Multi-partition loudspeaker control system

Abstract

The invention relates to the technical field of audio processing, in particular to a multi-partition loudspeaker control system which comprises an audio input processing module, a room geometric information acquisition module, an acoustic model construction module, a sound effect optimization module and a partition control module. Wherein the audio input processing module is used for receiving an external audio signal; the room geometric information acquisition module is used for acquiring geometric structure data and material attribute information of a room in real time; the acoustic model construction module is used for constructing a three-dimensional acoustic model of the room; the sound effect optimization module is used for optimizing the audio signal; and the partition control module is used for distributing the optimized audio signals to different loudspeaker partitions. According to the invention, the audio signal is intelligently optimized by combining the deep learning algorithm and the real-time environment data, the acoustic characteristics of the room are accurately adapted, and the audio synchronization of the multi-partition loudspeaker is realized through the low-delay communication protocol, so that the tone quality and the auditory experience of the user are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of audio processing, and in particular to a multi-zone loudspeaker control system. Background Art

[0002] With the continuous pursuit of audio effects in home entertainment, audio systems and commercial performances, more and more application scenarios require speaker systems that can provide accurate, balanced and high-quality audio output; in these application scenarios, the geometry of the room, building materials and environmental conditions have a significant impact on the propagation characteristics of the audio signal; in order to achieve the best audio effect, traditional audio systems often need to be manually adjusted or rely on simple sound enhancement technology, which cannot effectively solve the impact of the room's acoustic characteristics; for example, the room's acoustic characteristics (such as reflection, refraction, sound wave propagation path, etc.) will cause the sound in some areas to be distorted or too noisy, while some areas may not be heard clearly due to the attenuation of sound waves, which makes it impossible for traditional audio systems to fully utilize the acoustic characteristics of the space, thereby affecting the overall audio effect.

[0003] Although the existing multi-zone speaker control system can output audio for different zones, it usually lacks intelligent adaptation and optimization of the acoustic characteristics of the room; most of the systems in the existing technology use fixed sound processing algorithms or simplified room models, which are unable to analyze and adjust the acoustic characteristics of the room in real time, resulting in the audio signals of the zone speakers being inaccurate and unable to effectively achieve synchronization between the zone speakers; in addition, traditional sound optimization technologies mostly rely on manual adjustment, lack automated and intelligent processing capabilities, and fail to fully solve the problem of sound adjustment under different environmental conditions. Summary of the invention

[0004] Based on the above objectives, the present invention provides a multi-zone speaker control system.

[0005] A multi-zone speaker control system includes an audio input processing module, a room geometry information acquisition module, an acoustic model building module, a sound effect optimization module and a zone control module; wherein:

[0006] Audio input processing module: used to receive external audio signals and perform digital processing on them;

[0007] Room geometry information acquisition module: used to collect room geometry data and material property information in real time through integrated ultrasonic sensors and infrared imaging devices;

[0008] Acoustic model building module: Based on the data provided by the room geometry information acquisition module and combined with the deep learning algorithm, it builds a three-dimensional acoustic model of the room and generates the corresponding acoustic transfer function to describe the propagation characteristics of the audio signal in the room;

[0009] The acoustic model building module includes a feature extraction unit, a deep learning network unit, an environmental data acquisition unit and an acoustic transfer function generation unit; wherein:

[0010] Feature extraction unit: extracts the acoustic features of the room, including the sound wave propagation path, reflection and refraction information, based on the room geometry information and material property data;

[0011] Deep learning network unit: uses a deep learning algorithm of a recurrent neural network to train and build a three-dimensional acoustic model of the room based on the acoustic feature data provided by the feature extraction unit to simulate the propagation and interaction of sound waves in the room;

[0012] Environmental data collection unit: used to collect real-time environmental data in the room, including temperature and humidity;

[0013] Acoustic transfer function generation unit: Based on the three-dimensional acoustic model output by the deep learning network unit and the temperature and humidity data provided by the environmental data acquisition unit, the corresponding acoustic transfer function is calculated and generated to describe the propagation characteristics of the audio signal in the room, including the attenuation, reflection, refraction and propagation delay of the sound wave;

[0014] Sound optimization module: used to optimize the audio signal according to the acoustic transfer function generated by the acoustic model building module, so that the audio signal adapts to the acoustic characteristics of the room;

[0015] Partition control module: used to distribute the optimized audio signal output by the sound optimization module to different speaker partitions, and realize audio synchronization between the speakers of each partition through a low-latency communication protocol.

[0016] Optionally, the audio input processing module includes a receiving unit, an analog-to-digital conversion unit, a signal processing unit and a data interface unit; wherein:

[0017] Receiving unit: used to receive external audio signals through a wired or wireless interface, the wired interface includes an HDMI interface and an optical fiber interface, and the wireless interface includes Bluetooth and Wi-Fi;

[0018] The analog-to-digital conversion unit includes an analog-to-digital converter for converting the external audio signal received by the receiving unit into a digital signal;

[0019] Signal processing unit: used to perform noise reduction and signal enhancement processing on the digital audio signal converted by the analog-to-digital conversion unit to eliminate noise interference and enhance the clarity of the audio signal;

[0020] Data interface unit: used to transmit the digital audio signal processed by the signal processing unit to the sound optimization module.

