Dynamic ear pressure balance self-adaptive adjusting method and system and in-ear earphone

Abstract

The invention relates to the technical field of acoustic equipment and intelligent control, and particularly discloses a dynamic ear pressure balance self-adaptive adjustment method and system and an in-ear earphone. The system comprises an auditory meatus micro-pressure sensing module, an auditory perception state inferring module, a multi-modal adjustment decision module and a collaborative execution driving module, and a multi-level adjustment instruction is generated by monitoring auditory meatus pressure and inferring an auditory perception state in real time; and the micro electrically-actuated pressure release valve and the digital audio compensation processor are driven to cooperatively execute pressure balance and audio compensation, so that low-noise and self-adaptive ear pressure adjustment and auditory perception recovery are realized.

Description

Technical Field

[0001] This invention belongs to the field of acoustic equipment and intelligent control technology, specifically relating to a dynamic ear pressure balance adaptive adjustment method, system, and in-ear headphones. Background Technology

[0002] In the field of audio equipment and human-computer interaction, in-ear headphones are widely used due to their excellent noise isolation and portability. In-ear headphones create a closed acoustic environment by fitting snugly into the ear canal, thereby improving low-frequency response and isolating external noise, providing users with an immersive listening experience.

[0003] Among them, ear pressure balance adjustment technology for in-ear headphones is a key direction for improving wearing comfort and sound quality. This technology aims to relieve or eliminate the pressure on the eardrum caused by the pressure difference between the inside and outside of the ear canal. Its basic principle usually involves monitoring and actively or passively adjusting the pressure state of the sealed air cavity in the ear canal.

[0004] Existing technologies mainly achieve pressure balance through physical pressure relief holes or miniature air valves. However, such methods have limitations: physical pressure relief structures are prone to introducing wind noise when airflow passes through, interfering with the purity of audio signals and reducing fidelity; at the same time, the pressure relief process may lead to a decrease in acoustic sealing, resulting in slight sound leakage and affecting low-frequency performance.

[0005] Furthermore, when ear pressure imbalance occurs, the tympanic membrane deforms due to the pressure difference. This physiological state directly alters the ear's perception of sound. Even if pressure balance is restored through depressurization, the recovery of tympanic membrane shape and auditory perception is delayed, preventing the hearing state from returning to its optimal state synchronously. Therefore, achieving rapid and adaptive ear pressure balancing in complex usage scenarios, while simultaneously ensuring audio quality and auditory perception stability, has become a pressing technical challenge in this field. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the defects of existing in-ear headphones, such as wind noise interference, decreased acoustic sealing, and delayed recovery of tympanic membrane shape and auditory perception caused by the physical pressure relief structure, and to provide a dynamic ear pressure balance adaptive adjustment method, system and in-ear headphones.

[0007] The present invention provides a dynamic ear pressure balance adaptive adjustment system integrated inside the in-ear headphone body. Its core is to build a closed-loop control system that integrates ear canal pressure state perception, auditory perception state modeling, multimodal adjustment decision and collaborative execution.

[0008] The system includes an ear canal micro-pressure sensing module, an auditory perception state inference module, a multimodal adjustment decision module, and a collaborative execution driving module.

[0009] The ear canal micro-pressure sensing module is used to collect the absolute pressure value and pressure change rate of the sealed air cavity in the ear canal in real time.

[0010] The auditory perception state inference module is connected to the ear canal micro-pressure sensing module. Based on the received pressure data and the preset tympanic membrane deformation-auditory perception mapping model, it infers the current user's tympanic membrane deformation state and its impact on sound perception characteristics, and outputs a perception state index that characterizes the degree of auditory perception deviation.

[0011] The multimodal adjustment decision module is connected to both the ear canal micro-pressure sensing module and the auditory perception state inference module. It receives pressure data and perception state index, and generates specific adjustment commands based on a set of preset multi-level decision logic.

[0012] The collaborative execution drive module is connected to the multimodal adjustment decision module to receive adjustment commands and drive the corresponding physical actuators, which include a micro-electrically actuated pressure relief valve and a digital audio compensation processor.

