Audio equalization method based on cloud partition deployment and related equipment

Abstract

The invention relates to the technical field of audio signal processing, and discloses an audio equalization method based on cloud partition deployment and related equipment. The method comprises the following steps that: a cloud server acquires environmental acoustic response data and corresponding geographic area identifiers of multi-user equipment; frequency response feature vectors of the acoustic response data are extracted, acoustic clustering is carried out in combination with the geographic area identifiers, a plurality of area acoustic categories are generated, and each category corresponds to at least one geographic area; comparing the frequency response feature vector of each area acoustic category with a preset target frequency response curve, calculating frequency response deviation and generating a corresponding audio equalization parameter; and obtaining a geographic region identifier of the to-be-balanced equipment, matching the geographic region identifier with the region acoustic category to which the geographic region identifier belongs, and sending the audio balancing parameter corresponding to the category to the equipment. According to the invention, automatic equalization parameter generation and regional accurate distribution based on regional acoustic features are realized, and the tone quality consistency and suitability of audio equipment in different geographical environments are improved.

Description

Technical Field

[0001] This invention relates to the field of audio signal processing technology, and in particular to an audio equalization method and related equipment based on cloud-based partitioned deployment. Background Technology

[0002] With the deep integration of digital audio technology and cloud computing services, audio equalization technology has evolved from traditional hardware filters to intelligent optimization systems based on digital signal processing. Existing audio equalization methods typically rely on local embedded algorithms, acquiring acoustic response signals through the device's built-in microphone and performing sound field calibration based on preset frequency response curves or a fixed library of filter parameters. Some high-end audio equipment incorporates room acoustic measurement technology, using swept-frequency signals or pink noise to test the frequency response characteristics of a specific space, and calculating compensation filter coefficients using a minimum mean square error algorithm, thereby improving sound performance to some extent in that environment. These methods can achieve relatively ideal results in single-device, single-space application scenarios, providing users with a relatively optimized listening experience.

[0003] However, existing technological solutions exhibit significant limitations when facing global product deployment. The acoustic characteristics of buildings differ fundamentally across geographical regions. For example, North American homes often feature wood structures and carpeting, while Asian homes commonly use concrete walls and hard floors. This results in drastically different acoustic effects from the same equalization parameters in different regions. Traditional equalization parameters are typically stored in the device firmware as fixed configuration files, making it impossible to adjust them based on the user's actual geographical location and corresponding acoustic environment characteristics. This makes it difficult for manufacturers to provide suitable sound quality for different regional markets. Therefore, the current methods of generating equalization parameters cannot take into account the differences in acoustic environments across regions, leading to significant inconsistencies in the sound quality of the same device in different areas. Summary of the Invention

[0004] The main objective of this invention is to solve the problem that existing audio equalization parameter generation methods fail to generate regionalized parameters based on the acoustic environmental characteristics of different regions, resulting in inconsistent sound quality performance of the same device in different geographical regions.

[0005] The first aspect of this invention provides an audio equalization method based on cloud-based partitioned deployment. The method includes: acquiring acoustic response data of the environment collected by multiple user devices and the geographical region identifiers corresponding to each user device through a cloud server; extracting frequency response feature vectors from the acoustic response data and performing acoustic clustering on the frequency response feature vectors and the corresponding geographical region identifiers to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region; comparing the frequency response feature vectors of each regional acoustic category with a preset target frequency response curve to obtain a frequency response deviation, and generating audio equalization parameters based on the frequency response deviation; acquiring the geographical region identifier of the device to be equalized, determining the regional acoustic category to which the device to be equalized belongs, and sending the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized.

[0006] Optionally, in a first implementation of the first aspect of the present invention, the step of obtaining the acoustic response data of the environment collected by multiple user devices and the geographical area identifiers corresponding to each user device through a cloud server includes: receiving environmental response signals of the environment collected by the multiple user devices through their built-in microphones after the devices play a preset calibration signal through a cloud server; performing frequency domain transformation and noise filtering on the environmental response signals to generate purified acoustic response data; and parsing the location information data uploaded by each user device through a cloud server and converting the location information data into corresponding geographical area identifiers.

[0007] Optionally, in a second implementation of the first aspect of the present invention, the step of extracting the frequency response feature vector of the acoustic response data and performing acoustic clustering on the frequency response feature vector and the corresponding geographical region identifier to obtain multiple regional acoustic categories includes: performing a fast Fourier transform on the acoustic response data to obtain frequency domain data, calculating the energy value of each frequency point in the frequency domain data within a preset frequency range, arranging the energy values ​​of each frequency point in frequency order to generate a frequency response feature vector; calculating the comprehensive clustering distance between each user device based on the similarity distance between the frequency response feature vectors and the association weight of the corresponding geographical region identifier; and grouping and clustering each user device based on the comprehensive clustering distance to generate multiple regional acoustic categories.

[0008] Optionally, in a third implementation of the first aspect of the present invention, the step of calculating the comprehensive clustering distance between user devices based on the similarity distance between the frequency response feature vectors and the association weight of the corresponding geographical region identifier includes: calculating the Euclidean distance between the frequency response feature vectors of any two user devices to obtain the similarity distance; determining the geographical proximity between the two user devices based on the geographical region identifier of any two user devices, and determining the corresponding association weight based on the geographical proximity, wherein the higher the geographical proximity, the smaller the association weight; and multiplying the similarity distance between each user device by the corresponding association weight to obtain the comprehensive clustering distance between each user device.

[0009] Optionally, in a fourth implementation of the first aspect of the present invention, the step of comparing the frequency response feature vectors of each of the regional acoustic categories with a preset target frequency response curve to obtain a frequency response deviation, and generating audio equalization parameters based on the frequency response deviation, includes: statistically calculating the frequency response feature vectors within each of the regional acoustic categories to obtain the average frequency response value of each frequency band, and constructing a regional acoustic template for the corresponding regional acoustic category based on the average frequency response value of each frequency band; calculating the difference between each of the regional acoustic templates and the preset target frequency response curve in each frequency band to obtain the frequency response deviation of each regional acoustic category in each frequency band; determining frequency compensation parameters according to the frequency response deviation, and encapsulating the frequency compensation parameters as audio equalization parameters for the corresponding regional acoustic category.

[0010] Optionally, in a fifth implementation of the first aspect of the present invention, the step of statistically calculating the frequency response feature vectors within each of the said regional acoustic categories to obtain the average frequency response value of each frequency band, and constructing a regional acoustic template for the corresponding regional acoustic category based on the average frequency response value of each frequency band, includes: reorganizing the frequency response feature vectors within each regional acoustic category by frequency band to obtain feature value groups corresponding to each frequency band; averaging the feature value groups of each frequency band to obtain the average frequency response value of the corresponding frequency band; and arranging the average frequency response values ​​of each frequency band in frequency order to generate the regional acoustic template for the regional acoustic category.