[0021] Optionally, the room geometry information acquisition module includes an ultrasonic measurement unit and an infrared imaging unit; wherein:

[0022] Ultrasonic measurement unit: Through multiple integrated ultrasonic sensors, it transmits and receives ultrasonic signals to measure the distance and angle between objects in the room in real time, thereby obtaining the geometric structure data of the room;

[0023] Infrared imaging unit: Through the integrated infrared imaging device, the material property information of each surface in the room is captured, including surface material, reflectivity and sound absorption characteristics.

[0024] Optionally, the deep learning network unit includes:

[0025] The data input subunit is used to receive the room acoustic feature data provided by the feature extraction unit, including the feature data of the sound wave propagation path, reflection, and refraction, and convert it into a time series format suitable for recurrent neural network processing, wherein the data conversion includes converting the acoustic feature data into a series of vector representations and arranging them in time series order;

[0026] Recurrent Neural Network Subunit: A recurrent neural network of the long short-term memory network type is used to learn the propagation law of sound waves in the room, identify and extract potential time series patterns to simulate the propagation and interaction of sound waves in the room;

[0027] Model training subunit: Based on the input acoustic feature data, the weights of the long short-term memory network are iteratively updated through the back-propagation algorithm and the gradient descent optimization algorithm, and the prediction ability of the network is gradually optimized to generate a three-dimensional acoustic model of the room.

[0028] Optionally, the model training subunit includes:

[0029] Data initialization: Initialize all weight parameters and bias items in the recurrent neural network, set the initial value to a random value or a small constant value; and set the hyperparameters of the learning rate and number of iterations;

[0030] Forward propagation: Based on the input time series data and the weights and biases of the current network, forward propagation is performed through the recurrent neural network to obtain the output at each moment and calculate the predicted value and errors;

[0031] Calculate the loss function: Based on the predicted output and the actual target value y t , calculate the error at the current moment, that is, the loss function, the loss function is the mean square error; its calculation formula is: Among them, L t is the loss value at time t, y t is the actual value, is the predicted value;

[0032] Back propagation and weight update: According to the calculated loss function L t , the gradient is calculated using the back-propagation algorithm, and the weights and biases are updated by the gradient descent method;

[0033] Optimization process: By gradually updating the network's weight parameters, the model gradually learns to predict the room's acoustic characteristics and optimizes its output at each moment;

[0034] Generate a 3D acoustic model: Through the above training process, the final 3D acoustic model M is expressed as: M = f(W (T) , b (T) , x t ), where W (T) and b (T) are the weights and biases after training, x t is the input acoustic feature data, and f is the trained neural network model.

[0035] Optionally, the acoustic transfer function generating unit includes:

[0036] Environmental data integration subunit: used to receive real-time temperature and humidity data provided by the environmental data acquisition unit, and integrate the environmental data with the three-dimensional acoustic model data output by the deep learning network unit;

[0037] Acoustic characteristic adjustment subunit: Based on the temperature and humidity data in the room, analyze its influence on the sound wave propagation speed, sound wave attenuation and reflection characteristics, and adjust the relevant parameters in the three-dimensional acoustic model to reflect the actual situation of sound wave propagation under the current environmental conditions;

[0038] Transfer function calculation subunit: calculates the acoustic transfer function of each partition speaker according to the adjusted three-dimensional acoustic model data to describe the propagation characteristics of the audio signal in the room;

[0039] Transfer function output subunit: transmits the calculated acoustic transfer function to the sound optimization module for optimizing the frequency response and volume distribution of the audio signal.

[0040] Optionally, the acoustic characteristic adjustment subunit includes:

[0041] Data analysis: Based on the real-time temperature and humidity data provided by the environmental data acquisition unit, analyze its impact on the sound wave propagation velocity v, sound wave attenuation α and reflection characteristics; the specific analysis steps include:

[0042] The formula for calculating the sound wave propagation velocity v is: Among them, v0 is the standard sound speed, T is the current temperature; the formula of sound wave propagation speed v shows that the sound speed increases with the increase of temperature;

[0043] The calculation formula of the sound wave attenuation α is: α=α0(1+k·H), where α0 is the standard attenuation coefficient, k is the humidity influence coefficient, and H is the current humidity. The formula of the sound wave attenuation α indicates that an increase in humidity will lead to an increase in the sound wave attenuation.

[0044] The reflection coefficient R is adjusted according to the temperature and humidity data, and its calculation formula is: Where R0 is the standard reflection coefficient, T0 and H0 are the reference temperature and humidity respectively, and m and n are the adjustment coefficients of temperature and humidity;

[0045] Parameter adjustment: According to the above calculation results, the calculated sound velocity v is substituted into the sound velocity parameter in the three-dimensional acoustic model to ensure that the sound wave propagation speed is consistent with the actual environment; according to the calculated attenuation coefficient α, the sound wave attenuation parameter in the model is modified to describe the energy loss of the sound wave in the environment; according to the adjusted reflection coefficient R, the reflection characteristics of the material surface in the model are updated to ensure that the reflection effect of the sound wave is consistent with the actual environmental conditions.

