Adaptive frequency optimization system in radio and television microwave transmission

By designing an adaptive frequency optimization system in the radio and television microwave transmission system, collecting and processing audio signals in real time, building a gain compensation model and performing real-time optimization, the problems of frequency switching delay and sound quality stability in dynamic environments are solved, and efficient frequency resource optimization and audio and video signal transmission are achieved.

CN120223820AInactive Publication Date: 2025-06-27杨建梅
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202510416316.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-04-03
Publication Date
2025-06-27
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In a dynamic environment, existing frequency optimization technologies are difficult to perceive channel characteristics changes in real time, resulting in an increase in frequency switching delay or an increase in the risk of signal interruption. The difference in loudness parameters between different frequency channels lacks rapid calibration methods, which affects the stability of sound quality.

Method used

An adaptive frequency optimization system in broadcasting and television microwave transmission is designed. Through the initial audio acquisition and processing module, the audio acquisition and processing module calculates the loudness offset during frequency switching. The feature extraction module uses the fast Fourier transform algorithm to extract the frequency distribution feature vector. The model building module builds a gain compensation model and performs compensation calculations. The audio quality compensation module monitors in real time and processes the gain compensation value through the smoothing factor. The audio optimization module updates the gain compensation model based on the minimum mean square adaptive filtering algorithm.

Benefits of technology

Real-time calibration of frequency channel loudness differences and suppression of sound quality damage, improve sound quality stability and transmission efficiency, effectively respond to sudden interference and fast switching needs, and provide a high-quality audio and video transmission experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120223820A_ABST
    Figure CN120223820A_ABST
Patent Text Reader

Abstract

The invention relates to a self-adaptive frequency optimization system in radio and television microwave transmission, and the system comprises an initial audio collection and processing module which collects an audio signal of a current frequency channel in real time, and generates a first data set; the conversion audio acquisition processing module synchronously acquires a target channel signal according to the frequency switching signal, and calculates the loudness offset between two channels; a feature extraction module extracts frequency domain energy distribution feature vectors of the first data set and the second data set; the model construction module constructs a dynamic gain compensation model and generates a compensation value of a target channel; the tone quality compensation module performs smoothing processing on the compensation value to generate an optimized third data set; and the audio optimization module updates model parameters, dynamically adjusts a frequency domain analysis strategy, and outputs low-delay and high-stability radio and television transmission signals. According to the invention, the problems of high switching delay and degraded tone quality caused by static configuration in a traditional scheme are solved, and the transmission efficiency and user experience in a complex electromagnetic environment are remarkably improved through modular dynamic calibration and compensation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and particularly to an adaptive frequency optimization system in radio and television microwave transmission. Background Art

[0002] As one of the key technologies in the field of wireless communication, radio and television microwave transmission undertakes the important task of efficiently and stably transmitting audio and video signals to end users. In a complex electromagnetic environment and dynamic service requirements, how to achieve seamless switching between multiple frequency channels and signal quality optimization has always been the core research direction in the industry. Traditional frequency optimization methods usually rely on static spectrum allocation and preset parameter adjustment, aiming to reduce interference and improve transmission efficiency through fixed strategies. However, with the popularization of ultra-high-definition content and the upgrading of real-time requirements, the existing solutions have gradually exposed their limitations in dynamic scenarios, and there is an urgent need for a more intelligent optimization mechanism to balance transmission rate, sound quality stability, and resource utilization.

[0003] Currently, most frequency optimization technologies still adopt a static configuration mode based on empirical thresholds, which is difficult to cope with sudden interference or rapid switching requirements. For example, when encountering multipath fading or weather changes, the system cannot perceive the changes in channel characteristics in real time, resulting in an increase in switching delay or an increase in the risk of signal interruption. In addition, there is a lack of a means to quickly calibrate the loudness parameter differences between different frequency channels, and manual adjustment is likely to cause sound quality degradation or loudness jumps, seriously affecting the user experience. Especially in high-density urban environments or emergency broadcast scenarios, the existing solutions have insufficient adaptability and are difficult to achieve the dual goals of frequency switching and sound quality protection within a millisecond time window. Therefore, how to achieve real-time optimization of frequency resources and lossless transmission of audio and video signals in a dynamic environment has become a technical bottleneck that urgently needs to be broken through in the field of radio and television microwave transmission. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide an adaptive frequency optimization system in radio and television microwave transmission that can real-time calibrate the loudness differences between frequency channels and suppress sound quality damage.

[0005] The purpose of the present invention is achieved by the following solutions:

[0006] In a first aspect, the present invention provides an adaptive frequency optimization system in radio and television microwave transmission, and the system is configured with the following modules:

[0007] An initial audio acquisition and processing module, which is used to perform real-time acquisition and preprocessing on the current frequency channel based on a preset sampling frequency, and generate a first data set including original loudness parameters and channel state description values;

[0008] A conversion audio acquisition and processing module is used to acquire audio signals of a switching frequency channel based on a frequency switching signal, generate a second data set, and calculate the loudness of the second data set based on the first data set to obtain a loudness offset;

[0009] A feature extraction module is used to extract features from the first data set and the second data set based on the fast Fourier transform algorithm to generate a frequency distribution feature vector;

[0010] A model construction module is used to construct a gain compensation model based on the first data set, the loudness offset, and the frequency distribution feature vector, and perform compensation calculation on the loudness offset based on the gain compensation model to generate a gain compensation value for the switching frequency channel;

[0011] A sound quality compensation module is used to identify whether the gain compensation value exceeds a preset sound quality protection threshold. If it exceeds, the gain compensation value is processed and compensated in real time based on a smoothing factor to generate a third data set;

[0012] An audio optimization module is used to update the gain compensation model based on the least mean square adaptive filtering algorithm and the loudness change trend parameter of the third data set, and process the third data set based on the updated gain compensation model to generate a final audio data set, and the final audio data set is used to output an optimized radio and television microwave transmission signal.

[0013] In one embodiment, the initial audio acquisition and processing module of an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0014] S11: Synchronously acquire audio signals of the current frequency channel through a multi-channel analog-to-digital converter based on a preset sampling frequency;

[0015] S12: Calculate parameters of the audio signal based on an audio loudness standard algorithm to generate original loudness parameters of the current frequency channel, and calculate the signal-to-noise ratio and group delay characteristics of the current frequency channel based on the original loudness parameters to generate a channel state description value;

[0016] S13: Process the original loudness parameters and the channel state description value to generate a first data set.

[0017] In one embodiment, the conversion audio acquisition and processing module of an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0018] S21: When a frequency switching signal is detected, start a parallel analog-to-digital converter to acquire audio signals of the switching frequency channel, and generate a second data set containing the mean value of the loudness parameters and the state description value of the switching frequency channel;

[0019] S22: Calculate the difference between the mean value of the original loudness parameters in the first dataset and the mean value of the loudness parameters in the second dataset to generate an initial parameter offset;

[0020] S23: Normalize and correct the initial parameter offset according to the interference level in the channel status description value of the first dataset to generate a loudness offset.

[0021] In one embodiment, the feature extraction module of an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0022] S31: Divide the audio signals in the first dataset and the second dataset into a preset number of time segments, and perform a fast Fourier transform on each time segment to generate a frequency-domain energy distribution;

[0023] S32: Extract the energy ratios of the low-frequency band, the middle-frequency band, and the high-frequency band in the frequency-domain energy distribution to generate an initial feature vector with unnormalized frequency distribution;

[0024] S33: Perform a normalization process on the initial feature vector to generate a frequency distribution feature vector.