[0013] The ear canal micro-pressure sensing module specifically includes a high-precision microelectromechanical system (MEMS) pressure sensor array that synchronously monitors the hydrostatic pressure at multiple locations within the ear canal at a sampling frequency of 1000 times per second. The module integrates signal conditioning circuitry and an analog-to-digital converter to filter, reduce noise, and digitize the raw pressure signal, ultimately outputting a smooth absolute pressure value and the rate of pressure change calculated differentially.

[0014] The core of the auditory perception state inference module is the tympanic membrane deformation-auditory perception mapping model.

[0015] The model was trained using a large amount of prior physiological acoustic experimental data, and a nonlinear mapping relationship was established between ear canal pressure value, pressure change rate, tympanic membrane deformation prediction, and the resulting changes in auditory sensitivity in each frequency band.

[0016] The module's inference process is as follows: First, the received real-time pressure data is input into the tympanic membrane deformation-auditory perception mapping model to calculate a set of estimated tympanic membrane deformation parameters. Then, based on this set of deformation parameters, the model further calculates the perceived gain or attenuation of the human ear at several key frequency points in the range of 20Hz to 20000Hz under the current state.

[0017] Finally, by combining the perceived changes at these frequencies, a dimensionless perception state index ranging from 0 to 1 is calculated and output using a weighted summation algorithm. Here, 0 represents that auditory perception is in the best baseline state, and 1 represents that the perception deviation has reached the preset maximum acceptable threshold.

[0018] The multimodal regulation decision module has three preset decision thresholds: pressure balance threshold, perception compensation threshold, and emergency pressure relief threshold.

[0019] The decision-making logic of this module is as follows: First, the continuously monitored absolute pressure value in the ear canal is compared with the pressure balance threshold.

[0020] If the absolute pressure value exceeds the pressure balance threshold, then it is further determined whether the perception state index also exceeds the perception compensation threshold.

[0021] If the sensing state index does not exceed the sensing compensation threshold, the module generates a first type of adjustment command. This command only contains parameters for the opening degree and duration of the micro-electrically actuated pressure relief valve, aiming to prioritize the restoration of physical pressure balance.

[0022] If the perception state index also exceeds the perception compensation threshold, the module generates a second type of adjustment instruction. This instruction is a composite instruction, which includes not only the control parameters of the micro-electrically actuated pressure relief valve, but also the audio signal correction parameters sent to the digital audio compensation processor, requiring the simultaneous execution of pressure release and auditory perception compensation.

[0023] Secondly, if the detected absolute pressure value in the ear canal exceeds a higher emergency pressure relief threshold, the module generates a third type of adjustment command regardless of the perceived state index. This command forces the micro-electrically actuated pressure relief valve to perform a rapid pressure relief operation at its maximum opening to ensure user comfort and safety.

[0024] Finally, if the absolute pressure value is within the pressure balance threshold, but the perception state index exceeds the perception compensation threshold, the module generates a fourth type of adjustment instruction. This instruction only contains the audio signal correction parameters sent to the digital audio compensation processor, and implements perception correction in a purely electroacoustic manner.

[0025] The collaborative execution driver module drives the corresponding physical actuators according to the different types of adjustment instructions received.

[0026] When the command involves a miniature electrically actuated pressure relief valve, the drive module converts the opening and duration parameters in the command into a precise pulse width modulation signal, which controls the piezoelectric ceramic actuator to drive the valve core to perform micron-level displacement, thereby achieving controllable and low-noise gas release.

[0027] When the instruction involves the digital audio compensation processor, the driver module loads the audio signal correction parameters into the processor's digital filter core. Based on the perceived change in each frequency point provided by the auditory perception state inference module, the filter core performs real-time inverse compensation of the amplitude frequency response and phase frequency response of the audio signal to be output, so that the sound signal finally transmitted to the tympanic membrane can cancel the perceptual distortion caused by the deformation of the tympanic membrane.