[0011] Optionally, in a sixth implementation of the first aspect of the present invention, obtaining the geographic region identifier of the device to be equalized, determining the regional acoustic category to which the device to be equalized belongs, and sending the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized includes: receiving an equalization parameter request sent by the device to be equalized, and parsing the geographic location data in the equalization parameter request; converting the geographic location data into a geographic region identifier, and performing geographic region matching between the geographic region identifier and each regional acoustic category to determine the target acoustic category; reading the audio equalization parameters corresponding to the target acoustic category from the cloud database, and sending the audio equalization parameters to the device to be equalized.

[0012] A second aspect of the present invention provides an audio equalization device based on cloud-based partitioned deployment. The cloud-based audio equalization device includes: a data acquisition module, used to acquire acoustic response data of the environment collected by multiple user devices and the geographical region identifiers corresponding to each user device through a cloud server; a clustering analysis module, used to extract frequency response feature vectors from the acoustic response data and perform acoustic clustering on the frequency response feature vectors and the corresponding geographical region identifiers to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region; a parameter generation module, used to compare the frequency response feature vectors of each regional acoustic category with a preset target frequency response curve to obtain a frequency response deviation, and generate audio equalization parameters based on the frequency response deviation; and a distribution and push module, used to acquire the geographical region identifier of the device to be equalized, determine the regional acoustic category to which the device to be equalized belongs, and send the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized.

[0013] A third aspect of the present invention provides an audio equalizer based on cloud-based partition deployment, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the audio equalizer based on cloud-based partition deployment to perform the various steps of the aforementioned audio equalizer method based on cloud-based partition deployment.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the steps of the above-described cloud-based partitioned audio equalization method.

[0015] The above-described audio equalization method and related equipment are based on cloud-based partitioned deployment. In this embodiment of the invention, acoustic response data of the environment collected by multiple user devices and the geographical region identifiers corresponding to each user device are obtained through a cloud server; frequency response feature vectors of the acoustic response data are extracted, and acoustic clustering is performed on the frequency response feature vectors and the corresponding geographical region identifiers to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region; the frequency response feature vectors of each regional acoustic category are compared with a preset target frequency response curve to obtain the frequency response deviation, and audio equalization parameters are generated based on the frequency response deviation; the geographical region identifier of the device to be equalized is obtained, the regional acoustic category to which the device to be equalized belongs is determined, and the audio equalization parameters corresponding to the regional acoustic category are sent to the device to be equalized. This application constructs a regional acoustic template library by introducing a joint clustering mechanism of geographic region identifiers and acoustic feature vectors. It uniformly analyzes the environmental acoustic data of multiple users worldwide in the cloud, dynamically identifies the differences in architectural acoustic characteristics in different geographic regions, and realizes regional customization and precise distribution of audio equalization parameters. This solves the problem of inconsistent sound quality across regions caused by the fixed firmware parameters in existing audio equalization systems, effectively improves the acoustic adaptability and user experience consistency of the system in global deployment scenarios, and ensures that users in different geographic regions can obtain optimized sound quality performance that conforms to the characteristics of the local acoustic environment.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the first embodiment of the audio equalization method based on cloud-based partition deployment in this invention. Figure 2 This is a schematic diagram of an embodiment of an audio equalization device based on cloud-based partition deployment in this invention. Figure 3 This is a schematic diagram of an embodiment of an audio equalization device based on cloud-based partition deployment in this invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The terms "comprising" and "having," and any variations thereof, used in the embodiments of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0021] To facilitate understanding of this embodiment, the specific process of this embodiment is described below. Please refer to [link / reference]. Figure 1 The first embodiment of the audio equalization method based on cloud-based partition deployment in this invention includes: 101. Obtain acoustic response data of the environment collected by multiple user devices and the geographical area identifiers corresponding to each user device through the cloud server; The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0022] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0023] In this embodiment, the cloud server receives the environmental response signal of the surrounding environment after multiple user devices play a preset calibration signal and collects it through the built-in microphone; the environmental response signal is frequency domain transformed and noise filtered to generate purified acoustic response data; the cloud server parses the location information data uploaded by each user device and converts the location information data into the corresponding geographical area identifier.

[0024] In practical applications, the cloud server receives environmental response signals uploaded from multiple user devices via a network interface. When a user device activates the room calibration function, its built-in speaker plays a preset calibration signal, which is either pink noise or a swept-frequency signal, covering the audible range of 20Hz to 20kHz. The device's built-in microphone simultaneously records the acoustic feedback generated in the room after the calibration signal is played, forming the environmental response signal. The environmental response signal here refers to the actual sound waveform captured by the microphone after the calibration signal has undergone acoustic effects such as wall reflection, furniture absorption, and spatial reverberation in the room. For example, when playing a 100Hz low-frequency test sound in a living room with wooden floors, the energy recorded by the microphone will be significantly lower than the energy during playback due to the absorption of low frequencies by the wooden floor; this energy difference is fully recorded in the environmental response signal. The cloud server categorizes and stores the environmental response signals uploaded by each user device according to the device number. This step collects raw acoustic data from a real-world usage environment; differences in building structures in different regions are directly reflected in the frequency characteristics of the environmental response signal. The environmental response signal is then subjected to frequency domain transformation, specifically using the Fast Fourier Transform (FFT) algorithm to convert the time-domain waveform data into frequency-domain spectral data. The FFT is a mathematical transformation method that decomposes a sound waveform on the time axis into different frequency components, yielding energy distribution information for each frequency point within the 20Hz to 20kHz range. The server reads the transformed spectral data and identifies frequency components with abnormally low energy and narrow bandwidth as noise features. Noise filtering is achieved by setting an energy threshold; frequency components with energy levels more than 30dB below the average energy of the environmental response signal are identified as noise and set to zero (the 30dB threshold is based on the conventional signal-to-noise ratio requirements in acoustic measurements, ensuring that the retained acoustic response data has sufficient signal strength while effectively removing low-energy environmental interference). For example, if the average energy of an environmental response signal in the 3kHz band is -20dB, while at 50Hz it is only -55dB, a difference of 35dB exceeding the 30dB threshold, then the energy at 50Hz is set to zero to remove power frequency interference. The data after frequency domain transformation and noise filtering is the purified acoustic response data. Frequency domain transformation converts complex waveforms into quantifiable and analyzable frequency energy data. Noise filtering eliminates environmental interference such as air conditioning noise and traffic noise, ensuring that the acoustic response data reflects only the acoustic characteristics of the room itself. This is crucial for accurately extracting regional acoustic patterns. The cloud server then parses the location information data carried in the data packets uploaded by the user device. This location information data includes GPS latitude and longitude coordinates or network IP addresses. For GPS coordinate data, the server queries a built-in geographic information data table to match the corresponding country, province, and city name based on the latitude and longitude range. For example, latitude and longitude (34.05°N, 118.24°W) matches Los Angeles, California, USA.For IP address data, the server queries the IP address geolocation database to obtain the geographical location information of the IP address. After obtaining the geographical location information, the server generates standardized geographic region identifiers according to preset regional division rules. The geographic region identifiers adopt a multi-level hierarchical structure, including information at the levels of country, first-level administrative region, and city, with different levels separated by underscores. For example, "US_CA_LA" represents the Los Angeles area of ​​California, USA. Standardized geographic region identifiers provide a unified regional attribution label for subsequent acoustic clustering analysis, enabling the system to identify user devices from the same geographic region. This is a prerequisite for discovering regional acoustic patterns.