[0046] Optionally, the transfer function calculation subunit includes:

[0047] Acoustic parameter extraction: Based on the adjusted three-dimensional acoustic model data, the acoustic parameters of the sound wave propagation velocity v, the sound wave attenuation coefficient α and the reflection coefficient R corresponding to each partition speaker are extracted;

[0048] Frequency response calculation: For each partition speaker, calculate its frequency response H(f) at different frequencies based on the extracted acoustic parameters, where f represents the frequency;

[0049] Application of sound wave propagation model: Apply the sound wave propagation model, combined with the reflection coefficient R, to adjust the frequency response H(f) to reflect the multiple reflections and refractions of sound waves in the room; the adjusted frequency response formula is: Where H′(f) is the adjusted frequency response, τ r is the propagation delay time of the reflected sound wave;

[0050] Transfer function integration: Integrate the adjusted frequency response H′(f) of each zone speaker into the complete acoustic transfer function H total (f) is used to describe the propagation characteristics of audio signals throughout the room.

[0051] Optionally, the sound effect optimization module includes an audio signal processing unit, a dynamic range compression unit and an output signal generation unit; wherein:

[0052] Audio signal processing unit: used to receive the acoustic transfer function H generated by the acoustic model building module total (f), and process the input audio signal according to the acoustic characteristics of the room; specifically, use a bandpass filter to filter the audio signal; let the output audio signal of the filtering process be y(t), where the calculation formula is: y(t) = F -1 (H total (f)·F(x(t))), where x(t) is the input audio signal, F represents the Fourier transform, and F -1 represents inverse Fourier transform;

[0053] Dynamic range compression unit: used to adjust the dynamic range of the audio signal; by setting the compression threshold T and compression ratio R, the strength of the audio signal is dynamically adjusted according to the amplitude of the input audio signal. The formula is:

[0054]

[0055] Where x(t) is the input audio signal, y * (t) is the compressed audio signal, T is the compression threshold, and R is the compression ratio;

[0056] Output signal generation unit: used to convert the compressed audio signal y * (t) is converted into a format suitable for speaker playback, and the optimized audio signal s(t) is output through each partition speaker.

[0057] Optionally, the partition control module includes an audio signal distribution unit and an audio synchronization unit; wherein:

[0058] Audio signal distribution unit: used to receive the optimized audio signal s(t) output by the sound effect optimization module and distribute it to different speaker partitions according to the speaker partition settings;

[0059] Audio synchronization unit: used to ensure audio synchronization between speakers in each zone through a low-latency communication protocol to avoid time deviation between different speakers;

[0060] The audio synchronization unit includes a clock synchronization subunit and a low-delay communication protocol subunit; wherein:

[0061] Clock synchronization subunit: used to ensure time synchronization between speaker partitions according to the system global clock;

[0062] Low-latency communication protocol subunit: Uses low-latency communication protocols, including Wi-Fi, Bluetooth or wired protocols, to achieve real-time transmission of audio signals, and ensures that each speaker partition receives and plays audio signals at the same time through the timestamp and synchronization identifier of the data packet.

[0063] Beneficial effects of the present invention:

[0064] The present invention effectively solves the problem that traditional audio systems cannot intelligently adapt to the acoustic characteristics of the room by introducing deep learning algorithms and real-time environmental data acquisition technology; the system can perceive and analyze environmental factors such as the room's geometry, material properties, temperature and humidity in real time, and achieve precise optimization of audio signals by constructing an accurate three-dimensional acoustic model and generating a corresponding acoustic transfer function; this intelligent optimization not only improves the sound quality, but also can dynamically adjust the frequency response and volume distribution of the audio signal to ensure the best propagation effect of the audio signal under different room conditions.

[0065] The present invention realizes precise synchronization between multi-zone speakers through a low-latency communication protocol, avoiding the delay difference of audio signals between different zones; the optimization and synchronous distribution of audio signals can ensure that the speakers in each zone achieve the optimal timing and balance effect when outputting audio signals, further improving the overall performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0067] Figure 1 A schematic diagram of a multi-zone speaker control system according to an embodiment of the present invention;

[0068] Figure 2 A schematic diagram of an acoustic model construction module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0069] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0070] like Figure 1-Figure 2 As shown, a multi-zone speaker control system includes an audio input processing module, a room geometry information acquisition module, an acoustic model building module, a sound effect optimization module and a zone control module; wherein:

[0071] Audio input processing module: used to receive external audio signals, perform digital processing on them, and provide audio input for the sound effect optimization module;

[0072] Room geometry information acquisition module: used to collect room geometry data and material property information in real time through integrated ultrasonic sensors and infrared imaging devices, providing accurate data support for room acoustic modeling;

[0073] Acoustic model building module: Based on the data provided by the room geometry information acquisition module and combined with the deep learning algorithm, it builds a three-dimensional acoustic model of the room and generates the corresponding acoustic transfer function (RTF) to describe the propagation characteristics of the audio signal in the room;

[0074] The acoustic model building module includes a feature extraction unit, a deep learning network unit, an environmental data acquisition unit, and an acoustic transfer function generation unit; wherein:

[0075] Feature extraction unit: Extracts the acoustic features of the room, including the sound wave propagation path, reflection, and refraction information, based on the room geometry information and material property data, and uses the extracted features as input data for further training of the deep learning model;

[0076] Deep learning network unit: uses a deep learning algorithm of a recurrent neural network to train and build a three-dimensional acoustic model of the room based on the acoustic feature data provided by the feature extraction unit to simulate the propagation and interaction of sound waves in the room;