[0028] Furthermore, the valve body structure of the miniature electrically actuated pressure relief valve adopts a multi-layer microporous noise reduction design. The valve body contains at least two layers of staggered microporous plates, each with a through-hole diameter of no more than 50 μm, forming a meandering airflow channel between the plates. This structure allows gas to pass through slowly while utilizing viscous dissipation and acoustic impedance matching principles to suppress the peak energy spectrum of airflow noise to outside the 2000Hz to 5000Hz frequency range most sensitive to the human ear, and to keep its overall sound pressure level below 10 dB, thereby avoiding perceptible interference with audio signals.

[0029] Furthermore, the audio signal correction process of the digital audio compensation processor runs on a separate digital signal processing kernel. This kernel receives correction parameters from the multimodal adjustment decision module in real time and applies a dynamic equalization algorithm based on an infinite impulse response filter structure.

[0030] This algorithm can dynamically adjust the gain and quality factor of up to 10 independent parametric equalization bands within milliseconds based on changes in the perception state index, thereby achieving precise and rapid reshaping of the audio signal spectrum to match the auditory perception characteristics under the current tympanic membrane shape.

[0031] Furthermore, the system also integrates a user adaptive learning unit.

[0032] This unit continuously records pressure data, the types of control commands executed, and subsequent effect feedback data during historical control events.

[0033] The effect feedback is indirectly assessed by monitoring the rate of decline of pressure stability and perceived state index over a period of time after adjustment.

[0034] Based on this historical data, the user adaptive learning unit periodically uses the gradient descent algorithm to fine-tune the specific values ​​of the pressure balance threshold and the perception compensation threshold in the multimodal adjustment decision module, so that they gradually adapt to the individual ear canal physiological characteristics and pressure sensitivity of the user, and achieve personalized optimization of long-term wearing comfort.

[0035] The present invention also provides a dynamic ear pressure balance adaptive adjustment method, which is executed by the above-mentioned system and specifically includes the following steps: Step 1: Continuously monitor the absolute pressure value and pressure change rate in the ear canal using the ear canal micro-pressure sensing module.

[0036] Step 2: Using the auditory perception state inference module, based on the pressure data obtained in Step 1 and the preset tympanic membrane deformation-auditory perception mapping model, calculate and output the current perception state index.

[0037] Step 3: In the multimodal adjustment decision module, the real-time pressure data and the perceived state index are compared and analyzed with the preset multi-level decision thresholds, and the corresponding adjustment instructions are generated based on the comparison results.

[0038] Step 4: Through the collaborative execution drive module, the adjustment instructions generated in step 3 are parsed and executed to drive the micro electro-actuated pressure relief valve and / or digital audio compensation processor to perform corresponding physical adjustment or electroacoustic compensation operations.

[0039] Step 5: After the adjustment operation is executed, return to step 1 to continue monitoring and form a closed-loop control; at the same time, the user adaptive learning unit periodically optimizes and updates the decision threshold based on historical adjustment data.

[0040] The present invention also provides an in-ear headphone with dynamic ear pressure balance adaptive adjustment, including the above-mentioned dynamic ear pressure balance adaptive adjustment system.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a sensory-level control mechanism that transcends simple physical pressure balance by integrating ear canal micro-pressure sensing and auditory perception state inference. The system not only responds to changes in physical pressure within the ear canal but also proactively infers the potential impact of pressure differences on tympanic membrane morphology and auditory perception, and formulates adjustment strategies accordingly. When physical pressure imbalance has caused auditory perception deviation, the system can simultaneously initiate pressure release and audio-optical compensation, achieving comprehensive and rapid recovery from the physical environment to physiological perception. This solves the technical problem of delayed auditory perception recovery in traditional technologies, ensuring the immediacy and high quality of the listening experience.

[0042] 2. This invention uses a miniature electrically actuated pressure relief valve to replace the traditional passive physical pressure relief hole, and combines it with its unique multi-layer microporous sound-absorbing structure design to achieve controllability and low noise in the pressure relief process.

[0043] Electrical actuation allows for precise programming control of the timing, rate, and amount of pressure relief, avoiding sudden noise caused by airflow impact.