[0025] 102. Extract the frequency response feature vector of the acoustic response data, and perform acoustic clustering on the frequency response feature vector and the corresponding geographic region identifier to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographic region. In this embodiment, a Fast Fourier Transform is performed on the acoustic response data to obtain frequency domain data. The energy value of each frequency point within a preset frequency range is calculated, and the energy values ​​of each frequency point are arranged in frequency order to generate a frequency response feature vector. Based on the similarity distance between the frequency response feature vectors and the association weight of the corresponding geographical region identifier, the comprehensive clustering distance between each user device is calculated (wherein, calculating the comprehensive clustering distance between each user device based on the similarity distance between the frequency response feature vectors and the association weight of the corresponding geographical region identifier includes: calculating the Euclidean distance between the frequency response feature vectors of any two user devices to obtain the similarity distance; determining the geographical proximity between the two user devices based on the geographical region identifier of any two user devices, and determining the corresponding association weight based on the geographical proximity, wherein the higher the geographical proximity, the smaller the association weight; multiplying the similarity distance between each user device by the corresponding association weight to obtain the comprehensive clustering distance between each user device). Based on the comprehensive clustering distance, each user device is grouped and clustered to generate multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region.

[0026] In practical applications, the cloud server performs a Fast Fourier Transform (FFT) operation on the purified acoustic response data. The server reads the time-domain waveform sampling sequence of the acoustic response data and inputs it into a pre-defined FFT algorithm program for processing. The FFT is a mathematical transformation method that decomposes a time-domain sound signal into different frequency components, outputting frequency domain data. This frequency domain data contains complex representations of each frequency component within a preset frequency range of 20Hz to 20kHz. The server calculates the amplitude of the complex values ​​in the frequency domain data to obtain the energy intensity of each frequency component. The server divides the 20Hz to 20kHz frequency range into 128 frequency points, each representing a frequency position, with a frequency point interval of approximately 156Hz. The server extracts the energy values ​​corresponding to these 128 frequency points and arranges them into a one-dimensional array in ascending order of frequency; this array is the frequency response feature vector. The frequency response feature vector is a vector containing 128 numerical elements, each element representing the acoustic energy at the corresponding frequency point. For example, the first element of the frequency response feature vector for user device A has a value of 0.25, indicating that the device has 0.25 energy around 20Hz. The second element has a value of 0.18, indicating that the device has 0.18 energy around 176Hz, and so on up to the 128th element. Acoustic differences in different room environments lead to significant variations in the values ​​of the elements in the frequency response feature vector. Living rooms with wooden floors generally have lower low-frequency energy values, while rooms with concrete walls generally have higher mid-to-high frequency energy values. By converting the acoustic response data of each user device into a standardized 128-dimensional frequency response feature vector, a standardized digital representation of the room's acoustic characteristics is achieved, making the acoustic data of different devices comparable.

[0027] Then, the cloud server calculates the comprehensive clustering distance between each user device. The server first calculates the Euclidean distance between the frequency response feature vectors of any two user devices; this Euclidean distance is used as the similarity distance. Taking user device A and user device B as an example, the frequency response feature vector of device A is denoted as vector A, containing 128 elements (a1, a2, ..., a...). 128 Let the frequency response characteristic vector of device B be denoted as vector B, which contains 128 elements (b1, b2, ..., b...). 128 The server calculates the difference between corresponding elements of vectors A and B one by one, squaring the difference at the first position (a1-b1), the difference at the second position (a2-b2), and so on up to the difference at the 128th position (a... 128 -b 128The server sums these 128 squared values ​​and then takes the square root of the sum. The resulting value is the similarity distance. The smaller the similarity distance, the closer the two frequency response feature vectors are, and the more similar the acoustic characteristics of the corresponding two rooms are. Next, the geographical proximity is determined based on the geographical region identifiers of the two user devices. The geographical region identifiers are parsed to obtain the latitude and longitude coordinates of the device's location. The actual geographical distance D between the two devices is calculated using the spherical distance formula, in kilometers. The geographical proximity is calculated according to the formula: Geographical proximity = 1 / (1+D / D0), where D0 is a preset distance constant. The value of D0 is determined based on the geographical coverage of the system deployment and the clustering granularity requirements. When the system needs to identify acoustic differences between different areas within a city (such as the city center and suburbs), D0 is set to 50 kilometers; when the system mainly distinguishes acoustic differences between different countries or large geographical regions, D0 is set to 500 kilometers. For example, device A is located in downtown Los Angeles, and device B is located in the suburbs of Los Angeles, 30 kilometers apart. In a city-level deployment scenario, D0 = 50 kilometers, so the geographical proximity = 1 / (1+30 / 50) = 0.625. The geographical proximity value ranges from 0 to 1; the larger the value, the closer the two devices are geographically. The association weight is then determined based on the geographical proximity, using the mapping relationship: Association Weight = 2 - Geographical Proximity. This mapping ensures that the higher the geographical proximity, the smaller the association weight, because in subsequent calculations, the comprehensive clustering distance equals the similarity distance multiplied by the association weight. The smaller the association weight, the smaller the comprehensive clustering distance, indicating that the two devices are more likely to be classified into the same category. In the example above, with a geographical proximity of 0.625, the association weight = 2 - 0.625 = 1.375. The server multiplies the similarity distance of each pair of user devices by the corresponding association weight to obtain the comprehensive clustering distance. Specifically, the calculation process of the comprehensive clustering distance can be expressed by the following formula: Calculation of Similarity Distance The calculation formula is: , in, and These represent the values ​​of the i-th and j-th objects on the k-th feature, respectively. Geographic proximity calculation. The calculation formula is: , Where D is the actual distance between the two objects. It is a preset distance threshold. Association weight. The calculation formula is: Calculation of the overall cluster distance The calculation formula is: , in: This represents the combined clustering distance between user equipment i and user equipment j; This represents the energy value of the frequency response feature vector of user equipment i at the k-th frequency point; This represents the energy value of the frequency response feature vector of user equipment j at the k-th frequency point; the value of k ranges from 1 to 128, corresponding to 128 frequency bands; This represents the actual geographical distance between user equipment i and user equipment j, in kilometers. A preset distance constant is used, with a value ranging from 50 kilometers to 500 kilometers. Taking the aforementioned embodiment as an example: assuming the frequency response feature vectors of user equipment A and user equipment B are vector A(0.25, 0.18, ..., 0.32) and vector B(0.22, 0.20, ..., 0.35) respectively, and the two devices are located in different areas of the same city, with an actual geographical distance... =30 kilometers, in a city-level deployment scenario =50 kilometers. First, calculate the Euclidean distance: , Then calculate geographical proximity. The correlation weight is 1 / (1+30 / 50)=0.625. The final calculated cluster distance is: 2 - 0.625 = 1.375. The value is 0.156 × 1.375 = 0.2145. By multiplying and fusing acoustic similarity distance with geographic association weights, a dual constraint on acoustic features and geographic location is achieved. This avoids the geographic dispersion problem caused by simply clustering based on acoustic features, ensuring that devices within the same acoustic category in the same region have similar acoustic response characteristics while maintaining relative geographic concentration. This improves the representativeness of regional acoustic templates and the applicability of audio equalization parameters.

