A multi-layered linkage sound leakage suppression method and system for open earphones

Abstract

This application relates to the technical field of audio data processing, and provides a multi-layer linkage sound leakage suppression method and system for open-back headphones, including: obtaining an initial sound leakage suppression level; matching corresponding processing parameters according to the initial sound leakage suppression level; obtaining target audio data to be played, analyzing sound leakage adjustment data, performing left and right channel coordinated adjustment and time-domain expansion processing on the target audio data to obtain a corresponding frequency band energy ratio evaluation value; when the frequency band energy ratio evaluation value is greater than a preset threshold, determining an updated sound leakage suppression level based on the frequency band energy ratio evaluation value and matching the updated processing parameters; repeating the above steps for subsequent target audio data to be played until the frequency band energy ratio evaluation value is less than the preset threshold. By continuously adaptively adjusting the subsequent target audio data to be played, the problem of insufficient sound leakage suppression accuracy in the prior art is overcome.

Description

Technical Field

[0001] This invention relates to the field of audio data processing technology, and in particular to a multi-layer linkage method and system for suppressing sound leakage in open-back headphones. Background Technology

[0002] With the widespread use of open-back headphones in sports, fitness, office commutes, and daily entertainment, while ensuring ambient sound perception and wearing comfort, they inevitably experience some sound leakage due to the open-back speaker unit's acoustic structure. To reduce the impact of sound leakage on those around them, existing open-back headphones typically employ digital audio processing techniques such as left / right channel adjustment, frequency compensation, time-domain extension, or acoustic compensation to suppress sound leakage and improve playback performance in various usage scenarios.

[0003] However, most existing sound leakage suppression solutions adjust a single processing stage independently, lacking a collaborative mechanism between the stages. Furthermore, the processing parameters are usually configured in a fixed or independent manner, making it difficult to optimize multiple processing stages in a coordinated manner based on the sound leakage status. This results in inconsistent sound leakage suppression effects under different wearing conditions, playback content, and acoustic environments, leading to insufficient accuracy in sound leakage suppression. Summary of the Invention

[0004] This invention provides a multi-layer linkage sound leakage suppression method and system for open-back headphones, to solve the problem of insufficient accuracy in sound leakage suppression in the prior art.

[0005] In a first aspect, embodiments of this application provide a multi-layer linkage sound leakage suppression method for open-back headphones, comprising: obtaining an initial sound leakage suppression level; matching corresponding processing parameters according to the initial sound leakage suppression level, wherein the processing parameters include a folding ratio coefficient, a delay depth, a mixing ratio, and a gain coefficient; obtaining target audio data to be played; performing frequency band analysis on the target audio data; extracting high-frequency difference information corresponding to the left and right channels; and performing a weighted operation on the high-frequency difference information using the folding ratio coefficient in the processing parameters to obtain sound leakage adjustment data; performing left and right channel coordinated adjustment on the target audio data based on the sound leakage adjustment data to obtain first processed audio data; obtaining the difference signal between the first processed audio data and the target audio data; and using the difference signal as residual adjustment information corresponding to time-domain extension processing; and performing a weighted operation on the residual adjustment information using the delay depth and mixing ratio in the processing parameters. The process involves time-domain expansion processing, where residual adjustment information is superimposed onto the first processed audio data to obtain expanded audio data. This expanded audio data is then input into a preset acoustic propagation model to generate a corresponding compensation signal. The compensation signal is weighted using the gain coefficients in the processing parameters, and the expanded audio data is then compensated using the weighted compensation signal to obtain second processed audio data. The second processed audio data undergoes frequency band energy ratio evaluation to obtain a corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than a preset threshold, an updated leakage suppression level is determined based on the frequency band energy ratio evaluation value. Updated processing parameters are then matched based on the updated leakage suppression level. The above steps are repeated on subsequent target audio data to be played using the updated processing parameters until the frequency band energy ratio evaluation value is less than the preset threshold.

[0006] Secondly, embodiments of this application provide a multi-layer linkage sound leakage suppression system for open-back headphones, comprising: a level acquisition module, configured to acquire an initial sound leakage suppression level and match corresponding processing parameters according to the initial sound leakage suppression level, wherein the processing parameters include a folding ratio coefficient, a delay depth, a mixing ratio, and a gain coefficient; a frequency band analysis module, configured to acquire target audio data to be played, perform frequency band analysis on the target audio data, extract high-frequency difference information corresponding to the left and right channels, and perform weighted calculation on the high-frequency difference information using the folding ratio coefficient in the processing parameters to obtain sound leakage adjustment data; a collaborative processing module, configured to perform collaborative adjustment of the left and right channels of the target audio data based on the sound leakage adjustment data to obtain first processed audio data; and a time-domain expansion module, configured to acquire the difference signal between the first processed audio data and the target audio data, and use the difference signal as residual adjustment information corresponding to the time-domain expansion processing, and use the delay depth and mixing ratio in the processing parameters to adjust the residual adjustment... The audio data is processed by performing time-domain expansion on the segment information, and the residual adjustment information after time-domain expansion is superimposed on the first processed audio data to obtain expanded audio data. The expanded audio data is then input into a preset acoustic propagation model to generate a corresponding compensation signal. A weighted processing module is used to weight the compensation signal using the gain coefficient in the processing parameters, and then use the weighted compensation signal to compensate the expanded audio data to obtain second processed audio data. A level update module is used to evaluate the frequency band energy ratio of the second processed audio data to obtain a corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than a preset threshold, an updated leakage suppression level is determined based on the frequency band energy ratio evaluation value, and updated processing parameters are matched based on the updated leakage suppression level. A repetition execution module is used to repeatedly execute the above steps on the subsequent target audio data to be played using the updated processing parameters until the frequency band energy ratio evaluation value is less than the preset threshold.

[0007] In one approach provided by the aforementioned method and system, an initial leakage suppression level is obtained. Corresponding processing parameters are then matched based on this initial level. The folding ratio coefficient, delay depth, mixing ratio, and gain coefficient within these processing parameters are applied to multiple processing stages, including high-frequency differential information processing, left and right channel coordinated adjustment, time-domain extension processing, and compensation processing. This allows each processing stage to coordinate and collaborate around unified processing parameters. After completing leakage suppression processing, the frequency band energy ratio of the second-processed audio data is further evaluated. Based on the frequency band energy ratio evaluation value, the updated leakage suppression level is dynamically determined, and the updated processing parameters are matched. This allows for continuous adaptive adjustment of the target audio data to be played subsequently. This enables collaborative optimization of multiple processing stages based on the actual leakage state, improving the leakage suppression accuracy of open-back headphones under different playback content, wearing conditions, and usage environments. This overcomes the problem of insufficient leakage suppression accuracy in existing technologies while also ensuring the consistency and stability of audio playback effects. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram of the structure of an audio processing environment in one embodiment of the present invention; Figure 2 This is a schematic flowchart of a multi-layer linkage sound leakage suppression method for open-back headphones according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the process for generating the first audio data in one embodiment of the present invention; Figure 4 This is a schematic diagram of the process for high-frequency differential information extraction and leakage sound adjustment data production in one embodiment of the present invention; Figure 5 This is a schematic diagram of the updated processing parameter generation process in one embodiment of the present invention; Figure 6 This is a schematic diagram of the mapping table between sound leakage suppression level and processing parameter group in one embodiment of the present invention; Figure 7 This is a schematic diagram of a multi-layer linkage sound leakage suppression system for open-back headphones according to an embodiment of the present invention. Detailed Implementation

[0010] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0011] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. It should also be understood that, as used in this specification and the appended claims, the term "and / or" refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0012] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0013] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0014] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0015] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0016] To balance ambient sound perception and wearing comfort, current open-back headphones generally employ an open-back acoustic structure for audio playback. However, because open-back structures lack the acoustic isolation provided by a closed ear canal, some high-frequency sound energy emitted by the speaker propagates outwards, resulting in varying degrees of sound leakage depending on wearing conditions, playback content, and usage environment. Existing sound leakage suppression methods mostly adjust individual processing stages—left / right channel adjustment, frequency compensation, time domain expansion, or compensation processing—independently. The lack of a unified parameter linkage mechanism between these stages makes it difficult to maintain consistent adjustment levels. When wearing conditions or playback content change, some processing stages may be under-adjusted while others are over-adjusted, thus reducing the overall accuracy of sound leakage suppression.

[0017] To address the aforementioned issues, this application proposes a multi-layered linkage sound leakage suppression method for open-back headphones. This method obtains an initial sound leakage suppression level, matches corresponding processing parameters based on this level, and utilizes the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient within these parameters to apply to multiple processing stages, including high-frequency differential information processing, left and right channel coordinated adjustment, time-domain expansion processing, and compensation processing. Simultaneously, the processing parameters are dynamically updated based on the frequency band energy ratio evaluation value corresponding to the second processed audio data. This ensures that multiple processing stages are always coordinated and adjusted around unified processing parameters, thereby improving the sound leakage suppression accuracy of open-back headphones under different playback content, wearing conditions, and usage environments, and guaranteeing the stability of audio playback performance.

