is able to simulate the long-distance propagation and multiple reflections of sound waves in the entire spherical space. The advantage of this method is that it can simultaneously consider local and global acoustic characteristics to achieve accurate prediction of echo evolution in complex spaces.

[0050] Finally, a dynamic echo risk map is generated based on the prediction results. This map is a four-dimensional data structure that contains three-dimensional spatial coordinates and a time dimension. For each spatial point and time point, the map records the predicted echo intensity. For example, the entire process of an echo from its occurrence to its disappearance can be clearly seen in the map, including its intensity change and spatial movement trajectory. This visualization method intuitively shows the temporal and spatial distribution of the echo within the dome, providing a precise target for subsequent echo suppression.

[0051] It can be understood that this embodiment is able to handle the propagation of sound waves in complex geometric structures such as the spherical screen, and capture detail effects that are difficult to simulate with traditional methods. Secondly, by combining real-time audio features, this method can dynamically adapt to different sound source content and provide more accurate echo prediction. Thirdly, the use of graph neural networks greatly improves computational efficiency, making real-time echo prediction possible in complex environments. Finally, the generated dynamic echo risk map provides detailed guidance for subsequent directional echo suppression, which helps to achieve more accurate and effective acoustic environment control. The application of this method enables the spherical screen display system to provide a stable and clear auditory experience in complex and changing acoustic environments, greatly improving the overall performance of the system.

[0052] In one embodiment of the present invention, the use of a graph neural network algorithm to dynamically analyze the acoustic unit association network and predict the spatiotemporal evolution of the echo includes: constructing a graph neural network model, representing each acoustic unit as a node, and representing the sound wave propagation relationship between acoustic units as an edge; assigning a feature vector to each node, including the acoustic characteristics, current sound pressure level and historical echo data of the acoustic unit, as an input of the graph neural network; based on the node feature vector, designing and applying a graph convolution layer to capture the sound wave interaction between adjacent acoustic units to obtain updated node features; using the updated node features, constructing and applying a time convolution layer to process the timing characteristics of sound wave propagation and generate a timing-related node representation; applying an attention mechanism to the timing-related node representation to calculate the importance weight of each acoustic unit in the echo formation process; based on the importance weight and the timing-related node representation, predicting the echo intensity and phase of each acoustic unit at a future time point through iterative calculation of a multi-layer graph neural network; and generating a prediction of the spatiotemporal evolution of the echo of the entire dome space by combining the prediction results of each acoustic unit.

[0053] Specifically, building a graph neural network model is a key step in achieving echo prediction in complex dome spaces. First, each acoustic unit is represented as a node in the graph, and the sound wave propagation relationship between acoustic units is represented as an edge. For example, in a dome containing 100,000 acoustic units, the graph model will contain 100,000 nodes. The edge weights between adjacent acoustic units can be represented by the sound wave propagation time, which can be calculated by dividing the distance between units by the speed of sound. This representation method can intuitively reflect the acoustic topology of the dome space.

[0054] Assigning a feature vector to each node is the basis of the graph neural network input. The feature vector contains the acoustic properties of the acoustic unit (such as absorption coefficient, scattering coefficient), current sound pressure level, and historical echo data. Specifically, a 32-dimensional vector is used to represent these features, where the first 10 dimensions represent the absorption coefficient in different frequency bands, the next 10 dimensions represent the scattering coefficient, then the current sound pressure level, and the last 10 dimensions are the echo intensity in the past 10 time steps. This comprehensive feature representation enables the model to make full use of the static properties and dynamic state information of the acoustic unit.

[0055] Designing and applying graph convolutional layers based on node feature vectors is the key to capturing the acoustic wave interactions between adjacent acoustic units. The core operation of the graph convolutional layer is to aggregate the information of adjacent nodes. The specific implementation can be achieved using the following formula: ; in, Indicates that node i is in The characteristics of the layer, is the activation function (ReLU), For the The weight matrix of the layer, is the feature of node i’s neighboring node j at layer l, is the bias term, is the set of neighbor nodes of node i.

[0056] In this way, the new features of each acoustic unit not only contain its own information, but also integrate the information of surrounding units, effectively simulating the local propagation characteristics of sound waves.

[0057] Constructing and applying a temporal convolutional layer using updated node features is an important step in processing the temporal characteristics of sound wave propagation. The temporal convolutional layer uses a one-dimensional convolution operation to process the time series features of each node. For example, a one-dimensional convolution with a kernel size of 3 and a step size of 1 can be used to convolve the features of the last three time steps of each node to generate a time-dependent node representation. This operation can capture short-term temporal dependencies in the process of sound wave propagation.