[0028] The cloud server clusters user devices based on a comprehensive clustering distance, generating multiple regional acoustic categories. The server uses the K-Means clustering algorithm for this grouping operation. First, the server determines the number of clusters K based on the distribution of geographical region identifiers. For example, if there is user data from 20 different countries / regions, K is initially set to 20. The server randomly selects K user devices as initial cluster centers. Each cluster center is a reference point representing a category, and initially, each cluster center corresponds to a user device's frequency response feature vector. The server calculates the comprehensive clustering distance from all other user devices to these K cluster centers and assigns each user device to the category of the cluster center with the smallest comprehensive clustering distance. After one round of assignment, the server recalculates the average of the frequency response feature vectors of all devices within each category as the new cluster centers. The server repeats the iterative process of "calculating comprehensive clustering distance - assigning categories - updating cluster centers" until the category affiliation of all devices no longer changes in two consecutive iterations, at which point clustering convergence is complete. After clustering, the server obtains K regional acoustic categories, each containing several user devices. The server analyzes the distribution of geographic region identifiers for devices within each regional acoustic category, selecting the most frequently occurring geographic region identifier as the corresponding geographic region for that acoustic category. For example, if a regional acoustic category contains 180 devices, with 150 devices having the geographic region identifier "US_CA_LA", 20 "US_CA_SF", and 10 "US_NV", then the geographic region corresponding to this acoustic category is California, USA. Furthermore, a unique identifier is assigned to each regional acoustic category, and its list of included user devices and representative geographic region information are stored. Cluster analysis groups user devices with similar acoustic characteristics and geographical proximity into the same regional acoustic category. Each regional acoustic category reflects common acoustic features caused by environmental factors such as building structure and decoration materials within a specific geographic area, providing a scientific classification basis for generating targeted regional audio equalization parameters.

[0029] 103. Compare the frequency response feature vectors of each acoustic category with the preset target frequency response curve to obtain the frequency response deviation, and generate audio equalization parameters based on the frequency response deviation. In this embodiment, the frequency response feature vectors within each regional acoustic category are statistically calculated to obtain the average frequency response value of each frequency band. Based on the average frequency response value of each frequency band, a regional acoustic template for the corresponding regional acoustic category is constructed. (The process of statistically calculating the frequency response feature vectors within each regional acoustic category to obtain the average frequency response value of each frequency band and constructing the regional acoustic template for the corresponding regional acoustic category includes: reorganizing the frequency response feature vectors within each regional acoustic category by frequency band to obtain feature value groups corresponding to each frequency band; averaging the feature value groups of each frequency band to obtain the average frequency response value of the corresponding frequency band; and arranging the average frequency response values ​​of each frequency band in frequency order to generate the regional acoustic template for the regional acoustic category.) (The process involves) calculating the difference between the acoustic template of each region and the preset target frequency response curve in each frequency band to obtain the frequency response deviation of each region's acoustic category in each frequency band; determining the frequency compensation parameters based on the frequency response deviation, and encapsulating the frequency compensation parameters into audio equalization parameters for the corresponding region's acoustic category (wherein, determining the frequency compensation parameters based on the frequency response deviation and encapsulating the frequency compensation parameters into audio equalization parameters for the corresponding region's acoustic category includes: performing an inversion operation on the frequency response deviation to obtain the gain adjustment value for each frequency band; using the frequency value corresponding to each frequency band as the center frequency point of the filter, determining the quality factor of the preset filter, and generating frequency compensation parameters based on the filter center frequency point, quality factor, and gain adjustment value; and encapsulating the frequency compensation parameters into audio equalization parameters for the corresponding region's acoustic category). In addition, after sending the audio equalization parameters to the device to be equalized, the process also includes: receiving the equalization effect score uploaded by the device to be equalized, and extracting the satisfaction score and the corresponding regional acoustic category from the equalization effect score; accumulating and statistically analyzing the satisfaction score values ​​of each regional acoustic category to obtain the average satisfaction score and the number of score samples for each regional acoustic category; when the number of score samples reaches a preset threshold, adjusting and optimizing the regional acoustic template of the corresponding regional acoustic category based on the score data of the corresponding regional acoustic category, and regenerating the corresponding audio equalization parameters based on the adjusted regional acoustic template.

[0030] In practical applications, cloud servers perform statistical calculations on frequency response feature vectors within each regional acoustic category to construct regional acoustic templates. The server reads the frequency response feature vector data of all user devices within a specific regional acoustic category. These frequency response feature vectors are then recombined by frequency band. The specific recombining process is as follows: the server creates 128 data groups, each corresponding to a frequency band. Then, it iterates through each frequency response feature vector within the regional acoustic category, adding the element value at the first position of the vector to the first data group, the element value at the second position to the second data group, and so on, until the element value at the 128th position is added to the 128th data group. After recombining, each data group contains the energy values ​​of all devices within the regional acoustic category at the same frequency point; this data group is the feature value group corresponding to that frequency band. For example, if a regional acoustic category contains 200 user devices, the recombined feature value group for the first frequency band contains 200 values, which are the energy values ​​measured by the 200 devices at a frequency of 20Hz. The characteristic values ​​of each frequency band are then grouped and averaged. All values ​​within a group are summed, and the sum is divided by the number of values ​​in that group to obtain the average frequency response value for that frequency band. Taking the first frequency band as an example, assuming the sum of 200 energy values ​​is 48, the average frequency response value is 48 ÷ 200 = 0.24. The server calculates the average frequency response values ​​for all 128 frequency bands sequentially. These 128 average values ​​are arranged in ascending order of frequency, forming a new vector. This vector is the regional acoustic template for the acoustic category of that area. The regional acoustic template is a statistical representation of the acoustic environment of numerous user devices within a geographical area. The value of each frequency band reflects the typical energy level of that area at the corresponding frequency. For example, in California, the average frequency response value of its regional acoustic template in the 3kHz to 8kHz frequency band is significantly higher than in other frequency bands, reflecting the enhanced mid-to-high frequency reflection characteristics caused by wooden floors and glass windows. By statistically averaging individual data, the regional acoustic template filters out the influence of individual special environments and extracts the common patterns of the architectural acoustic environment in the region, making the audio equalization parameters generated based on the template applicable to most usage scenarios in the region.