[0044] The multi-layer microporous noise reduction structure uses acoustic design to shift the energy spectrum of the pressure relief airflow noise out of the human ear's sensitive frequency range and reduce its overall sound pressure level to an extremely low level. This fundamentally eliminates the interference of the pressure relief process on the purity of the audio signal, ensuring the fidelity and immersiveness of the sound.

[0045] 3. This invention introduces a decision threshold adjustment mechanism based on adaptive learning optimization using historical user data. By continuously learning the individual user's ear pressure change patterns and adjustment effect feedback, the system can dynamically fine-tune key threshold parameters in its decision logic, making them more aligned with the user's physiological characteristics and subjective feelings. This personalized adaptation capability allows the system to continuously optimize its adjustment performance over time, providing highly customized and continuously optimized wearing comfort and listening experience for different users, thus enhancing the product's intelligence and user stickiness. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the overall technical architecture of the dynamic ear pressure balance adaptive adjustment system proposed in this invention; Figure 2 This is a schematic diagram of the core principle framework of the tympanic membrane deformation-auditory perception mapping model in this invention; Figure 3 This is a schematic diagram of the multi-level decision logic framework of the multimodal adjustment decision module in this invention; Figure 4 This is a schematic diagram of the driving and collaborative control framework of the collaborative execution driving module for the physical execution mechanism in this invention; Figure 5 This is a schematic diagram of the threshold optimization and personalized adaptation logic framework of the user adaptive learning unit in this invention. Detailed Implementation

[0047] Please refer to the attached document. Figure 1 This paper details the technical implementation of a dynamic ear pressure balance adaptive adjustment system. This system is integrated within the in-ear headphone body, forming a complete closed-loop control system. The system consists of an ear canal micro-pressure sensing module, an auditory perception state inference module, a multimodal adjustment decision module, a collaborative execution drive module, and a user adaptive learning unit. All modules are connected via a high-speed digital bus to ensure real-time and reliable data interaction.

[0048] The ear canal micro-pressure sensing module is responsible for directly sensing the pressure state of the sealed air cavity inside the ear canal. The core of this module is a high-precision microelectromechanical system (MEMS) pressure sensor array.

[0049] The array consists of three independent miniature pressure sensing units, which are fixed to the inner wall of the earphone shell in an equilateral triangle layout to ensure that the static pressure distribution at different spatial locations in the ear canal can be monitored simultaneously.

[0050] Each miniature pressure sensing unit operates based on the piezoresistive effect, with a sensing element measuring 500μm × 500μm and a thickness of 20μm. The module integrates a dedicated signal conditioning circuit, which includes a low-pass filter with a cutoff frequency of 100Hz to suppress high-frequency noise interference caused by head movements or voice vibrations.

[0051] The conditioned analog signal is digitized by a 16-bit precision analog-to-digital converter, with a sampling frequency fixed at 1000 times per second.

[0052] The digitized pressure data is first processed by a three-point moving average filter to further improve the signal-to-noise ratio. Then, the module's embedded processor calculates the arithmetic mean of the readings from the three sensor units in real time and outputs it as the absolute ear canal pressure value at the current moment. The typical range of this value is 90 kPa to 110 kPa, with a resolution of 1 Pa.

[0053] Meanwhile, the processor calculates the pressure change rate based on the absolute pressure values ​​of 10 consecutive sampling points using a first-order backward difference algorithm. The calculation formula is: the pressure change rate equals the current absolute pressure value minus the absolute pressure value of the previous 10 sampling points, and then divided by 10 times the sampling time interval of 0.01s.

[0054] The calculated pressure change rate is expressed in Pa / s, and the output range is typically between -1000 and 1000. Finally, the smoothed absolute pressure value and the calculated pressure change rate are transmitted in real time to the auditory perception state inference module via an integrated circuit bus protocol.

[0055] Please refer to the attached document. Figure 2 The core function of the auditory perception state inference module is to infer the user's tympanic membrane deformation state and its impact on auditory perception based on the input pressure data.