[0077] Environmental data acquisition unit: used to collect real-time environmental data in the room, including temperature and humidity, to compensate for the impact of changes in environmental conditions on the propagation characteristics of sound waves;

[0078] Acoustic transfer function generation unit: Based on the three-dimensional acoustic model output by the deep learning network unit and the temperature and humidity data provided by the environmental data acquisition unit, the corresponding acoustic transfer function is calculated and generated to describe the propagation characteristics of the audio signal in the room, including the attenuation, reflection, refraction and propagation delay of the sound wave;

[0079] Sound optimization module: used to optimize the audio signal based on the acoustic transfer function generated by the acoustic model building module, so that the audio signal adapts to the acoustic characteristics of the room and ensures that the sound output by the speakers in each partition is clear, balanced and distortion-free;

[0080] Partition control module: used to distribute the optimized audio signals output by the sound optimization module to different speaker partitions, and achieve audio synchronization between speakers in each partition through a low-latency communication protocol to ensure the consistency and coordination of multi-partition audio output.

[0081] The audio input processing module includes a receiving unit, an analog-to-digital conversion unit, a signal processing unit and a data interface unit; wherein:

[0082] Receiving unit: used to receive external audio signals through a wired or wireless interface, the wired interface includes an HDMI interface and an optical fiber interface, and the wireless interface includes Bluetooth and Wi-Fi;

[0083] The analog-to-digital conversion unit includes an analog-to-digital converter (ADC), which is used to convert the external audio signal received by the receiving unit into a digital signal;

[0084] Signal processing unit: used to perform noise reduction and signal enhancement processing on the digital audio signal converted by the analog-to-digital conversion unit to eliminate noise interference and enhance the clarity of the audio signal;

[0085] Data interface unit: used to transmit the digital audio signal processed by the signal processing unit to the sound effect optimization module, ensuring that the sound effect optimization module receives high-quality audio input; through the design of the above-mentioned audio input processing module, the system can efficiently and accurately receive and digitize external audio signals, and provide high-quality digital audio signals to the sound effect optimization module through processing steps such as filtering, noise reduction and signal enhancement.

[0086] The room geometry information acquisition module includes an ultrasonic measurement unit and an infrared imaging unit; wherein:

[0087] Ultrasonic measurement unit: Through multiple integrated ultrasonic sensors, it transmits and receives ultrasonic signals to measure the distance and angle between objects in the room in real time, thereby obtaining the geometric structure data of the room;

[0088] Infrared imaging unit: Through the integrated infrared imaging device, the material property information of each surface in the room is captured, including surface material, reflectivity and sound absorption characteristics; and this data is transmitted to the acoustic model construction module; through the design of the above-mentioned room geometry information acquisition module, the geometric dimensions of the room and the property information of each surface material can be accurately and real-time collected; the ultrasonic sensor unit ensures the accurate measurement of the room's geometric structure, and the infrared imaging unit further supplements the identification of the room material through the thermal conductivity characteristics of the material, ensuring that the constructed acoustic model has high precision and high reliability.

[0089] The deep learning network unit includes:

[0090] Data input subunit: used to receive the room acoustic feature data provided by the feature extraction unit, including the feature data of the sound wave propagation path, reflection, and refraction, and convert it into a time series format suitable for recurrent neural network (RNN) processing, wherein the data conversion includes converting the acoustic feature data into a series of vector representations and arranging them in time series order so as to be input into the recurrent neural network unit for further processing;

[0091] Recurrent Neural Network Subunit: A recurrent neural network of the long short-term memory (LSTM) network type is used to learn the propagation law of sound waves in the room, identify and extract potential time series patterns to simulate the propagation and interaction of sound waves in the room;

[0092] The above-mentioned long short-term memory network processes and learns data through the following steps:

[0093] Input gate: receives the current input x t and the hidden state h at the previous moment t-1 , and obtain the candidate value of the input by weighted summation The formula is: Among them, W x and W h is the weight matrix of the input and the hidden state at the previous moment, b c is the bias term, tanh is the hyperbolic tangent activation function, is the candidate cell state at the current moment;

[0094] Forget gate: controls the cell state C at the previous moment t-1 How much is retained? The formula is: f t =σ(W f x t +U f ht-1+b f ), where σ is the sigmoid activation function, W f and U f is the weight matrix of the forget gate, b f is the bias term, f t is the output of the forget gate, indicating the part to be forgotten;

[0095] Output gate: determines the output h at the current moment t , by combining the current cell state and hidden state, the formula is: h t =o t tanh(C t ), where o t is the control signal of the output gate, C t is the current cell state, tanh is the activation function;

[0096] Cell state update: Combine the input gate and forget gate to update the cell state C t , the formula is: Among them, i t is the output of the input gate, C t-1 is the cell state at the previous moment, is the candidate cell state at the current moment; the time series pattern of sound wave propagation in the room is identified through the trained LSTM network, including the reflection, refraction, propagation delay and other characteristics of the sound wave.