[0031] The difference between the acoustic template for each region and the preset target frequency response curve in each frequency band is then calculated to obtain the frequency response deviation. The preset target frequency response curve represents the energy distribution standard that each frequency band should have under an ideal acoustic environment. The server determines the preset target frequency response curve based on the hardware parameters of the user's equipment (in addition, a uniform target frequency response curve can be set for all equipment). The server queries the equipment model to be generated for equalization parameters from the equipment information database and obtains the hardware parameter information of that model, including the number of equalizer frequency bands, speaker unit type, power amplifier output power, and other technical specifications. Different preset target frequency response curves are used for different equipment types. For high-end equipment equipped with multi-unit speaker systems, the server adopts a flat target curve that maintains energy balance across the entire frequency band, with consistent target values ​​across all frequency bands in the 20Hz to 20kHz range. For entry-level equipment using a single speaker unit, due to the limitations of the speaker's physical characteristics, its reproduction capability at extremely low and high frequencies is weak. The server adopts a target curve that appropriately reduces the expected values ​​in the low and high frequency bands to match the target curve with the actual capabilities of the equipment. The server compares the regional acoustic template with the preset target frequency response curve for the corresponding device type, band by band, and performs difference calculation for each band. The difference calculation operation is as follows: read the average frequency response value of the regional acoustic template in a certain frequency band and the target value of the preset target frequency response curve in that band, and subtract the target value from the average frequency response value to obtain the frequency response deviation for that band. For example, if the average frequency response value of a regional acoustic template in the 100Hz band is 0.38, and the target value of the preset target frequency response curve in the 100Hz band is 0.30, then the frequency response deviation for the 100Hz band is 0.38 - 0.30 = 0.08. A positive frequency response deviation indicates that the actual energy in that band is higher than the ideal level, and a negative number indicates that the actual energy is lower than the ideal level. The server performs difference calculations on all 128 frequency bands, obtaining a deviation array containing 128 frequency response deviation values. Frequency response deviation quantifies the specific differences between the actual acoustic environment and the ideal standard in each frequency band, and clarifies which frequency bands need to be enhanced, which frequency bands need to be attenuated, and the specific adjustment range, providing a quantitative basis for generating accurate frequency compensation parameters.

[0032] The frequency compensation parameters are then determined based on the frequency response deviation and encapsulated as audio equalization parameters. The server performs an inversion operation on the frequency response deviation of each frequency band in the deviation array. Inversion involves multiplying the value by -1, turning a positive value into a negative value and a negative value into a positive value. For example, if the frequency response deviation of a certain frequency band is 0.08, the inversion operation yields -0.08. The value obtained after the inversion operation is the gain adjustment value for that frequency band. The gain adjustment value represents the amount of energy adjustment that needs to be applied to that frequency band during audio playback. When expressed in decibels (dB), a negative value represents attenuation, and a positive value represents enhancement. In the example above, the actual energy of this frequency band is higher than the target of 0.08, so the gain adjustment value is -0.08, indicating that the energy of this frequency band needs to be attenuated. The server performs an inversion operation on the frequency response deviations of all 128 frequency bands, obtaining 128 gain adjustment values. In some optimized embodiments, the server can further consider the consistency characteristics of device responses within a regional acoustic category based on the basic inversion operation, and adaptively correct the gain adjustment value to improve the applicability of the audio equalization parameters to different devices within the region. Specifically, the server first inverts the frequency response deviation according to the aforementioned inversion operation method to obtain the basic gain adjustment value for each frequency band. Its formula is: ,in, This represents the base gain adjustment value at frequency f; This represents the average frequency response of the acoustic template at frequency f. This represents the target value of the preset target frequency response curve at frequency f. Then, the server can introduce an adaptive correction mechanism to calculate the corrected frequency compensation gain. : ,in, This represents the frequency compensation gain at frequency f after adaptive correction. This is a frequency-dependent adjustment factor used to reflect the differences in auditory sensitivity across different frequency bands: 0.8 in the low-frequency band (20Hz-200Hz), 1.0 in the mid-frequency band (200Hz-4kHz), and 0.9 in the high-frequency band (4kHz-20kHz). This represents the coefficient of variation of the energy value of all devices within the acoustic category of this region at frequency f, and is calculated as follows: ,in, Standard deviation, The mean; This is a variability adjustment factor, ranging from 0.1 to 0.3, with a preferred value of 0.2. Coefficient of variation. The calculation process is as follows: The server reads the energy values ​​of all user devices within the acoustic category of the area at frequency f, and calculates the arithmetic mean of these energy values ​​as the mean. Calculate the standard deviation of these energy values ​​relative to the mean. Dividing the two yields the coefficient of variation. The coefficient of variation reflects the degree of consistency in the response of different devices within this region to this frequency band. A smaller value indicates higher consistency. A larger value indicates a higher degree of dispersion. Taking the aforementioned embodiment as an example: for a certain acoustic category in a region at frequency f=100Hz, the average frequency response value of the regional acoustic template is... The target value of the preset target frequency response curve The standard deviation of energy values ​​of 200 devices in this region at 100Hz mean 100Hz is considered a low frequency range. Substitute into the formula to calculate: The results show that, considering the inter-device differences (coefficient of variation 0.158) in the 100Hz frequency band, the system automatically and appropriately reduces the compensation strength from the basic -0.08 to -0.062 to balance the compatibility of devices with different acoustic environments within the region. The technical effect of this adaptive correction mechanism is as follows: by introducing the coefficient of variation CV(f), when the device response consistency within a certain frequency band is high (CV(f) is small), the correction term [1-β×CV(f)] approaches 1, maintaining a strong compensation strength. This ensures that the frequency compensation gain is not only based on the regional average deviation but also considers the differences in auditory sensitivity across different frequency bands and the degree of acoustic response dispersion among devices within the region. When the device response dispersion within a certain frequency band is large (CV(f) is large, i.e., high variability), the correction term [1-β×CV(f)] decreases, reducing the compensation strength and avoiding overcompensation for some devices that could lead to sound quality degradation. When the variability is small, more aggressive compensation can be implemented, improving the overall sound quality. Meanwhile, the introduction of the frequency-dependent adjustment coefficient α(f) allows for a relatively conservative compensation strategy in the low and high frequency bands, while a standard compensation strategy is used in the mid-frequency band. This aligns with the human ear's sensitivity to different frequency bands and improves the robustness of the audio equalization parameters. It should be noted that the above adaptive correction mechanism is an optional optimized implementation. In the basic implementation, the server can directly use the basic gain adjustment value. As the final frequency compensation gain, that is, without performing The adaptive correction of the formula can also achieve effective audio equalization.