[0056] This module contains a built-in tympanic membrane deformation-auditory perception mapping model. This model is a nonlinear regression model trained with a large amount of prior physiological acoustic experimental data. Its input dimension is 2, namely the absolute pressure value of the ear canal and the rate of pressure change, and its output dimension is 11, including 3 tympanic membrane deformation prediction parameters and the perceived gain or attenuation of 8 key frequencies. The model's training data comes from the results of tympanic membrane laser vibration measurement and psychoacoustic hearing threshold test of more than 100 subjects under controlled pressure conditions. The module's inference process is divided into 3 consecutive stages.

[0057] In the first stage, the received real-time absolute pressure value and pressure change rate are normalized and preprocessed to fall within the input range of -1 to 1 required by the model.

[0058] The preprocessed data is fed into the first layer of the mapping model, which contains 32 neurons and uses the hyperbolic tangent function as the activation function to extract stress features.

[0059] The second layer contains 16 neurons and outputs three tympanic membrane deformation prediction parameters: tympanic membrane apex displacement, tympanic membrane mean curvature change, and tympanic membrane tension coefficient change. These parameters are all expressed in dimensionless form.

[0060] In the second stage, these three deformation parameters are used as inputs and fed into the third layer of the model. This layer also contains 16 neurons to calculate the perceptual changes of the human ear at eight preset key frequencies.

[0061] These eight frequency points cover the full audio range from 20Hz to 20000Hz, specifically 125Hz, 500Hz, 1000Hz, 2000Hz, 4000Hz, 8000Hz, 12000Hz and 16000Hz.

[0062] In the third stage, the module performs a weighted summation of the perceived changes at eight frequency points to calculate the final perception state index. The weighting coefficients are pre-set based on the importance of each frequency point in human auditory sensitivity, with the 1000Hz to 4000Hz band having the highest weight, accounting for 60% of the total. The formula for calculating the perception state index is: The perception state index equals the absolute value of the perception change at 0.1 × 125 Hz plus the absolute value of the perception change at 0.1 × 500 Hz plus the absolute value of the perception change at 0.15 × 1000 Hz plus the absolute value of the perception change at 0.2 × 2000 Hz plus the absolute value of the perception change at 0.2 × 4000 Hz plus the absolute value of the perception change at 0.1 × 8000 Hz plus the absolute value of the perception change at 0.1 × 12000 Hz plus the absolute value of the perception change at 0.05 × 16000 Hz.

[0063] The calculated perception state index is a dimensionless value between 0 and 1, where 0 represents that auditory perception is in the best baseline state and 1 represents that the perception deviation has reached the preset maximum acceptable threshold.

[0064] The index is updated every 10ms and synchronously transmitted to the multimodal adjustment decision module via a serial peripheral interface.

[0065] Please refer to the attached document. Figure 3 The multimodal regulation decision module is responsible for generating precise regulation commands based on the input absolute pressure value, pressure change rate, and sensing state index. This module has three preset decision threshold levels: pressure balance threshold, sensing compensation threshold, and emergency pressure relief threshold.

[0066] The initial default values ​​for these thresholds are set by the system manufacturer based on extensive experimental data. The pressure balance threshold is typically ±200 Pa of absolute pressure change, the sensing compensation threshold is typically 0.3 of the sensing state index, and the emergency pressure relief threshold is typically ±500 Pa of absolute pressure change. The module's decision logic is a continuously running loop with a period of 5 ms.

[0067] Within each decision cycle, the module executes four judgment steps sequentially.

[0068] The first step is to determine whether the absolute value of the current absolute pressure exceeds the emergency pressure relief threshold. If it does, a third-type adjustment command is immediately generated.

[0069] The third type of adjustment command is the highest priority command. It forces the micro-electrically actuated pressure relief valve to perform a rapid pressure relief operation at its maximum opening. The pressure relief valve opening parameter is set to 100%, the duration parameter is set to 500ms, and this command does not contain any audio compensation parameters.

[0070] The second step is to determine whether the absolute pressure value exceeds the emergency pressure relief threshold if it does not exceed the pressure balance threshold.

