Sound field parameter calibration method and sound system

Abstract

This application provides a sound field parameter calibration method and an audio system. In this method, by utilizing sensors and a pre-trained AI model, the operating parameters of the speakers in the audio system can be adjusted without manual operation, thus automating the sound field calibration and significantly improving its efficiency. Furthermore, the AI ​​model can accurately predict target values ​​for sound field response parameters that match the current environment and user location information by learning from preset data. Compared to manual calibration, this reduces deviations caused by individual differences, enabling the system to accurately determine parameters such as sound image direction, sound pressure level, and relative arrival time delay, thereby improving calibration accuracy.

Description

Technical Field

[0001] This application relates to the field of home audio processing technology, and in particular to a sound field parameter calibration method and an audio system. Background Technology

[0002] In the field of audio processing, sound field calibration technology is crucial for improving the sound quality and user experience of audio systems. Traditional sound field calibration methods mainly rely on manual intervention, requiring professionals to perform complex procedures, including repeated adjustments to the physical positions of speakers, measurement and compensation of frequency response characteristics, and manual optimization of environmental acoustic parameters. This manual calibration method is not only time-consuming and labor-intensive, but also has a high technical threshold for ordinary users, making it difficult to widely apply in homes or everyday scenarios.

[0003] Therefore, improving the calibration efficiency of sound field parameters is an urgent problem to be solved. Summary of the Invention

[0004] This application provides a sound field parameter calibration method and an audio system. By utilizing sensors and a pre-trained AI model, the operating parameters of the speakers in the audio system can be adjusted without manual operation, thus automating the sound field calibration and significantly improving its efficiency. The sound field parameter calibration method provided in this application is implemented in the following manner: The method acquires the current data collected by the sensors of an audio system at the current moment; the method is applied to an audio system, which includes multiple speakers and sensors. Based on the currently collected data from the sensor, the target environment features of the user's current environment and the user's target location information in the current environment are obtained. The target location information includes the user's coordinates and orientation. The user's orientation is used to indicate the angle between the plane where the user's face is located and the reference plane. The target environmental features and the target location information are input into the target sound field model to obtain the target values ​​of the sound field response parameters. The target sound field model is obtained by training an initial artificial intelligence (AI) model based on a first training sample. The first training sample includes multiple sample environmental features, multiple sample location information, and multiple reference values ​​of the sound field response parameters. The sound field response parameters include at least one of relative arrival delay, sound image direction, and sound pressure level. Based on the target values ​​of the sound field response parameters, the operating parameters of the plurality of loudspeakers are adjusted, including at least one of gain coefficient, delay compensation, offset angle, and electrical power.

[0005] In the above technical solution, by utilizing sensors and a pre-trained AI model, the operating parameters of the speakers in the audio system can be adjusted without manual operation, thus automating sound field calibration and significantly improving its efficiency. Furthermore, the AI ​​model can accurately predict target values ​​for sound field response parameters that match the current environment and user location information by learning from preset data. Compared to manual calibration, this reduces deviations caused by individual differences, enabling the system to accurately determine parameters such as sound image direction, sound pressure level, and relative arrival delay, thereby improving calibration accuracy.

[0006] In addition, this method takes the environmental characteristics of the user's environment as the input of the model, that is, it takes into account the influence of complex indoor environmental factors on the sound field characteristics. This enables the method to accurately analyze and adjust for different environments, overcoming the limitation of poor adaptability of existing technologies in complex environments.

[0007] In some possible implementations, before acquiring the target environmental features of the user's current environment and the user's target location information in the current environment based on the currently collected data from the sensor, the method further includes: Obtain the historical data collected by the sensor at the previous moment; Based on the historical data and the current data, the user's displacement distance is obtained; The step of obtaining the target environmental features of the user's current environment and the user's target location information in the current environment based on the currently collected data from the sensor includes: When the displacement distance is greater than or equal to a distance threshold, the target environmental features of the user's current environment and the user's target location information in the current environment are obtained based on the currently collected data from the sensor.

[0008] In the above technical solution, by calculating the displacement between the current moment and the previous moment, it can be determined whether the user's position has changed. If the user's position has not changed, the main control speaker can use the operating parameters determined in the previous moment, thereby reducing the amount of calculation and further reducing energy consumption.

[0009] In some possible implementations, after obtaining the user's displacement distance based on the historical data and the current data, the method further includes: If the displacement distance is less than the distance threshold, the currently collected data is discarded.

[0010] Because the sensor collects a large amount of data, it will occupy a lot of storage space. By calculating the displacement between the current moment and the previous moment, it can be determined whether the user's position has changed. If the displacement is less than the distance threshold, it is considered that the user's position has not changed. This means that the data collected at the current moment is similar to the data collected at the previous moment. If the current data is also saved, there will be two duplicate data in the storage space. Therefore, in this case, the current data can be discarded, and in this case, there is no need to adjust the speaker's operating parameters.

[0011] In some possible implementations, the sound field response parameters include sound image direction, the operating parameters include gain coefficient, and adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: A target coordinate system is established based on the user's coordinates and orientation, with the user's coordinates as the origin and the user's orientation as the positive X-axis direction. Based on the user's coordinates and the coordinates of each speaker, determine the angle between each speaker and the positive X-axis direction; Based on the sound image direction, a target loudspeaker group is determined from the plurality of loudspeakers. The target loudspeaker group includes a first loudspeaker and a second loudspeaker. The first loudspeaker has a first angle relative to the positive X-axis direction that is smaller than the sound image direction, and the second loudspeaker has a second angle relative to the positive X-axis direction that is larger than the sound image direction. Based on the sound image direction, the first included angle, the second included angle, and the first calculation formula, determine the first gain coefficient of the first loudspeaker and the second gain coefficient of the second loudspeaker; The first calculation formula is as follows: ; ; in, For the first The gain coefficient of each speaker, For the direction of sound and image, The first included angle, The second included angle is i, which is 1 or 2.

[0012] The above technical solution can dynamically allocate the gain coefficient according to the relative position of the target sound image direction between the two speakers, so that the audio system can provide users with accurately positioned sound images, improve the accuracy of sound image positioning in multi-speaker environments, and make the sound field parameter calibration more refined.

[0013] In some possible implementations, the sound field response parameters further include sound pressure level, and the operating parameters further include electrical power. Adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: The sound power distribution ratio between the first speaker and the second speaker is determined based on the gain coefficient of each speaker; Based on the first conversion formula, the target value of the sound pressure level, and the target value of the reference sound pressure, the target value of the effective sound pressure is determined, wherein the effective sound pressure is the sum of the effective sound pressure of the first loudspeaker and the effective sound pressure of the second loudspeaker; Based on the target value of the effective sound pressure, the sound power distribution ratio, and the second conversion formula, calculate the first sound power of the first loudspeaker and the second sound power of the second loudspeaker; The electrical power of the first loudspeaker is determined by multiplying the first conversion efficiency between the acoustic power and electrical power of the first loudspeaker with the first acoustic power; and the electrical power of the second loudspeaker is determined by multiplying the second conversion efficiency between the acoustic power and electrical power of the second loudspeaker with the second acoustic power. The first conversion formula is as follows: ; in, The effective sound pressure level, The reference sound pressure level is... Sound pressure level; The second conversion formula is as follows: ; Among them, P i Let L be the acoustic power of the i-th loudspeaker. i The effective sound pressure level of the i-th loudspeaker is... Air characteristic impedance; Let i be the distance between the coordinates of the i-th speaker and the coordinates of the user, where i is 1 or 2.

[0014] Through the above technical solution, this application can not only achieve precise positioning of the sound image direction during the sound field calibration process, but also achieve fine-grained control of the sound pressure level perceived by the user. By introducing sound pressure level as a sound field response parameter and electrical power as an operating parameter, and combining the sound power distribution ratio, the conversion formula between sound pressure level and effective sound pressure, and the conversion formula between effective sound pressure and sound power, a complete mapping from the user's desired sound pressure level target to the actual electrical power output of the loudspeakers is achieved. This allows the audio system to dynamically adjust the electrical power of each loudspeaker according to the user's specific needs and environmental conditions, thereby accurately achieving the target sound pressure level at the user's location.