[0058] The attention mechanism is applied to the time-series related node representations to calculate the importance weight of each acoustic unit in the echo formation process. The core idea of ​​the attention mechanism is to let the model focus on the most relevant information. The specific implementation can use the self-attention mechanism, and the calculation formula is: ; Where Q, K, and V are query, key, and value matrices, respectively, which are obtained by linear transformation of the time series representation of the nodes. is the dimension of the key vector.

[0059] In this way, the model can adaptively adjust the degree of attention paid to different acoustic units and highlight the influence of important areas.

[0060] Based on the importance weights and time-related node representations, the echo intensity and phase of each acoustic unit at future time points are predicted through iterative calculations of multi-layer graph neural networks. This process can be seen as a combination of multiple graph convolutions and time convolutions. For example, 3 layers of graph convolution and 2 layers of time convolution can be stacked, each followed by a fully connected layer, and finally an output layer is used to predict the echo intensity and phase for the next 5 time steps. This deep structure can capture complex spatiotemporal dependencies and improve the accuracy of predictions.

[0061] Finally, the prediction results of each acoustic unit are combined to generate a prediction of the spatiotemporal evolution of the echo in the entire dome space. This step integrates the prediction results of all acoustic units into a unified four-dimensional data structure, where three dimensions represent spatial coordinates and one dimension represents time. For example, for a 30-meter diameter dome, a 100×100×50×10 tensor can be generated, representing the echo prediction of a 100×100×50 spatial grid in the next 10 time steps.

[0062] This echo prediction method based on graph neural network has multiple advantages. First, it can effectively handle the propagation of sound waves in complex geometric structures such as dome screens, and capture detail effects that are difficult to simulate with traditional methods. Second, by combining graph convolution and temporal convolution, this method simultaneously considers the spatial and temporal characteristics of sound wave propagation, providing more comprehensive and accurate predictions. Third, the introduction of the attention mechanism enables the model to adaptively focus on important areas, improving the pertinence and efficiency of predictions. Finally, the end-to-end learning characteristics of this method enable it to learn complex echo patterns from a large amount of historical data, and the prediction performance will continue to improve with the increase in usage time. The application of this method enables the dome screen display system to accurately predict and control echoes in complex acoustic environments, greatly improving the system's acoustic performance and user experience.

[0063] Please continue reading Figure 2 , calculating real-time cancellation acoustic wave parameters for canceling potential echoes according to the dynamic echo risk map; Specifically, the system extracts the predicted echo intensity and phase information for each spatiotemporal point from the dynamic echo risk map. For example, in a 30-meter diameter dome, there may be 100×100×50 spatial grid points, each of which contains echo prediction data for the next 10 time steps. The system then uses an inverse filtering algorithm to calculate the parameters of the canceling sound wave. Specifically, for each predicted echo, a canceling sound wave with equal amplitude but opposite phase is calculated. This process can be achieved through a fast Fourier transform (FFT), which converts the time domain signal to the frequency domain, adjusts the amplitude and phase, and then converts it back to the time domain through an inverse FFT. Next, the system considers the propagation time of the sound wave from the sound source to the target location and compensates the time delay of the canceling sound wave. This delay time can be calculated by dividing the distance from the sound source to the target point by the speed of sound. Finally, the system also needs to consider the acoustic transfer function of the dome space and pre-equalize the canceling sound wave to ensure that the canceling sound wave has the desired characteristics when it arrives at the target location. The advantage of this method is that it can generate precise cancellation sound waves for each potential echo point, rather than simply performing global noise suppression. Through this fine control, the system can effectively eliminate echoes at specific locations and times without affecting the original audio quality, greatly improving the acoustic performance of the dome display system.

[0064] Please continue reading Figure 2 , based on the real-time cancellation sound wave parameters, accurately generating time-space directional cancellation sound waves in the spherical screen; See also Figure 5 and Figure 6 , Figure 5 A schematic diagram of an embodiment of a curved LED display unit board 1 of a spherical screen display system of the present invention; Figure 6 for Figure 5 The local enlarged view of A in the middle; it should be noted that Figure 5 A plurality of display units 2 are provided on the arc-shaped LED display unit board 1 for jointly displaying the content to be displayed, and the sound-permeable holes 3 are provided between the display units 2 .

[0065] It should be noted that in the present application, the sound-permeable hole 3 can not only be used to allow sound waves to pass through, but also can be used to dissipate the heat generated by the display unit 2 or other heat to achieve cooling. In addition, in some embodiments, sound-absorbing cotton can be arranged on the side of the curved LED display unit board 1 facing away from the user to achieve more effective sound reduction.