[0071] If it exceeds the threshold, then it is further determined whether the current perception state index exceeds the perception compensation threshold.

[0072] If the perception state index does not exceed the perception compensation threshold, the module generates the first type of adjustment instruction.

[0073] The first type of regulation command is only for physical pressure balance and includes the control parameters of the miniature electrically actuated pressure relief valve.

[0074] The opening degree of the pressure relief valve is calculated proportionally based on the extent to which it exceeds the pressure balance threshold. The calculation formula is: the opening percentage equals the absolute pressure value minus the pressure balance threshold, then the absolute value is divided by the pressure balance threshold, and finally multiplied by a proportional coefficient of 0.5, with the result limited to between 5% and 50%.

[0075] The duration is fixed at 300ms. If the perception state index also exceeds the perception compensation threshold, the module generates a second type of adjustment command.

[0076] The second type of adjustment command is a composite command, which includes both the pressure relief valve control parameters calculated above and the audio signal correction parameters sent to the digital audio compensation processor.

[0077] The audio signal correction parameters directly reference the perceptual changes at eight key frequency points provided by the auditory perception state inference module, but with the signs reversed for subsequent inverse compensation.

[0078] Third, if the absolute pressure value is within the pressure balance threshold, but the perception state index exceeds the perception compensation threshold, the module generates a fourth type of adjustment command.

[0079] The fourth type of adjustment instruction contains only audio signal correction parameters. Its generation logic is exactly the same as the audio part in the second type of instruction, and it is designed to implement perception correction in a purely electroacoustic manner.

[0080] Fourth, if both the absolute pressure value and the perception state index are below their corresponding thresholds, the module remains idle and does not generate any adjustment commands. All generated commands are accompanied by a timestamp and command type identifier and are sent to the collaborative execution driver module via the parallel bus.

[0081] Please refer to the attached document. Figure 4 The collaborative execution drive module is responsible for parsing and executing the adjustment instructions from the multimodal adjustment decision module, driving the corresponding physical actuators. This module contains two independent drive subunits, corresponding to the micro-electrically actuated pressure relief valve and the digital audio compensation processor, respectively.

[0082] When the received instruction relates to the miniature electrically actuated pressure relief valve, the pressure relief valve control subunit of the drive module begins to operate.

[0083] This subunit contains a pulse width modulation (PWM) signal generator with a frequency of 20kHz and a resolution of 12 bits. The subunit converts the pressure relief valve opening percentage parameter carried in the instruction into the corresponding PWM duty cycle.

[0084] For example, a 50% opening corresponds to a 50% duty cycle. This pulse width modulation signal is amplified and then drives a piezoelectric ceramic actuator.

[0085] Piezoelectric ceramic actuators generate micron-level displacement under the action of an electric field. The amount of displacement is linearly related to the applied voltage, thereby precisely controlling the opening degree of the valve core that is mechanically coupled to it.

[0086] The valve core displacement is fed back in a closed loop through an integrated linear variable differential transformer to ensure that the opening control accuracy is within ±2%.

[0087] The duration parameter carried in the instruction is implemented through a timer inside the module. Upon reaching the set time, the pulse width modulation signal output automatically stops, causing the pressure relief valve to close. The valve body structure of the miniature electrically actuated pressure relief valve employs a multi-layer microporous noise reduction design.

[0088] The valve body contains three layers of staggered 316L stainless steel microporous plates.

[0089] Each microplate layer is 100 μm thick, with the diameter of the through-holes strictly controlled at 50 μm and a porosity of 30%.

[0090] The spacing between adjacent microplates is 200 μm, forming a meandering airflow channel.

[0091] When gas passes through this structure, due to fluid viscosity and the eddy shedding effect at the micropore edges, the acoustic energy of the airflow noise is mainly dissipated in the infrasound and ultrasonic frequency bands.

[0092] Tests have shown that this structure can suppress the total sound pressure level of broadband noise generated by the depressurized airflow to below 8dB, and its energy spectrum peak appears above 8000Hz, avoiding the 2000Hz to 5000Hz frequency band that the human ear is most sensitive to, thus avoiding perceptible interference to audio signals.