[0097] Model training subunit: Based on the input acoustic feature data, the weights of the long short-term memory network are iteratively updated through the back-propagation algorithm and the gradient descent optimization algorithm, and the prediction ability of the network is gradually optimized to generate a three-dimensional acoustic model of the room; by using the LSTM network to perform deep learning on the acoustic features of the room, the time series propagation law of sound waves in a complex room can be effectively captured; the LSTM network can identify long-term dependencies in the acoustic feature data to ensure the accurate construction of the three-dimensional acoustic model; this process greatly improves the accuracy of the room acoustic model and provides more reliable data support for the subsequent generation of acoustic transfer functions.

[0098] The model training subunit includes:

[0099] Data initialization: Initialize all weight parameters and bias items in the recurrent neural network, set the initial value to a random value or a small constant value; and set the hyperparameters of the learning rate and number of iterations;

[0100] Forward propagation: Based on the input time series data and the weights and biases of the current network, forward propagation is performed through the recurrent neural network to obtain the output at each moment and calculate the predicted value and errors;

[0101] Calculate the loss function: Based on the predicted output and the actual target value y t , calculate the error at the current moment, that is, the loss function, the loss function is the mean square error (MSE); its calculation formula is: Among them, L t is the loss value at time t, y t is the actual value, is the predicted value;

[0102] Back propagation and weight update: According to the calculated loss function L t , the back propagation algorithm is used to calculate the gradient, and the weights and biases are updated by the gradient descent method; specifically, the partial derivatives of the loss function with respect to each weight and bias are calculated first, and the update formula is: Among them, η is the learning rate, T is the number of iterations, and W (n) and b (n) are the weight and bias in the nth iteration, respectively. and are the gradients of weights and biases respectively; the weights are then updated through multiple iterations until the loss function reaches a minimum or the set stopping condition is met;

[0103] Optimization process: By gradually updating the network's weight parameters, the model gradually learns to predict the room's acoustic characteristics and optimizes its output at each moment. The training process continuously adjusts the weights through back-propagation and gradient descent algorithms, ultimately achieving the goal of optimizing the room's acoustic model.

[0104] Generate a 3D acoustic model: Through the above training process, the final 3D acoustic model M is expressed as: M = f(W (T) , b (T) , x t ), where W (T) and b (T) are the weights and biases after training, x t is the input acoustic feature data, and f is the trained neural network model; through the steps of the above model training unit, the system can gradually optimize the weights in the recurrent neural network through the back propagation algorithm and the gradient descent method, and enhance the network's ability to learn the room's acoustic characteristics; the minimization of the loss function ensures the accuracy of the network, and the three-dimensional acoustic model finally obtained through iteration can efficiently and accurately describe the propagation characteristics of sound waves in the room, providing accurate data support for subsequent sound effect optimization and speaker partition control.

[0105] The acoustic transfer function generation unit includes:

[0106] Environmental data integration subunit: used to receive real-time temperature and humidity data provided by the environmental data acquisition unit, and integrate the environmental data with the three-dimensional acoustic model data output by the deep learning network unit;

[0107] Acoustic characteristic adjustment subunit: Based on the temperature and humidity data in the room, analyze its influence on the sound wave propagation speed, sound wave attenuation and reflection characteristics, and adjust the relevant parameters in the three-dimensional acoustic model to reflect the actual situation of sound wave propagation under the current environmental conditions;

[0108] Transfer function calculation subunit: calculates the acoustic transfer function of each partition speaker according to the adjusted three-dimensional acoustic model data to describe the propagation characteristics of the audio signal in the room;

[0109] Transfer function output subunit: The calculated acoustic transfer function is transmitted to the sound optimization module to optimize the frequency response and volume distribution of the audio signal; by introducing the environmental data integration subunit and the acoustic characteristic adjustment subunit in the acoustic transfer function generation unit, the system can dynamically respond to real-time temperature and humidity changes in the room and accurately adjust the parameters of the three-dimensional acoustic model; this design ensures that the acoustic transfer function can accurately reflect the sound wave propagation characteristics under the current environmental conditions, thereby optimizing the clarity and balance of the sound output.

[0110] The acoustic characteristic adjustment subunit includes:

[0111] Data analysis: Based on the real-time temperature and humidity data provided by the environmental data acquisition unit, analyze its impact on the sound wave propagation velocity v, sound wave attenuation α and reflection characteristics; the specific analysis steps include:

[0112] The formula for calculating the sound wave propagation velocity v is: Where v0 is the standard speed of sound (about 343 m / s) and T is the current temperature (degrees Celsius). The formula for the speed of sound wave v shows that the speed of sound increases with increasing temperature.

[0113] The calculation formula of the sound wave attenuation α is: α=α0(1+k·H), where α0 is the standard attenuation coefficient, k is the humidity influence coefficient, and H is the current humidity (percentage); the formula of the sound wave attenuation α indicates that an increase in humidity will lead to an increase in sound wave attenuation;

[0114] The reflection coefficient R is adjusted according to the temperature and humidity data, and its calculation formula is: Among them, R0 is the standard reflection coefficient, T0 and H0 are the reference temperature and humidity respectively, m and n are the adjustment coefficients of temperature and humidity. This formula is used to reflect the impact of environmental changes on the reflection characteristics of the material;

[0115] Parameter adjustment: According to the above calculation results, the calculated sound velocity v is substituted into the sound velocity parameter in the three-dimensional acoustic model to ensure that the sound wave propagation speed is consistent with the actual environment; according to the calculated attenuation coefficient α, the sound wave attenuation parameter in the model is modified to describe the energy loss of the sound wave in the environment; according to the adjusted reflection coefficient R, the reflection characteristics of the material surface in the model are updated to ensure that the reflection effect of the sound wave is consistent with the actual environmental conditions; through the design of the above-mentioned acoustic characteristic adjustment subunit, the system can accurately analyze and adjust the sound wave propagation speed, attenuation and reflection characteristics based on the real-time collected temperature and humidity data.