[0025] In one embodiment, the model construction module of an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0026] S41: Based on the original loudness parameters, the loudness offset, and the frequency distribution feature vector in the first dataset, construct a polynomial regression model to generate a gain compensation model;

[0027] S42: Based on the gain compensation model, perform a supplementary calculation on the loudness offset and the frequency distribution feature vector to generate an initial compensation value. The calculation formula for the initial compensation value is as follows:

[0028]

[0029] Where, G initial is the initial compensation value, ΔR is the loudness offset, F i is the i-th element of the frequency distribution feature vector (i = 1, 2, 3), corresponding to the energy ratios of the low-frequency, middle-frequency, and high-frequency bands respectively, k0 is the constant term, k1 is the coefficient of the gain compensation model, and w i is the first-order term coefficient corresponding to the element of the frequency distribution feature vector F i ;

[0030] S43: Fuse and calculate the initial compensation value with the gain coefficient in the channel status description value to generate a gain compensation value for the switching frequency channel. The calculation formula for the gain compensation value is:

[0031]

[0032] Among them, G target is the gain compensation value, T is the gain coefficient in the channel state description value, and T ref is the reference gain coefficient, R ref is the reference loudness value, and β is the fusion weight factor.

[0033] In one embodiment, the sound quality compensation module of an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0034] S51: Identify whether the gain compensation value exceeds the preset sound quality protection threshold. If it exceeds, perform a weighted average calculation on the gain compensation value based on the historical gain compensation value and the smoothing factor to generate a smoothed gain value;

[0035] S52: Based on the smoothed gain value, adjust the amplitude of the audio signal in the second data set through a digital signal processor to generate a compensated audio signal;

[0036] S53: Integrate and store the compensated audio signal with the real-time calculated loudness change trend parameter to generate a third data set.

[0037] In one embodiment, the audio optimization module of an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0038] S61: Based on the loudness change trend parameter in the third data set, calculate the model error signal using the least mean square adaptive filtering algorithm to generate a weight update amount;

[0039] S62: Iteratively adjust the weight parameters of the gain compensation model according to the weight update amount to generate an updated gain compensation model. The update formula for the weight parameters is:

[0040]

[0041] k new = k old + μ * ∈ * ΔR norm

[0042] Among them, and respectively represent the new value and the old value before update of the frequency-related weight parameter w i in the gain compensation model, k new and k old are respectively the new value and the old value before update of the loudness-related parameter k in the gain compensation model, μ is the learning rate, F i is the i-th element of the frequency distribution feature vector, and ΔR normis the normalized loudness offset;

[0043] S63: Perform secondary compensation calculation on the audio signals of the third data set based on the updated gain compensation model to generate an optimized frequency-domain energy distribution;

[0044] S64: Convert the optimized frequency-domain energy distribution into a time-domain signal to generate a final audio data set.

[0045] In a second aspect, the present invention provides an adaptive frequency optimization method in radio and television microwave transmission. This method is applied to an adaptive frequency optimization system in radio and television microwave transmission provided in any of the above embodiments. The method includes the following steps:

[0046] S710: Perform real-time acquisition and preprocessing on the current frequency channel based on a preset sampling frequency to generate a first data set containing original loudness parameters and channel status description values;

[0047] S720: Acquire the audio signals of the switched frequency channel based on a frequency switching signal to generate a second data set, and perform loudness calculation on the second data set based on the first data set to obtain a loudness offset;

[0048] S730: Extract features from the first data set and the second data set based on the fast Fourier transform algorithm to generate a frequency distribution feature vector;

[0049] S740: Construct a gain compensation model based on the first data set, the loudness offset, and the frequency distribution feature vector, and perform compensation calculation on the loudness offset based on the gain compensation model to generate a gain compensation value for the switched frequency channel;

[0050] S750: Identify whether the gain compensation value exceeds a preset sound quality protection threshold. If it exceeds, process and perform real-time compensation on the gain compensation value based on a smoothing factor to generate a third data set;

[0051] S760: Update the gain compensation model based on the least mean square adaptive filtering algorithm and the loudness change trend parameters of the third data set, and process the third data set based on the updated gain compensation model to generate a final audio data set. The final audio data set is used to output an optimized radio and television microwave transmission signal.

[0052] In a third aspect, the present application provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements any of the above adaptive frequency optimization methods in radio and television microwave transmission.

[0053] Fourthly, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements any one of the above-mentioned adaptive frequency optimization methods in radio and television microwave transmission.

[0054] In summary, an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention uses an initial audio acquisition and processing module to perform real-time acquisition and preprocessing on the current frequency channel at a preset sampling frequency, generating a first data set containing original loudness parameters and channel status description values. When the frequency is switched, the conversion audio acquisition and processing module acquires the audio signal of the switched frequency channel and calculates the loudness offset based on the first data set to ensure the sound quality stability of the audio signal during frequency switching. The feature extraction module uses the fast Fourier transform algorithm to perform frequency-domain analysis on the data set and extracts the frequency distribution feature vector. The model construction module constructs a gain compensation model based on the data set, loudness offset, and frequency distribution feature vector, and performs compensation calculation on the loudness offset to generate the gain compensation value for the switched frequency channel. The sound quality compensation module monitors in real time whether the gain compensation value exceeds a preset sound quality protection threshold. If it exceeds, it is processed through a smoothing factor to avoid sound quality deterioration or loudness jump. The audio optimization module dynamically updates the gain compensation model based on the least mean square adaptive filtering algorithm and loudness change trend parameters to ensure that the output radio and television microwave transmission signal has high fidelity and stability. Through real-time acquisition, feature extraction, model construction, and optimization, the present invention realizes the dynamic optimization of audio signals in radio and television microwave transmission. This system can effectively cope with sudden interference and rapid switching requirements, improve sound quality stability, and at the same time optimize resource utilization, providing users with a high-quality audio and video transmission experience.

[0055] Under complex electromagnetic environments and dynamic service requirements, this system effectively solves the limitations of traditional frequency optimization methods in dynamic scenarios by real-time optimizing frequency resources and audio and video signals, improving the efficiency and reliability of radio and television microwave transmission, and providing users with higher-quality audio and video services.

[0056] For better understanding and implementation, the present invention will be described in detail below with reference to the accompanying drawings. Description of the Drawings

[0057] Figure 1 It is a schematic structural diagram of an adaptive frequency optimization system in radio and television microwave transmission provided by an embodiment of the present application;

[0058] Figure 2 It is a schematic flowchart of generating a gain compensation value in the model construction module of an adaptive frequency optimization system in radio and television microwave transmission provided by an embodiment of the present application;

[0059] Figure 3Schematic diagram of the process of generating the final audio dataset in the audio optimization module of an adaptive frequency optimization system in radio and television microwave transmission provided by an embodiment of the present application;

[0060] Figure 4 Schematic diagram of the process of an adaptive frequency optimization method in radio and television microwave transmission provided by another embodiment of the present application. Detailed implementation manners

[0061] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Preferred embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the understanding of the disclosure of the invention more thorough and comprehensive.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the description of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0063] Please refer to Figure 1 , which shows an adaptive frequency optimization system in radio and television microwave transmission provided by an embodiment of the present application. In this embodiment, the system is exemplified by being applied to a terminal. It can be understood that the system can also be applied to a server and can also be applied to a system including a terminal and a server, and is realized through the interaction between the terminal and the server. This embodiment is applicable to multi-platform advertisement content delivery, cross-media distribution of film and television content, VR / AR multi-device compatibility, and multi-terminal display of e-commerce live broadcasts. As Figure 1 shown, the present invention provides an adaptive frequency optimization system 100 in radio and television microwave transmission. The system is configured with the following modules:

[0064] An initial audio acquisition and processing module 110, configured to perform real-time acquisition and preprocessing on the current frequency channel based on a preset sampling frequency, and generate a first dataset including original loudness parameters and channel status description values.