[0066] In one embodiment of the present invention, based on the real-time cancellation sound wave parameters, accurately generating time-space directional cancellation sound waves in the spherical screen, including: determining the distribution structure of the sound-permeable holes 3 on the spherical screen display screen, each sound-permeable hole 3 corresponds to at least one acoustic unit; based on the acoustic unit association network, establishing a mapping relationship between the sound-permeable holes 3 and the acoustic unit; according to the real-time cancellation sound wave parameters and the mapping relationship, calculating the sound wave characteristics that each sound-permeable hole 3 needs to output; using the acoustic impedance characteristics of the sound-permeable holes 3, modulating the sound waves passing through each sound-permeable hole 3 to achieve directional propagation of the sound waves; by controlling the sound wave phase difference of adjacent sound-permeable holes 3, forming a controllable sound wave interference structure; based on the sound wave interference structure, forming a local cancellation sound field at the target acoustic unit position; and integrating multiple local cancellation sound fields to construct the overall time-space directional cancellation sound waves in the spherical screen.

[0067] Specifically, a precise geometric model of the dome surface is obtained through three-dimensional scanning technology, and then the position and size of each sound-permeable hole 3 are marked on the model. For example, on a hemispherical dome with a diameter of 30 meters, there are about 10,000 evenly distributed sound-permeable holes 3, each with a hole diameter of 5 mm. Next, the dome space (the space here refers to the space on the side of the dome close to the user) is divided into acoustic units, and each sound-permeable hole 3 is established with the acoustic units around it. Usually, one sound-permeable hole 3 corresponds to multiple acoustic units, forming a one-to-many mapping.

[0068] Establishing the mapping relationship between the sound-permeable hole 3 and the acoustic unit based on the acoustic unit association network is the key to optimizing the sound wave output. This step uses the node connection information in the graph neural network to assign a set of acoustic units within the influence range to each sound-permeable hole 3. For example, a sound-permeable hole 3 may be associated with 100 acoustic units within a range of 1 meter around it. This mapping relationship takes into account the propagation characteristics of the sound wave and ensures the accuracy of subsequent sound wave control.

[0069] The key to achieving a directional sound field is to calculate the sound wave characteristics that each sound-permeable hole 3 needs to output based on the real-time cancellation sound wave parameters and mapping relationship. This step uses the inverse problem solving technology to calculate the sound wave amplitude, phase and frequency that each sound-permeable hole 3 should output based on the target cancellation sound field and the position of the sound-permeable hole 3. For example, if a cancellation sound field needs to be formed 5 meters away from the surface of the dome, the system will calculate the precise sound wave parameters that each relevant sound-permeable hole 3 needs to output.

[0070] The acoustic impedance characteristics of the sound-transmitting holes 3 are used to modulate the sound waves passing through each sound-transmitting hole 3 to achieve directional propagation of the sound waves. This step takes into account the effects of the geometric shape and material properties of the sound-transmitting holes 3 on the propagation of the sound waves. By adjusting the frequency response of the input signal, the sound wave attenuation and phase change caused by the sound-transmitting holes 3 are compensated to ensure that the output sound waves have the expected characteristics. For example, for a circular sound-transmitting hole 3 with a diameter of 5 mm, the system will pre-equalize the input signal according to its acoustic impedance characteristics.

[0071] Controlling the phase difference of the sound waves between adjacent sound-permeable holes 3 to form a controllable sound wave interference structure is a key technology for achieving precise sound field control. This step uses the principle of wavefront synthesis to form the desired sound wave interference structure in space by precisely controlling the phase of the sound waves output by each sound-permeable hole 3. For example, by adjusting the phase difference between adjacent sound-permeable holes 3, a sound beam pointing in a specific direction can be formed, or an acoustic focus point can be formed at a specific position.

[0072] Based on the sound wave interference structure, forming a local cancellation sound field at the target acoustic unit position is the core of achieving accurate echo suppression. This step comprehensively considers the sound waves output by multiple sound-permeable holes 3 to form a sound field with the same amplitude but opposite phase as the predicted echo at the target position. For example, if an echo with an intensity of 60dB is predicted 5 meters from the center of the dome, the system will output sound waves through the surrounding sound-permeable holes 3 to form a 60dB, opposite-phase cancellation sound field at that position.

[0073] Finally, multiple local cancellation sound fields are combined to construct the overall space-time directional cancellation sound waves in the dome. This step integrates all local cancellation sound fields to form a global cancellation sound field covering the entire dome space. The system needs to balance the cancellation requirements of different areas to ensure the continuity and smoothness of the global sound field. For example, when dealing with multiple scattered echo hotspots, the system will coordinate the local cancellation sound fields to avoid mutual interference between them.

[0074] This precise sound field control method based on the sound-permeable holes 3 has several advantages. First, it makes full use of the inherent structural characteristics of the spherical screen display without the need for additional acoustic equipment. Second, by precisely controlling the output of each sound-permeable hole 3, highly directional sound field control is achieved, greatly improving the accuracy of echo suppression. Third, by utilizing the principle of acoustic wave interference, this method can form the desired sound field distribution in a complex three-dimensional space and adapt to various echo patterns. Finally, the flexibility and scalability of this method enables it to cope with spherical screen display systems of different sizes and shapes, providing an effective solution for improving the acoustic performance of large immersive display environments.