[0018] This invention provides a multi-layer linkage sound leakage suppression method for open-back headphones, which can be applied to, for example... Figure 1 The audio processing environment shown includes a terminal device and an audio processing device. The terminal device and the audio processing device interact with each other via a wired bus or wireless communication. The terminal device acquires the target audio data to be played and performs corresponding sound leakage suppression processing. The audio processing device stores processing parameters, filter coefficients, and filter coefficients corresponding to the acoustic propagation model, and works with the terminal device to control audio playback.

[0019] In practical applications, after receiving the target audio data to be played, the terminal device matches the corresponding processing parameters according to the currently set initial sound leakage suppression level, adjusts the high-frequency differential information of the left and right channels in the target audio data, and sequentially completes left and right channel coordinated adjustment, time domain expansion processing, compensation processing, and frequency band energy ratio evaluation. When the frequency band energy ratio evaluation value exceeds the preset threshold, the sound leakage suppression level is adjusted according to the frequency band energy ratio evaluation value, and the corresponding processing parameters are rematched, so that subsequent target audio data to be played will continue to undergo sound leakage suppression processing with the updated processing parameters, thereby enabling the sound leakage suppression effect to be dynamically adjusted according to changes in the playback status.

[0020] The terminal device can be any one of a smartphone, tablet computer, laptop computer, personal computer, portable audio player, smartwatch, or portable wearable device; the audio processing device can be integrated inside the open-back headphones or integrated into the digital audio processing module in the terminal device, and the corresponding audio processing program is executed by a digital signal processor (DSP) or a central processing unit to complete the method corresponding to the embodiments of this application.

[0021] In one embodiment, such as Figure 2 As shown, a multi-layer linkage sound leakage suppression method for open-back headphones is provided, comprising the following steps: S10: Obtain the initial sound leakage suppression level, and match the corresponding processing parameters according to the initial sound leakage suppression level. The processing parameters include the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient.

[0022] The terminal device pre-establishes a correspondence between sound leakage suppression levels and processing parameters, and stores this correspondence in its local memory. Sound leakage suppression levels include at least four levels: off, level one, level two, level three, and level four. Each sound leakage suppression level uniquely corresponds to a set of processing parameters, and each set of processing parameters includes a folding ratio coefficient, delay depth, mixing ratio, and gain coefficient. When the terminal device starts an audio playback task, it reads the currently set initial sound leakage suppression level, finds the corresponding processing parameters in the correspondence based on the initial sound leakage suppression level, and loads them into the runtime cache as the current processing parameters. This ensures that subsequent audio processing processes, such as high-frequency differential information processing, left and right channel coordinated adjustment, time-domain expansion processing, and compensation processing, all utilize the same set of processing parameters to complete the audio processing.

[0023] For example, during the manufacturing process, based on the speaker size, the installation distance between the speaker and the ear canal, the target frequency response, and the preset leakage suppression target for open-back headphones, multiple leakage suppression levels are calibrated to establish a one-to-one correspondence between leakage suppression levels and processing parameters. When the initial leakage suppression level is set to level two, the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient corresponding to level two are read from memory and used as the current processing parameters for leakage suppression processing of the current audio frame. When the initial leakage suppression level changes during playback, the processing parameters associated with the corresponding level are reread to ensure that multiple processing stages corresponding to the same leakage suppression level always use consistent parameter configurations.

[0024] S20: Obtain the target audio data to be played, perform frequency band analysis on the target audio data, extract the high-frequency differential information corresponding to the left and right channels, and use the folding ratio coefficient in the processing parameters to perform weighted calculation on the high-frequency differential information to obtain the leakage adjustment data.

[0025] The terminal device acquires the left and right channel digital audio signals corresponding to the target audio data to be played. Based on the sampling rate of the target audio data, it selects a matching digital filter to perform frequency band analysis on the difference signal between the left and right channel signals. A second-order high-pass filter is then used to extract high-frequency components above a preset cutoff frequency as high-frequency differential information. The preset cutoff frequency is pre-set based on the acoustic leakage frequency range of the open-back headphones, for example, any frequency value among 2kHz, 2.5kHz, or 3kHz. After extracting the high-frequency differential information, the folding ratio coefficient in the current processing parameters is used to perform a weighted calculation on the high-frequency differential information. This results in leakage adjustment data generated according to the adjustment intensity corresponding to the current leakage suppression level. The larger the folding ratio coefficient, the greater the adjustment amplitude of the generated leakage adjustment data.

[0026] When the sampling rate of the target audio data is 48kHz, a second-order high-pass filter coefficient corresponding to 48kHz is selected to perform high-pass filtering on the difference signal between the left and right channel signals, retaining only high-frequency components above 2.5kHz as high-frequency difference information. If the folding ratio coefficient in the current processing parameters is 0.35, the extracted high-frequency difference information is multiplied by 0.35 to obtain the corresponding leakage adjustment data. This ensures that subsequent left and right channel coordinated adjustment only targets the high-frequency difference components that are prone to sound leakage, while maintaining the original playback effect of the low-frequency signal.

[0027] S30: Based on the leakage adjustment data, the left and right channels of the target audio data are coordinated to obtain the first processed audio data.

[0028] The terminal device reads the left and right channel signals from the target audio data and calls the leakage adjustment data generated in step S20 to perform synchronous and coordinated adjustment on the left and right channels. Specifically, the leakage adjustment data is subtracted from the left channel signal, and the same leakage adjustment data is superimposed on the right channel signal, causing the corresponding high-frequency differential components of the left and right channels to adjust synchronously in opposite directions, thereby reducing the high-frequency difference amplitude between the left and right channels. After completing the left and right channel adjustment, the updated left channel signal and the updated right channel signal are recombined to form the first processed audio data.

[0029] By using synchronized adjustment of the left and right channels, rather than adjusting them independently, it ensures that the left and right channels always change in sync, keeping the sound image center stable and avoiding stereo positioning shifts caused by adjusting only one channel. At the same time, since sound leakage mainly originates from the significant high-frequency differences between the left and right channels, reducing these differences synchronously can decrease the high-frequency sound energy radiated outwards by open-back headphones.

[0030] like Figure 3 As shown, when the amplitude of the left channel signal at a certain sampling point in the target audio data is 0.42 and the amplitude of the right channel signal is 0.36, and the leakage adjustment data generated in step S20 is 0.03, 0.03 is subtracted from the left channel signal to obtain the updated left channel signal 0.39; 0.03 is superimposed on the right channel signal to obtain the updated right channel signal 0.39. After completing the adjustment of the left and right channels corresponding to all sampling points, the updated left and right channels are recombined to obtain the first processed audio data, so that the high-frequency difference between the left and right channels is reduced, while the overall audio playback content remains consistent.

[0031] S40: Obtain the difference signal between the first processed audio data and the target audio data, and use the difference signal as the residual adjustment information corresponding to the time-domain expansion processing. Use the delay depth and mixing ratio in the processing parameters to perform time-domain expansion processing on the residual adjustment information. Superimpose the time-domain expanded residual adjustment information onto the first processed audio data to obtain the expanded audio data. Input the expanded audio data into the preset acoustic propagation model to generate the corresponding compensation signal.

[0032] The terminal device compares the amplitude differences between the corresponding sampling points of the first processed audio data and the target audio data, and defines the difference signal between the two as residual adjustment information. The residual adjustment information reflects the audio components that actually change during the coordinated adjustment of the left and right channels. Therefore, the residual adjustment information mainly includes the high-frequency information that has been reduced or redistributed by the left and right channel adjustment.

[0033] The terminal device writes residual adjustment information into a circular delay buffer. The buffer depth of the circular delay buffer is configured according to the delay depth in the processing parameters, ensuring that each frame of residual adjustment information can participate in subsequent processing according to the set delay length. Subsequently, multiple historical residual adjustment information is read from multiple equally spaced tap positions in the circular delay buffer and averaged to create a time-diffusion effect for the residual adjustment information corresponding to different time positions. After the averaging operation is completed, a preset all-pass filter is used to perform diffusion filtering on the average residual adjustment information.

[0034] Among them, the all-pass filter adopts the first-order or second-order all-pass filter commonly used in the field of digital signal processing. Its amplitude-frequency response remains constant, and only the phase relationship corresponding to each frequency component is changed, so that the residual adjustment information can form a more uniform diffusion effect in the time dimension without changing its frequency amplitude.

[0035] After the diffusion filtering process is completed, the residual adjustment information after diffusion is superimposed onto the left and right channel signals in the first processed audio data according to the mixing ratio in the processing parameters to obtain the expanded audio data.

[0036] Subsequently, the expanded audio data is input into the acoustic propagation model. The acoustic propagation model employs a finite impulse response (FIR) filter based on the acoustic leakage transfer function corresponding to the open-back headphone structure. The FIR filter is implemented using existing FIR filters from the field of digital signal processing, and its filter order is determined based on the impulse response length corresponding to the actual leakage path of the open-back headphone, such as 64th, 96th, or 128th order. The default coefficients corresponding to the filter are obtained through acoustic testing of the headphone prototype and stored in the firmware; calibration coefficients are obtained by measuring the actual acoustic leakage response of different headphones during the production line calibration phase and written to non-volatile memory. When calibration coefficients are detected in the non-volatile memory, they are used as the filter coefficients corresponding to the FIR filter; otherwise, the default coefficients are used to complete the convolution operation to obtain the compensation signal.