[0015] In some possible implementations, the sound field response parameters further include a relative arrival delay, the operating parameters further include delay compensation, and adjusting the operating parameters of the plurality of loudspeakers according to a target value of the sound field response parameters includes: The first speaker is designated as the reference speaker, and the electrical delay of the first speaker is set to zero; The delay compensation for the electrical delay of the second speaker is determined based on the difference between the first distance and the second distance, and the speed of sound. The first distance is the distance between the first speaker and the user, and the second distance is the distance between the second speaker and the user.

[0016] Through the above technical solution, when calibrating the sound field parameters, not only are spatial and energy parameters such as sound image direction and sound pressure level considered, but also the relative arrival time delay is introduced as a key sound field response parameter, with its target value set to zero. By determining the first loudspeaker as the reference loudspeaker and setting its electrical delay to zero, and by accurately calculating and compensating for the electrical delays of other loudspeakers based on the distance difference between the loudspeaker and the user and the speed of sound, this application can effectively eliminate the time difference in the arrival of sound waves from different loudspeakers at the user's location, avoiding phase distortion caused by inconsistent sound wave arrival times.

[0017] In some possible implementations, the operating parameters further include an offset angle, and adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: For a target speaker, the distance between the target speaker and the user is determined based on the coordinates of the target speaker and the coordinates of the user, as well as the direction vector of the target speaker pointing towards the user; the target speaker is either the first speaker or the second speaker; According to the third conversion formula, the value of the directional attenuation parameter of the target loudspeaker is determined. The directional attenuation parameter is used to indicate the degree of sound pressure attenuation when deviating from the main axis direction of the target loudspeaker. Substituting the value of the directional attenuation parameter into the directional function of the target loudspeaker yields the target value of the deviation angle of the target loudspeaker. The directional function represents the mapping relationship between the deviation angle and the directional attenuation parameter. The deviation angle is related to the pitch angle and the azimuth angle. The third conversion formula is as follows: ; Among them, P i Let L be the acoustic power of the i-th loudspeaker. i The effective sound pressure level of the i-th loudspeaker is... Air characteristic impedance; Let be the distance between the coordinates of the i-th speaker and the coordinates of the user, and D be the directional attenuation parameter of the i-th speaker, where i is 1 or 2.

[0018] By using the above technical solution, the sound pressure attenuation caused by directivity is calculated and compensated based on the user's deviation angle relative to the speaker, so that the sound field calibration result is more consistent with the user's actual listening experience. This can effectively avoid problems such as uneven sound pressure and sound image drift, thereby providing users with a more stable and high-quality listening experience.

[0019] In some possible implementations, adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: The sound field response parameters are input into the parameter adjustment model to obtain the operating parameters of the multiple loudspeakers. The parameter adjustment model is obtained by training the initial adjustment model based on the second training sample. The second training sample includes multiple preset values ​​of the sound field response parameters and multiple reference values ​​of the operating parameters.

[0020] In the above technical solution, adjusting the model's output speaker operating parameters through this parameter adjustment can improve the calculation efficiency of the operating parameters. Calculating the speaker's operating parameters individually is quite cumbersome; calculating them through a parameter adjustment model simplifies the process.

[0021] In some possible implementations, after adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters, the method further includes: Audio is played through the multiple speakers; Obtain the user's satisfaction rating for the audio system; If the satisfaction score is less than the score threshold, the user's feature information is obtained, and the target sound field model is updated based on the user's feature information. The feature information includes the user's age and the user's perceived difference threshold for changes in the sound field response parameters.

[0022] In the above technical solution, by introducing a user feedback mechanism after initial calibration, the sound field calibration process can be dynamically adjusted based on the user's actual auditory experience. When the user is dissatisfied with the calibration effect, personalized characteristic information such as age and sensitivity to sound field response parameters can be actively collected, and this information can be used to update the target sound field model. This iterative optimization process allows the target sound field model to learn and adapt to the specific user's auditory preferences, thereby generating sound field response parameters that better meet the user's expectations in subsequent calibration processes. This effectively solves the problem that traditional calibration methods may not be able to fully meet the user's subjective auditory experience, achieving personalization and adaptability in sound field calibration.

[0023] This application also provides an audio system, which includes a plurality of speakers and sensors, wherein: The sensor is configured to acquire the currently collected data at the current moment; The target loudspeaker includes a control unit and an execution unit. The control unit is configured to acquire, based on the currently acquired data from the sensor, the target environmental features of the user's current environment and the user's target location information in the current environment. The target location information includes the user's coordinates and orientation, and the user's orientation is used to indicate the angle between the plane where the user's face is located and a reference plane. The target environmental features and the target location information are input into the target sound field model to obtain the target values ​​of the sound field response parameters. The target sound field model is obtained by training an initial artificial intelligence (AI) model based on a first training sample. The first training sample includes multiple sample environmental features, multiple sample location information, and multiple reference values ​​of the sound field response parameters. The sound field response parameters include at least one of relative arrival delay, sound image direction, and sound pressure level. Based on the target values ​​of the sound field response parameters, the operating parameters of the execution units of the plurality of loudspeakers are adjusted. The operating parameters include at least one of gain coefficient, delay compensation, offset angle, and electrical power. The execution unit includes an amplifier and / or a motor. The target speaker is one of the plurality of speakers. Attached Figure Description

[0024] Figure 1 This is a structural block diagram of an audio system disclosed in an embodiment of this application; Figure 2 A flowchart illustrating an example of a sound field parameter calibration method disclosed in an embodiment of this application; Figure 3 This is a structural block diagram of a target sound field model disclosed in an embodiment of this application; Figure 4 This is a flowchart illustrating another example of the sound field parameter calibration method disclosed in the embodiments of this application; Figure 5 This is a flowchart illustrating another example of the sound field parameter calibration method disclosed in the embodiments of this application; Figure 6 This is a flowchart illustrating another example of the sound field parameter calibration method disclosed in the embodiments of this application. Detailed Implementation

[0025] The technical solutions in this application will now be described with reference to the accompanying drawings.

[0026] To facilitate a clear description of the technical solutions in the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with essentially the same function and effect. For example, "first instruction" and "second instruction" are used to distinguish different user instructions and do not limit their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0027] It should be noted that, in this application, the words "exemplarily" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplarily" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of words such as "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0028] Furthermore, the terms "comprising" and "having," and any variations thereof, in the embodiments and drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0029] In the field of audio processing, sound field calibration technology is crucial for improving the sound quality and user experience of audio systems. For an example, please refer to... Figure 1 This is a schematic diagram of an example audio system. Figure 1 As shown, the audio system 100 includes a plurality of speakers 110 and at least one sensor 120. Figure 1 Taking multiple sensors 120 as an example, multiple speakers 110 and multiple sensors 120 can be arranged in a surround effect at multiple different locations. When a user is within the surround area, a stereo effect can be provided to the user. The speakers 110 can be audio players or other devices with audio playback functions, and the sensors 120 can be infrared sensors, cameras, or radar devices. Multiple sensors 120 can be sensors of the same type, such as all being cameras, or they can be sensors of different types. For example, some of the sensors 120 may be cameras, and the remaining sensors may be infrared sensors. Examples are not listed here.

[0030] In the audio system 100, multiple speakers 110 and multiple sensors 120 can communicate with each other. Alternatively, only the speaker 110 that acts as the master control device can communicate with the other speakers and sensors. For example, there are three speakers, labeled as speaker A, speaker B, and speaker C. Speaker A can communicate with speaker B and speaker C respectively. For example, speaker A can send the values ​​of its respective operating parameters to speaker B and speaker C, so that speaker B and speaker C operate according to the values ​​of the operating parameters sent by speaker A. In this case, speaker A is the master control speaker in the audio system 100, and speaker B and speaker C cannot communicate. This application does not limit the operating mode of the multiple speakers 110 in the audio system 100.

[0031] Of course, the audio system 100 may also include other devices, such as a central processing unit or input devices, etc., without limitation.

[0032] by Figure 1 Taking the audio system 100 shown as an example in a home theater scenario, since the wall material and furniture placement will affect the sound wave propagation in a home theater scenario, it is necessary to adjust the operating parameters of the audio system in order to provide users with a stereo effect.

[0033] Traditional sound field calibration methods rely heavily on manual intervention, requiring professionals to perform complex procedures, including repeated adjustments to the physical position of speakers, measurement and compensation of frequency response characteristics, and manual optimization of environmental acoustic parameters. This manual calibration method is not only time-consuming and labor-intensive, but also presents a high technical barrier for ordinary users, making it difficult to widely apply in homes or everyday scenarios.