[0075] In one embodiment of the present invention, the acoustic impedance characteristics of the sound-permeable holes 3 are used to modulate the sound waves passing through each sound-permeable hole 3 to achieve directional propagation of the sound waves, including: classifying the sound-permeable holes 3 on the spherical screen display screen, and grouping similar sound-permeable holes 3 according to geometric parameters and material properties; establishing a representative acoustic impedance model for each group of sound-permeable holes 3; constructing an overall acoustic characteristic mapping of the spherical screen display screen based on the representative acoustic impedance models of each group; designing an adaptive sound wave modulation algorithm using the overall acoustic characteristic mapping; inputting the offset sound wave signal to be output into the adaptive sound wave modulation algorithm; controlling the propagation characteristics of the sound wave through each group of sound-permeable holes 3 according to the adaptively modulated signal; and achieving directional propagation of the sound wave by adjusting the relationship between the sound wave phase and amplitude between different groups of sound-permeable holes 3.

[0076] Specifically, the geometric parameters of each sound-transmitting hole 3, including diameter, depth and shape factor, are obtained through precise measurement. For example, on a hemispherical ball screen with a diameter of 30 meters, there may be 10,000 sound-transmitting holes 3, each with a diameter between 3 and 8 mm and a depth between 2 and 5 mm. At the same time, the material properties of the area where the sound-transmitting holes 3 are located are measured, such as density, elastic modulus and acoustic impedance. Then, a clustering algorithm, such as K-means or hierarchical clustering, is used to group the sound-transmitting holes 3 with similar geometric parameters and material properties. This classification method may classify 10,000 sound-transmitting holes 3 into 50 to 100 different groups.

[0077] Establishing a representative acoustic impedance model for each group of sound-transmitting holes 3 is the basis for accurately simulating sound wave propagation. For each group, a representative sound-transmitting hole 3 is selected, and a detailed acoustic impedance model is constructed using finite element analysis software, such as COMSOL Multiphysics. This model takes into account the geometric shape, material properties and surrounding structure of the sound-transmitting hole 3. The model output includes frequency response function, phase response and acoustic energy transmission efficiency. For example, for a typical cylindrical sound-transmitting hole 3 with a diameter of 5 mm and a depth of 3 mm, its impedance model may show a flat frequency response in the range of 1 kHz to 10 kHz, but a resonance peak appears at higher frequencies.

[0078] Based on each set of representative acoustic impedance models, an overall acoustic property map of the dome display is constructed. This step uses an interpolation algorithm, such as Kriging, to extend the discrete impedance model to the entire dome surface. The result is a continuous three-dimensional acoustic property map that describes the propagation characteristics of sound waves at any location on the dome surface. This map can predict the attenuation, phase change, and frequency response of sound waves passing through any acoustic perforation 3.

[0079] Utilizing the overall acoustic characteristic mapping, designing an adaptive sound wave modulation algorithm is the core of achieving precise sound field control. This algorithm is based on digital signal processing technology and dynamically adjusts the input signal according to the target sound field and the current environment. Specifically, the algorithm first calculates the sound wave propagation path from the sound source to the target position, and then pre-distorts the signal according to the acoustic characteristics of each sound-permeable hole 3 on the path. This includes amplitude equalization, phase compensation, and time delay adjustment. For example, if the goal is to form a 1kHz pure tone sound field 5 meters from the center of the dome, the algorithm will calculate the precise signal that each relevant sound-permeable hole 3 needs to output, including amplitude, phase, and time delay.

[0080] The canceling sound wave signal to be output is input into the adaptive sound wave modulation algorithm. After processing, the system controls the sound wave output of each group of sound-transmitting holes 3 according to the modulated signal. This step uses a digital-to-analog converter and a power amplifier to convert the processed digital signal into an analog electrical signal for driving the sound-transmitting holes 3. Each group of sound-transmitting holes 3 may be equipped with an independent signal processing channel to achieve more precise control.

[0081] Finally, by adjusting the phase and amplitude relationship of the sound waves between different groups of sound-permeable holes 3, directional propagation of the sound waves is achieved. This step uses the principle of wavefront synthesis to form the desired sound wave interference structure in space by precisely controlling the phase difference and amplitude ratio between three groups of adjacent sound-permeable holes. For example, by applying a linear phase difference in the horizontal direction, a horizontally directional sound beam can be formed; by adjusting the amplitude distribution, the width of the sound beam and the sidelobe level can be controlled.