[0037] When the delay depth in the processing parameters is configured to 64 sampling points, the continuously generated residual adjustment information is sequentially written into a circular delay buffer of length 64. Historical residual adjustment information corresponding to the 0th, 16th, 32nd, and 48th tap positions is read, and the average of the four sets of historical residual adjustment information is calculated. Then, a first-order all-pass filter is used to complete the diffusion filtering process. If the mixing ratio in the current processing parameters is 0.30, it is superimposed on the left and right channels of the first processed audio data at a ratio of 30%, forming the expanded audio data. The expanded audio data is input into a 96th-order finite impulse response filter and then convolved to obtain a compensation signal, enabling the compensation signal to simulate the actual acoustic leakage propagation process corresponding to open-back headphones.

[0038] S50: The compensation signal is weighted using the gain coefficient in the processing parameters, and the expanded audio data is compensated using the weighted compensation signal to obtain the second processed audio data.

[0039] The terminal device reads the gain coefficient from the processing parameters and applies it to the compensation signal to adjust its amplitude. The gain coefficient is set to a negative gain coefficient, ensuring that the compensation signal and the expanded audio data are superimposed as inverse signals. After weighting, the weighted compensation signal and the expanded audio data are subtracted one-to-one according to their sampling points to obtain the second processed audio data.

[0040] The specific value of the negative gain coefficient is determined by the corresponding sound leakage suppression level, with different sound leakage suppression levels corresponding to different compensation intensities. The gain coefficient can be expressed as linear gain or decibel gain. When expressed as decibel gain, it is converted into linear gain using existing decibel conversion methods in digital audio processing for use in weighted calculations.

[0041] When the amplitude of a certain sampling point of the compensation signal is 0.12, and the gain coefficient in the current processing parameters is -0.45, the compensation signal is multiplied by -0.45 to obtain a weighted compensation signal of -0.054. If the amplitude of the sampling point corresponding to the expanded audio data is 0.38, then the compensation processing corresponding to 0.38-(-0.054) is performed to obtain the amplitude of the sampling point corresponding to the second processed audio data of 0.434. After completing the compensation processing corresponding to all sampling points, the second processed audio data is obtained, so that the compensation signal can offset the leakage sound energy actually generated by the open-back headphones according to the compensation intensity corresponding to the current leakage suppression level, improve the overall leakage suppression effect, and at the same time ensure that the audio playback content remains stable.

[0042] S60: Evaluate the frequency band energy ratio of the second processed audio data to obtain the corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than the preset threshold, determine the updated sound leakage suppression level based on the frequency band energy ratio evaluation value, and match the updated processing parameters based on the updated sound leakage suppression level.

[0043] The terminal device acquires the second processed audio data, divides the audio data into frequency bands, and calculates the low-frequency energy envelope and high-frequency energy envelope respectively. The low-frequency energy envelope corresponds to the energy change at each sampling point within a preset low-frequency range, and the high-frequency energy envelope corresponds to the energy change at each sampling point within a preset high-frequency range. Frequency band division can be accomplished using digital filtering methods, such as using finite impulse response bandpass filters or infinite impulse response bandpass filters in the field of digital signal processing to extract the low-frequency and high-frequency signals respectively. Then, the corresponding energy is calculated based on the square of the amplitude at each sampling point, and the corresponding energy envelope is generated using the sliding window averaging method.

[0044] The terminal device calculates a frequency band energy ratio assessment value based on the ratio between the low-frequency energy envelope and the high-frequency energy envelope. This assessment value reflects the proportion of high-frequency leakage components in the current second-processed audio data relative to the overall audio energy. A higher frequency band energy ratio assessment value indicates a higher proportion of high-frequency leakage components in the current audio, requiring further improvement in leakage suppression.

[0045] The preset threshold is pre-set based on the target sound leakage index corresponding to the open-back headphones. When establishing the preset threshold, actual sound leakage data corresponding to multiple different sound leakage suppression levels can be collected in a standard anechoic environment, and the range of frequency band energy ratio evaluation values ​​corresponding to meeting the target sound leakage requirements can be statistically analyzed. The average value or weighted average value obtained from the statistics is then written into the memory as the preset threshold.

[0046] The terminal device compares the calculated frequency band energy ratio assessment value with a preset threshold. When the frequency band energy ratio assessment value is greater than the preset threshold, the updated sound leakage suppression level is determined according to the preset level adjustment strategy.

[0047] The preset level adjustment strategy is pre-established in memory, employing a one-to-one correspondence between frequency band energy ratio evaluation value intervals and leakage suppression levels. The establishment method is as follows: based on a large number of audio sample test results under different leakage suppression levels, the actual leakage effect corresponding to different frequency band energy ratio evaluation values ​​is statistically analyzed, and multiple continuous evaluation intervals are divided, each uniquely corresponding to a leakage suppression level. When the frequency band energy ratio evaluation value falls into the corresponding interval, the updated leakage suppression level is determined.

[0048] After determining the updated sound leakage suppression level, the corresponding folding ratio coefficient, delay depth, mixing ratio, and gain coefficient are read from the pre-established processing parameter correspondence based on the updated sound leakage suppression level, and then combined to form the updated processing parameters.

[0049] For example, the terminal device uses a finite impulse response bandpass filter to extract low-frequency signals in the 200Hz–1000Hz range and high-frequency signals above 2500Hz, respectively, and calculates the energy envelope for each frequency band. If the current average energy envelope value for the low-frequency band is 0.82 and the average energy envelope value for the high-frequency band is 0.36, the calculated corresponding frequency band energy ratio evaluation value is approximately 0.439. If the preset threshold is set to 0.35, since 0.439 is greater than 0.35, the corresponding relationship is queried according to the preset level adjustment strategy to determine that the leakage suppression level is adjusted from level two to level three, and the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient corresponding to level three are read to form the updated processing parameters.

[0050] S70: Repeat the above steps for the target audio data to be played later using the updated processing parameters until the frequency band energy ratio evaluation value is less than the preset threshold.

[0051] The terminal device writes the updated processing parameters into the current audio processing configuration and keeps the updated processing parameters active. When a new audio frame or new target audio data enters the audio processing flow, the updated processing parameters are directly called to re-execute the processing procedures corresponding to steps S20 to S60, so that the new target audio data completes the sound leakage suppression processing according to the updated folding ratio coefficient, delay depth, mixing ratio, and gain coefficient.

[0052] Since the current audio frame has already completed playback buffering, in order to avoid playback interruption or transient distortion caused by reprocessing the already calculated audio data during playback, this embodiment applies the updated processing parameters to the target audio data to be played later, instead of reprocessing the already output second processed audio data.

[0053] The terminal device continuously executes steps S20 to S60 on the target audio data to be played. When the recalculated frequency band energy ratio evaluation value is less than the preset threshold, the audio playback continues with the currently updated processing parameters. If the recalculated frequency band energy ratio evaluation value is still greater than the preset threshold, the new leakage suppression level is re-determined according to the preset level adjustment strategy, and the processing parameters are updated again until the frequency band energy ratio evaluation value is less than the preset threshold.

[0054] Throughout the entire process, the initial sound leakage suppression level is only used to determine the processing parameters for the first audio processing, while subsequent processing parameters are dynamically updated based on the frequency band energy ratio evaluation value calculated in real time, so that the sound leakage suppression intensity can be continuously adjusted according to the actual playback status.

[0055] When the frequency band energy ratio evaluation value obtained in step S60 is 0.439, the updated processing parameters correspond to a level 3 sound leakage suppression level. After the next frame of target audio data enters the processing flow, the terminal device directly uses the corresponding folding ratio coefficient, delay depth, mixing ratio, and gain coefficient at level 3 to complete the left and right channel coordinated adjustment, time domain expansion processing, and compensation processing. The recalculated frequency band energy ratio evaluation value decreases to 0.31. Since 0.31 is less than the preset threshold of 0.35, the current processing parameters are maintained to continue processing the subsequent target audio data to be played, and the sound leakage suppression level adjustment is no longer performed. When the subsequent playback content changes and causes the frequency band energy ratio evaluation value to exceed the preset threshold again, the processing parameters are updated again according to the preset level adjustment strategy.

[0056] This embodiment provides a multi-layer linkage sound leakage suppression method for open-back headphones. It obtains an initial sound leakage suppression level, matches corresponding processing parameters based on this level, and uses a folding ratio coefficient in the processing parameters to adjust high-frequency differential information. It then uses delay depth and mixing ratio in the processing parameters to perform temporal expansion processing corresponding to residual adjustment information, and uses gain coefficients in the processing parameters to perform compensation processing. Finally, it dynamically determines the updated sound leakage suppression level based on the frequency band energy ratio evaluation value corresponding to the second processed audio data, and matches the updated processing parameters to the target audio data to be played. This ensures that the left and right channel coordinated adjustment, temporal expansion processing, and compensation processing are always linked and controlled around the same set of processing parameters, forming a closed-loop adjustment mechanism based on the frequency band energy ratio evaluation result. This allows for dynamic adjustment of the sound leakage suppression intensity according to the actual playback state, improving the accuracy of sound leakage suppression for open-back headphones under different playback content, different wearing conditions, and different usage environments. Simultaneously, it maintains the stability of audio playback effects and avoids inconsistencies caused by independent adjustments of multiple processing stages.