[0033] Next, frequency compensation parameters are generated, which consist of three elements: the filter center frequency, the quality factor, and the gain adjustment value (in embodiments employing adaptive correction, the corrected frequency compensation gain is used). In the basic embodiment, a basic gain adjustment is used. The filter's center frequency is defined by the frequency value corresponding to each frequency band: 20Hz for band 1, 176Hz for band 2, and so on up to 20kHz for band 128. The center frequency of the filter specifies the target frequency position for the equalization adjustment. The server determines the quality factor for each filter. The quality factor is a parameter describing the frequency selectivity of the filter, reflecting the width of the frequency range over which the filter operates. For example, in this scheme, the server can use preset quality factor values, setting a quality factor of 1.0 for low-frequency bands with larger frequency intervals and a quality factor of 1.5 for high-frequency bands with closer frequency intervals, ensuring that the adjustment range of each frequency band adapts to the frequency distribution characteristics. Based on the filter's center frequency, quality factor, and gain adjustment value, a set of frequency compensation parameters is generated for each frequency band. The frequency compensation parameters are stored in key-value pair format, containing three fields: "frequency," "Q value," and "gain." For example, the frequency compensation parameters for the 100Hz band are {Frequency: 100Hz, Q value: 1.0, Gain: -0.08}. In an embodiment using adaptive correction, the frequency compensation parameters for the 100Hz band are {Frequency: 100Hz, Q value: 1.0, Gain: -0.062}. The server organizes the frequency compensation parameters for 128 frequency bands according to JSON or XML data format and encapsulates them into an audio equalization parameter file for that region's acoustic category. The audio equalization parameter file contains a complete frequency adjustment scheme. After the device's digital signal processor reads this file, it can directly configure the various filter parameters of the equalizer and process the audio signal in real time. A compensatory adjustment strategy is implemented through inversion operations, that is, attenuating whichever frequency band is too high and enhancing whichever frequency band is too low, so that the frequency response of the audio output approaches the preset target curve, thereby improving the sound balance.

[0034] In addition, after applying audio equalization parameters to the device, users can subjectively evaluate the sound effect through a mobile application. The application interface displays a rating slider or star rating option, and users select a satisfaction rating value based on their listening experience, ranging from 1 to 5 points, where 1 point represents very poor effect, 3 points represents average effect, and 5 points represents very good effect. After the user confirms the rating, the device generates an equalization effect rating data packet. The data packet contains two core pieces of information: the satisfaction rating value and the regional acoustic category identifier currently used by the device. For example, if a user rates the sound effect as 4 points, the device records the satisfaction rating value as 4, and simultaneously reads the regional acoustic category identifier "US_CA_LA" stored in the system, encapsulating both pieces of information into a data packet. The device uploads the data packet to a cloud server via a network connection. After receiving the data packet, the server parses and extracts the satisfaction rating value and regional acoustic category identifier, and writes the rating record into the rating data table for the corresponding regional acoustic category. The rating data table is managed by regional acoustic category, and each record contains a rating value, an upload timestamp, and a unique device identifier. User satisfaction ratings are subjective feedback on the actual performance of audio equalization parameters. This real user experience data provides a basis for subsequent evaluation to determine whether equalization parameters need optimization. The cloud server then accumulates and statistically analyzes the satisfaction ratings for each acoustic category in different regions. The server periodically executes statistical tasks, reading all historical rating records from the rating data tables of each acoustic category in different regions. Two statistical indicators are calculated for each acoustic category in different regions: the average satisfaction rating and the number of rating samples. The average satisfaction rating is calculated by summing the satisfaction ratings of all records in the acoustic category in that region and then dividing by the total number of records. For example, if the acoustic category in the "US_CA_LA" region has 800 records in its rating data table, and the sum of all ratings is 3200, then the average satisfaction rating is 3200 ÷ 800 = 4.0. The number of rating samples is the total number of rating records received for that acoustic category in that region; in the example above, the number of rating samples is 800. The server updates the calculation results to the statistical information table for each acoustic category in different regions, which records the real-time evaluation status of each acoustic category in different regions. The server sets a preset threshold for the number of rating samples. This threshold is determined based on the regional user base and statistical significance requirements, and in actual deployments, the preset threshold ranges from 500 to 5000 samples. For example, for regions with a large user base, the preset threshold is set to 2000 samples; for regions with a small user base, the preset threshold is set to 500 samples. By continuously accumulating and statistically analyzing user satisfaction data for each region, the server can monitor the actual effect of audio equalization parameters and identify acoustic categories in regions with low satisfaction that require improvement. Then, when the number of rating samples reaches the preset threshold, the regional acoustic template is adjusted and optimized. Once the server detects that the number of rating samples for a certain acoustic category in a region has reached the preset threshold, the optimization process is initiated.The server reads the average satisfaction rating and rating distribution for the acoustic category in the region, analyzing the overall trend of user feedback. By extracting samples with low ratings, the device identifiers corresponding to these low-rated samples are recorded. Then, the server retrieves the historical acoustic response data uploaded by these devices and recalculates their frequency response feature vectors. The server compares the frequency response feature vectors of the low-rated devices with the current regional acoustic template, identifying frequency bands with significant differences. For example, it finds that the average energy of the low-rated device group in the 5kHz to 6kHz frequency band is 0.42, while the current regional acoustic template has a value of 0.35 in this band, indicating that the current template does not fully reflect the acoustic environment characteristics of this group of users. The server adjusts the regional acoustic template, aligning the template values ​​in the significantly different frequency bands towards the statistical mean of the low-rated device group. The adjustment method involves a weighted average of the original template value and the mean of the low-rated devices, giving the mean of the low-rated devices a higher weight, making the adjusted template value closer to the actual needs of this group of users. The server performs the above adjustment operation on each frequency band that needs adjustment, generating the adjusted regional acoustic template. The server re-executes the aforementioned complete process of difference calculation, inversion operation, and frequency compensation parameter generation based on the adjusted regional acoustic template to obtain optimized audio equalization parameters. The server then updates the parameter library for that region's acoustic category with the optimized audio equalization parameters, replacing the original version. Subsequent requests for audio equalization parameters from devices in that region will receive the optimized new version. Through an iterative optimization mechanism based on user satisfaction feedback, the audio equalization parameters can continuously adapt to changes in users' actual listening preferences, constantly improving user satisfaction with the sound effect and avoiding the problem of a disconnect between purely algorithm-generated solutions and user subjective experience.

[0035] 104. Obtain the geographical region identifier of the device to be equalized, determine the regional acoustic category to which the device to be equalized belongs, and send the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized.

[0036] In this embodiment, an equalization parameter request sent by the device to be equalized is received, and the geographic location data in the equalization parameter request is parsed; the geographic location data is converted into a geographic region identifier, and the geographic region identifier is matched with the acoustic category of each region to determine the target acoustic category; the audio equalization parameters corresponding to the target acoustic category are read from the cloud database, and the audio equalization parameters are sent to the device to be equalized (wherein, reading the audio equalization parameters corresponding to the target acoustic category from the cloud database and sending the audio equalization parameters to the device to be equalized includes: obtaining the audio equalization parameters corresponding to the target acoustic category from the cloud database, compressing the audio equalization parameters to generate compressed audio equalization parameters; dividing the compressed audio equalization parameters into multiple data segments, and generating a checksum corresponding to each data segment; encapsulating each data segment and the corresponding checksum into a parameter data packet, and sending it to the device to be equalized).