[0082] This method based on the grouping and adaptive modulation of the sound-transmitting holes 3 has several advantages. First, the modeling and control of complex systems are simplified through grouping, which greatly reduces the computational complexity and makes real-time processing possible. Second, the use of representative models not only ensures accuracy, but also avoids the huge workload of detailed modeling of each sound-transmitting hole 3. Third, the overall acoustic characteristic mapping provides a continuous acoustic description, allowing the system to adapt to spherical screens of different sizes and shapes. Finally, the introduction of the adaptive modulation algorithm enables the system to dynamically respond to environmental changes and different sound field requirements, greatly improving the flexibility and adaptability of the system. This method provides an efficient, accurate and scalable sound field control solution for the spherical screen display system, effectively improving the overall acoustic performance of the system.

[0083] In one embodiment of the present invention, the core formula of the adaptive sound wave modulation algorithm is: ; in, represents the output directional sound wave field, is the angular frequency, and Represent the azimuth and elevation angles respectively, N is the total number of three groups of sound-permeable holes, is the input cancellation sound wave signal, is the weight function of the i-th group of sound-transmitting holes 3, used to adjust the contribution of the group of sound-transmitting holes 3, is the transfer function of the i-th group of sound-transmitting holes 3, describing the propagation characteristics of the sound wave after passing through the group of sound-transmitting holes 3, and Σ represents the sum of the contributions of all groups.

[0084] The calculation formula of the weight function is: ; in, is an amplitude modulation factor, used to adjust the output intensity of the i-th group of sound-transmitting holes 3; is a phase modulation factor used to adjust the output phase of the i-th group of sound-transmitting holes 3; j is an imaginary unit.

[0085] Transfer Function The calculation formula is: ; in, is the frequency response function of the i-th group of sound-permeable holes 3; is the directivity function of the i-th group of sound-permeable holes 3; is the wave number, which is equal to ω / c, where c is the speed of sound; is the distance from the sound source to the center of the i-th group of sound-permeable holes 3; is the angle between the incident sound wave and the normal of the i-th group of sound-transmitting holes 3.

[0086] The algorithm is optimized through iteration To minimize the objective function: ; in, is the desired directional sound field distribution, and ||·|| represents the Euclidean norm.

[0087] The algorithm achieves precise control of the output sound field by dynamically adjusting weights and phases, thereby achieving the purpose of directional propagation of sound waves.

[0088] In one embodiment of the present invention, the sound-transmitting holes 3 on the spherical screen are classified, and similar sound-transmitting holes 3 are grouped together according to geometric parameters and material properties, including: obtaining geometric parameters of each sound-transmitting hole 3, including aperture, depth, shape factor and spatial position coordinates; measuring the material acoustic properties of the area where each sound-transmitting hole 3 is located, including acoustic impedance, sound absorption coefficient and sound wave propagation speed; constructing a multidimensional feature space, each dimension corresponding to a geometric parameter or material property; mapping each sound-transmitting hole 3 to a point in the multidimensional feature space; applying a clustering algorithm to group the points in the feature space; calculating the consistency of the acoustic response of the sound-transmitting holes 3 in each group, and evaluating the classification effect; based on the consistency evaluation result, dynamically adjusting the clustering parameters, and iteratively optimizing the classification result; according to the final clustering result, the sound-transmitting holes 3 with similar geometric parameters and material properties are grouped together to complete the classification.

[0089] Specifically, a 3D laser scanner is used to comprehensively scan the dome display screen to obtain an accurate 3D model of the surface. For example, there are about 10,000 sound-transmitting holes 3 on a hemispherical dome with a diameter of 30 meters. For each sound-transmitting hole 3, an image processing algorithm is used to extract the aperture, depth and shape factor from the scan data. The aperture is usually between 3 and 8 mm, the depth is between 2 and 5 mm, and the shape factor describes the roundness or ellipticity of the hole. At the same time, the spatial position coordinates of each sound-transmitting hole 3 are recorded for subsequent analysis of its distribution characteristics.

[0090] Measuring the material acoustic properties of each sound-permeable hole 3 is a key step to ensure the accuracy of the sound wave propagation model. The acoustic impedance is measured using an acoustic impedance tube, which describes the material's resistance to sound wave propagation. The sound absorption coefficient is measured using the reverberation chamber method, which represents the material's ability to absorb sound energy. The sound wave propagation speed is determined using the pulse-echo method. These measurements are performed on multiple representative areas of the dome surface, and the material properties of each sound-permeable hole 3 are then inferred using interpolation methods.

[0091] Constructing a multidimensional feature space is a preparatory work for data dimensionality reduction and classification. The characteristics of each sound-permeable hole 3 are represented by a multidimensional vector, and each component of the vector corresponds to a geometric parameter or material property. For example, a 7-dimensional feature space can be constructed, including aperture, depth, shape factor, X coordinate, Y coordinate, acoustic impedance and sound absorption coefficient. This representation method converts the physical characteristics of the sound-permeable hole 3 into points in a mathematical space, which is convenient for subsequent data analysis and processing.