[0116] The transfer function calculation subunit includes:

[0117] Acoustic parameter extraction: Based on the adjusted three-dimensional acoustic model data, the acoustic parameters of the sound wave propagation velocity v, the sound wave attenuation coefficient α and the reflection coefficient R corresponding to each partition speaker are extracted;

[0118] Frequency response calculation: For each partition speaker, calculate its frequency response H(f) at different frequencies based on the extracted acoustic parameters, where f represents the frequency; the frequency response calculation formula is: Where Y(f) is the frequency domain representation of the loudspeaker output signal, X(f) is the frequency domain representation of the input signal, A(f) is the gain coefficient, τ is the propagation delay time, α(f) is the frequency-dependent attenuation coefficient, and j is the imaginary unit;

[0119] Application of sound wave propagation model: Apply the sound wave propagation model, combined with the reflection coefficient R, to adjust the frequency response H(f) to reflect the multiple reflections and refractions of sound waves in the room; the adjusted frequency response formula is: Where H′(f) is the adjusted frequency response, τ r is the propagation delay time of the reflected sound wave;

[0120] Transfer function integration: Integrate the adjusted frequency response H′(f) of each zone speaker into the complete acoustic transfer function H total (f) is used to describe the propagation characteristics of the audio signal throughout the room; the integration formula is: Where N is the total number of zone speakers, H i ′(f) is the adjusted frequency response of the i-th partition speaker; through the design of the transfer function calculation subunit, the system can accurately calculate the acoustic transfer function of each partition speaker based on the adjusted three-dimensional acoustic model data and real-time environmental data; by extracting key acoustic parameters, calculating frequency response, applying sound wave propagation model and integrating transfer function, the propagation characteristics of the audio signal in the room are accurately described.

[0121] The sound effect optimization module includes an audio signal processing unit, a dynamic range compression unit, and an output signal generation unit; wherein:

[0122] Audio signal processing unit: used to receive the acoustic transfer function H generated by the acoustic model building module total (f), and process the input audio signal according to the acoustic characteristics of the room; specifically, a bandpass filter is used to filter the audio signal to enhance or suppress the audio components of the specified frequency band to adapt to the frequency characteristics of the room; let the output audio signal of the filtering process be y(t), where the calculation formula is: y(t) = F -1 (H total (f)·F(x(t))), where x(t) is the input audio signal, F represents the Fourier transform, and F -1 represents inverse Fourier transform;

[0123] Dynamic range compression unit: used to adjust the dynamic range of the audio signal to ensure that the audio signal is not over-amplified or compressed when played in the room and adapt to the reflection characteristics of the room; by setting the compression threshold T and compression ratio R, the intensity of the audio signal is dynamically adjusted according to the amplitude of the input audio signal. The formula is:

[0124]

[0125] Where x(t) is the input audio signal, y * (t) is the compressed audio signal, T is the compression threshold, and R is the compression ratio. Through the design of the sound optimization module, the audio signal can be accurately adjusted based on the acoustic transfer function generated by the acoustic model building module to adapt it to the acoustic characteristics of the room.

[0126] Output signal generation unit: used to convert the compressed audio signal y * (t) is converted into a format for speaker playback, and the optimized audio signal s(t) is output through the speakers in each partition. Through the design of the sound optimization module, the system can accurately adjust the frequency response and dynamic range of the audio signal based on the acoustic transfer function generated by the acoustic model building module, ensuring that the audio signal adapts to the acoustic characteristics of the room.

[0127] The partition control module includes an audio signal distribution unit and an audio synchronization unit; wherein:

[0128] Audio signal distribution unit: used to receive the optimized audio signal s(t) output by the sound effect optimization module and distribute it to different speaker partitions according to the speaker partition settings;

[0129] Audio synchronization unit: used to ensure audio synchronization between speakers in each zone through a low-latency communication protocol to avoid time deviation between different speakers;

[0130] The audio synchronization unit includes a clock synchronization subunit and a low-latency communication protocol subunit; wherein:

[0131] Clock synchronization subunit: used to ensure time synchronization between speaker partitions according to the system global clock;

[0132] Low-latency communication protocol subunit: adopts low-latency communication protocols, including Wi-Fi, Bluetooth or wired protocols, to achieve real-time transmission of audio signals, and ensures that each speaker partition receives and plays audio signals at the same time through the timestamp and synchronization identifier of the data packet, thereby eliminating audio asynchrony caused by communication delay; through the design of the partition control module, the system can efficiently and accurately distribute the optimized audio signals to different speaker partitions, and ensure the audio synchronization between the speakers in each partition.