[0037] In practical applications, the cloud server receives equalization parameter requests from the devices to be equalized and parses the geographic location data within them. After completing initial configuration or triggering room calibration by the user, the device sends an equalization parameter request to the cloud server. This request is a data message encapsulated using the HTTP protocol, requesting audio equalization parameters from the device. It includes information such as the device identifier, device model, and geographic location data. The geographic location data describes the device's location, including latitude and longitude coordinates obtained from a GPS module or the device's IP address. Upon receiving the request, the server parses the message content, reads the geographic location data fields, and identifies the data type. For GPS coordinates, the server extracts the longitude and latitude values; for IP addresses, the server extracts the complete address string. For example, if a device's request contains geographic location data of (34.05°N, 118.24°W), the server parses it to obtain latitude 34.05 and longitude -118.24. This parsing provides a basis for matching the appropriate regional acoustic category. The geographic location data is then converted into a geographic region identifier, and geographic region matching is performed to determine the target acoustic category. For GPS coordinates, the server queries a geocoding data table, which stores the correspondence between latitude and longitude ranges and administrative divisions. The server compares the device coordinates with the coordinate ranges in the table, finds the record containing the coordinates, obtains the country, state, province, and city information, and combines them according to a standard format to generate a geographic region identifier. For example, (34.05°N, 118.24°W) matches to Los Angeles, California, USA, and is converted to "US_CA_LA". For IP addresses, the server queries the IP address geolocation database to obtain the geographic location and generates a geographic region identifier. The server compares the device's geographic region identifier with the geographic region identifiers of each regional acoustic category one by one. When they match completely, the target acoustic category is determined. For example, if the device's geographic region identifier is "US_CA_LA", the server finds the category with the same identifier in the regional acoustic category table and determines it as the target acoustic category. When the device's geographic region identifier cannot be precisely matched, the server calculates the geographical distance between the device and the location represented by each regional acoustic category, selects the closest regional acoustic category as the target acoustic category, or uses the default globally universal audio equalization parameters. For example, the system automatically matches the "EU_FR" region parameter for French users, and reverts to a global template when a match cannot be found for specific locations such as airports. This geographic region matching mechanism helps find the most suitable regional acoustic category for each device based on its geographical environment.

[0038] The cloud server then reads the audio equalization parameters corresponding to the target acoustic category from the database, compresses and segments them before sending them to the device to be equalized. Based on the target acoustic category identifier, the server queries the audio equalization parameter table to retrieve the corresponding parameter file, which stores frequency compensation parameters for 128 frequency bands in JSON format. The file is compressed using the GZIP algorithm to identify and encode data repetition patterns. For example, a 200KB file can be compressed to 70KB, a compression ratio of 65%, significantly reducing the amount of data transmitted. The server divides the compressed file into multiple data segments, each 10KB in size, reading the file bytes sequentially, forming a segment of 10KB each. The 70KB file is divided into 7 data segments. Data segmentation cuts large files into smaller data blocks, facilitating batch transmission and resumeable interrupted transmission. A checksum is generated for each segment, and a 32-bit checksum is calculated using the CRC32 algorithm. The checksum is used to verify data integrity; the receiver can detect transmission errors by recalculating and comparing the checksum. The server encapsulates each fragment and its corresponding checksum into a parameter data packet. The data packet contains the fragment number, content, checksum, and total number of fragments. Seven parameter data packets are sent to the device sequentially. Upon receiving each data packet, the device verifies its integrity using the checksum. After receiving all packets, the data packets are merged and decompressed to restore the complete parameter file. Compression reduces bandwidth consumption and transmission time, fragmented transmission supports resuming interrupted transmissions after network outages, and the checksum ensures data transmission accuracy, preventing parameter errors from affecting audio quality.

[0039] In this embodiment of the invention, acoustic response data of the environment collected by multiple user devices and the geographical region identifiers corresponding to each user device are obtained through a cloud server; frequency response feature vectors of the acoustic response data are extracted, and acoustic clustering is performed on the frequency response feature vectors and the corresponding geographical region identifiers to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region; the frequency response feature vectors of each regional acoustic category are compared with a preset target frequency response curve to obtain the frequency response deviation, and audio equalization parameters are generated based on the frequency response deviation; the geographical region identifier of the device to be equalized is obtained, the regional acoustic category to which the device to be equalized belongs is determined, and the audio equalization parameters corresponding to the regional acoustic category are sent to the device to be equalized. This application constructs a regional acoustic template library by introducing a joint clustering mechanism of geographic region identifiers and acoustic feature vectors. It uniformly analyzes the environmental acoustic data of multiple users worldwide in the cloud, dynamically identifies the differences in architectural acoustic characteristics in different geographic regions, and realizes regional customization and precise distribution of audio equalization parameters. This solves the problem of inconsistent sound quality across regions caused by the fixed firmware parameters in existing audio equalization systems, effectively improves the acoustic adaptability and user experience consistency of the system in global deployment scenarios, and ensures that users in different geographic regions can obtain optimized sound quality performance that conforms to the characteristics of the local acoustic environment.

[0040] The audio equalization method based on cloud-based partition deployment in the embodiments of the present invention has been described above. The audio equalization device based on cloud-based partition deployment in the embodiments of the present invention will be described below. Please refer to [link / reference]. Figure 2 One embodiment of the audio equalization device based on cloud-based partition deployment in this invention includes: The data acquisition module 201 is used to acquire acoustic response data of the environment collected by multiple user devices and the geographical area identifier corresponding to each user device through the cloud server. Clustering analysis module 202 is used to extract the frequency response feature vector of the acoustic response data, and perform acoustic clustering on the frequency response feature vector and the corresponding geographical region identifier to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region. The parameter generation module 203 is used to compare the frequency response feature vector of each acoustic category in the region with the preset target frequency response curve to obtain the frequency response deviation, and generate audio equalization parameters based on the frequency response deviation. The distribution and push module 204 is used to obtain the geographical area identifier of the device to be equalized, determine the regional acoustic category to which the device to be equalized belongs, and send the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized.

[0041] In this embodiment of the invention, acoustic response data of the environment collected by multiple user devices and the geographical region identifiers corresponding to each user device are obtained through a cloud server; frequency response feature vectors of the acoustic response data are extracted, and acoustic clustering is performed on the frequency response feature vectors and the corresponding geographical region identifiers to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region; the frequency response feature vectors of each regional acoustic category are compared with a preset target frequency response curve to obtain the frequency response deviation, and audio equalization parameters are generated based on the frequency response deviation; the geographical region identifier of the device to be equalized is obtained, the regional acoustic category to which the device to be equalized belongs is determined, and the audio equalization parameters corresponding to the regional acoustic category are sent to the device to be equalized. This application constructs a regional acoustic template library by introducing a joint clustering mechanism of geographic region identifiers and acoustic feature vectors. It uniformly analyzes the environmental acoustic data of multiple users worldwide in the cloud, dynamically identifies the differences in architectural acoustic characteristics in different geographic regions, and realizes regional customization and precise distribution of audio equalization parameters. This solves the problem of inconsistent sound quality across regions caused by the fixed firmware parameters in existing audio equalization systems, effectively improves the acoustic adaptability and user experience consistency of the system in global deployment scenarios, and ensures that users in different geographic regions can obtain optimized sound quality performance that conforms to the characteristics of the local acoustic environment.