[0092] Mapping each acoustic hole 3 to a point in the multidimensional feature space is an important step in data visualization and analysis. For example, a 5 mm diameter, 3 mm deep, circular (shape factor 1), located at coordinates (10, 20, 15), with an acoustic impedance of 4000 , a sound-permeable hole 3 with a sound absorption coefficient of 0.3 can be represented as a point (5, 3, 1, 10, 20, 4000, 0.3) in the feature space. This mapping allows complex physical properties to be compared and analyzed in mathematical space.

[0093] Applying a clustering algorithm to group points in the feature space is the core step of acoustic hole 3 classification. Commonly used algorithms include K-means, DBSCAN, or hierarchical clustering. For example, using the K-means algorithm, 50 categories can be preset, and the algorithm will iteratively group similar points together until the convergence condition is reached. This method can effectively group acoustic holes 3 with similar characteristics into the same group.

[0094] Calculating the consistency of the acoustic response of the sound-transmitting holes 3 in each group is an important means of evaluating the classification effect. Perform acoustic simulation on the sound-transmitting holes 3 in each group and calculate their sound pressure level responses at different frequencies. Then calculate the standard deviation of the response within the group. The smaller the standard deviation, the better the classification effect. For example, if the standard deviation of the sound pressure level response of all the sound-transmitting holes 3 in a group within the range of 100Hz-10kHz is less than 1dB, the consistency of the group is considered to be good.

[0095] Based on the consistency evaluation results, dynamically adjusting clustering parameters and iteratively optimizing classification results are the key to improving classification accuracy. If the consistency of some groups is poor, the number of clusters can be increased or the feature weights can be adjusted. For example, if it is found that the depth parameter has a greater impact on the acoustic response, its weight in the feature vector can be increased. Then rerun the clustering algorithm until the preset consistency threshold or the upper limit of the number of iterations is reached.

[0096] Finally, based on the final clustering results, the sound-transmitting holes 3 with similar geometric parameters and material properties are grouped together to complete the classification. This step maps the clustering results in the mathematical space back to the physical space and assigns a category label to each sound-transmitting hole 3. For example, 10,000 sound-transmitting holes 3 may eventually be divided into 50-100 different categories.

[0097] This classification method of acoustic holes 3 based on multi-dimensional feature analysis and iterative optimization has several advantages. First, it takes into account the comprehensive characteristics of acoustic holes 3, including geometric and material properties, ensuring the comprehensiveness of the classification. Second, the use of multi-dimensional feature space allows complex physical properties to be accurately compared and analyzed mathematically. Third, the application of clustering algorithms enables automated classification, greatly improving efficiency. Finally, iterative optimization based on the consistency of acoustic response ensures the practicality of the classification results, making subsequent sound field control more accurate and efficient. This method provides a solid foundation for the acoustic optimization of the dome display system, which is conducive to more accurate echo suppression and sound field control.

[0098] Please continue reading Figure 2 , monitor the cancellation effect in real time, and correct the real-time acoustic model according to the monitoring results.

[0099] Specifically, this process is achieved through microphone arrays strategically arranged inside the dome. For example, in a hemispherical dome with a diameter of 30 meters, 32 microphone arrays can be evenly distributed, each containing 4 omnidirectional microphones. These microphones continuously collect real-time acoustic data, including sound pressure level, spectral distribution, and sound field directivity. The collected data is processed by fast Fourier transform and converted into frequency domain information. The system compares these measured data with the cancellation effect predicted by the acoustic model and calculates the error value. If the error in a certain area exceeds a preset threshold, such as 3dB, the system triggers the model correction procedure. The correction process uses adaptive filtering algorithms, such as the minimum mean square error (LMS) algorithm, to dynamically adjust the model parameters according to the measured error. This real-time monitoring and correction mechanism has several advantages: first, it can adapt to dynamic changes in the environment, such as changes in temperature, humidity, or personnel distribution; second, it compensates for the deviation between the model prediction and the actual effect, improving the long-term stability of the system; finally, through continuous optimization, the system performance will continue to improve over time, ensuring that the dome display always maintains the best acoustic performance.

[0100] The above describes the echo suppression method of the ball screen display system in the embodiment of the present invention. The following describes the echo suppression device of the ball screen display system in the embodiment of the present invention. Figure 3 An embodiment of the echo suppression device of the ball screen display system in the embodiment of the present invention includes: The model building module 201 is used to obtain the structural data and acoustic parameters of the dome display system and build an initial acoustic model; The model updating module 202 is used to collect real-time acoustic data inside the dome, dynamically update the initial acoustic model, and obtain a real-time acoustic model; A risk mapping module 203, for calculating the sound wave propagation path based on the real-time acoustic model and current audio output characteristics, and generating a dynamic echo risk map identifying potential echo locations, strengths, times and propagation trends; A parameter calculation module 204, used to calculate real-time cancellation sound wave parameters for canceling potential echoes according to the dynamic echo risk map; A sound wave generation module 205, for accurately generating a time-space directional cancellation sound wave in the spherical screen based on the real-time cancellation sound wave parameters; The effect monitoring module 206 is used to monitor the cancellation effect in real time and calibrate the real-time acoustic model according to the monitoring result.