[0133] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0134] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-zone speaker control system, characterized in that: It includes an audio input processing module, a room geometry information acquisition module, an acoustic model building module, a sound effect optimization module and a partition control module; among which: Audio input processing module: used to receive external audio signals and perform digital processing on them; Room geometry information acquisition module: used to collect room geometry data and material property information in real time through integrated ultrasonic sensors and infrared imaging devices; Acoustic model building module: Based on the data provided by the room geometry information acquisition module and combined with the deep learning algorithm, it builds a three-dimensional acoustic model of the room and generates the corresponding acoustic transfer function to describe the propagation characteristics of the audio signal in the room; The acoustic model building module includes a feature extraction unit, a deep learning network unit, an environmental data acquisition unit and an acoustic transfer function generation unit; wherein: Feature extraction unit: extracts the acoustic features of the room, including the sound wave propagation path, reflection and refraction information, based on the room geometry information and material property data; Deep learning network unit: uses a deep learning algorithm of a recurrent neural network to train and build a three-dimensional acoustic model of the room based on the acoustic feature data provided by the feature extraction unit to simulate the propagation and interaction of sound waves in the room; Environmental data collection unit: used to collect real-time environmental data in the room, including temperature and humidity; Acoustic transfer function generation unit: Based on the three-dimensional acoustic model output by the deep learning network unit and the temperature and humidity data provided by the environmental data acquisition unit, the corresponding acoustic transfer function is calculated and generated to describe the propagation characteristics of the audio signal in the room, including the attenuation, reflection, refraction and propagation delay of the sound wave; Sound optimization module: used to optimize the audio signal according to the acoustic transfer function generated by the acoustic model building module, so that the audio signal adapts to the acoustic characteristics of the room; Partition control module: used to distribute the optimized audio signal output by the sound optimization module to different speaker partitions, and realize audio synchronization between the speakers of each partition through a low-latency communication protocol.

2. A multi-zone speaker control system according to claim 1, characterized in that: The audio input processing module includes a receiving unit, an analog-to-digital conversion unit, a signal processing unit and a data interface unit; wherein: Receiving unit: used to receive external audio signals through a wired or wireless interface, the wired interface includes an HDMI interface and an optical fiber interface, and the wireless interface includes Bluetooth and Wi-Fi; The analog-to-digital conversion unit includes an analog-to-digital converter for converting the external audio signal received by the receiving unit into a digital signal; Signal processing unit: used to perform noise reduction and signal enhancement processing on the digital audio signal converted by the analog-to-digital conversion unit to eliminate noise interference and enhance the clarity of the audio signal; Data interface unit: used to transmit the digital audio signal processed by the signal processing unit to the sound optimization module.

3. A multi-zone speaker control system according to claim 1, characterized in that: The room geometry information acquisition module includes an ultrasonic measurement unit and an infrared imaging unit; wherein: Ultrasonic measurement unit: Through multiple integrated ultrasonic sensors, it transmits and receives ultrasonic signals to measure the distance and angle between objects in the room in real time, thereby obtaining the geometric structure data of the room; Infrared imaging unit: Through the integrated infrared imaging device, the material property information of each surface in the room is captured, including surface material, reflectivity and sound absorption characteristics.

4. A multi-zone speaker control system according to claim 1, characterized in that: The deep learning network unit comprises: The data input subunit is used to receive the room acoustic feature data provided by the feature extraction unit, including the feature data of the sound wave propagation path, reflection, and refraction, and convert it into a time series format suitable for recurrent neural network processing, wherein the data conversion includes converting the acoustic feature data into a series of vector representations and arranging them in time series order; Recurrent Neural Network Subunit: A recurrent neural network of the long short-term memory network type is used to learn the propagation law of sound waves in the room, identify and extract potential time series patterns to simulate the propagation and interaction of sound waves in the room; Model training subunit: Based on the input acoustic feature data, the weights of the long short-term memory network are iteratively updated through the back-propagation algorithm and the gradient descent optimization algorithm, and the prediction ability of the network is gradually optimized to generate a three-dimensional acoustic model of the room.

5. A multi-zone speaker control system according to claim 4, characterized in that: The model training subunit comprises: Data initialization: Initialize all weight parameters and bias items in the recurrent neural network, set the initial value to a random value or a small constant value; and set the hyperparameters of the learning rate and number of iterations; Forward propagation: Based on the input time series data and the weights and biases of the current network, forward propagation is performed through the recurrent neural network to obtain the output at each moment and calculate the predicted value and errors; Calculate the loss function: Based on the predicted output and the actual target value y t , calculate the error at the current moment, that is, the loss function, the loss function is the mean square error; its calculation formula is: Among them, L t is the loss value at time t, y t is the actual value, is the predicted value; Back propagation and weight update: According to the calculated loss function L t , the gradient is calculated using the back-propagation algorithm, and the weights and biases are updated by the gradient descent method; Optimization process: By gradually updating the network's weight parameters, the model gradually learns to predict the room's acoustic characteristics and optimizes its output at each moment; Generate a 3D acoustic model: Through the above training process, the final 3D acoustic model M is expressed as: M = f(W (T) , b (T) , x t ), where W (T) and b (T) are the weights and biases after training, x t is the input acoustic feature data, and f is the trained neural network model.