[0034] Therefore, improving the calibration efficiency of sound field parameters is an urgent problem to be solved.

[0035] To address this, a sound field parameter calibration method is proposed for use in an audio system, which includes multiple loudspeakers and sensors. The method includes: Obtain the current data collected by the sensor at the current moment; Based on the currently collected data from the sensor, information about the user's current environment and the user's target location in the current environment is obtained. The target location information includes the user's coordinates and orientation. The user's orientation is used to indicate the angle between the plane where the user's face is located and the reference plane. The current environment and the target location information are input into the target sound field model to obtain the target values ​​of the sound field response parameters. The target sound field model is obtained by training the initial AI model based on the first training samples. The first training samples include multiple sample environments, multiple sample location information, and multiple reference values ​​of the sound field response parameters. The sound field response parameters include at least one of the following parameters: relative arrival delay, sound image direction, and sound pressure level. Based on the target values ​​of the sound field response parameters, the operating parameters of the plurality of loudspeakers are adjusted, including at least one of gain coefficient, delay compensation, offset angle, and electrical power.

[0036] In the above technical solution, the user's current environment and location are first obtained. Then, the environment and location are input into an AI model, which obtains the target values ​​for the sound field parameters. Based on these target values, the speaker's operating parameters are adjusted. This method, utilizing sensors and a pre-trained AI model, can adjust the speaker's operating parameters in an audio system without manual intervention, automating sound field calibration and significantly improving its efficiency. Furthermore, the AI ​​model can accurately predict the target values ​​of the sound field response parameters that match the current environment and user location information by learning from preset data. Compared to manual calibration, this reduces deviations caused by individual differences, enabling the system to accurately determine parameters such as sound image direction, sound pressure level, and relative arrival time delay, thereby improving calibration accuracy.

[0037] The sound field parameter calibration method provided in some embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0038] like Figure 2 The diagram shows a flowchart of a sound field parameter calibration method provided in some embodiments of this application. This method is applied to, for example... Figure 1 In the aforementioned audio system 100, the method includes the following steps: Step S201: Obtain the current data collected by the sensor at the current moment.

[0039] The sensor 120 can periodically collect data according to a preset sampling period, and the current time is any sampling time of the sensor 120.

[0040] If multiple sensors 120 are of the same type, they collect the same type of data. For example, if multiple sensors 120 are all cameras, and these cameras can be set at different locations in the current environment, they can acquire multiple images from different angles. If multiple sensors 120 are of different types, they collect different types of data. For example, if multiple sensors 120 include cameras and millimeter-wave radar, they can acquire image data and point cloud data. Examples are not listed here.

[0041] It is understood that the method in this embodiment can be executed by a speaker in the audio system 100. For example, if multiple speakers 110 in the audio system 100 can communicate with each other, the method in this embodiment can be executed by any one of the speakers 110 in the audio system 100; if only the master speaker in the audio system 100 can communicate with other speakers, the method can be executed by the master speaker in the audio system 100. Alternatively, when the audio system 100 also includes other devices, such as a central processing unit, the method can also be executed by the central processing unit in the audio system 100, without limitation. For ease of explanation, Figure 1 Taking speaker B in the audio system 100 as the main control speaker as an example, the following explanation will use the main control speaker in the audio system 100 as the execution subject.

[0042] As an example, once the sensor 120 has collected data, it can send the collected data to the speaker 110.

[0043] Step S202: Based on the currently collected data from the sensor, obtain the target environment features of the user's current environment and the user's target location information in the current environment.

[0044] This method uses data collected by sensors in the audio system to obtain the target environmental features of the user's current environment and the user's target location information in that environment. The target location information includes the user's coordinates and orientation, which are used to describe the user's listening position in the physical space of the current environment. The user's coordinates can be understood as the user's two-dimensional or three-dimensional coordinates in the current environment. The user's orientation can be understood as the angle between the user's face and a reference surface. The reference surface used to mark the user's orientation can be the surface of an object in the current environment. For example, the plane where the television is located in the living room can be used as the reference surface. The user's orientation can be facing the television or having their back to the television, etc. Alternatively, the reference surface can also be an absolute direction. For example, the user's orientation can be facing north or southeast, etc., without any restrictions.

[0045] The user's current environment refers to the physical space in which the user is located when performing sound field parameter calibration. The physical space may include various factors that affect sound wave transmission, such as size, shape, wall material, furniture layout, etc. Wall material and furniture layout will affect the emission and attenuation of sound waves. Therefore, the target environment characteristics can include the attribute characteristics of the physical space itself, such as size and shape characteristics, as well as characteristics used to characterize sound wave transmission in the physical space, such as reverberation time, sound wave reflection characteristics, or sound wave attenuation characteristics.

[0046] As an example, when the audio system 100 includes multiple sensors 120, to improve the accuracy of the acquired environmental features and location information, the data collected by the multiple sensors 120 can be preprocessed, for example, by filtering or noise reduction. Then, the preprocessed data can be used to acquire the target environmental features and target location information. As another example, if the multiple sensors 120 are of different types, such as including cameras and radar, then the data collected by the different types of sensors 120 can be normalized.

[0047] In some embodiments, the user's coordinates, orientation, size and shape characteristics of the target environment, etc., can be acquired through data collected by a camera or radar. Sound wave reflection or attenuation characteristics can be acquired through an ultrasonic sensor. For example, ultrasonic sensors at different locations sequentially emit a specific pulse, and then acquire the echo signal corresponding to each specific pulse. The reverberation time, sound wave reflection characteristics, or sound wave attenuation characteristics of the current environment are determined based on the signal amplitude and direction of arrival of these multiple echo signals. The specific processing method is similar to that in related technologies and will not be elaborated further here.

[0048] Step S203: Input the target environment features and the target location information into the target sound field model to obtain the target values ​​of the sound field response parameters.

[0049] The sound field response parameters include at least one of the following: relative arrival delay, sound image direction, and sound pressure level; relative arrival delay is the relative time delay for the sound to reach the user's ear; sound image direction is the perceived direction of the sound in space; and sound pressure level is the loudness of the sound.

[0050] The target sound field model is obtained by training an initial artificial intelligence (AI) model using first training samples. The AI ​​model can be a neural network, which theoretically can approximate any continuous function, thus enabling it to learn arbitrary mappings. The neural networks mentioned in this application can include various types, such as deep neural networks (DNN), convolutional neural networks (CNN), recurrent neural networks (RNN), residual networks, neural networks using the transformer model, or other neural networks.

[0051] like Figure 3 As shown, the convolutional neural network (CNN) 200 may include an input layer 210, convolutional / pooling layers 220 (where pooling layers are optional), and a neural network layer 230. Figure 3 The convolutional / pooling layer 220 shown may include layers such as 221 and 226 in examples. In one implementation, layer 221 is a convolutional layer, layer 222 is a pooling layer, layer 223 is a convolutional layer, layer 224 is a pooling layer, layer 225 is a convolutional layer, and layer 226 is a pooling layer; in another implementation, layers 221 and 222 are convolutional layers, layer 223 is a pooling layer, layers 224 and 225 are convolutional layers, and layer 226 is a pooling layer. That is, the output of the convolutional layer can be used as the input of a subsequent pooling layer, or as the input of another convolutional layer to continue the convolution operation. Taking convolutional layer 221 as an example, convolutional layer 221 may include many convolution operators, which are also called convolution kernels. A convolution operator can essentially be a weight matrix, which is usually predefined. Taking image processing as an example, different weight matrices extract different features from the image. The weight values ​​in these weight matrices need to be obtained through a lot of training in practical applications. The weight matrices formed by the weight values ​​obtained through training can extract information from the input data, thereby helping the convolutional neural network 200 to make correct predictions.

[0052] After processing by the convolutional / pooling layers 220, the convolutional neural network 200 is still insufficient to output the required information. As mentioned earlier, the convolutional / pooling layers 220 only extract features and reduce the parameters introduced by the input image. However, to generate the final output information (the required class information or other relevant information), the convolutional neural network 200 needs to utilize neural network layers 230 to generate one or a set of required class numbers of output. Therefore, neural network layers 230 can include multiple hidden layers (such as...). Figure 3As shown in layers 231, 232 to 23n) and output layer 240, the parameters contained in these multi-layer hidden layers can be pre-trained based on relevant training data for specific task types, such as image recognition, image classification, image super-resolution reconstruction, etc. After the multiple hidden layers in neural network layer 230, which is the last layer of the entire convolutional neural network 200, is the output layer 240. The output layer 240 has a loss function similar to the classification cross-entropy, which is specifically used to calculate the prediction error. Once the forward propagation of the entire convolutional neural network 200 is completed, the backpropagation will begin to update the weight values ​​and biases of the aforementioned layers in order to reduce the loss of the convolutional neural network 200 and the error between the result output by the convolutional neural network 200 through the output layer and the ideal result.