[0101] above Figure 3 The echo suppression device of the mid-spherical screen display system in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The spherical screen display system in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0102] Figure 4 3 is a structural diagram of a spherical screen display system provided by an embodiment of the present invention. The spherical screen display system 300 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage device terminals) storing application programs 333 or data 332. Among them, the memory 320 and the storage medium 330 may be temporary storage or permanent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the spherical screen display system 300. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, and execute a series of instruction operations in the storage medium 330 on the spherical screen display system 300 to implement the steps of the echo suppression method of the spherical screen display system.

[0103] The ball screen display system 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 4 The illustrated spherical screen display system structure does not constitute a limitation on the spherical screen display system provided by the present invention, and may include more or fewer components than illustrated, or a combination of certain components, or a different arrangement of components.

[0104] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the echo suppression method of the spherical screen display system.

[0105] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0106] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0107] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. An echo suppression method for a spherical screen display system, characterized in that: include: Obtain the structural data and acoustic parameters of the dome display system and build an initial acoustic model; Collecting real-time acoustic data inside the dome, dynamically updating the initial acoustic model, and obtaining a real-time acoustic model; Based on the real-time acoustic model and current audio output characteristics, the sound wave propagation path is calculated to generate a dynamic echo risk map that identifies the potential echo location, intensity, time and propagation trend; Calculating real-time cancellation acoustic wave parameters for canceling potential echoes according to the dynamic echo risk map; Based on the real-time cancellation sound wave parameters, accurately generate time-space directional cancellation sound waves in the spherical screen; The cancellation effect is monitored in real time, and the real-time acoustic model is corrected according to the monitoring result.

2. The echo suppression method of the spherical screen display system according to claim 1, characterized in that: The method of calculating the sound wave propagation path based on the real-time acoustic model and the current audio output characteristics and generating a dynamic echo risk map identifying the potential echo position, intensity, time and propagation trend comprises: Discretize the internal space of the dome into multiple acoustic units; For each of the acoustic units, calculating the propagation characteristics of the sound wave in the acoustic unit based on the real-time acoustic model; In combination with the time-frequency characteristics of the current audio output, an echo risk assessment is performed on each of the acoustic units; Based on the spatial relationship between the acoustic units, establishing an acoustic unit association network; Using a graph neural network algorithm, the acoustic unit association network is dynamically analyzed to predict the spatiotemporal evolution of the echo; Based on the prediction results, a dynamic echo risk map including echo location, intensity, time and propagation trend is generated.

3. The echo suppression method of the spherical screen display system according to claim 2, characterized in that: The method of using a graph neural network algorithm to dynamically analyze the acoustic unit association network and predict the spatiotemporal evolution of the echo includes: Construct a graph neural network model, represent each acoustic unit as a node, and represent the sound wave propagation relationship between acoustic units as edges; Assign a feature vector to each node, which contains the acoustic characteristics, current sound pressure level and historical echo data of the acoustic unit, as the input of the graph neural network; Based on the node feature vector, a graph convolution layer is designed and applied to capture the acoustic wave interaction between adjacent acoustic units to obtain updated node features; Using the updated node features, construct and apply a temporal convolution layer to process the temporal characteristics of sound wave propagation and generate a temporally related node representation; Applying an attention mechanism to the time-series related node representations to calculate the importance weight of each acoustic unit in the echo formation process; Based on the importance weights and the time-series related node representations, the echo intensity and phase of each acoustic unit at a future time point are predicted through iterative calculation of a multi-layer graph neural network; The prediction results of each acoustic unit are integrated to generate a prediction of the spatiotemporal evolution of the echo in the entire dome space.

4. The echo suppression method of the spherical screen display system according to claim 2, characterized in that: The method of accurately generating a time-space directional cancellation sound wave in the spherical screen based on the real-time cancellation sound wave parameters includes: Determine the distribution structure of the sound-transmitting holes on the spherical screen display screen, each sound-transmitting hole corresponds to at least one acoustic unit; Based on the acoustic unit association network, a mapping relationship between the sound-permeable holes and the acoustic units is established; Calculating the sound wave characteristics that each sound-permeable hole needs to output according to the real-time sound wave cancellation parameters and the mapping relationship; The acoustic impedance characteristics of the sound-permeable holes are used to modulate the sound waves passing through each sound-permeable hole to achieve directional propagation of the sound waves. By controlling the phase difference of the sound waves between adjacent sound-permeable holes, a controllable sound wave interference structure is formed; Based on the acoustic wave interference structure, a local cancellation sound field is formed at the position of the target acoustic unit; By integrating multiple local cancellation sound fields, the overall space-time directional cancellation sound waves in the spherical screen are constructed.