[0065] Specifically, the initial audio acquisition and processing module 110 performs real-time digital acquisition of the analog signal of the current frequency channel through a high-speed analog-to-digital converter (ADC) at a preset sampling frequency (such as 44.1 kHz or 96 kHz). The selection of the sampling frequency needs to satisfy the Nyquist sampling theorem to ensure the integrity and aliasing-free of the signal spectrum. In the preprocessing stage, a digital filter (such as a Butterworth filter or a Chebyshev filter) is used to perform band-pass filtering on the acquired signal to remove out-of-band noise and interference components, and at the same time, automatic gain control (AGC) is performed to stabilize the dynamic range of the signal amplitude. The original loudness parameter is extracted through a loudness calculation algorithm (such as the ITU-R BS.1770-4 standard), and the channel state description values include the signal-to-noise ratio (SNR), bit error rate (BER), multipath fading coefficient, etc., which are calculated in real time by the signal quality evaluation module. Preferably, the preprocessing algorithm can adopt adaptive filtering technology to dynamically adjust the filter coefficients to cope with changes in environmental noise. The loudness parameter calculation adopts a segmented cumulative average algorithm, and the channel state description value comprehensively evaluates the signal quality through multi-dimensional feature extraction (such as short-time energy, zero-crossing rate, sub-band power).

[0066] Specifically, the data acquisition device may include a radio frequency front-end module (RF Front-End), an analog-to-digital converter (ADC), a digital signal processor (DSP), and an embedded controller. The RF front-end is responsible for signal down-conversion and amplification, the ADC converts the analog signal into a digital signal, the DSP runs the preprocessing algorithm, and the embedded controller is responsible for parameter configuration and status monitoring. The software part uses a real-time operating system (RTOS) to ensure low latency in data acquisition and processing.

[0067] The switched audio acquisition and processing module 120 is used to acquire the audio signal of the switched frequency channel based on the frequency switching signal, generate a second data set, and perform loudness calculation on the second data set based on the first data set to obtain a loudness offset.

[0068] Specifically, when the conversion audio acquisition and processing module 120 detects a frequency switching signal, it will immediately acquire the audio signal of the switched frequency channel to generate a second data set. The conversion audio acquisition and processing module 120 can adopt synchronous sampling technology to ensure that the signals of the new frequency channel can be consistent with the signals of the original channel in terms of time and phase. The synchronous sampling is achieved through precise clock control, and the clock signal is provided by a highly stable crystal oscillator to ensure the stability and accuracy of the sampling frequency. The acquired audio signals are then stored in the buffer, waiting for further processing. At the same time, based on the previously generated first data set, the system calculates the loudness of the second data set. The loudness calculation can adopt the international standard loudness measurement algorithm, such as the ITU-R BS.1770-4 standard, which can accurately reflect the human ear's perception of sound loudness. By comparing the loudness parameters in the two data sets, the system can calculate the loudness offset, that is, the volume difference between the two frequency channels. Preferably, in terms of hardware, the audio acquisition unit adopts a high-precision ADC, and its sampling frequency is strictly synchronized with the current frequency channel to ensure the phase consistency of the signals. In terms of software, the audio signal acquisition algorithm adopts a dynamic buffer management mechanism, which can effectively cope with the rapid acquisition and storage requirements of burst signals, ensuring the integrity and real-time performance of the data.

[0069] The feature extraction module 130 is used to extract features from the first data set and the second data set based on the fast Fourier transform algorithm to generate a frequency distribution feature vector.

[0070] Specifically, in order to deeply analyze the spectral characteristics of the signal, the feature extraction module 130 extracts features from the first data set and the second data set based on the Fast Fourier Transform (FFT) algorithm. The FFT algorithm is an efficient frequency-domain analysis tool that can quickly convert a time-domain signal into a frequency-domain signal, thereby revealing the frequency distribution characteristics of the signal. In this process, the feature extraction module 130 first performs windowing processing on the first data set and the second data set to reduce the spectral leakage effect. Preferably, the window function can use a Hanning window, which can effectively balance the main lobe width and sidelobe attenuation to ensure the accuracy of spectral analysis. The data after windowing processing is sent to the FFT processing unit, which is implemented based on an embedded GPU and can perform real-time FFT calculations on 1024-point data to generate a frequency distribution feature vector. The feature vector contains key information such as the amplitude spectrum, phase spectrum, and energy distribution of the signal, which is crucial for the subsequent construction of the gain compensation model. The feature extraction process can also adopt a sliding window technique, with the window length set to 10 milliseconds and the step size to 5 milliseconds, so as to improve the real-time performance of processing while ensuring the analysis accuracy. In terms of hardware, the FFT processing unit is connected to the feature extraction module 130 through a high-speed PCIe bus to ensure the efficiency and low latency of data transmission. In terms of software implementation, the feature extraction algorithm adopts a parallel computing architecture, supports multi-threaded processing, and can make full use of the computing power of the embedded GPU to achieve fast processing of large-scale data.

[0071] The model construction module 140 is used to construct a gain compensation model based on the first data set, the loudness offset, and the frequency distribution feature vector, and perform compensation calculations on the loudness offset based on the gain compensation model to generate the gain compensation value for the switching frequency channel.

[0072] Specifically, the construction of the gain compensation model can be achieved through machine learning algorithms or mathematical optimization models to map the input data to the target output for realizing the compensation calculation of the loudness offset. The training process of the model needs to introduce historical data and real-time data, and continuously adjust the model parameters through iterative optimization algorithms to improve the accuracy and robustness of the model. Subsequently, based on the gain compensation model, the model construction module 140 performs compensation calculations on the loudness offset to generate the gain compensation value for the switching frequency channel. The calculation of the gain compensation value needs to consider the dynamic range of the signal, the frequency response characteristics, and the requirements for sound quality protection to ensure that the compensated audio signal meets the industry standards in terms of sound quality and stability. The gain compensation model adopts an adaptive neural network structure and dynamically adjusts the weights through the Backpropagation algorithm to adapt to the non-linear mapping relationship between different frequency channels.

[0073] The sound quality compensation module 150 is used to identify whether the gain compensation value exceeds a preset sound quality protection threshold. If it exceeds, the gain compensation value is processed and compensated in real time based on a smoothing factor to generate a third data set.