[0053] The first training sample is used to train the AI ​​model. The first training sample includes multiple sample environment features, multiple sample location information, and multiple reference values ​​for the sound field response parameters. Each first training sample includes multiple sets of data. Each set of data includes at least one environmental feature of a preset environment, one sample location information, and multiple reference values. These multiple reference values ​​are the values ​​of the sound field response parameters corresponding to an audio signal that meets preset conditions in the preset environment and at the sample location. The preset conditions may be: distortion rate less than or equal to a distortion rate threshold; deviation of the sound image direction from the sound image direction set during recording less than or equal to a preset angle; reverberation duration less than or equal to a duration threshold, etc. The audio signal that meets the preset conditions can be understood as the ideal audio signal in the preset environment and at the sample location, or as the audio signal that provides the best listening experience in the preset environment and at the sample location. These multiple reference values ​​can be obtained offline using room acoustic modeling software based on ray tracing or wave acoustic methods, or by placing speakers and microphone arrays in a real room, playing and recording signals reflected by the room, and then analyzing the results. These methods will not be elaborated upon here.

[0054] During model training, a loss function can be defined. The loss function describes the difference between the model's output value and the target output value. This application does not limit the specific form of the loss function. The model training process involves adjusting the model parameters to make the loss function value less than a threshold, or to make the loss function value meet the target requirements. When the difference between the AI ​​model's output value for each sample in the first training sample and the corresponding reference value is less than the threshold, it indicates that the AI ​​model training is complete, i.e., the target sound field model is obtained.

[0055] After acquiring the target environmental features and target location information of the user's environment, these features and location information are input into the target sound field model. The output of the target sound field model is the target value of the sound field response parameter. The specific sound field response parameter value output by the target sound field model is determined by the reference values ​​in the first training sample. For example, if each set of data in the first training sample includes three sound field response parameter values, then the output of the target sound field model is the target value of these three sound field response parameters; if each set of data in the first training sample includes only one sound field response parameter value, then the output of the target sound field model is the target value of that single sound field response parameter. No restriction is imposed here.

[0056] Step S204: Adjust the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters.

[0057] Once the target values ​​for the sound field response parameters are obtained, the operating parameters of multiple loudspeakers can be adjusted based on these target values. These operating parameters include at least one of the following: gain coefficient, delay compensation, offset angle, and electrical power. As an example, adjusting the operating parameters of multiple loudspeakers can include adjusting the operating parameters of each individual loudspeaker, or it can include adjusting the operating parameters of some loudspeakers. For example, first, select relevant loudspeakers, adjust the operating parameters of those relevant loudspeakers, and then turn off the remaining loudspeakers.

[0058] In some embodiments, when the sound field response parameter is the sound image direction, the gain coefficient of the loudspeaker can be adjusted according to the target value of the sound image direction; in some embodiments, when the sound field response parameter is the sound pressure level, the power and / or offset angle of the loudspeaker can be adjusted according to the target value of the sound pressure level; in some embodiments, when the sound field response parameter is the relative arrival delay, the delay compensation of the loudspeaker can be adjusted according to the target value of the arrival delay. The above sound field response parameters can also be used in combination, which will not be described one by one here.

[0059] Once the main control speaker determines the operating parameters of each speaker, it can send these parameters to the corresponding speakers, which then operate according to those parameters. It's important to note that each speaker may include a control unit and an execution unit. The control unit can be understood as a controller, and the execution unit as a motor or amplifier, etc. The speaker's control unit executes the above process, obtaining the operating parameters of each execution unit, and then controls the execution units to operate according to those parameters. For example, the gain coefficient can be understood as an amplifier's operating parameter, and the offset angle and electrical power can be understood as a motor's operating parameters, etc. The motor can be a synchronous motor or an asynchronous motor, etc., without limitation.

[0060] The following example will provide a more detailed explanation of the above technical solution: Imagine a home environment where user A is watching a movie through a sound system in their living room. This sound system is equipped with multiple speakers and a sensor array.

[0061] First, the sensor acquires information about user A's current environment and target location within that environment. For example, multiple wide-angle cameras in the sensor array continuously monitor the living room area, obtaining images of user A from multiple angles and sending them to the main speaker. The main speaker uses image processing technology to identify user A's three-dimensional coordinates (X1, Y1, Z1) and their orientation angle, confirming they are facing the television. Simultaneously, the sensor array also detects echo signals in the living room by emitting specific audio signals and sends these echo signals to the main speaker. The main speaker extracts features from the echo signals to obtain environmental characteristics of the living room, such as size, shape, reverberation time at three-dimensional coordinates (X1, Y1, Z1), and sound wave attenuation.

[0062] Then, the main control speaker vectorizes the three-dimensional coordinates (X1, Y1, Z1) of user A, the orientation angle facing the TV, the size and shape characteristics of the living room, the reverberation time at the three-dimensional coordinates (X1, Y1, Z1), and the sound wave attenuation characteristics to obtain a feature vector. The feature vector is then input into the target sound field model to obtain the output result of the target sound field model. For example, the sound image direction should be 30 degrees to the right, the sound pressure level should be 75dB, and the sound emitted by each speaker should have a relative arrival delay of 0ms at user A.

[0063] Finally, the main control loudspeaker adjusts each speaker in the audio system in real time based on the target values ​​of the sound field response parameters output by the model. For example, to achieve a sound image direction of 30 degrees to the right, the gain coefficient of the right speaker is adjusted to be slightly higher than that of the left speaker according to the preset gain allocation rules. To achieve a sound pressure level of 75dB, the electrical power output of all speakers is adjusted. Since the target relative arrival delay is 0ms, the necessary electrical delay compensation is calculated based on the distance of each speaker from user A and the speed of sound to ensure that the sound from all speakers reaches user A's listening position simultaneously. Through these adjustments, the audio system can provide user A with an adjusted sound field experience that matches their position and orientation.

[0064] In the above technical solution, by utilizing sensors and a pre-trained AI model, the operating parameters of the speakers in the audio system can be adjusted without manual operation, thus automating sound field calibration and significantly improving its efficiency. Furthermore, the AI ​​model can accurately predict target values ​​for sound field response parameters that match the current environment and user location information by learning from preset data. Compared to manual calibration, this reduces deviations caused by individual differences, enabling the system to accurately determine parameters such as sound image direction, sound pressure level, and relative arrival delay, thereby improving calibration accuracy.

[0065] In addition, this method takes the environmental characteristics of the user's environment as the input of the model, that is, it takes into account the influence of complex indoor environmental factors on the sound field characteristics. This enables the method to accurately analyze and adjust for different environments, overcoming the limitation of poor adaptability of existing technologies in complex environments.

[0066] In the above-described embodiments of this application, a target value for the sound field response parameters is obtained through a target AI model, and the operating parameters of the loudspeakers are adjusted based on these values. However, when the sound field response parameters include the sound image direction, how to accurately map the target value of the sound image direction to the gain coefficients of the multiple loudspeakers to achieve accurate perception of the sound image in a specific direction by the user is a challenge. To address this, this application further proposes a control scheme for the sound image direction.

[0067] The sound field response parameters include the sound image direction, the operating parameters include the gain coefficient, and adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: Step 11: Establish a target coordinate system based on the user's coordinates and orientation. The target coordinate system has the user's coordinates as its origin and the user's orientation as the positive X-axis. This target coordinate system aims to provide a user-centered reference frame for subsequent acoustic localization calculations. It can be a two-dimensional planar coordinate system or a three-dimensional spatial coordinate system, depending on the system's requirements for acoustic localization accuracy.

[0068] Step 12: Based on the user's coordinates and the coordinates of each speaker, determine the angle between each speaker and the positive X-axis. This is achieved through geometric calculations, such as using the arctangent function based on the speaker and user coordinates, thereby quantifying the geometric position of each speaker relative to the user's perceived direction and clarifying the relative orientation of the speakers within the user's listening area.

[0069] Step 13: Determine the target loudspeaker group from the plurality of loudspeakers according to the sound image direction.