[0042] above Figure 2The audio equalization device based on cloud partition deployment in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The audio equalization device based on cloud partition deployment in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0043] Figure 3 This is a schematic diagram of a cloud-based partitioned audio equalizer 300 according to an embodiment of the present invention. The cloud-based partitioned audio equalizer 300 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 310 (e.g., one or more processors) and a memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) for storing applications 333 or data 332. The memory 320 and storage media 330 can be temporary or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the cloud-based partitioned audio equalizer 300. Furthermore, the processor 310 may be configured to communicate with the storage media 330 and execute the series of instruction operations in the storage media 330 on the cloud-based partitioned audio equalizer 300.

[0044] The cloud-based, zone-deployed audio equalizer 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 3 The illustrated structure of a cloud-based partitioned audio equalizer does not constitute a limitation on cloud-based partitioned audio equalizers, which may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0045] The present invention also provides an audio equalization device based on cloud-based partition deployment. The computer device includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs each step of the audio equalization method based on cloud-based partition deployment in the above embodiments.

[0046] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the various steps of the cloud-based partitioned audio equalization method.

[0047] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0048] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0049] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0050] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An audio equalization method based on cloud-based partitioned deployment, characterized in that, The cloud-based partitioned audio equalization method includes: Acoustic response data of the environment collected by multiple user devices and the geographical area identifiers corresponding to each user device are obtained through a cloud server. Extract the frequency response feature vector of the acoustic response data, and perform acoustic clustering on the frequency response feature vector and the corresponding geographic region identifier to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographic region. The frequency response feature vectors of each acoustic category in the region are compared with the preset target frequency response curve to obtain the frequency response deviation, and audio equalization parameters are generated based on the frequency response deviation. Obtain the geographical region identifier of the device to be equalized, determine the regional acoustic category to which the device to be equalized belongs, and send the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized.

2. The audio equalization method based on cloud-based partitioned deployment according to claim 1, characterized in that, The process of acquiring acoustic response data of the environment collected by multiple user devices and the geographical area identifiers corresponding to each user device through a cloud server includes: The system receives preset calibration signals from multiple user devices via a cloud server and then collects environmental response signals from the surrounding environment through a built-in microphone. The environmental response signal is subjected to frequency domain transformation and noise filtering to generate purified acoustic response data. The cloud server parses the location information data uploaded by each user device and converts the location information data into corresponding geographic area identifiers.

3. The audio equalization method based on cloud-based partitioned deployment according to claim 1, characterized in that, The frequency response feature vector of the acoustic response data is extracted, and acoustic clustering is performed on the frequency response feature vector and the corresponding geographical region identifier to obtain multiple regional acoustic categories, including: The acoustic response data is subjected to a fast Fourier transform to obtain frequency domain data, and the energy value of each frequency point in the frequency domain data within a preset frequency range is calculated. The energy values ​​of each frequency point are arranged in frequency order to generate a frequency response feature vector. The comprehensive clustering distance between each user device is calculated based on the similarity distance between the frequency response feature vectors and the association weight of the corresponding geographical region identifiers. Based on the comprehensive clustering distance, each user device is grouped and clustered to generate multiple regional acoustic categories.

4. The audio equalization method based on cloud-based partitioned deployment according to claim 3, characterized in that, The step of calculating the comprehensive clustering distance between user devices based on the similarity distance between the frequency response feature vectors and the association weight of the corresponding geographical region identifiers includes: Calculate the Euclidean distance between the frequency response feature vectors of any two user devices to obtain the similarity distance; Based on the geographic region identifiers of any two user devices, determine the geographic proximity between the two user devices, and determine the corresponding association weight based on the geographic proximity, wherein the higher the geographic proximity, the smaller the association weight; The similarity distance between each user device is multiplied by the corresponding association weight to obtain the comprehensive clustering distance between each user device.

5. The audio equalization method based on cloud-based partitioned deployment according to claim 1, characterized in that, The step of comparing the frequency response feature vectors of each acoustic category in the region with the preset target frequency response curve to obtain the frequency response deviation, and generating audio equalization parameters based on the frequency response deviation, includes: Statistical calculations are performed on the frequency response feature vectors within each of the aforementioned regional acoustic categories to obtain the average frequency response value of each frequency band. Based on the average frequency response value of each frequency band, a regional acoustic template for the corresponding regional acoustic category is constructed. Calculate the difference between the acoustic template of each region and the preset target frequency response curve in each frequency band to obtain the frequency response deviation of each region acoustic category in each frequency band. Based on the frequency response deviation, frequency compensation parameters are determined, and these parameters are encapsulated as audio equalization parameters for the corresponding acoustic category of the region.

6. The audio equalization method based on cloud-based partitioned deployment according to claim 5, characterized in that, The step of statistically calculating the frequency response feature vectors within each of the aforementioned regional acoustic categories to obtain the average frequency response value for each frequency band, and constructing a regional acoustic template for the corresponding regional acoustic category based on the average frequency response value for each frequency band, includes: The frequency response feature vectors within each acoustic category are reorganized by frequency band to obtain the feature value groupings corresponding to each frequency band. The characteristic values ​​of each frequency band are grouped and averaged to obtain the average frequency response value of the corresponding frequency band. The average frequency response values ​​of each frequency band are arranged in frequency order to generate the regional acoustic template for the aforementioned regional acoustic category.

7. The audio equalization method based on cloud-based partitioned deployment according to claim 1, characterized in that, The step of obtaining the geographical region identifier of the device to be equalized, determining the regional acoustic category to which the device to be equalized belongs, and sending the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized includes: Receive the balancing parameter request sent by the device to be balanced, and parse the geographical location data in the balancing parameter request; The geographic location data is converted into geographic region identifiers, and the geographic region identifiers are matched with the acoustic categories of each region to determine the target acoustic category; The audio equalization parameters corresponding to the target acoustic category are read from the cloud database and sent to the device to be equalized.

8. An audio equalizer based on cloud-based partitioned deployment, characterized in that, The cloud-based partitioned audio equalizer includes: The data acquisition module is used to acquire acoustic response data of the environment collected by multiple user devices and the geographical area identifiers corresponding to each user device through the cloud server. The clustering analysis module is used to extract the frequency response feature vector of the acoustic response data, and perform acoustic clustering on the frequency response feature vector and the corresponding geographical region identifier to obtain multiple regional acoustic categories, wherein each regional acoustic category corresponds to at least one geographical region. The parameter generation module is used to compare the frequency response feature vector of each acoustic category in the region with the preset target frequency response curve to obtain the frequency response deviation, and generate audio equalization parameters based on the frequency response deviation. The distribution and push module is used to obtain the geographical region identifier of the device to be equalized, determine the regional acoustic category to which the device to be equalized belongs, and send the audio equalization parameters corresponding to the regional acoustic category to the device to be equalized.

9. An audio equalizer based on cloud-based partitioned deployment, characterized in that, The cloud-based partitioned audio equalizer includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the cloud-based partitioned audio equalization device to perform the steps of the cloud-based partitioned audio equalization method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the steps of the audio equalization method based on cloud partition deployment as described in any one of claims 1-7.

Images