6. The multi-zone speaker control system according to claim 1, characterized in that: The acoustic transfer function generating unit comprises: Environmental data integration subunit: used to receive real-time temperature and humidity data provided by the environmental data acquisition unit, and integrate the environmental data with the three-dimensional acoustic model data output by the deep learning network unit; Acoustic characteristic adjustment subunit: Based on the temperature and humidity data in the room, analyze its influence on the sound wave propagation speed, sound wave attenuation and reflection characteristics, and adjust the relevant parameters in the three-dimensional acoustic model to reflect the actual situation of sound wave propagation under the current environmental conditions; Transfer function calculation subunit: calculates the acoustic transfer function of each partition speaker according to the adjusted three-dimensional acoustic model data to describe the propagation characteristics of the audio signal in the room; Transfer function output subunit: transmits the calculated acoustic transfer function to the sound optimization module for optimizing the frequency response and volume distribution of the audio signal.

7. A multi-zone speaker control system according to claim 6, characterized in that: The acoustic characteristic adjustment subunit comprises: Data analysis: Based on the real-time temperature and humidity data provided by the environmental data acquisition unit, analyze its impact on the sound wave propagation velocity v, sound wave attenuation α and reflection characteristics; the specific analysis steps include: The formula for calculating the sound wave propagation velocity v is: Among them, v0 is the standard sound speed, T is the current temperature; the formula of sound wave propagation speed v shows that the sound speed increases with the increase of temperature; The calculation formula of the sound wave attenuation α is: α=α0(1+k·H), where α0 is the standard attenuation coefficient, k is the humidity influence coefficient, and H is the current humidity. The formula of the sound wave attenuation α indicates that an increase in humidity will lead to an increase in the sound wave attenuation. The reflection coefficient R is adjusted according to the temperature and humidity data, and its calculation formula is: Where R0 is the standard reflection coefficient, T0 and H0 are the reference temperature and humidity respectively, and m and n are the adjustment coefficients of temperature and humidity; Parameter adjustment: According to the above calculation results, the calculated sound velocity v is substituted into the sound velocity parameter in the three-dimensional acoustic model to ensure that the sound wave propagation speed is consistent with the actual environment; according to the calculated attenuation coefficient α, the sound wave attenuation parameter in the model is modified to describe the energy loss of the sound wave in the environment; according to the adjusted reflection coefficient R, the reflection characteristics of the material surface in the model are updated to ensure that the reflection effect of the sound wave is consistent with the actual environmental conditions.

8. A multi-zone speaker control system according to claim 7, characterized in that: The transfer function calculation subunit comprises: Acoustic parameter extraction: Based on the adjusted three-dimensional acoustic model data, the acoustic parameters of the sound wave propagation velocity v, the sound wave attenuation coefficient α and the reflection coefficient R corresponding to each partition speaker are extracted; Frequency response calculation: For each partition speaker, calculate its frequency response H(f) at different frequencies based on the extracted acoustic parameters, where f represents the frequency; Application of sound wave propagation model: Apply the sound wave propagation model, combined with the reflection coefficient R, to adjust the frequency response H(f) to reflect the multiple reflections and refractions of sound waves in the room; the adjusted frequency response formula is: Where H′(f) is the adjusted frequency response, τ r is the propagation delay time of the reflected sound wave; Transfer function integration: Integrate the adjusted frequency response H′(f) of each zone speaker into the complete acoustic transfer function H total (f) is used to describe the propagation characteristics of audio signals throughout the room.

9. The multi-zone speaker control system according to claim 8, characterized in that: The sound effect optimization module includes an audio signal processing unit, a dynamic range compression unit and an output signal generation unit; wherein: Audio signal processing unit: used to receive the acoustic transfer function H generated by the acoustic model building module total (f), and process the input audio signal according to the acoustic characteristics of the room; specifically, use a bandpass filter to filter the audio signal; let the output audio signal of the filtering process be y(t), where the calculation formula is: y(t) = F -1 (H total (f)·F(x(t))), where x(t) is the input audio signal, F represents the Fourier transform, and F -1 represents inverse Fourier transform; Dynamic range compression unit: used to adjust the dynamic range of the audio signal; by setting the compression threshold T and compression ratio R, the strength of the audio signal is dynamically adjusted according to the amplitude of the input audio signal. The formula is: Where x(t) is the input audio signal, y * (t) is the compressed audio signal, T is the compression threshold, and R is the compression ratio; Output signal generation unit: used to convert the compressed audio signal y * (t) is converted into a format suitable for speaker playback, and the optimized audio signal s(t) is output through each partition speaker.

10. The multi-zone speaker control system according to claim 9, characterized in that: The partition control module includes an audio signal distribution unit and an audio synchronization unit; wherein: Audio signal distribution unit: used to receive the optimized audio signal s(t) output by the sound effect optimization module and distribute it to different speaker partitions according to the speaker partition settings; Audio synchronization unit: used to ensure audio synchronization between speakers in each zone through a low-latency communication protocol to avoid time deviation between different speakers; The audio synchronization unit includes a clock synchronization subunit and a low-delay communication protocol subunit; wherein: Clock synchronization subunit: used to ensure time synchronization between speaker partitions according to the system global clock; Low-latency communication protocol subunit: Uses low-latency communication protocols, including Wi-Fi, Bluetooth or wired protocols, to achieve real-time transmission of audio signals, and ensures that each speaker partition receives and plays audio signals at the same time through the timestamp and synchronization identifier of the data packet.

Images