[0070] The target loudspeaker group includes a first loudspeaker and a second loudspeaker. The first loudspeaker has a first angle relative to the positive X-axis direction smaller than the sound image direction, and the second loudspeaker has a second angle relative to the positive X-axis direction larger than the sound image direction. The target loudspeaker group can be determined by iterating through all loudspeakers and finding two loudspeakers whose angles are respectively smaller than and larger than the target sound image direction. The loudspeakers in the target loudspeaker group can be adjacent or non-adjacent. The number of first loudspeakers and second loudspeakers is unlimited; there can be one or more. If there are multiple first loudspeakers and second loudspeakers, then any first loudspeaker has a first angle relative to the positive X-axis direction smaller than the sound image direction, and any second loudspeaker has a second angle relative to the positive X-axis direction larger than the sound image direction.

[0071] Step 14: Determine the first gain coefficient of the first loudspeaker and the second gain coefficient of the second loudspeaker based on the sound image direction, the first included angle, the second included angle, and the first calculation formula; The first calculation formula, denoted as formula (1) and formula (2), is shown below: (1) (2) in, For the first The gain coefficient of each speaker, For the direction of sound and image, The first included angle, The second included angle is i, which is 1 or 2.

[0072] For example, suppose the audio system includes three speakers: speaker A, speaker B, and speaker C. Speaker A is the main speaker, and the user is located in the center of the room, facing forward. Speaker A first obtains the user's precise coordinates and orientation information through sensors. For example, the user's coordinates are (0,0), and the orientation is the positive X-axis. Based on this, a target coordinate system is established. Speaker A is located to the user's left front, with coordinates (-1, 1); speaker B is located to the user's right front, with coordinates (1, 1); and speaker C is located to the user's right rear, with coordinates (1, -1).

[0073] Speaker A calculates the angle between each speaker and the positive X-axis of the user's coordinate system. For example, speaker A has an angle θ_A of 135 degrees relative to the positive X-axis, speaker B has an angle θ_B of 45 degrees relative to the positive X-axis, and speaker C has an angle θ_C of 315 degrees relative to the positive X-axis. If the target sound image direction θ_"target" is 60 degrees, since the angles between loudspeaker A and loudspeaker C and the positive X-axis are both greater than the target sound image direction, then loudspeaker A and loudspeaker B can be selected from loudspeaker A and loudspeaker C to form the target loudspeaker group. For example, if loudspeaker A and loudspeaker B are selected to form the target loudspeaker group, substituting the corresponding angles of loudspeaker A and loudspeaker B into the first calculation formula, we get g_B = sin(135-60) / sin(135-45) = sin(75) / sin(90) ≈ 0.966 / 1 = 0.966; g_A = sin(60-45) / sin(135-45) = sin(15) / sin(90) ≈ 0.259 / 1 = 0.259. Therefore, the gain coefficient g_B of loudspeaker B is determined to be 0.966, the gain coefficient g_A of loudspeaker A is determined to be 0.259, and loudspeaker C is in the off state. Then speaker A adjusts its gain coefficient and sends corresponding control commands to speaker B and speaker C respectively, so that speaker B and speaker C work in the corresponding states.

[0074] The above technical solution can dynamically allocate the gain coefficient according to the relative position of the target sound image direction between the two speakers, so that the audio system can provide users with accurately positioned sound images, improve the accuracy of sound image positioning in multi-speaker environments, and make the sound field parameter calibration more refined.

[0075] However, in practical applications, simply adjusting parameters such as the gain coefficient to control the sound image direction may not be sufficient to accurately control the overall sound pressure level perceived by the user, especially when multiple speakers work together. How to finely distribute the sound power to achieve the expected sound pressure level and convert it into operable electrical power output is a challenge. In this regard, this application further proposes a precise control scheme for electrical power.

[0076] The sound field response parameters also include sound pressure level, and the operating parameters also include electrical power. Adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: Step 21: Determine the sound power distribution ratio between the first speaker and the second speaker based on the gain coefficient of each speaker; The gain coefficient here can be the gain coefficient calculated according to the previous example, or it can be the gain coefficient preset by the speaker; there is no restriction here. Sound pressure level (SPL) represents the intensity of sound and can reflect the loudness of sound perceived by a user at a specific location. Electrical power refers to the amount of electrical energy required to drive the speaker to produce sound, which directly affects the speaker's sound power output. By precisely controlling the electrical power, fine-tuning of the SPL can be achieved.

[0077] Step 22: Determine the target value of the effective sound pressure based on the first conversion formula, the target value of the sound pressure level, and the target value of the reference sound pressure.

[0078] The first conversion formula, denoted as formula (3), is as follows: (3) in, The effective sound pressure level, The reference sound pressure level is... Sound pressure level; The effective sound pressure is the sum of the effective sound pressure of the first loudspeaker and the effective sound pressure of the second loudspeaker. The reference sound pressure is preset and is usually taken as 20 μPa = 2 × 10^(-5) Pa.

[0079] Step 23: Calculate the first sound power of the first loudspeaker and the second sound power of the second loudspeaker based on the target value of the effective sound pressure, the sound power distribution ratio, and the second conversion formula; The second conversion formula is labeled as formula (4) as follows: (4) Among them, P i Let L be the acoustic power of the i-th loudspeaker. i The effective sound pressure level of the i-th loudspeaker is... Air characteristic impedance; Let i be the distance between the coordinates of the i-th speaker and the coordinates of the user, where i is 1 or 2.

[0080] First, based on the sound power distribution ratio, for example, if the sound power distribution ratio between speaker A and speaker B is 1:2, then the sound power of speaker A can be represented by x, and the sound power of speaker B by 2x. Substituting these values ​​into the second conversion formula yields the effective sound pressure level of each speaker. Since the target value of the effective sound pressure level is the sum of the effective sound pressure levels of the two speakers, and the sum of the effective sound pressure levels of the speakers is known, the value of x can be calculated, thus obtaining the first sound power of speaker A and the second sound power of speaker B.

[0081] Step 24: Determine the electrical power of the first loudspeaker by multiplying the first conversion efficiency between the acoustic power and electrical power of the first loudspeaker with the first acoustic power; and determine the electrical power of the second loudspeaker by multiplying the second conversion efficiency between the acoustic power and electrical power of the second loudspeaker with the second acoustic power. The conversion efficiency between the acoustic power and electrical power of each loudspeaker is a known parameter. If the first loudspeaker and the second loudspeaker have the same model, then the conversion efficiency between the acoustic power and electrical power of the first loudspeaker and the second loudspeaker will be the same. If the first loudspeaker and the second loudspeaker have different models, then the conversion efficiency between the acoustic power and electrical power of the first loudspeaker and the second loudspeaker may also be different. This is not a limitation.

[0082] The following is a specific example to illustrate this.

[0083] Assuming the target sound pressure level of the audio system is 75 dB, and based on the adjustment of the sound image direction, the gain coefficients of the first and second loudspeakers are 0.8 and 0.6, respectively. The main control loudspeaker can then determine the sound power distribution ratio based on these gain coefficients, for example, through normalization. For instance, the sound power distribution ratio of the first loudspeaker can be set to 0.8 / (0.8+0.6) ≈ 0.57, and the sound power distribution ratio of the second loudspeaker to 0.6 / (0.8+0.6) ≈ 0.43. Next, using the first conversion formula and the target sound pressure level L_p of 75 dB, the total effective sound pressure can be calculated as 20 × 10^(75 / 20) μPa. Then, based on the target total effective sound pressure value and the previously determined sound power distribution ratio, combined with the second conversion formula, the required sound powers P_1 and P_2 for the first and second loudspeakers, respectively, can be calculated. For example, if the distance d_1 between the first speaker and the user is 2 meters, the distance d_2 between the second speaker and the user is 2.5 meters, and the air characteristic impedance ρc is 413 Pa·s / m, then P_1 and P_2 can be solved iteratively or by solving simultaneous equations based on the distribution ratio and the total effective sound pressure. Finally, assuming the electroacoustic conversion efficiency of the first speaker is 0.01 (i.e., 1%) and the electroacoustic conversion efficiency of the second speaker is 0.008 (i.e., 0.8%), then dividing the calculated P_1 by 0.01 gives the electrical power required by the first speaker, and dividing P_2 by 0.008 gives the electrical power required by the second speaker. These electrical power values ​​are then sent to the speaker amplifier to drive the speakers to emit sound of corresponding intensity.