[0074] Specifically, to ensure sound quality protection, the sound quality compensation module 150 needs to identify whether the gain compensation value exceeds a preset sound quality protection threshold. The preset sound quality protection threshold is set based on industry standards (such as the sound quality specifications of the European Broadcasting Union EBU) and user experience requirements, and a reasonable range is determined through experiments and data analysis. If the gain compensation value exceeds the preset threshold, the gain compensation value is processed and compensated in real time based on a smoothing factor to generate a third data set. The introduction of the smoothing factor can be optimized through mathematical models and algorithms to ensure that the compensated audio signal maintains a smooth transition in loudness changes, avoiding sound quality deterioration or loudness jumps. The smoothing process can adopt the Exponentially Weighted Moving Average (EWMA) algorithm, and by dynamically adjusting the smoothing coefficient, ensure that the loudness change at the moment of signal switching conforms to the human ear's auditory perception characteristics.

[0075] The audio optimization module 160 is used to update the gain compensation model based on the least mean square adaptive filtering algorithm and the loudness change trend parameters of the third data set, and process the third data set based on the updated gain compensation model to generate a final audio data set, which is used to output an optimized radio and television microwave transmission signal.

[0076] Specifically, the audio optimization module 160 uses the real-time loudness change trend parameters in the third data set to construct an adaptive filtering iterator (Least Mean Square, LMS) with the least mean square error as the criterion, and dynamically corrects the weight coefficients of the gain compensation model through an error feedback mechanism. During the model update process, a sliding time window strategy can be adopted to balance the contribution of historical data and new input data, ensuring the fast convergence and steady-state accuracy of the model when the environment changes suddenly. The optimized model performs secondary compensation on the third data set to eliminate residual linear distortion and non-stationary interference, generating a final audio data set that meets the target loudness tolerance. Preferably, the processed signal can be injected into the microwave transmission link through a bitstream recombination and synchronization encapsulation module, and channel coding and error correction verification information can be embedded at the same time to ensure the robustness of the signal during transmission. The final output signal is radiated through a power amplifier and a directional antenna array to complete end-to-end adaptive frequency optimization and sound quality protection. It should be noted that the generation of the final audio data set requires a multi-level verification mechanism, including loudness consistency verification, frequency response verification, and dynamic range verification, to ensure that the output signal meets industry standards and user requirements in multiple dimensions.

[0077] In summary, an adaptive frequency optimization system in radio and television microwave transmission provided by the present invention uses an initial audio acquisition and processing module to perform real-time acquisition and preprocessing of the current frequency channel at a preset sampling frequency, generating a first data set containing original loudness parameters and channel status description values. When the frequency is switched, the conversion audio acquisition and processing module acquires the audio signal of the switched frequency channel and calculates the loudness offset based on the first data set to ensure the sound quality stability of the audio signal during frequency switching. The feature extraction module performs frequency domain analysis on the data set using the fast Fourier transform algorithm to extract frequency distribution feature vectors. The model construction module constructs a gain compensation model based on the data set, loudness offset, and frequency distribution feature vectors, and performs compensation calculation on the loudness offset to generate a gain compensation value for the switched frequency channel. The sound quality compensation module monitors in real time whether the gain compensation value exceeds a preset sound quality protection threshold. If it exceeds, it is processed through a smoothing factor to avoid sound quality deterioration or loudness jump. The audio optimization module dynamically updates the gain compensation model based on the least mean square adaptive filtering algorithm and loudness change trend parameters to ensure that the output radio and television microwave transmission signal has high fidelity and stability. The present invention realizes the dynamic optimization of audio signals in radio and television microwave transmission through real-time acquisition, feature extraction, model construction, and optimization. This system can effectively cope with sudden interference and rapid switching requirements, improve sound quality stability, and optimize resource utilization at the same time, providing users with a high-quality audio and video transmission experience.

[0078] Under complex electromagnetic environments and dynamic service requirements, this system effectively solves the limitations of traditional frequency optimization methods in dynamic scenarios by optimizing frequency resources and audio and video signals in real time, improving the efficiency and reliability of radio and television microwave transmission, and providing users with higher-quality audio and video services.

[0079] In one embodiment, the initial audio acquisition and processing module 110 of an adaptive frequency optimization system 100 in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0080] S11: Synchronously acquire the audio signal of the current frequency channel through a multi-channel analog-to-digital converter based on a preset sampling frequency.

[0081] Specifically, to achieve the adaptive optimization of frequency channels, the initial audio acquisition and processing module 110 needs to synchronously acquire the audio signals of the current frequency channel through a multi-channel analog-to-digital converter. The synchronization mechanism of the multi-channel analog-to-digital converter (ADC) is controlled by an accurate clock signal to ensure that the sampling moments of each channel are strictly aligned, so as to avoid phase deviation and frequency offset caused by asynchronous sampling. The selection of the preset sampling frequency needs to comprehensively consider the signal bandwidth, the highest frequency component, and the requirements of the Nyquist sampling theorem to ensure that the acquired digital signal can completely restore the original audio signal. The quantization accuracy of the ADC needs to be optimized according to the signal dynamic range and noise floor, and usually 16-bit or 24-bit quantization is adopted to balance the data volume and signal fidelity.

[0082] S12: Calculate the parameters of the audio signal based on the audio loudness standard algorithm to generate the original loudness parameters of the current frequency channel, and calculate the signal-to-noise ratio and group delay characteristics of the current frequency channel based on the original loudness parameters to generate a channel state description value.

[0083] Specifically, the initial audio acquisition and processing module 110 calculates the parameters of the acquired audio signal through the audio loudness standard algorithm to generate the original loudness parameters of the current frequency channel. The audio loudness standard algorithm adopts the ITU-R BS.1770 standard recommended by the International Telecommunication Union (ITU). This standard accurately measures the loudness level of the audio signal through weighted filtering, true peak detection, and integral operation. The calculation of the original loudness parameters needs to consider the auditory characteristics of the human ear. By simulating the sensitivity differences of the human ear to different frequencies through the A-weighting or C-weighting filter, it is ensured that the measurement results are consistent with the subjective auditory perception. In addition, based on the original loudness parameters, the signal-to-noise ratio (SNR) and group delay characteristics of the current frequency channel also need to be calculated. The calculation of the signal-to-noise ratio evaluates the purity and transmission quality of the signal by measuring the ratio of the signal power to the noise power. The calculation of the group delay characteristics analyzes the propagation delay differences of the signal at different frequency components to ensure the phase consistency of the audio signal and avoid the deterioration of the sound quality caused by group delay distortion.

[0084] The generation of the channel state description value needs to comprehensively consider parameters such as the signal-to-noise ratio, bit error rate (BER), multipath fading characteristics, and frequency selective fading of the signal. The channel state description value generates a quantization index through real-time monitoring and analysis of these parameters to evaluate the transmission performance and stability of the current frequency channel. For example, the calculation of the signal-to-noise ratio can adopt the sliding window technique, and by dynamically adjusting the window size, the short-term changes and long-term trends of the signal can be captured. The calculation of the bit error rate needs to combine the error correction ability of forward error correction coding (FEC) to evaluate the bit error situation in actual transmission. The multipath fading characteristics evaluate the propagation characteristics of the signal in a complex electromagnetic environment by measuring the power spectral density and delay spread of the signal

[0085] S13: Process the original loudness parameter and the channel status description value to generate a first data set.

[0086] Specifically, the initial audio acquisition and processing module 110 integrates data from different sources to generate a unified data format, and integrates the original loudness parameter and the channel status description value into a structured data set. The data fusion adopts a weighted average algorithm to ensure the integrity and consistency of the data. The first data set is stored in floating-point format and contains key information such as loudness parameters, signal-to-noise ratio, and group delay, providing data support for subsequent frequency optimization decisions. Preferably, the generation of the first data set can use timestamp alignment technology to ensure that the original loudness parameter and the channel status description value are strictly synchronized in time for subsequent processing and analysis. In addition, the first data set also needs to contain auxiliary information such as sampling time, frequency channel identifier, and environmental parameters (such as temperature, humidity, etc.) to support more comprehensive signal analysis and optimization decisions.