5. The echo suppression method of the spherical screen display system according to claim 4, characterized in that: The method of utilizing the acoustic impedance characteristics of the sound-transmitting holes to modulate the sound waves passing through each sound-transmitting hole to achieve directional propagation of the sound waves includes: Classify the sound-transmitting holes on the spherical screen and group similar sound-transmitting holes into one group according to geometric parameters and material properties; Establish a representative acoustic impedance model for each group of sound-permeable holes; Based on each group of representative acoustic impedance models, the overall acoustic characteristic mapping of the dome display screen is constructed; Designing an adaptive acoustic wave modulation algorithm using the overall acoustic property mapping; Inputting the to-be-outputted cancelling sound wave signal into the adaptive sound wave modulation algorithm; According to the adaptively modulated signal, the propagation characteristics of the sound waves through each group of sound-permeable holes are controlled; By adjusting the phase and amplitude relationship of the sound waves between different groups of sound-permeable holes, the directional propagation of the sound waves can be achieved.

6. The echo suppression method of the spherical screen display system according to claim 5, characterized in that: The core formula of the adaptive sound wave modulation algorithm is: ； in, represents the output directional sound wave field, is the angular frequency, and Represent the azimuth and elevation angles respectively, N is the total number of sound-permeable hole groups, is the input cancellation sound wave signal, is the weight function of the i-th group of sound-permeable holes, is the transfer function of the i-th group of sound-permeable holes, and Σ represents the sum of the contributions of all groups.

7. The echo suppression method of the spherical screen display system according to claim 5, characterized in that: The sound-transmitting holes on the spherical screen are classified, and similar sound-transmitting holes are grouped together according to geometric parameters and material properties, including: Obtaining geometric parameters of each sound-transmitting hole, including aperture, depth, shape factor and spatial position coordinates; Measure the acoustic properties of the material in the area where each sound-permeable hole is located, including acoustic impedance, sound absorption coefficient and sound wave propagation speed; Construct a multidimensional feature space, where each dimension corresponds to a geometric parameter or material property; Mapping each sound-permeable hole to a point in the multidimensional feature space; Apply clustering algorithms to group points in feature space; Calculate the consistency of acoustic responses of the sound-permeable holes in each group to evaluate the classification effect; Based on the consistency evaluation results, clustering parameters are dynamically adjusted and classification results are iteratively optimized; According to the final clustering results, the sound-permeable holes with similar geometric parameters and material properties are grouped together to complete the classification.

8. The echo suppression method of the spherical screen display system according to claim 1, characterized in that: The real-time acoustic data inside the dome is collected, and the initial acoustic model is dynamically updated to obtain a real-time acoustic model, including: Collect real-time acoustic data through the acoustic sensor network pre-installed in the dome display system; Preprocessing the real-time acoustic data to eliminate noise and extract effective information; Compare the preprocessed real-time acoustic data with the initial acoustic model to identify changes in the acoustic environment; Based on the identified changes, determining acoustic model parameters that need to be updated; Using real-time acoustic data to update and calculate the determined model parameters; Integrate the updated parameters into the initial acoustic model to form a temporary updated model; Verifying the temporarily updated model and evaluating its adaptability to the current acoustic environment; According to the verification results, if the performance of the temporary updated model is better than the current model, the temporary updated model is adopted as the new real-time acoustic model; otherwise, the current model is retained and the reasons for the update failure are recorded for subsequent optimization reference.

9. An echo suppression device for a spherical screen display system, characterized in that: include: Model building module, used to obtain the structural data and acoustic parameters of the dome display system and build an initial acoustic model; A model updating module, used for collecting real-time acoustic data inside the dome, dynamically updating the initial acoustic model, and obtaining a real-time acoustic model; A risk mapping module, for calculating the sound wave propagation path based on the real-time acoustic model and current audio output characteristics, and generating a dynamic echo risk map identifying the potential echo location, intensity, time and propagation trend; A parameter calculation module, used for calculating real-time cancellation sound wave parameters for canceling potential echoes according to the dynamic echo risk map; A sound wave generation module, used for accurately generating a time-space directional cancellation sound wave in the spherical screen based on the real-time cancellation sound wave parameters; The effect monitoring module is used to monitor the cancellation effect in real time and calibrate the real-time acoustic model according to the monitoring result.

10. A ball screen display system, characterized in that: The spherical screen display system comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory to enable the spherical screen display system to execute the steps of the echo suppression method for a spherical screen display system as described in any one of claims 1-8.

Images