[0084] Through the above technical solution, this application can not only achieve precise positioning of the sound image direction during the sound field calibration process, but also achieve fine-grained control of the sound pressure level perceived by the user. By introducing sound pressure level as a sound field response parameter and electrical power as an operating parameter, and combining the sound power distribution ratio, the conversion formula between sound pressure level and effective sound pressure, and the conversion formula between effective sound pressure and sound power, a complete mapping from the user's desired sound pressure level target to the actual electrical power output of the loudspeakers is achieved. This allows the audio system to dynamically adjust the electrical power of each loudspeaker according to the user's specific needs and environmental conditions, thereby accurately achieving the target sound pressure level at the user's location.

[0085] In some embodiments, if the sound emitted by different speakers arrives at the user's location at different times, it may cause problems such as blurred sound images and phase distortion, affecting the accuracy of sound field calibration and user experience. In this regard, this application further proposes a precise control scheme for relative arrival delay.

[0086] The sound field response parameters also include relative arrival delay, and the operating parameters also include delay compensation. Adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: Step 31: Determine the first speaker as the reference speaker and set the electrical delay of the first speaker to zero; Step 32: Determine the delay compensation of the electrical delay of the second speaker based on the difference between the first distance and the second distance, and the speed of sound. The first distance is the distance between the first speaker and the user, and the second distance is the distance between the second speaker and the user.

[0087] The relative arrival delay refers to the time difference between sound waves emitted by different speakers reaching the user's listening position. In a multi-speaker system, the arrival time of sound waves varies depending on the distance between the speakers and the user. Assume the audio system contains multiple speakers, and a target coordinate system has been established based on the user's coordinates and orientation, defining a target speaker group containing a first speaker and a second speaker. To ensure simultaneous sound arrival at the user's position, the first speaker is designated as the reference speaker, and its electrical delay is set to zero. The distance between the first speaker and the user (e.g., a first distance of 2.5 meters) and the distance between the second speaker and the user (e.g., a second distance of 3.2 meters) are then measured or calculated. Assuming the speed of sound is 343 meters per second, the difference between the first and second distances is 3.2 meters - 2.5 meters = 0.7 meters. Therefore, the electrical delay that the second speaker needs to compensate for is 0.7 meters / 343 meters per second ≈ 0.00204 seconds (i.e., 2.04 milliseconds). The main speaker applies this 2.04-millisecond electrical delay to the audio signal path of the second speaker. In this way, even though the second speaker is farther from the user than the first speaker, by introducing a precise electrical delay, it is ensured that the sound emitted from both speakers can reach the user's location simultaneously, thus avoiding the problem of inconsistent sound wave arrival times caused by distance differences.

[0088] Through the above technical solution, when calibrating the sound field parameters, not only are spatial and energy parameters such as sound image direction and sound pressure level considered, but also the relative arrival time delay is introduced as a key sound field response parameter, with its target value set to zero. By determining the first loudspeaker as the reference loudspeaker and setting its electrical delay to zero, and by accurately calculating and compensating for the electrical delays of other loudspeakers based on the distance difference between the loudspeaker and the user and the speed of sound, this application can effectively eliminate the time difference in the arrival of sound waves from different loudspeakers at the user's location, avoiding phase distortion caused by inconsistent sound wave arrival times.

[0089] Since the sound field radiation of a loudspeaker is directional, the sound pressure will attenuate when the user deviates from the direction of the loudspeaker's main axis. This may cause the calibration results of the sound field parameters to deviate from the user's actual hearing. In response, this application further proposes a control scheme for the deviation angle.

[0090] The operating parameters also include the deviation angle, and adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: Step 41: For the target speaker, determine the distance between the target speaker and the user, and the direction vector of the target speaker pointing towards the user, based on the coordinates of the target speaker and the coordinates of the user; the target speaker is either the first speaker or the second speaker; The target loudspeaker can be any one of multiple loudspeakers, or any one of the target loudspeaker groups.

[0091] Step 42: Determine the value of the directional attenuation parameter of the target loudspeaker according to the third conversion formula. The directional attenuation parameter is used to indicate the degree of sound pressure attenuation when deviating from the main axis direction of the target loudspeaker. The third conversion formula is denoted as formula (5) as follows: (5) Among them, P i Let L be the acoustic power of the i-th loudspeaker. i The effective sound pressure level of the i-th loudspeaker is... Air characteristic impedance; Let be the distance between the coordinates of the i-th speaker and the coordinates of the user, and D be the directional attenuation parameter of the i-th speaker, where i is 1 or 2.

[0092] The sound power and effective sound pressure of the loudspeaker can be calculated using the method described in the previous example. Once the effective sound pressure and sound power of each loudspeaker are obtained, they can be substituted into formula (5) to calculate the value of the directivity attenuation parameter of each loudspeaker.

[0093] Step 43: Substitute the value of the directional attenuation parameter into the directional function of the target loudspeaker to obtain the target value of the deviation angle of the target loudspeaker. The directional function is used to represent the mapping relationship between the deviation angle and the directional attenuation parameter. The deviation angle is related to the pitch angle and the azimuth angle.

[0094] The directivity function is a function that represents the mapping relationship between the offset angle and the directivity attenuation parameter. This function describes the sound pressure attenuation law of the loudspeaker at different offset angles. Its function is to calculate the corresponding directivity attenuation based on the offset angle. The directivity function of each loudspeaker can be obtained by fitting the measurement data through a mathematical model (such as Gaussian function, cosine function, etc.) and stored in the main control loudspeaker. Thus, after obtaining the value of the directivity attenuation parameter, the main control loudspeaker directly substitutes this value into the directivity function of the corresponding loudspeaker to obtain the target value of the offset angle of the corresponding loudspeaker. The directivity functions of different loudspeaker models may be different, while the directivity functions of loudspeakers with the same signal may be the same; this is not a limitation.

[0095] Suppose the target speaker group requiring parameter adjustment includes a first speaker and a second speaker. First, for the first speaker, the main control speaker calculates the distance between the first speaker and the user based on its installation coordinates and the user's current coordinates, determining the direction vector of the first speaker pointing towards the user. For example, if the main axis of the first speaker is directly forward, and the user is located at a 30-degree angle to its right, the deviation angle is 30 degrees. Next, using pre-stored effective sound pressure level and sound power of the first speaker, combined with a third conversion formula, the directivity attenuation parameter D at this 30-degree deviation angle is calculated. This directivity attenuation parameter D is then substituted into the directivity function of the first speaker, which may be a lookup table based on measurement data or a mathematical model (such as cos(α)^n), to obtain the target deviation angle value of the first speaker at the current deviation angle. This target deviation angle value will guide the system on how to adjust the output of the first speaker to compensate for the sound pressure attenuation caused by deviating from the main axis direction. A similar process is performed for the second speaker.

[0096] By using the above technical solution, the sound pressure attenuation caused by directivity is calculated and compensated based on the user's deviation angle relative to the speaker, so that the sound field calibration result is more consistent with the user's actual listening experience. This can effectively avoid problems such as uneven sound pressure and sound image drift, thereby providing users with a more stable and high-quality listening experience.

[0097] In the above embodiments, the operating parameters of the loudspeakers need to be calculated separately, which is a rather cumbersome process. To simplify the process, a parameter adjustment model can be set in the main loudspeaker. This parameter adjustment model is obtained by training an initial adjustment model based on a second training sample, which includes multiple preset values ​​of the sound field response parameters and multiple reference values ​​of the operating parameters. In this way, the operating parameters of the loudspeakers can be output through this parameter adjustment model. For example, by inputting the sound field response parameters into the parameter adjustment model, the operating parameters of the multiple loudspeakers can be obtained, thus improving the calculation efficiency of the operating parameters.

[0098] The specific form and training method of the parameter adjustment model are similar to those of the aforementioned target sound field model, and will not be repeated here. The second training sample will be described below. This sample contains pairs of data points, each pair consisting of a set of preset sound field response parameter values ​​and a corresponding set of reference operating parameter values. These preset and reference values ​​can be obtained in various ways, such as through high-precision acoustic simulation, extensive experimental measurements in a controlled environment, or manual calibration and recording by experienced acoustic engineers. The quality and quantity of the second training sample directly affect the learning effect and final adjustment accuracy of the parameter adjustment model.