[0087] In one embodiment, the converted audio acquisition and processing module 120 of the adaptive frequency optimization system 100 provided by the present invention for radio and television microwave transmission is used to perform the following steps:

[0088] S21: When a frequency switching signal is detected, start a parallel analog-to-digital converter to collect the audio signal of the switched frequency channel, and generate a second data set containing the mean loudness parameter and the status description value of the switched frequency channel.

[0089] Specifically, when a frequency switching signal is detected, the converted audio acquisition and processing module 120 starts a parallel analog-to-digital converter to collect the audio signal of the switched frequency channel. The start of the parallel analog-to-digital converter is based on a fast response mechanism to ensure seamless acquisition of the audio signal at the moment of frequency switching, avoiding signal loss or interruption. The sampling frequency of the parallel analog-to-digital converter needs to be consistent with the sampling frequency of the current frequency channel to ensure data consistency and comparability. The converted audio acquisition and processing module 120 calculates the parameters of the collected audio signal through the audio loudness standard algorithm to generate the mean loudness parameter of the switched frequency channel, and calculates the status description value of the switched frequency channel based on the mean loudness parameter to generate a second data set. The calculation of the status description value needs to comprehensively consider parameters such as signal-to-noise ratio, bit error rate, multipath fading characteristics, and frequency selective fading. By real-time monitoring and analyzing these parameters, a quantization index is generated to evaluate the transmission performance and stability of the switched frequency channel.

[0090] S22: Calculate the difference between the mean value of the original loudness parameter in the first data set and the mean value of the loudness parameter in the second data set to generate an initial parameter offset.

[0091] Specifically, based on the spatio-temporal alignment result of the first data set and the second data set, the conversion audio acquisition and processing module 120 extracts the mean sequence of loudness parameters of the current channel and the switched channel respectively, and eliminates the phase deviation caused by transmission delay through the dynamic time warping algorithm to achieve accurate time-domain matching of the two-channel data. Subsequently, the conversion audio acquisition and processing module 120 performs frame-by-frame difference calculation on the matched loudness parameter sequence to generate an initial parameter offset matrix. Preferably, to suppress instantaneous noise interference, median filtering and moving average filtering can be used to jointly smooth the offset, and the historical offset trend prediction value is introduced to correct the current calculation result. The finally generated initial parameter offset matrix contains two-dimensional time-frequency correlation characteristics, providing multi-dimensional input for subsequent correction.

[0092] S23: Normalize and correct the initial parameter offset according to the interference level in the channel state description value of the first data set to generate a loudness offset.

[0093] Specifically, the conversion audio acquisition and processing module 120 establishes a non-linear mapping relationship between the interference intensity and the offset correction coefficient based on the interference level parameter in the channel state description value of the first data set. The interference level is comprehensively determined according to the background noise power, co-channel interference density, and multipath fading depth, and is divided into three level intervals: low, medium, and high. The conversion audio acquisition and processing module 120 designs different correction strategies for different interference levels: linear proportional scaling can be used in low-interference scenarios to retain the original offset trend; frequency-domain selective weighting can be enabled in medium-interference scenarios to suppress the offset components in the interference overlapping frequency bands; a logarithmic compression function can be introduced in high-interference scenarios to limit the amplitude of extreme offsets and superimpose the trend prediction value to offset systematic biases. During the correction process, the conversion audio acquisition and processing module 120 real-time verifies the correction effect through a closed-loop feedback mechanism. If the residual exceeds the limit, the model parameter dynamic reconfiguration is triggered. The finally generated loudness offset is subjected to time-frequency joint smoothing processing to form a standardized output.

[0094] In one embodiment, the feature extraction module 130 of the adaptive frequency optimization system 100 provided by the present invention for radio and television microwave transmission is used to perform the following steps:

[0095] S31: Divide the audio signals in the first data set and the second data set into a preset number of time segments, and perform fast Fourier transform on each time segment to generate the frequency-domain energy distribution.

[0096] Specifically, the feature extraction module 130 divides the audio signals in the first data set and the second data set into frames according to preset time segments, and can use a fixed-length Hanning window function to window and segment the continuous signal. The window length is set to 20ms to ensure the effective capture of the transient characteristics of the audio. At the same time, a 50% overlap rate is used between adjacent windows to reduce spectrum leakage. Subsequently, the feature extraction module 130 applies a fast Fourier transform (FFT) to each time segment, calculates the complex spectrum, and takes its modulus square to generate a power spectral density to characterize the energy distribution of the signal in the frequency domain. All time segments are synchronously processed through a multi-threaded parallel computing framework to generate frequency domain energy distribution data containing a time-frequency energy matrix, whose dimension is the product of the number of time segments and the number of frequency points, forming the basic input for subsequent feature extraction.

[0097] S32: extracting energy proportions of low-frequency bands, mid-frequency bands, and high-frequency bands of the frequency domain energy distribution, and generating an initial unnormalized eigenvector of the frequency distribution.

[0098] Specifically, the division of low frequency band, medium frequency band and high frequency band is based on the bandwidth and frequency characteristics of the signal. Usually, the low frequency band covers 0 to 1 / 3 of the signal bandwidth, the medium frequency band covers 1 / 3 to 2 / 3, and the high frequency band covers 2 / 3 to the upper limit of the signal bandwidth. The calculation of the energy proportion is achieved by integrating the energy of each frequency band and dividing it by the total energy to ensure that the sum of the energy proportions of each frequency band is 1. Generate an initial feature vector whose frequency distribution is not normalized, which contains three elements, corresponding to the energy proportions of the low frequency band, the medium frequency band and the high frequency band. Specifically, the feature extraction module 130 performs frequency band accumulation on the power spectral density of each time segment and calculates the proportion of the energy of each frequency band to the total energy of the whole frequency band. The energy proportion of the low frequency band reflects the fundamental frequency and resonance peak structure of the audio signal, the medium frequency band represents the speech clarity and the harmonic components of the instrument, and the high frequency band reflects the brightness of the timbre and the noise characteristics. The energy proportions of the three frequency bands are arranged in time series to form an initial feature vector, whose dimension is the number of time segments × 3. According to the characteristics of broadcast signals, the sub-band energy fluctuation rate is additionally introduced as a supplementary feature to calculate the standard deviation of the energy of each frequency band between adjacent time segments, thereby enhancing the ability to describe signal stability.

[0099] S33: Normalize the initial feature vector to generate a frequency distribution feature vector.

[0100] Specifically, the feature extraction module 130 normalizes each element (energy ratio of low, medium, and high frequency bands) in the initial feature vector to a preset range (such as 0 - 1). The normalization method can adopt maximum normalization, min - max normalization, or Z - score normalization to ensure the consistency of the numerical range of the feature vector, facilitating subsequent model processing. The normalized feature vector is used as the frequency distribution feature vector to describe the frequency distribution characteristics of the audio signal. The dimension and numerical range of the feature vector are optimized to ensure efficient processing in subsequent models.