[0057] In one embodiment, such as Figure 4 As shown, in step S20, the target audio data to be played is acquired, frequency band analysis is performed on the target audio data, high-frequency difference information corresponding to the left and right channels is extracted, and the high-frequency difference information is weighted using the folding ratio coefficient in the processing parameters to obtain the leakage adjustment data. Specifically, the steps include the following: S21: Obtain the sampling rate corresponding to the target audio data to be played, and match the corresponding target filter coefficient from multiple pre-stored filter coefficients according to the sampling rate. The multiple filter coefficients correspond one-to-one with multiple different sampling rates.

[0058] The system acquires the target audio data to be played and reads the corresponding audio format information, which includes at least the sampling rate, quantization bit width, and number of channels. The terminal device parses the sampling rate from the audio format information and queries a pre-established filter coefficient mapping table based on the sampling rate to match the corresponding target filter coefficients.

[0059] The filter coefficient mapping table is pre-stored in memory, and the mapping relationship is established using the sampling rate as the index. Different sampling rates correspond to a set of independently designed filter coefficients. Each set of filter coefficients is obtained using the filter design methods corresponding to second-order infinite impulse response (IIR) filters in the existing digital signal processing field. For example, the filter coefficients are calculated using the bilinear transform method combined with the Butterworth high-pass filter design method, so that the cutoff frequency and filtering characteristics remain consistent under different sampling rate conditions.

[0060] When establishing filter coefficients, the preset cutoff frequency is digitally normalized based on the Nyquist frequency corresponding to each sampling rate. Then, the corresponding second-order infinite impulse response filter coefficients are calculated using the bilinear transform method. The calculated filter coefficients are then categorized and stored according to the sampling rate to form a filter coefficient mapping table. For example, multiple sets of filter coefficients corresponding to 44.1kHz, 48kHz, 96kHz, and 192kHz can be established, maintaining a one-to-one correspondence between each set of filter coefficients and the corresponding sampling rate.

[0061] After the terminal device completes the sampling rate parsing, it directly looks up the filter coefficient mapping table based on the sampling rate, reads the corresponding target filter coefficients, and loads them into the current audio processing cache for high-frequency differential information extraction corresponding to the current target audio data.

[0062] For example, the target audio data to be played uses PCM encoding format, and the audio file header records a sampling rate of 48kHz, a quantization bit width of 24bit, and a stereo format. The terminal device parses the audio file header to obtain a sampling rate of 48kHz, retrieves the corresponding filter coefficients for 48kHz from the filter coefficient mapping table, and reads the numerator and denominator coefficients of the corresponding second-order infinite impulse response high-pass filter as the target filter coefficients. If another segment of target audio data corresponds to a sampling rate of 96kHz, then another set of target filter coefficients corresponding to 96kHz is read, ensuring consistent high-frequency extraction performance under different sampling rate conditions without recalculating the filter coefficients.

[0063] S22: Obtain the difference signal between the left channel signal and the right channel signal in the target audio data, and use the target filter coefficients to perform second-order high-pass filtering on the difference signal to obtain high-frequency difference information of the corresponding high-frequency bands of the left and right channels. The frequency corresponding to the high-frequency difference information is higher than the preset cutoff frequency.

[0064] The terminal device reads the left and right channel signals from the target audio data in the order of sampling points, and performs difference calculation on the left and right channel sample values ​​corresponding to the same sampling time to form a difference signal between the left and right channels. The difference signal reflects the amplitude difference between the left and right channels at the same time position, and it contains difference information across the entire frequency range between the left and right channels.

[0065] The terminal device inputs the difference signal into the second-order high-pass filter corresponding to the target filter coefficient obtained in step S21 for filtering processing, retaining only the frequency components corresponding to the preset cutoff frequency, and filtering out the low-frequency difference components below the preset cutoff frequency, thereby obtaining the high-frequency difference information of the corresponding high-frequency bands of the left and right channels.

[0066] The preset cutoff frequency is pre-set based on the actual leakage frequency range of the open-back headphones. When establishing the preset cutoff frequency, acoustic leakage tests are performed on the open-back headphones, and the variation of leakage sound pressure level at different frequencies is statistically analyzed to determine the frequency range where sound leakage is mainly concentrated. The lower limit of this frequency range is then used as the preset cutoff frequency. When test results show that the proportion of sound leakage above 2.5kHz is relatively high, 2.5kHz is set as the preset cutoff frequency, ensuring that the high-pass filter primarily retains the high-frequency components that are prone to leakage.

[0067] Since the common audio components existing in both the left and right channels do not cause significant spatial differences, after high-pass filtering of the difference signals between the left and right channels, only the signal components that are significantly different between the left and right channels and are located in the high-frequency range are retained. This allows the subsequent sound leakage adjustment to mainly target the frequency range where open-back headphones are prone to sound leakage, reducing the impact on normal playback content.

[0068] When the left and right channel signals in a continuous audio frame form two sampling sequences, the terminal device calculates the difference between the left and right channel signals point by point according to the sampling order, forming the corresponding difference signal sequence. If the preset cutoff frequency is set to 2.5kHz, the difference signal is filtered using the second-order high-pass filter corresponding to 48kHz obtained in step S21. This filters out low-frequency differences below 2.5kHz, retaining only the high-frequency difference information corresponding to the frequency band above 2.5kHz, so that the obtained high-frequency difference information mainly reflects the high-frequency differences between the left and right channels that are prone to spatial leakage.

[0069] S23: Use the folding ratio coefficient in the processing parameters to perform weighted calculation on the high-frequency differential information to obtain the weighted high-frequency differential information, which can be used as leakage adjustment data.

[0070] The terminal device reads the folding ratio coefficient in the current processing parameters and performs a linear weighting operation on the high-frequency differential information using the folding ratio coefficient to obtain the weighted high-frequency differential information. The weighted high-frequency differential information is then determined as the leakage sound adjustment data.

[0071] The folding ratio coefficient is used to control the degree of compression of high-frequency differences between the left and right channels, and its value is determined by the processing parameters obtained in step S10. Different initial leakage suppression levels correspond to different folding ratio coefficients. The higher the leakage suppression level, the larger the corresponding folding ratio coefficient, and the smaller the proportion of high-frequency differential information retained, further reducing the high-frequency differences between the left and right channels.

[0072] The terminal device performs weighted calculations point by point according to the sampling points, so that the high-frequency differential information corresponding to each sampling point is scaled according to a uniform folding ratio coefficient, forming a continuous leakage adjustment data sequence. Since the folding ratio coefficient directly affects the high-frequency differential information, it does not change the common audio components contained in the left and right channels, but only adjusts the corresponding high-frequency differences between the left and right channels, thereby reducing the high-frequency sound energy radiated outward during playback of open-back headphones.

[0073] For example, in step S22, a high-frequency differential information sequence corresponding to an audio frame is obtained. The folding ratio coefficient in the current processing parameters is 0.35. The terminal device then multiplies each sampling point in the high-frequency differential information sequence by 0.35 to form a corresponding weighted high-frequency differential information sequence. This sequence is then used as leakage sound adjustment data in step S30 for left and right channel coordinated adjustment. When a folding ratio coefficient of 0.55 is subsequently obtained based on the updated leakage sound suppression level, the leakage sound adjustment data is regenerated using 0.55, further reducing the high-frequency difference between the left and right channels to meet higher leakage sound suppression requirements.

[0074] In this embodiment, by matching the target filter coefficients according to the sampling rate of the target audio data, the high-frequency differential information extraction is completed using the matching filtering characteristics under different sampling rate conditions. By performing second-order high-pass filtering on the difference signal between the left and right channel signals, the extracted high-frequency differential information is concentrated in the high-frequency range where open-back headphones are prone to sound leakage. By using the folding ratio coefficient in the processing parameters to perform weighted calculations on the high-frequency differential information, leakage adjustment data corresponding to the current leakage suppression level is generated, so that the leakage adjustment intensity can be kept consistent with the current processing parameters. This improves the accuracy of high-frequency differential information extraction and leakage adjustment data generation, and ensures that subsequent left and right channel coordinated adjustment can be linked and controlled around the unified processing parameters.

[0075] In one embodiment, step S30 involves applying the leakage adjustment data to the left and right channel signals according to the sign direction of the leakage adjustment data, performing inverse coordinated adjustment on the left and right channel signals to reduce the high-frequency difference information between the left and right channels, and obtaining the first processed audio data. This specifically includes the following steps: Obtain the left and right channel signals from the target audio data; subtract the leakage adjustment data from the left channel signal to obtain the updated left channel signal; superimpose the leakage adjustment data onto the right channel signal to obtain the updated right channel signal; combine the updated left and right channel signals to obtain the first processed audio data.