[0099] In the above embodiments, the operating parameters of each speaker are calculated using the target sound field model in the main speaker, thereby ensuring that each speaker operates with appropriate parameters and providing the user with a better audiovisual experience. However, the calculation of the target sound field model consumes electrical energy. Therefore, to reduce energy consumption, flowcharts of another example of sound field parameter calibration are provided in some embodiments of this application, such as... Figure 4 As shown.

[0100] exist Figure 4 In this process, the sound field parameter calibration method includes the following steps: Step S401: Obtain the current data collected by the sensor at the current moment.

[0101] Step S401 is similar to step S201, and will not be described again here.

[0102] Step S402: Obtain the historical data collected by the sensor at the previous moment; The main speaker can store the data collected by the sensor, so the main speaker can read the historical data collected at the previous moment from the memory.

[0103] Step S403: Obtain the user's displacement distance based on the historical data and the current data. Specifically, the displacement distance of the user at the current moment compared to the previous moment can be calculated using the Euclidean distance formula, which will not be elaborated here.

[0104] Step S404: If the displacement distance is greater than or equal to the distance threshold, based on the current data collected by the sensor, obtain the target environment features of the current environment in which the user is located and the target location information of the user in the current environment.

[0105] Step S405: Input the target environment features and the target location information into the target sound field model to obtain the target values ​​of the sound field response parameters.

[0106] Step S406: Adjust the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters.

[0107] Steps S404 to S406 are similar to steps 202 to 204, and will not be described again here.

[0108] Understandably, since the sensor can periodically collect data according to a preset cycle, upon receiving the data uploaded by the sensor at any given time, the operating parameters of multiple speakers can be determined based on the uploaded data, and the collected data and speaker operating parameters can be saved. By calculating the displacement between the current moment and the previous moment, it can be determined whether the user's position has changed. If the displacement is greater than or equal to a distance threshold, the user's position can be considered to have changed, and the speaker operating parameters need to be recalculated based on the user's current position. Alternatively, the angle between the user's orientation at the previous moment and the orientation at the current moment can be determined. If the angle is greater than a preset angle, the calculation of the operating parameters can also be retried.

[0109] In this way, if the user's position does not change, the main control speaker can use the operating parameters determined at the previous moment, thereby reducing the amount of calculation and further reducing energy consumption.

[0110] Because the amount of data collected by the sensor is relatively large, it will occupy a lot of storage space. To reduce the storage space occupied, a flowchart of another example of sound field parameter calibration is provided in some embodiments of this application, such as... Figure 5 As shown.

[0111] exist Figure 5 In this process, the sound field parameter calibration method includes the following steps: Step S501: Obtain the current data collected by the sensor at the current moment.

[0112] Step S502: Obtain the historical data collected by the sensor at the previous moment; Step S503: Obtain the user's displacement distance based on the historical data and the current data. Steps 501 to 503 are similar to steps 401 to 403, and will not be repeated here.

[0113] Step S504: If the displacement distance is less than the distance threshold, discard the currently collected data.

[0114] By calculating the displacement between the current moment and the previous moment, it can be determined whether the user's position has changed. If the displacement is less than the distance threshold, it is considered that the user's position has not changed, which means that the data collected at the current moment is similar to the data collected at the previous moment. If the current data is also saved, there will be two duplicate data in the storage space. Therefore, in this case, the current data can be discarded, and in this case, there is no need to adjust the speaker's operating parameters.

[0115] After acquiring the data for the next moment, if it is necessary to compare whether the user's position has changed in the next moment, the sampling data in the storage space that is closest to the sampling time of the next moment can be used as reference data. Based on the reference data and the data acquired in the next moment, it can be determined whether the user's position has changed.

[0116] In some embodiments of this application, acquiring environmental and location information through sensors and calibrating sound field parameters using a target sound field model can achieve preliminary sound field optimization. However, in practical applications, even calibration based on precise environmental and location information may not perfectly match each user's unique auditory preferences. Therefore, in other embodiments of this application, a flowchart of another example of sound field parameter calibration is provided, such as... Figure 6 As shown.

[0117] Step S601: Obtain the current data collected by the sensor at the current moment.

[0118] Step S602: Based on the current data collected by the sensor, obtain the target environment features of the user's current environment and the user's target location information in the current environment.

[0119] Step S603: Input the target environment features and the target location information into the target sound field model to obtain the target values ​​of the sound field response parameters.

[0120] Step S604: Adjust the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters.

[0121] Steps 601 to 604 are similar to steps 201 to 204, and will not be repeated here.

[0122] Step 605: Play audio through the target speaker group; It can play preset test audio to the target speaker group, such as audio containing specific frequency scans, vocals, or music clips, or it can play regular audio content that the user is currently listening to, such as music, movie soundtracks, or game sound effects.

[0123] Step 606: Obtain the user's satisfaction rating for the audio system; Satisfaction levels can be obtained in various ways, such as providing rating options (e.g., 1-5 star ratings, satisfied / dissatisfied buttons) through a user interface (e.g., a mobile app, remote control, or the audio system's built-in display), or through a voice interaction system where users verbally express their satisfaction, with the main speaker performing voice recognition and emotion analysis. When the obtained satisfaction score is lower than a preset threshold, it indicates that the current sound field calibration effect has not met the user's expectations, triggering subsequent optimization processes. The satisfaction threshold can be a specific numerical value, such as below 3 stars in a 1-5 star rating system, indicating dissatisfaction, or a Boolean value, such as when the user explicitly selects the "dissatisfied" option.

[0124] Step 607: If the satisfaction score is less than the score threshold, obtain the user's feature information and update the target sound field model according to the user's preference information.

[0125] The feature information includes the user's age and the threshold for the user's perceived difference in the sound field response parameters.

[0126] In this scenario, the main speaker acquires the user's characteristic information. This characteristic information consists of attribute data related to the individual user, which may influence the user's subjective preference for the sound field and provide a basis for personalized optimization. The characteristic information may include the user's age, gender, listening habits, music preferences, etc., and specifically includes the user's age and the user's perceived difference threshold for changes in sound field response parameters. The perceived difference threshold for changes in sound field response parameters can be understood as the smallest change in the sound field response parameters that the user can perceive, or the smallest step size that the user can perceive when the sound field response parameters are adjusted. Age information helps the system understand the differences in sound perception among users of different age groups. For example, older adults (over 60 years old) may have a larger perceived difference threshold for high-frequency sounds, while middle-aged and young adults (15-40 years old) have a smaller perceived difference threshold for high-frequency sounds. The smaller the perceived difference threshold, the more sensitive the user is to this sound field response parameter, meaning that the user can perceive even small changes; conversely, the larger the perceived difference threshold, the less sensitive the user is to this sound field response parameter, meaning that the user can only perceive larger changes. This information can be obtained through manual input by users in the audio system settings, through questionnaires, or through analysis of user behavior during long-term use of the audio system. Finally, the main speaker updates the target sound field model based on this user preference information. This update process utilizes personalized user feedback and feature information to fine-tune or retrain the original target sound field model, making it better suited to specific user preferences, thereby providing more personalized sound field calibration. Preference information can be used as additional input features, combined with the original training data, to perform incremental learning or transfer learning on the target AI model, or, based on user satisfaction feedback, to adjust the weights or hyperparameters related to the sound field response parameters in the target sound field model to optimize the output results.

[0127] This application's solution introduces a user feedback mechanism after initial calibration, enabling the sound field calibration process to be dynamically adjusted based on the user's actual auditory experience. When the user is dissatisfied with the calibration results, personalized user characteristics, such as age and perceived difference thresholds for changes in sound field response parameters, can be actively collected, and this information can be used to update the target sound field model. This iterative optimization process allows the target sound field model to learn and adapt to the specific user's auditory preferences, thereby generating sound field response parameters that better meet the user's expectations in subsequent calibration processes. This effectively solves the problem that traditional calibration methods may not fully satisfy the user's subjective auditory experience, achieving personalized and adaptive sound field calibration.