[0101] In one embodiment, as Figure 2 shown, the model construction module 140 of an adaptive frequency optimization system 100 in radio and television microwave transmission provided by the present invention is used to perform the following steps:

[0102] S41: Based on the original loudness parameters, loudness offset, and frequency distribution feature vector in the first dataset, construct a polynomial regression model to generate a gain compensation model.

[0103] Specifically, the model input includes the time - domain loudness parameter sequence, cross - channel loudness offset matrix, and the normalized frequency - domain feature vector. The least - squares method is used to fit the polynomial regression equation to describe the non - linear mapping relationship between the input features and the target compensation value. During the model training process, the Ridge Regression optimization algorithm can be used to suppress overfitting, and the regularization coefficient is dynamically adjusted through cross - validation to ensure the generalization ability of the model in a complex electromagnetic environment. After training, the model coefficients (including the coefficients of the first - order term, second - order term, and cross - term) are solidified into the storage unit of the embedded processor for real - time compensation calculation.

[0104] S42: Based on the gain compensation model, perform supplementary calculations on the loudness offset and frequency distribution feature vector to generate an initial compensation value.

[0105] Specifically, the calculation formula for the initial compensation value is as follows:

[0106]

[0107] where G initial is the initial compensation value, ΔR is the loudness offset, F i is the i - th element (i = 1, 2, 3) of the frequency distribution feature vector, corresponding to the energy ratio of the low, medium, and high frequency bands respectively, k0 is the constant term, k1 is the coefficient of the gain compensation model, w i is the first - order term coefficient corresponding to the element of the frequency distribution feature vector F i ; the model parameters k0, k1, w i and w ijObtained through training with historical data to ensure that the initial compensation value can accurately reflect the dynamic characteristics of the audio signal.

[0108] During the calculation process, for the non-linear distortion characteristics in the high-frequency band, a piecewise linear approximation strategy can be introduced: when the proportion of high-frequency energy exceeds the threshold, quadratic term compensation is automatically enabled to suppress the harmonic intermodulation effect; at the same time, dynamic quantization step adjustment is adopted for the low-frequency band to ensure the balance between compensation accuracy and calculation efficiency. The calculation result of the initial compensation value is temporarily stored in a circular buffer for subsequent calls by the fusion module.

[0109] S43: Fuse and calculate the initial compensation value with the gain coefficient in the channel status description value to generate the gain compensation value for the switching frequency channel.

[0110] Specifically, the calculation formula for the gain compensation value is:

[0111]

[0112] Where, G target is the gain compensation value, T is the gain coefficient in the channel status description value, T ref is the reference gain coefficient, R ref is the reference loudness value, and β is the fusion weight factor.

[0113] The above-mentioned adaptive frequency optimization system in radio and television microwave transmission can achieve precise gain compensation for audio signals during frequency switching by constructing a polynomial regression model and combining the loudness offset and frequency distribution feature vector. First, through the construction of the polynomial regression model, the system can capture the complex non-linear relationship between the loudness offset and the frequency distribution feature vector, thereby improving the adaptability and accuracy of the compensation model. Second, the calculation of the initial compensation value comprehensively considers the interaction between the loudness offset and the frequency distribution feature vector to ensure that the compensation value can dynamically reflect the characteristic changes of the audio signal. Finally, by fusing and calculating the initial compensation value with the gain coefficient in the channel status description value, the system can achieve adaptive gain compensation under different channel conditions, ensuring the sound quality stability and transmission efficiency of the audio signal, so as to quickly respond to the frequency switching demand in a dynamic environment, effectively reduce the sound quality degradation and loudness jump phenomena, and at the same time improve the robustness and user experience of the system, especially suitable for high-density urban environments and emergency broadcast scenarios.

[0114] In one embodiment, the sound quality compensation module 150 of the adaptive frequency optimization system 100 provided by the present invention for radio and television microwave transmission is used to perform the following steps:

[0115] S51: Identify whether the gain compensation value exceeds the preset sound quality protection threshold. If it exceeds, calculate the weighted average of the gain compensation value based on the historical gain compensation value and the smoothing factor to generate a smoothed gain value.

[0116] Specifically, the sound quality compensation module 150 monitors the time-domain change curve of the gain compensation value in real time and makes multi-dimensional comparisons with the preset sound quality protection threshold. Among them, the sound quality protection threshold includes a static threshold and a dynamic threshold: the static threshold sets an absolute value constraint (such as ±3dB) based on the loudness tolerance standard of the International Telecommunication Union (ITU-R BS.645); the dynamic threshold generates an adaptive threshold according to the statistical characteristics of the historical compensation data (such as the mean and variance within a sliding window). When the gain compensation value exceeds any threshold, start the dynamic smoothing mechanism: extract the historical gain compensation sequence within the most recent 100 ms, and use the Exponentially Weighted Moving Average (EWMA) algorithm to calculate the smoothed gain value. Specifically, the smoothing factor is dynamically adjusted according to the channel coherence time. For example, it is set to 0.8 in a fast time-varying channel (coherence time < 10 ms) to enhance real-time performance, and set to 0.3 in a steady-state channel to improve stability. For continuous over-limit scenarios (such as continuous 5-frame over-limit), enable the emergency attenuation mode, reset the smoothing factor to 0.95 and superimpose a non-linear compression function to forcibly suppress the compensation amplitude. Among them, the calculation formula for the smoothed gain value is:

[0117] G smooth = αG target +(1 - α)G history

[0118] where G smooth is the current smoothed gain value, α is the smoothing factor, and G history is the smoothed gain value of the previous frame.

[0119] S52: Based on the smoothed gain value, adjust the amplitude of the audio signal of the second data set through a digital signal processor to generate a compensated audio signal.

[0120] Specifically, the sound quality compensation module 150 adopts a sub-band gain control strategy, decomposes the audio signal into low, middle, and high frequency bands, and applies band-specific gain coefficients respectively. The low-frequency band gain is directly mapped by the smoothed gain value, and the middle and high-frequency band gains are superimposed with frequency tilt compensation (such as +1dB per octave to improve speech clarity). During the amplitude adjustment process, the sound quality compensation module 150 compensates for the harmonic distortion introduced by the non-linear amplifier through pre-distortion correction, and uses a limiter to ensure that the peak-to-peak value of the output signal does not exceed 90% of the ADC range. The compensated audio signal is reconstructed into a time-domain waveform through an anti-aliasing filter to generate a compensated audio signal that conforms to the target loudness curve.

[0121] S53: Integrate and store the compensated audio signal with the loudness change trend parameter calculated in real time to generate a third data set.

[0122] Specifically, the loudness change trend parameter is obtained by monitoring the loudness change of the compensated audio signal in real time and can be calculated using a moving average algorithm to ensure the smoothness and reliability of the trend parameter. The sound quality compensation module 150 performs time alignment on the third data set through timestamp alignment technology to ensure that the compensated audio signal and the loudness change trend parameter are strictly synchronized in time. The third data set will be used for subsequent model updates and signal optimization to ensure that the system can continuously adapt to dynamic environmental changes and maintain high-quality audio transmission performance. Through smoothing processing and real-time adjustment, the sound quality compensation module 150 can effectively protect the sound quality during the frequency switching process, reduce loudness jumps and sound quality degradation phenomena, and improve the system's adaptability and user experience.

[0123] In one embodiment, as Figure 3 shown, the audio optimization module 160 of the adaptive frequency optimization system 100 provided by the present invention for radio and television microwave transmission is used to perform the following steps:

[0124] S61: Based on the loudness change trend parameter of the third data set, use the least mean square adaptive filtering algorithm to calculate the model error signal and generate a weight update amount.