[0076] The terminal device reads the left and right channel signals from the target audio data according to the sampling order of the audio frames, and aligns the leakage adjustment data generated in step S23 with the left and right channel signals point by point according to the corresponding sampling points. Since the leakage adjustment data and the left and right channel signals have the same sampling rate, sampling length and sampling order, a unique correspondence can be established for each sampling point.

[0077] The terminal device sequentially performs coordinated adjustment processing for the left and right channels according to the sampling points. It subtracts the leakage adjustment data from the corresponding sample value of the left channel signal, while simultaneously adding the same leakage adjustment data to the corresponding sample value of the right channel signal, thus redistributing the original high-frequency differences between the left and right channels. Throughout the adjustment process, the left and right channels corresponding to the same sampling point use the same amplitude but opposite adjustment methods to maintain a basically consistent overall energy between the left and right channels, while reducing the degree of high-frequency differences between them.

[0078] After all sampling points are adjusted, the updated left channel signal and the updated right channel signal are recombined according to the original channel order to form a new stereo audio frame, which is then used as the first audio data to be processed.

[0079] For example, a target audio data corresponds to an audio frame, where both the left and right channel signals contain 480 sampling points. Step S23 obtains the leakage adjustment data corresponding to the 480 sampling points. The terminal device sequentially reads the 1st to the 480th sampling points. At the i-th sampling point, the corresponding leakage adjustment data is subtracted from the left channel signal, and the same leakage adjustment data is superimposed on the right channel signal. After processing all sampling points, the left and right channels are recombined to obtain the corresponding first processed audio data.

[0080] In this embodiment, by using leakage adjustment data to synchronously adjust the left and right channel signals, the high-frequency differences that are prone to sound leakage between the left and right channels are redistributed. While maintaining the continuity of the overall playback content, the spatial differences between the left and right channels are reduced, improving the leakage suppression effect of open-back headphones, and ensuring that subsequent time-domain extension processing continues to use the unified first processed audio data.

[0081] In one embodiment, step S40 involves acquiring the difference signal between the first processed audio data and the target audio data as residual adjustment information. The residual adjustment information is then subjected to time-domain expansion processing using the delay depth and mixing ratio in the processing parameters. The time-domain expanded residual adjustment information is then superimposed onto the first processed audio data to obtain expanded audio data. This expanded audio data is then input into a preset acoustic propagation model to generate a corresponding compensation signal. Specifically, the steps include the following: S41: Write the residual adjustment information into a ring-shaped delay buffer with a preset depth, where the preset depth is the delay depth in the processing parameters. Read the corresponding historical residual adjustment information from multiple tap positions set at equal intervals in the ring-shaped delay buffer, and perform an average calculation on all the historical residual adjustment information to obtain the average residual adjustment information.

[0082] The terminal device calculates the difference between the first processed audio data and the target audio data point by point according to the sampling points to obtain residual adjustment information. Since the first processed audio data is obtained by co-adjusting the left and right channels based on the target audio data, the difference between the two can reflect the adjustment component that actually changes during the co-adjustment process of the left and right channels. This adjustment component is used as residual adjustment information for subsequent time-domain extension processing.

[0083] The terminal device establishes a circular delay buffer, which is implemented using a circular queue structure. Its buffer capacity is determined by the delay depth in the processing parameters, and the delay depth is represented by the number of sampling points, such as 256, 512, 1024, or 2048 sampling points. Each time a new residual adjustment information sample value is received, it is written to the current write pointer position and the write pointer is updated. When the write pointer reaches the end of the buffer, it returns to the beginning of the buffer to continue writing.

[0084] Multiple tap positions are established in a circular delay buffer, with each tap position evenly spaced along the length of the buffer. When the delay depth is 1024 sampling points and four tap positions are set, each tap position corresponds to a distance of 256, 512, 768, and 1024 sampling points from the current write pointer, respectively. The terminal device reads the historical residual adjustment information corresponding to each tap position and uses an arithmetic averaging algorithm from the existing digital signal processing field to calculate the average residual adjustment information by averaging all historical residual adjustment information.

[0085] For example, if the delay depth in the processing parameters is 1024 sampling points, four equally spaced tap positions are set in the circular delay buffer, and the corresponding four segments of historical residual adjustment information are read respectively. The arithmetic mean operation is performed on the four sets of historical residual adjustment information for each sampling point to obtain the corresponding average residual adjustment information.

[0086] S42: The average residual adjustment information is diffused and filtered using a preset all-pass filter to obtain the diffused residual adjustment information. The diffused residual adjustment information is then superimposed onto the left and right channel signals in the first processed audio data according to the same mixing ratio in the processing parameters to obtain the expanded audio data.

[0087] The terminal device inputs the average residual regulation information into a preset all-pass filter for diffusion filtering. The all-pass filter is implemented using a first-order or second-order all-pass filter in the existing field of digital signal processing. Its amplitude frequency response remains constant, while its phase frequency response varies with frequency, thereby changing the time distribution of the residual regulation information and making the residual regulation information form a continuous diffusion effect.

[0088] The filter coefficients corresponding to the preset all-pass filter are pre-set according to the expected diffusion time. For example, the feedback coefficient can be set according to the diffusion time so that the residual regulation information after diffusion covers multiple consecutive sampling points, reducing the energy abrupt change caused by the concentration of residual regulation information at local sampling points.

[0089] The terminal device reads the mixing ratio in the processing parameters and superimposes the diffused residual adjustment information onto the left and right channel signals in the first processed audio data according to a uniform mixing ratio, so that the diffused residual adjustment information acts on both the left and right channels simultaneously to form the expanded audio data.

[0090] For example, if the mixing ratio in the processing parameters is 0.25, the terminal device will multiply the diffused residual adjustment information by 0.25 and then superimpose it onto the left and right channel signals in the first processed audio data, so that the expanded audio data increases the time diffusion range of the residual adjustment information while maintaining the overall audio continuity.

[0091] S43: Input the expanded audio data into a preset acoustic propagation model, where the acoustic propagation model is a finite impulse response (FIR) filter established based on the acoustic leakage transfer function corresponding to the open-back headphone structure. The filter coefficients of the FIR filter include default coefficients pre-stored in the firmware and calibration coefficients written to the non-volatile memory through production line calibration; when calibration coefficients are stored in the non-volatile memory, the calibration coefficients are used as the filter coefficients of the FIR filter; when calibration coefficients are not stored in the non-volatile memory, the default coefficients are used as the filter coefficients of the FIR filter.

[0092] After the open-back headphones have completed their structural design, during the production line calibration phase, the headphones are placed in a standard acoustic testing fixture. A measurement microphone is positioned at a predetermined measurement location on the outside of the headphones, and a preset test signal, such as a linear sweep signal, MLS sequence, or pulse test signal, is played. The acoustic leakage transfer function is calculated based on the relationship between the signal played by the headphones and the signal acquired by the measurement microphone.

[0093] Using the obtained acoustic leakage transfer function, a finite impulse response (FIR) filter is constructed using existing digital signal processing design methods (such as the window function method or frequency sampling method), and the corresponding filter coefficients are generated. During mass production, the calibration filter coefficients for each earphone are written into non-volatile memory; simultaneously, default filter coefficients applicable to the same model of earphone are pre-saved in the firmware.

[0094] The terminal device reads the calibration information from the non-volatile memory. When the calibration coefficients are detected, they are loaded as the filtering coefficients of the finite impulse response filter. When no calibration coefficients are detected, the default coefficients in the firmware are read directly to complete the filter establishment.

[0095] For example, a certain model of open-back headphones is calibrated at the factory to obtain the coefficients of a 128th-order finite impulse response (FIR) filter and written to non-volatile memory. During playback, if the terminal device detects the existence of the corresponding calibration coefficients in the non-volatile memory, it directly loads the 128th-order filter coefficients to establish the FIR filter. If the headphones have not undergone production line calibration, the default 128th-order filter coefficients in the firmware are read to establish the FIR filter.

[0096] S44: If calibration coefficients exist, they are used first. The expanded audio data is then convolved using the determined finite impulse response filter to obtain the compensation signal.

[0097] The terminal device first determines whether calibration coefficients exist. If calibration coefficients exist, the finite impulse response filter is built using the calibration coefficients; if no calibration coefficients exist, the finite impulse response filter is built using default coefficients.

[0098] The terminal device inputs the expanded audio data into the established finite impulse response filter, and uses the discrete convolution algorithm in the field of existing digital signal processing to perform convolution operation point by point. This allows the expanded audio data to be multiplied and accumulated with the filter coefficients corresponding to the finite impulse response filter to obtain the compensation signal.

[0099] Using calibration coefficients can make the finite impulse response filter more consistent with the actual acoustic propagation path of the corresponding open-back headphones. Therefore, prioritizing the use of calibration coefficients can improve the consistency between the compensation signal and the actual sound leakage.

[0100] For example, if the expanded audio data contains 480 sampling points and the finite impulse response filter has an order of 128, the terminal device will perform convolution calculations for the 480 sampling points sequentially according to the discrete convolution algorithm, and finally generate the corresponding compensation signal, which will be sent to step S50 to continue the compensation processing.