[0128] Other embodiments of this application also provide an audio system, the audio system including a plurality of speakers and sensors, wherein: The sensor is configured to acquire the currently collected data at the current moment; The target loudspeaker includes a control unit and an execution unit. The control unit is configured to acquire, based on the currently acquired data from the sensor, the target environmental features of the user's current environment and the user's target location information in the current environment. The target location information includes the user's coordinates and orientation, and the user's orientation is used to indicate the angle between the plane where the user's face is located and a reference plane. The target environmental features and the target location information are input into the target sound field model to obtain the target values ​​of the sound field response parameters. The target sound field model is obtained by training an initial artificial intelligence (AI) model based on a first training sample. The first training sample includes multiple sample environmental features, multiple sample location information, and multiple reference values ​​of the sound field response parameters. The sound field response parameters include at least one of relative arrival delay, sound image direction, and sound pressure level. Based on the target values ​​of the sound field response parameters, the operating parameters of the execution units of the plurality of loudspeakers are adjusted. The operating parameters include at least one of gain coefficient, delay compensation, offset angle, and electrical power. The execution unit includes an amplifier and / or a motor. The target speaker is one of the plurality of speakers.

[0129] It should be noted that the above-mentioned audio system structure diagram is different from... Figure 1 The audio system 100 shown is similar, but the steps performed by the sensors and speakers are as follows: Figure 2 , Figures 4-6 The steps shown will not be repeated here.

[0130] It should be noted that the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0131] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0132] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for calibrating sound field parameters, characterized in that, Applied to an audio system, the audio system including multiple speakers and sensors, the method includes: Obtain the current data collected by the sensor at the current moment; Obtain the historical data collected by the sensor at the previous moment; Based on the historical data and the current data, the user's displacement distance is obtained; When the displacement distance is greater than or equal to the distance threshold, the target environment features of the current environment in which the user is located and the target position information of the user in the current environment are obtained based on the current data collected by the sensor. The target position information includes the user's coordinates and orientation, and the user's orientation is used to indicate the angle between the plane in which the user's face is located and the reference plane. The target environmental features and the target location information are input into the target sound field model to obtain the target values ​​of the sound field response parameters. The target sound field model is obtained by training an initial artificial intelligence (AI) model based on a first training sample. The first training sample includes multiple sample environmental features, multiple sample location information, and multiple reference values ​​of the sound field response parameters. The sound field response parameters include at least one of relative arrival delay, sound image direction, and sound pressure level. Based on the target values ​​of the sound field response parameters, the operating parameters of the plurality of loudspeakers are adjusted, including at least one of gain coefficient, delay compensation, offset angle, and electrical power.

2. The method according to claim 1, characterized in that, After obtaining the user's displacement distance based on the historical data and the current data, the method further includes: If the displacement distance is less than the distance threshold, the currently collected data is discarded.

3. The method according to claim 1, characterized in that, The sound field response parameters include the sound image direction, the operating parameters include the gain coefficient, and adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: A target coordinate system is established based on the user's coordinates and orientation, with the user's coordinates as the origin and the user's orientation as the positive X-axis direction. Based on the user's coordinates and the coordinates of each speaker, determine the angle between each speaker and the positive X-axis direction; Based on the sound image direction, a target loudspeaker group is determined from the plurality of loudspeakers. The target loudspeaker group includes a first loudspeaker and a second loudspeaker. The first loudspeaker has a first angle relative to the positive X-axis direction that is smaller than the sound image direction, and the second loudspeaker has a second angle relative to the positive X-axis direction that is larger than the sound image direction. Based on the sound image direction, the first included angle, the second included angle, and the first calculation formula, determine the first gain coefficient of the first loudspeaker and the second gain coefficient of the second loudspeaker; The first calculation formula is as follows: ； ； in, For the first The gain coefficient of each speaker, For the direction of sound and image, The first included angle, The second included angle is i, which is 1 or 2.

4. The method according to claim 1, characterized in that, The sound field response parameters also include sound pressure level, and the operating parameters also include electrical power. Adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: The sound power distribution ratio of the first speaker and the second speaker is determined based on the gain coefficient of each speaker. Based on the first conversion formula, the target value of the sound pressure level, and the target value of the reference sound pressure, the target value of the effective sound pressure is determined, wherein the effective sound pressure is the sum of the effective sound pressure of the first loudspeaker and the effective sound pressure of the second loudspeaker; Based on the target value of the effective sound pressure, the sound power distribution ratio, and the second conversion formula, calculate the first sound power of the first loudspeaker and the second sound power of the second loudspeaker; The electrical power of the first loudspeaker is determined by multiplying the first conversion efficiency between the acoustic power and electrical power of the first loudspeaker with the first acoustic power; and the electrical power of the second loudspeaker is determined by multiplying the second conversion efficiency between the acoustic power and electrical power of the second loudspeaker with the second acoustic power. The first conversion formula is as follows: ； in, The effective sound pressure level, The reference sound pressure level is... Sound pressure level; The second conversion formula is as follows: ; Among them, P i Let L be the acoustic power of the i-th loudspeaker. i The effective sound pressure level of the i-th loudspeaker is... Air characteristic impedance; Let i be the distance between the coordinates of the i-th speaker and the coordinates of the user, where i is 1 or 2.

5. The method according to claim 3 or 4, characterized in that, The sound field response parameters also include relative arrival delay, and the operating parameters also include delay compensation. Adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: The first speaker is designated as the reference speaker, and the electrical delay of the first speaker is set to zero; The delay compensation for the electrical delay of the second speaker is determined based on the difference between the first distance and the second distance, and the speed of sound. The first distance is the distance between the first speaker and the user, and the second distance is the distance between the second speaker and the user.

6. The method according to claim 4, characterized in that, The operating parameters also include the deviation angle, and adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: For a target speaker, the distance between the target speaker and the user is determined based on the coordinates of the target speaker and the coordinates of the user, as well as the direction vector of the target speaker pointing towards the user; the target speaker is either the first speaker or the second speaker; According to the third conversion formula, the value of the directional attenuation parameter of the target loudspeaker is determined. The directional attenuation parameter is used to indicate the degree of sound pressure attenuation when deviating from the main axis direction of the target loudspeaker. Substituting the value of the directional attenuation parameter into the directional function of the target loudspeaker yields the target value of the deviation angle of the target loudspeaker. The directional function represents the mapping relationship between the deviation angle and the directional attenuation parameter. The deviation angle is related to the pitch angle and the azimuth angle. The third conversion formula is as follows: ； Among them, P i Let L be the acoustic power of the i-th loudspeaker. i The effective sound pressure level of the i-th loudspeaker is... Air characteristic impedance; Let be the distance between the coordinates of the i-th speaker and the coordinates of the user, and D be the directional attenuation parameter of the i-th speaker, where i is 1 or 2.

7. The method according to claim 1, characterized in that, The step of adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters includes: The sound field response parameters are input into the parameter adjustment model to obtain the operating parameters of the multiple loudspeakers. The parameter adjustment model is obtained by training the initial adjustment model based on the second training sample. The second training sample includes multiple preset values ​​of the sound field response parameters and multiple reference values ​​of the operating parameters.

8. The method according to claim 1, characterized in that, After adjusting the operating parameters of the plurality of loudspeakers according to the target value of the sound field response parameters, the method further includes: Audio is played through the multiple speakers; Obtain the user's satisfaction rating for the audio system; If the satisfaction score is less than the score threshold, the user's feature information is obtained, and the target sound field model is updated based on the user's feature information. The feature information includes the user's age and the user's perceived difference threshold for changes in the sound field response parameters.

9. A sound system, characterized in that, The audio system includes multiple speakers and sensors, wherein: The sensor is configured to acquire current data at the current moment and historical data at the previous moment; The target loudspeaker includes a control unit and an execution unit. The control unit is configured to acquire the user's displacement distance based on the historical acquisition data and the current acquisition data; and, if the displacement distance is greater than or equal to a distance threshold, acquire the target environment features of the user's current environment and the user's target position information in the current environment based on the current acquisition data from the sensor. The target position information includes the user's coordinates and orientation, and the user's orientation is used to indicate the angle between the plane where the user's face is located and a reference plane. The target environmental features and the target location information are input into the target sound field model to obtain the target values ​​of the sound field response parameters. The target sound field model is obtained by training an initial artificial intelligence (AI) model based on a first training sample. The first training sample includes multiple sample environmental features, multiple sample location information, and multiple reference values ​​of the sound field response parameters. The sound field response parameters include at least one of relative arrival delay, sound image direction, and sound pressure level. Based on the target values ​​of the sound field response parameters, the operating parameters of the execution units of the plurality of loudspeakers are adjusted. The operating parameters include at least one of gain coefficient, delay compensation, offset angle, and electrical power. The execution unit includes an amplifier and / or a motor. The target speaker is one of the plurality of speakers.

Images