[0125] Specifically, the least mean square adaptive filtering algorithm minimizes the error between the model output and the actual signal by adjusting the filter coefficients in real time, thereby realizing the adaptive optimization of the model. The error signal is calculated by comparing the difference between the model output and the actual loudness change trend parameter to ensure that the error signal can accurately reflect the prediction deviation of the model. The weight update amount is generated based on the error signal and the learning rate, and the model parameters are iteratively adjusted to improve the accuracy and robustness of the model. Among them, the calculation formula of the error signal ∈ is:

[0126] ∈ = R rarget -R estimated

[0127] where R target is the target loudness value in the third data set, and R estimated is the loudness value predicted by the gain compensation model.

[0128] S62: Iteratively adjust the weight parameters of the gain compensation model according to the weight update amount to generate an updated gain compensation model.

[0129] Specifically, the update formula for the weight parameter is:

[0130]

[0131] k new= k old + μ * ∈ * ΔR norm

[0132] wherein, and respectively represent the new value and the old value of the frequency - related weight parameter w in the gain compensation model, k i after update and before update, k new and k old respectively are the new value and the old value of the loudness - related parameter k in the gain compensation model after update and before update, μ is the learning rate, F i is the i - th element of the frequency distribution feature vector, ΔR norm is the normalized loudness offset, which is achieved by mapping the loudness offset to a standardized range to ensure the consistency and comparability of parameter updates.

[0133] S63: Perform secondary compensation calculation on the audio signals of the third data set based on the updated gain compensation model to generate an optimized frequency - domain energy distribution.

[0134] Specifically, the secondary compensation calculation further optimizes the loudness and frequency characteristics of the audio signals by applying the updated model parameters, ensuring the balance and stability of the signals in different frequency bands. The optimized frequency - domain energy distribution improves the signal transmission efficiency and sound quality by adjusting the energy proportion of each frequency band. Among them, the secondary compensation calculation formula is:

[0135] G optimized = G initial + ΔG update

[0136] wherein, G optimized is the optimized gain compensation value, G initial is the initial compensation value, and ΔG update is the compensation increment calculated based on the weight update amount. The optimized frequency - domain energy distribution is calculated by the following formula:

[0137]

[0138] wherein, E optimized is the optimized frequency - domain energy distribution, and E original is the original frequency - domain energy distribution. This process ensures that the spectral characteristics of the audio signals are further optimized.

[0139] S64: Convert the optimized frequency - domain energy distribution into a time - domain signal to generate the final audio data set.

[0140] Specifically, the audio optimization module 160 converts the optimized frequency-domain energy distribution into a time-domain signal to generate the final audio dataset. The conversion process can adopt the Inverse Fast Fourier Transform (IFFT) algorithm to restore the frequency-domain signal to the time-domain signal. The implementation of the IFFT algorithm needs to ensure calculation accuracy and real-time performance. By optimizing algorithm parameters and hardware acceleration, the processing efficiency is improved. The final audio dataset needs to meet the comprehensive requirements of sound quality stability, transmission efficiency, and user experience, and is used to output the optimized radio and television microwave transmission signal. The generation of the final audio dataset also requires a multi-level verification mechanism, including loudness consistency verification, frequency response verification, and dynamic range verification, to ensure that the output signal meets industry standards and user requirements in multiple dimensions.

[0141] The above-provided adaptive frequency optimization system in radio and television microwave transmission can continuously optimize the gain compensation model through the adaptive filtering algorithm and the iterative update mechanism, improve the transmission quality and stability of the audio signal, and ensure that the optimization effects of the signal in different stages are retained and enhanced through secondary compensation calculation and the conversion from the frequency domain to the time domain. Ultimately, high-fidelity and high-stability radio and television microwave transmission are achieved.

[0142] Based on the same inventive concept, as Figure 4 shown, the present invention provides an adaptive frequency optimization method in radio and television microwave transmission. This method is applied to an adaptive frequency optimization system in radio and television microwave transmission provided in any of the above embodiments. The method includes the following steps:

[0143] S710: Based on a preset sampling frequency, perform real-time acquisition and preprocessing on the current frequency channel to generate a first dataset containing original loudness parameters and channel status description values;

[0144] S720: Based on the frequency switching signal, collect the audio signal of the switched frequency channel to generate a second dataset, and calculate the loudness of the second dataset based on the first dataset to obtain the loudness offset;

[0145] S730: Based on the Fast Fourier Transform algorithm, extract features from the first dataset and the second dataset to generate a frequency distribution feature vector;

[0146] S740: Based on the first dataset, the loudness offset, and the frequency distribution feature vector, construct a gain compensation model, and perform compensation calculation on the loudness offset based on the gain compensation model to generate the gain compensation value of the switched frequency channel;

[0147] S750: Identify whether the gain compensation value exceeds the preset sound quality protection threshold. If it exceeds, process and perform real-time compensation on the gain compensation value based on the smoothing factor to generate a third dataset;

[0148] S760: Update the gain compensation model based on the least mean square adaptive filtering algorithm and the loudness change trend parameters of the third dataset, and process the third dataset based on the updated gain compensation model to generate the final audio dataset, which is used to output the optimized radio and television microwave transmission signal.

[0149] In summary, an adaptive frequency optimization method in radio and television microwave transmission provided by the present invention performs real-time acquisition and preprocessing on the current frequency channel through a preset sampling frequency to generate a first dataset containing original loudness parameters and channel status description values. When the frequency is switched, the audio signal of the switched frequency channel is collected, and the loudness offset is calculated based on the first dataset to ensure the sound quality stability of the audio signal during frequency switching. The fast Fourier transform algorithm is used to perform frequency domain analysis on the dataset to extract the frequency distribution feature vector. A gain compensation model is constructed based on the dataset, loudness offset, and frequency distribution feature vector, and the loudness offset is compensated and calculated to generate the gain compensation value of the switched frequency channel. The gain compensation value is monitored in real time to determine whether it exceeds the preset sound quality protection threshold. If it exceeds, it is processed through a smoothing factor to avoid sound quality deterioration or loudness jump. The gain compensation model is dynamically updated based on the least mean square adaptive filtering algorithm and the loudness change trend parameters to ensure that the output radio and television microwave transmission signal has high fidelity and stability. The present invention realizes the dynamic optimization of the audio signal in radio and television microwave transmission through real-time acquisition, feature extraction, model construction, and optimization. The system can effectively cope with sudden interference and rapid switching requirements, improve the sound quality stability, and at the same time optimize the resource utilization rate, providing users with a high-quality audio and video transmission experience.

[0150] Under complex electromagnetic environments and dynamic service requirements, this method effectively solves the limitations of traditional frequency optimization methods in dynamic scenarios by real-time optimizing frequency resources and audio and video signals, improving the efficiency and reliability of radio and television microwave transmission, and providing users with higher quality audio and video services.

[0151] In one embodiment, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the above-mentioned adaptive frequency optimization method in radio and television microwave transmission.

[0152] In one embodiment, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the above-mentioned adaptive frequency optimization method in radio and television microwave transmission.

[0153] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0154] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present disclosure solution. Those of ordinary skill in the art can understand and implement it without creative effort.