[0101] In this embodiment, residual adjustment information is generated by using the difference between the first processed audio data and the target audio data. This residual adjustment information is then combined with a ring delay buffer, an all-pass filter, and a mixing ratio to complete time-domain expansion processing, enabling continuous diffusion of the residual adjustment information in the time dimension. By establishing a finite impulse response filter based on the acoustic leakage transfer function corresponding to the open-back headphone structure and combining it with calibration coefficients obtained from production line calibration to generate a compensation signal, the compensation signal can more accurately simulate the actual sound leakage path of the open-back headphone, thereby improving the accuracy of subsequent compensation processing and the overall sound leakage suppression effect.

[0102] In one embodiment, step S50, which involves weighting the compensation signal using the gain coefficient in the processing parameters and then using the weighted compensation signal to compensate the expanded audio data to obtain the second processed audio data, specifically includes the following steps: Obtain the gain coefficient from the processing parameters. The gain coefficient is a negative gain coefficient. Use the gain coefficient to perform a weighted operation on the compensation signal to obtain a weighted compensation signal. Subtract the weighted compensation signal from the expanded audio data to obtain the second processed audio data.

[0103] The terminal device reads the gain coefficient from the current processing parameters and performs a weighted calculation on the compensation signal point by point according to the gain coefficient. The gain coefficient uses a negative gain coefficient, and its value is preset according to different leakage suppression levels and stored in the corresponding processing parameter group. For example, different negative gain coefficients such as -0.15, -0.25, -0.35, -0.45, and -0.55 can be used to correspond to different leakage suppression levels.

[0104] By weighting the compensation signal with a negative gain coefficient, the phase direction of the compensation signal is made opposite to the direction of the leakage sound wave actually generated by the open-back headphones, thus forming a reverse compensation signal to cancel out the leakage sound wave. The terminal device multiplies the compensation signal with the negative gain coefficient point by point according to the sampling point sequence to obtain the weighted compensation signal.

[0105] After weighting, the terminal device performs a sample-by-sample subtraction operation between the expanded audio data and the weighted compensation signal. The expanded audio data and the weighted compensation signal maintain the same sampling rate, sampling length, and sampling order, with each sampling point establishing a unique correspondence. Since the weighted compensation signal already includes negative gain adjustment, the subtraction operation allows the reverse acoustic energy corresponding to the compensation signal to act on the expanded audio data, creating mutually canceling sound pressure distributions in the direction of leakage from the open-back headphones, resulting in the second processed audio data.

[0106] Throughout the entire compensation process, only the gain of the compensation signal is adjusted, without changing the original playback content corresponding to the expanded audio data, thus keeping the playback content continuous and reducing sound leakage in the corresponding direction of the open-back headphones.

[0107] The current gain coefficient corresponding to the leakage suppression level is -0.40. Step S44 yields a compensation signal with 480 sampling points. The terminal device multiplies each of these 480 sampling points by -0.40 to form a weighted compensation signal, which is then subtracted from the expanded audio data point by point to obtain the second processed audio data. When the leakage suppression level is increased, the corresponding gain coefficient is adjusted to -0.55, resulting in a larger compensation amplitude for the compensation signal, thus enabling the open-back headphones to produce a stronger leakage cancellation effect.

[0108] In this embodiment, the compensation signal is weighted using the gain coefficient in the processing parameters, and a negative gain method is used to form an inverse compensation signal. The weighted compensation signal is then used to compensate the expanded audio data, so that the compensation signal can more accurately cancel the leakage sound waves generated by the open-back headphones, improve the accuracy of the compensation processing, and further reduce the sound leakage of the open-back headphones.

[0109] In one embodiment, after step S50, the following step is further included: S501: Obtain the spectral information corresponding to the second processed audio data, and divide it into multiple frequency bands according to a preset frequency band division rule. Determine the psychoacoustic masking threshold corresponding to each frequency band based on the spectral energy corresponding to each frequency band. Compare the spectral energy corresponding to each frequency band with the corresponding psychoacoustic masking threshold to determine the target compensation weight corresponding to each frequency band. Adjust the compensation gain of the second processed audio data corresponding to each frequency band according to the target compensation weight to obtain the psychoacoustic optimized target audio data.

[0110] In this step, after acquiring the second processed audio data, a Fast Fourier Transform (FFT) is performed on the second processed audio data to obtain the corresponding spectrum information, and the data is divided into multiple frequency bands according to a preset frequency band division rule.

[0111] The preset frequency band division rule is established based on the human ear's hearing frequency resolution. It can be established using the Bark frequency band division rule, the ERB (Equivalent Rectangular Bandwidth) frequency band division rule, or the Mel frequency band division rule. In this embodiment, the Bark frequency band division rule is used to divide 20Hz to 20kHz into multiple continuous frequency bands, and the spectral energy corresponding to each frequency band is counted.

[0112] Subsequently, the corresponding psychoacoustic masking threshold is calculated based on the spectral energy of each frequency band.

[0113] The psychoacoustic masking threshold is established based on existing psychoacoustic models and can be calculated using MPEG Psychoacoustic Model I or MPEG Psychoacoustic Model II. During the calculation, the energy of the current frequency band is used as the masking source energy, and the psychoacoustic masking threshold for each frequency band is calculated by combining the frequency spread patterns between adjacent frequency bands.

[0114] Then, the spectral energy corresponding to each frequency band is compared with the corresponding psychoacoustic masking threshold.

[0115] If the compensation energy corresponding to a certain frequency band is lower than the corresponding psychoacoustic masking threshold, the current compensation gain remains unchanged; if the compensation energy corresponding to a certain frequency band is higher than the corresponding psychoacoustic masking threshold, the corresponding target compensation weight is calculated according to the proportion of spectral energy exceeding the psychoacoustic masking threshold.

[0116] The target compensation weight can be determined according to the following relationship: W i =m i n(1,T i / E i ); Among them, W i Indicates the first i The target compensation weight corresponding to each frequency band; T i Indicates the first i The psychoacoustic masking threshold corresponding to each frequency band; E i Indicates the first i The spectral energy corresponding to each frequency band.

[0117] Then, the compensation gain of the corresponding frequency band is adjusted according to the target compensation weight of each frequency band, and the data is re-synthesized into psychoacoustic optimized target audio data.

[0118] For example, when open-back headphones play music, the spectral energy corresponding to a certain high-frequency band is... 18dB, while the corresponding psychoacoustic masking threshold is 22dB.

[0119] Since the compensation energy has exceeded the psychoacoustic masking threshold, the calculated target compensation weight is 0.82, and the compensation gain for this frequency band is reduced by 18%. The spectral energy of the other frequency band is lower than the corresponding psychoacoustic masking threshold, so the original compensation gain remains unchanged.

[0120] After completing all frequency band adjustments, the psychoacoustic optimized target audio data is recovered through inverse fast fourier transform (IFFT), thereby hiding more of the compensation signal within the human ear's auditory masking area, reducing perceptible distortion while maintaining sound leakage suppression capability.

[0121] In this embodiment, by establishing a psychoacoustic masking threshold and dynamically adjusting the compensation gain corresponding to each frequency band based on the psychoacoustic masking threshold, the compensation energy is preferentially distributed in frequency bands that are difficult for the human ear to perceive. This reduces auditory distortion caused by the compensation signal while ensuring the sound leakage suppression effect, thereby improving the naturalness and comfort of audio playback from open-back headphones.

[0122] In one embodiment, after step S501, the following steps are further included: S502: Obtain the time-domain waveform corresponding to the psychoacoustic optimized target audio data, extract the envelope curve corresponding to the time-domain waveform, and determine the corresponding transient change information based on the envelope curve; compare the transient change information with a preset transient protection threshold, and when the transient change information is greater than the preset transient protection threshold, limit the compensation gain of the corresponding time period according to the preset transient protection strategy to obtain the output audio data after transient protection.

[0123] In this step, after obtaining the psychoacoustic optimized target audio data, an analytical signal is constructed using the Hilbert transform, and the envelope curve of the corresponding time-domain waveform is calculated based on the analytical signal.

[0124] Subsequently, transient change information is calculated based on the rate of change of the envelope amplitude corresponding to the continuous sampling points.

[0125] The transient change information can be represented by the rate of change of the envelope per unit time, i.e.: R = |A(n) A(n 1) | / Δt; Where A(n) represents the envelope amplitude corresponding to the current sampling point; A(n 1) represents the envelope amplitude corresponding to the previous sampling point; Δt represents the time interval between two sampling points; R represents transient change information.

[0126] Then, the transient change information is compared with a preset transient protection threshold.

[0127] The preset transient protection threshold is established based on the audio type.

[0128] For example, the setting can be 6dB / ms for voice, 10dB / ms for music, and 12dB / ms for game sound effects, and stored in a pre-established transient protection parameter table.

[0129] If the transient change information exceeds the corresponding preset transient protection threshold, the current compensation gain is limited according to the preset transient protection strategy. The preset transient protection strategy may employ an Attack / Release dynamic gain control algorithm.

[0130] Specifically, during the transient event, the compensation gain is rapidly reduced according to a preset Attack time constant, and after the transient ends, the compensation gain is gradually restored according to a preset Release time constant, thereby preventing the compensation signal from causing clipping of the transient waveform or weakening of the sense of aggression. Finally, the output audio data after transient protection is output.