[0155] As mentioned above, the above are only the specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of various changes or substitutions, and these should all be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An adaptive frequency optimization system in microwave transmission of radio and television, characterized in that: include: An initial audio acquisition and processing module, used to perform real-time acquisition and preprocessing of the current frequency channel based on a preset sampling frequency, and generate a first data set including original loudness parameters and channel state description values; A conversion audio acquisition processing module is used to acquire the audio signal of the switching frequency channel based on the frequency switching signal to generate a second data set, and perform loudness calculation on the second data set based on the first data set to obtain a loudness offset; A feature extraction module, used to extract features from the first data set and the second data set based on a fast Fourier transform algorithm to generate a frequency distribution feature vector; A model building module, configured to build a gain compensation model based on the first data set, the loudness offset and the frequency distribution feature vector, and to perform compensation calculation on the loudness offset based on the gain compensation model to generate a gain compensation value for a switching frequency channel; A sound quality compensation module, used to identify whether the gain compensation value exceeds a preset sound quality protection threshold, and if so, process and compensate the gain compensation value in real time based on a smoothing factor to generate a third data set; The audio optimization module is used to update the gain compensation model based on the least mean square adaptive filtering algorithm and the loudness change trend parameter of the third data set, and process the third data set based on the updated gain compensation model to generate a final audio data set, and the final audio data set is used to output an optimized radio and television microwave transmission signal.

2. The adaptive frequency optimization system according to claim 1, characterized in that: The initial audio acquisition processing module is used to perform the following steps: S11: Based on a preset sampling frequency, synchronously collect the audio signal of the current frequency channel through a multi-channel analog-to-digital converter; S12: performing parameter calculation on the audio signal based on an audio loudness standard algorithm to generate an original loudness parameter of a current frequency channel, and calculating a signal-to-noise ratio and a group delay characteristic of the current frequency channel based on the original loudness parameter to generate a channel state description value; S13: Processing the original loudness parameter and the channel state description value to generate a first data set.

3. The adaptive frequency optimization system according to claim 1, characterized in that: The conversion audio acquisition processing module is used to perform the following steps: S21: when a frequency switching signal is detected, starting a parallel analog-to-digital converter to collect an audio signal of the switching frequency channel, and generating a second data set containing a loudness parameter mean value and a state description value of the switching frequency channel; S22: performing a difference calculation on an original loudness parameter mean in the first data set and a loudness parameter mean in the second data set to generate an initial parameter offset; S23: performing normalization correction on the initial parameter offset according to the interference level in the channel state description value of the first data set to generate a loudness offset.

4. The adaptive frequency optimization system according to claim 1, characterized in that: The feature extraction module is used to perform the following steps: S31: dividing the audio signals in the first data set and the second data set into a preset number of time segments, and performing a fast Fourier transform on each time segment to generate a frequency domain energy distribution; S32: extracting energy proportions of the low frequency band, the middle frequency band, and the high frequency band of the frequency domain energy distribution, and generating an unnormalized initial eigenvector of the frequency distribution; S33: normalizing the initial feature vector to generate a frequency distribution feature vector.

5. The adaptive frequency optimization system according to claim 1, characterized in that: The model building module is used to perform the following steps: S41: constructing a polynomial regression model based on the original loudness parameters in the first data set, the loudness offset and the frequency distribution feature vector to generate a gain compensation model; S42: Performing a supplementary calculation on the loudness offset and the frequency distribution feature vector based on the gain compensation model to generate an initial compensation value. The calculation formula of the initial compensation value is as follows: Among them, G initial is the initial compensation value, ΔR is the loudness offset, F i is the i-th element of the frequency distribution eigenvector (i=1, 2, 3), corresponding to the energy proportion of the low frequency, medium frequency, and high frequency bands, respectively. k0 is a constant term, k1 is the coefficient of the gain compensation model, and w i is the frequency distribution eigenvector F i The coefficient of the first-order term corresponding to the element; S43: The initial compensation value is combined with the gain coefficient in the channel state description value to generate a gain compensation value of the switching frequency channel. The calculation formula of the gain compensation value is: Among them, G target is the gain compensation value, T is the gain coefficient in the channel status description value, T ref is the reference gain factor, R ref is the reference loudness value, and β is the fusion weight factor.

6. The adaptive frequency optimization system according to claim 1, characterized in that: The sound quality compensation module is used to perform the following steps: S51: Identify whether the gain compensation value exceeds a preset sound quality protection threshold, and if so, perform weighted average calculation on the gain compensation value based on historical gain compensation values ​​and a smoothing factor to generate a smoothed gain value; S52: Based on the smoothing gain value, adjusting the audio signal amplitude of the second data set by a digital signal processor to generate a compensated audio signal; S53: Integrate and store the compensated audio signal and the loudness change trend parameter calculated in real time to generate a third data set.

7. The adaptive frequency optimization system according to any one of claims 1 to 6, characterized in that: The audio optimization module is used to perform the following steps: S61: Based on the loudness change trend parameter of the third data set, a model error signal is calculated using a least mean square adaptive filtering algorithm to generate a weight update amount; S62: Iteratively adjust the weight parameters of the gain compensation model according to the weight update amount to generate an updated gain compensation model, wherein the update formula of the weight parameters is: k new =k old +μ*∈*ΔR norm in, and They represent the frequency-dependent weight parameters w in the gain compensation model. i The new value after updating and the old value before updating, k new and k old are the new value of the loudness-related parameter k in the gain compensation model after update and the old value before update, μ is the learning rate, F i is the i-th element of the frequency distribution eigenvector, ΔR norm is the normalized loudness offset; S63: performing secondary compensation calculation on the audio signal of the third data set based on the updated gain compensation model to generate an optimized frequency domain energy distribution; S64: Convert the optimized frequency domain energy distribution into a time domain signal to generate the final audio data set.

8. An adaptive frequency optimization method in microwave transmission of radio and television, applied to the adaptive frequency optimization system in microwave transmission of radio and television according to any one of claims 1 to 7, characterized in that: include: S710: Perform real-time acquisition and preprocessing on the current frequency channel based on a preset sampling frequency to generate a first data set including original loudness parameters and channel state description values; S720: collecting an audio signal of the switching frequency channel based on the frequency switching signal to generate a second data set, and performing loudness calculation on the second data set based on the first data set to obtain a loudness offset; S730: Perform feature extraction on the first data set and the second data set based on a fast Fourier transform algorithm to generate a frequency distribution feature vector; S740: constructing a gain compensation model based on the first data set, the loudness offset and the frequency distribution feature vector, and performing compensation calculation on the loudness offset based on the gain compensation model to generate a gain compensation value for the switching frequency channel; S750: Identify whether the gain compensation value exceeds a preset sound quality protection threshold, and if so, process and compensate the gain compensation value in real time based on a smoothing factor to generate a third data set; S760: The gain compensation model is updated based on the least mean square adaptive filtering algorithm and the loudness change trend parameter of the third data set, and the third data set is processed based on the updated gain compensation model to generate a final audio data set, and the final audio data set is used to output an optimized radio and television microwave transmission signal.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Citation Information

Cited By

  • Self-adaptive power amplifier nonlinear correction system based on independent architecture

    CN120415342A

  • Light follow-up sound based on data analysis and control system thereof

    CN120568253A

  • Multi-channel bio-electricity signal acquisition system and acquisition method

    CN121188649A

  • A multi-channel bioelectric signal acquisition system and acquisition method

    CN121188649B