[0131] For example, when open-back headphones play drum beats or gunshot sounds, the detected envelope change rate reaches 14dB / ms, which is higher than the preset transient protection threshold of 10dB / ms for music.

[0132] At this point, the compensation gain is reduced by 30% according to the Attack time constant of 2ms. After the drum beat ends, the compensation gain is gradually restored according to the Release time constant of 25ms. Finally, the output audio data after transient protection is output, so that the transient impact is maintained, while avoiding the drum beat becoming dull or clipping distortion due to sound leakage compensation.

[0133] In this embodiment, by detecting audio transient change information in real time and dynamically limiting the compensation gain according to the preset transient protection strategy, the compensation processing is kept consistent with the audio transient characteristics. While ensuring the ability to suppress sound leakage, the transient response speed is maintained, transient distortion and clipping are reduced, and the sound quality stability and listening performance of open-back headphones when playing high dynamic audio are improved.

[0134] In one embodiment, such as Figure 5 and Figure 6 As shown, in step S60, the frequency band energy ratio is evaluated on the second processed audio data to obtain the corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than a preset threshold, the updated sound leakage suppression level is determined based on the frequency band energy ratio evaluation value. Based on the updated sound leakage suppression level, the updated processing parameters are obtained by matching from multiple pre-stored processing parameter groups. Specifically, the steps include the following: S61: Extract the low-frequency energy envelope and high-frequency energy envelope from the second processed audio data respectively, and obtain the corresponding frequency band energy ratio evaluation value based on the ratio between the low-frequency energy envelope and the high-frequency energy envelope.

[0135] The terminal device inputs the second processed audio data into a low-frequency filter and a high-frequency filter for frequency band separation. The low-frequency filter extracts the audio signal corresponding to a preset low-frequency range, and the high-frequency filter extracts the audio signal corresponding to a preset high-frequency range. Both the low-frequency and high-frequency filters are implemented using finite impulse response (FIR) or infinite impulse response (IOR) filters, which are common in digital signal processing.

[0136] After frequency band separation is completed, the terminal device calculates the energy envelopes corresponding to the low-frequency and high-frequency signals respectively. The energy envelope is calculated using the root mean square (RMS) algorithm in the field of digital signal processing. By setting a fixed-length analysis window, the root mean square operation is performed on all sampling points within the window to obtain the energy envelope of the corresponding window.

[0137] After obtaining the low-frequency and high-frequency energy envelopes respectively, the terminal device calculates the frequency band energy ratio assessment value based on the ratio between the low-frequency and high-frequency energy envelopes. As the proportion of energy in the high-frequency band gradually increases, the frequency band energy ratio assessment value changes accordingly, thus reflecting the current high-frequency leakage level of the open-back headphones.

[0138] For example, the second processing audio data is processed according to a 20ms analysis window. The low-frequency filter extracts the audio signal corresponding to the range of 200Hz to 1000Hz, and the high-frequency filter extracts the audio signal corresponding to the range of 2500Hz to 8000Hz. The low-frequency energy envelope and the high-frequency energy envelope are calculated using the moving root mean square algorithm, respectively. Finally, the frequency band energy ratio evaluation value is calculated based on the ratio between the two.

[0139] S62: Compare the frequency band energy ratio assessment value with the preset threshold. When the frequency band energy ratio assessment value is greater than the preset threshold, determine the updated sound leakage suppression level according to the preset level adjustment strategy.

[0140] The terminal device inputs the frequency band energy ratio evaluation value into the level judgment module and compares it with a preset threshold. The preset threshold is determined in advance based on the actual sound leakage test results of the open-back headphones. By statistically analyzing the frequency band energy ratio evaluation values ​​corresponding to different levels of sound leakage, the threshold range that can meet the target sound leakage requirements is determined, and this threshold is written into the device parameter configuration.

[0141] The preset level adjustment strategy establishes a pre-defined correspondence between leakage suppression levels and frequency band energy ratio assessment values. When establishing the preset level adjustment strategy, the frequency band energy ratio assessment value is divided into multiple continuous intervals, each interval corresponding to a leakage suppression level. When the frequency band energy ratio assessment value is within the first threshold interval, the current leakage suppression level is maintained; when the frequency band energy ratio assessment value enters the second threshold interval, the leakage suppression level is increased by one level; when it enters the third threshold interval, the leakage suppression level is increased by two levels.

[0142] The terminal device queries the preset level adjustment strategy based on the current frequency band energy ratio assessment value range, and determines the updated sound leakage suppression level.

[0143] For example, the preset level adjustment strategy stipulates that: when the frequency band energy ratio assessment value is less than 0.28, the current leakage suppression level is maintained; when it is between 0.28 and 0.36, it is increased by one level; when it is between 0.36 and 0.45, it is increased by two levels; and when it is greater than 0.45, it is directly adjusted to the highest leakage suppression level. When the current frequency band energy ratio assessment value is 0.39, the terminal device determines that the updated leakage suppression level is increased by two levels.

[0144] S63: The updated folding ratio coefficient, updated delay depth, updated mixing ratio, and updated gain coefficient corresponding to the updated leakage suppression level are obtained from multiple pre-stored processing parameter groups based on the updated leakage suppression level, and are used as the updated processing parameters.

[0145] The terminal device queries a pre-established processing parameter group mapping table based on the updated noise suppression level. The processing parameter group mapping table is indexed according to the noise suppression level, with each noise suppression level corresponding to a set of processing parameters. Each set of processing parameters includes the updated folding ratio coefficient, the updated delay depth, the updated mixing ratio, and the updated gain coefficient.

[0146] The parameter set is established during the product development stage based on the acoustic test results corresponding to different leakage levels of open-back headphones. By adjusting the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient, leakage tests are conducted on different parameter combinations. Parameter sets that meet the corresponding leakage suppression effect are selected and stored in the device memory after establishing a one-to-one correspondence according to the leakage suppression level.

[0147] The terminal device reads the corresponding processing parameter group according to the updated sound leakage suppression level, and loads the updated folding ratio coefficient, updated delay depth, updated mixing ratio, and updated gain coefficient as new processing parameters, so that the next audio processing will re-execute the entire sound leakage suppression process according to the new processing parameters.

[0148] For example, if the updated sound leakage suppression level is level 4, the terminal device queries the processing parameter group mapping table, reads the folding ratio coefficient of 0.55, the delay depth of 1024 sampling points, the mixing ratio of 0.30, and the gain coefficient of -0.50 corresponding to level 4, and updates the above parameters as the current processing parameters for sound leakage suppression processing of the target audio data to be played later.

[0149] In this embodiment, by extracting the low-frequency energy envelope and high-frequency energy envelope corresponding to the second processed audio data, the frequency band energy ratio evaluation value is calculated, and the updated leakage suppression level is dynamically determined according to the preset level adjustment strategy. Then, the corresponding processing parameters are matched according to the updated leakage suppression level, so that the folding ratio coefficient, delay depth, mixing ratio and gain coefficient can be updated synchronously. This allows the leakage suppression intensity to be adaptively adjusted according to the current leakage situation, thereby improving the stability and accuracy of leakage suppression in open-back headphones.

[0150] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0151] In one embodiment, a multi-layer linkage sound leakage suppression system for open-back headphones is provided, which corresponds one-to-one with the multi-layer linkage sound leakage suppression method for open-back headphones described in the above embodiments. For example... Figure 7 As shown, this multi-layer linkage sound leakage suppression system for open-back headphones includes a level acquisition module, a frequency band analysis module, a collaborative processing module, a time domain extension module, a weighted processing module, a level update module, and a repetitive execution module. Detailed descriptions of each functional module are as follows: The level acquisition module is used to obtain the initial sound leakage suppression level and match the corresponding processing parameters based on the initial sound leakage suppression level. The processing parameters include the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient.

[0152] The frequency band analysis module is used to acquire the target audio data to be played, perform frequency band analysis on the target audio data, extract the high-frequency differential information corresponding to the left and right channels, and use the folding ratio coefficient in the processing parameters to perform weighted calculation on the high-frequency differential information to obtain the leakage adjustment data.

[0153] The collaborative processing module is used to perform collaborative adjustment of the left and right channels of the target audio data based on the leakage adjustment data to obtain the first processed audio data.

[0154] The time-domain extension module is used to acquire the difference signal between the first processed audio data and the target audio data, and use the difference signal as the residual adjustment information corresponding to the time-domain extension processing. The residual adjustment information is time-domain extended using the delay depth and mixing ratio in the processing parameters. The time-domain extended residual adjustment information is superimposed on the first processed audio data to obtain the extended audio data. The extended audio data is then input into a preset acoustic propagation model to generate the corresponding compensation signal.

[0155] The weighted processing module is used to weight the compensation signal using the gain coefficient in the processing parameters, and then use the weighted compensation signal to compensate the expanded audio data to obtain the second processed audio data.

[0156] The level update module is used to evaluate the frequency band energy ratio of the second processed audio data and obtain the corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than the preset threshold, the updated sound leakage suppression level is determined according to the frequency band energy ratio evaluation value, and the updated processing parameters are obtained by matching the updated sound leakage suppression level.

[0157] The repeat execution module is used to repeatedly execute the above steps on the target audio data to be played later using the updated processing parameters until the frequency band energy ratio evaluation value is less than the preset threshold.

[0158] It should be noted that the information interaction and execution process between the above systems / modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0159] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0160] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 this application, and should all be included within the protection scope of this application.

Claims

1. A multi-layer linkage sound leakage suppression method for open-back headphones, characterized in that, include: Obtain an initial sound leakage suppression level, and match corresponding processing parameters based on the initial sound leakage suppression level, wherein the processing parameters include folding ratio coefficient, delay depth, mixing ratio, and gain coefficient; The target audio data to be played is acquired, frequency band analysis is performed on the target audio data, high-frequency differential information corresponding to the left and right channels is extracted, and the high-frequency differential information is weighted using the folding ratio coefficient in the processing parameters to obtain leakage adjustment data. Based on the leakage adjustment data, the left and right channels of the target audio data are coordinated to obtain the first processed audio data. The difference signal between the first processed audio data and the target audio data is obtained, and the difference signal is used as the residual adjustment information corresponding to the time-domain expansion processing. The residual adjustment information is subjected to time-domain expansion processing using the delay depth and mixing ratio in the processing parameters. The residual adjustment information after time-domain expansion processing is superimposed on the first processed audio data to obtain the expanded audio data. The expanded audio data is then input into a preset acoustic propagation model to generate the corresponding compensation signal. The compensation signal is weighted using the gain coefficient in the processing parameters, and the expanded audio data is compensated using the weighted compensation signal to obtain the second processed audio data. The second processed audio data is evaluated for frequency band energy ratio to obtain the corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than a preset threshold, the updated sound leakage suppression level is determined based on the frequency band energy ratio evaluation value, and the updated processing parameters are matched based on the updated sound leakage suppression level. The above steps are repeated on the target audio data to be played later using the updated processing parameters until the frequency band energy ratio evaluation value is less than the preset threshold.

2. The method as described in claim 1, characterized in that, The steps of acquiring the target audio data to be played, performing frequency band analysis on the target audio data, extracting the high-frequency difference information corresponding to the left and right channels, and using the folding ratio coefficient in the processing parameters to perform weighted calculations on the high-frequency difference information to obtain the leakage adjustment data include: Obtain the sampling rate corresponding to the target audio data to be played, and match the corresponding target filter coefficient from a pre-stored set of filter coefficients according to the sampling rate; The difference signal between the left channel signal and the right channel signal in the target audio data is obtained, and the difference signal is subjected to second-order high-pass filtering using the target filter coefficients to obtain high-frequency difference information of the corresponding high-frequency bands of the left and right channels. The high-frequency differential information is weighted using the folding ratio coefficient in the processing parameters to obtain weighted high-frequency differential information, which is then used as leakage adjustment data.

3. The method as described in claim 1, characterized in that, The step of performing left and right channel coordinated adjustment on the target audio data based on the leakage adjustment data to obtain the first processed audio data includes: The leakage adjustment data is applied to the left channel signal and the right channel signal according to the sign direction, and the left channel signal and the right channel signal are adjusted in opposite directions to reduce the high-frequency difference information between the left and right channels, so as to obtain the first processed audio data.

4. The method as described in claim 1, characterized in that, The steps of obtaining the difference signal between the first processed audio data and the target audio data, using the difference signal as residual adjustment information corresponding to time-domain expansion processing, performing time-domain expansion processing on the residual adjustment information using the delay depth and mixing ratio in the processing parameters, and superimposing the time-domain expanded residual adjustment information onto the first processed audio data to obtain expanded audio data include: The residual adjustment information is written into a ring-shaped delay buffer of a preset depth. The corresponding historical residual adjustment information is read from multiple tap positions that are equally spaced in the ring-shaped delay buffer. The average residual adjustment information is obtained by averaging all the historical residual adjustment information. The average residual adjustment information is subjected to diffusion filtering to obtain diffused residual adjustment information. The diffused residual adjustment information is then superimposed onto the left and right channel signals of the first processed audio data according to the same mixing ratio in the processing parameters to obtain expanded audio data.

5. The method as described in claim 4, characterized in that, The step of inputting the expanded audio data into a preset acoustic propagation model to generate a corresponding compensation signal includes: The expanded audio data is input into a preset acoustic propagation model, wherein the acoustic propagation model is a finite impulse response filter established based on the acoustic leakage transfer function corresponding to the open headphone structure. The expanded audio data is convolved using the finite impulse response filter to obtain a compensation signal.

6. The method as described in claim 1, characterized in that, The step of weighting the compensation signal using the gain coefficient in the processing parameters, and then using the weighted compensation signal to compensate the expanded audio data to obtain the second processed audio data includes: Obtain the gain coefficient from the processing parameters, and use the gain coefficient to perform a weighted operation on the compensation signal to obtain a weighted compensation signal; The expanded audio data is subtracted from the weighted compensation signal to obtain the second processed audio data.

7. The method as described in claim 6, characterized in that, After weighting the compensation signal using the gain coefficient in the processing parameters, and then using the weighted compensation signal to compensate the expanded audio data to obtain the second processed audio data, the process further includes: The spectrum information corresponding to the second processed audio data is obtained and divided into multiple frequency bands according to a preset frequency band division rule. The psychoacoustic masking threshold corresponding to each frequency band is determined based on the spectrum energy corresponding to each frequency band. The spectral energy corresponding to each frequency band is compared with the corresponding psychoacoustic masking threshold to determine the target compensation weight corresponding to each frequency band. The compensation gain of each frequency band corresponding to the second processed audio data is adjusted according to the target compensation weights to obtain psychoacoustic optimized target audio data.

8. The method as described in claim 7, characterized in that, After adjusting the compensation gain of the second processed audio data corresponding to each frequency band according to each of the target compensation weights to obtain the psychoacoustic optimized target audio data, the method further includes: Obtain the time-domain waveform corresponding to the psychoacoustic optimized target audio data, extract the envelope curve corresponding to the time-domain waveform, and determine the corresponding transient change information based on the envelope curve; The transient change information is compared with a preset transient protection threshold. When the transient change information is greater than the preset transient protection threshold, the compensation gain for the corresponding time period is limited according to the preset transient protection strategy to obtain the output audio data after transient protection.

9. The method as described in claim 1, characterized in that, The steps of evaluating the frequency band energy ratio of the second processed audio data to obtain a corresponding frequency band energy ratio evaluation value, determining an updated sound leakage suppression level based on the frequency band energy ratio evaluation value when the evaluation value is greater than a preset threshold, and matching the updated processing parameters based on the updated sound leakage suppression level include: Extract the low-frequency energy envelope and high-frequency energy envelope from the second processed audio data, and obtain the corresponding frequency band energy ratio evaluation value based on the ratio between the low-frequency energy envelope and the high-frequency energy envelope; The frequency band energy ratio evaluation value is compared with a preset threshold. When the frequency band energy ratio evaluation value is greater than the preset threshold, the updated sound leakage suppression level is determined according to the preset level adjustment strategy. The updated folding ratio, updated delay depth, updated mixing ratio, and updated gain coefficient, which are matched with the updated sound leakage suppression level from a plurality of pre-stored processing parameter groups, are used as updated processing parameters.

10. A multi-layer linkage sound leakage suppression system for open-back headphones, characterized in that, include: The level acquisition module is used to acquire the initial sound leakage suppression level and match the corresponding processing parameters according to the initial sound leakage suppression level. The processing parameters include the folding ratio coefficient, delay depth, mixing ratio, and gain coefficient. The frequency band analysis module is used to acquire the target audio data to be played, perform frequency band analysis on the target audio data, extract the high-frequency differential information corresponding to the left and right channels, and use the folding ratio coefficient in the processing parameters to perform weighted calculation on the high-frequency differential information to obtain leakage adjustment data. The collaborative processing module is used to perform left and right channel collaborative adjustment on the target audio data based on the leakage adjustment data to obtain the first processed audio data; The time-domain extension module is used to acquire the difference signal between the first processed audio data and the target audio data, and use the difference signal as the residual adjustment information corresponding to the time-domain extension processing. The residual adjustment information is time-domain extended using the delay depth and mixing ratio in the processing parameters. The residual adjustment information after time-domain extension processing is superimposed on the first processed audio data to obtain the extended audio data. The extended audio data is then input into a preset acoustic propagation model to generate the corresponding compensation signal. The weighted processing module is used to weight the compensation signal using the gain coefficient in the processing parameters, and to use the weighted compensation signal to compensate the expanded audio data to obtain the second processed audio data. The level update module is used to evaluate the frequency band energy ratio of the second processed audio data to obtain the corresponding frequency band energy ratio evaluation value. When the frequency band energy ratio evaluation value is greater than a preset threshold, the updated sound leakage suppression level is determined according to the frequency band energy ratio evaluation value, and the updated processing parameters are obtained by matching the updated sound leakage suppression level. The repeat execution module is used to repeatedly execute the above steps on the target audio data to be played later using the updated processing parameters until the frequency band energy ratio evaluation value is less than the preset threshold.

Images