[0093] When a received instruction relates to the digital audio compensation processor, the audio compensation driver subunit of the driver module is activated. This subunit communicates with the digital audio compensation processor via a high-speed audio serial bus.

[0094] The digital audio compensation processor runs on a separate digital signal processing core with a clock frequency of 400MHz and supports 32-bit floating-point operations.

[0095] The driver subunit packages the audio signal correction parameters carried in the instruction—namely, the inverse sensing changes at eight key frequency points—into a parameter data packet and writes it to the configuration register of the digital signal processing kernel via direct memory access. The digital signal processing kernel then runs a dynamic equalization algorithm based on an infinite impulse response filter structure in real time.

[0096] The algorithm includes 10 independent parametric equalization bands with center frequencies set at 63Hz, 125Hz, 250Hz, 500Hz, 1000Hz, 2000Hz, 4000Hz, 8000Hz, 12000Hz, and 16000Hz. Each parametric equalizer allows independent adjustment of gain and quality factor. Gain adjustment ranges from ±12dB in 0.1dB steps. Quality factor adjustment ranges from 0.1 to 10.

[0097] Once the new correction parameters are loaded, the dynamic equalization algorithm completes the coefficient update for all 10 parametric equalization bands within 1 ms. Based on the inverse sensing changes received at 8 frequency points, the algorithm calculates the target gain values ​​for the 10 parametric equalization bands through linear interpolation and extrapolation.

[0098] The quality factor is adaptively adjusted according to the steepness of the change, and a higher quality factor is used for frequency bands with drastic changes to achieve more accurate compensation.

[0099] The updated filter coefficients are immediately applied to the processor's real-time audio pipeline to inversely compensate for the amplitude and phase frequency responses of the audio signal to be output, so that the sound signal finally transmitted to the tympanic membrane can cancel the perceptual distortion caused by tympanic membrane deformation and ensure the consistency of the listening experience.

[0100] The user adaptive learning unit is responsible for optimizing the system's decision thresholds based on historical data, achieving personalized adaptation. Please refer to the appendix. Figure 5 This unit continuously operates on a low-power auxiliary microprocessor.

[0101] It continuously monitors and records historical adjustment events via the system bus. Each adjustment event record contains five key fields: event timestamp, absolute pressure value before adjustment, perceived state index before adjustment, type of adjustment command executed, and pressure stability index and perceived state index decline rate within 60 seconds after adjustment.

[0102] The pressure stability index is defined as the standard deviation of the absolute pressure values ​​within 30 seconds after the adjustment operation ends. The perceived state index decline rate is defined as the average slope of the perceived state index decline within 10 seconds after the adjustment operation ends. A threshold optimization cycle is initiated every 100 valid adjustment event records accumulated by the user adaptive learning unit.

[0103] The optimization process employs the gradient descent algorithm, with the goal of minimizing subsequent regulatory interventions while maintaining the stress stability index and the perceived state index at a good level.

[0104] Specifically, the algorithm defines a loss function that is a weighted sum of the adjustment frequency, the average pressure stability index, and the average perceived state index fall rate.

[0105] The algorithm fine-tunes these two thresholds by calculating the gradient of the loss function with respect to the pressure balance threshold and the perception compensation threshold, and then moving in the opposite direction of the gradient. The step size of each adjustment is limited to within ±5% of the initial value to prevent over-adjustment.

[0106] For example, if the system finds that for a particular user, initiating pressure relief when the pressure change is small actually leads to larger subsequent pressure fluctuations, the learning unit will slowly increase the value of the pressure balance threshold to reduce unnecessary adjustments.

[0107] After the optimized new threshold passes the security check, it is written into the non-volatile memory of the multimodal adjustment decision module, overwriting the old threshold setting.

[0108] This process is carried out periodically, allowing the system to gradually adapt to the individual ear canal physiological characteristics and pressure sensitivity of each user, achieving personalized optimization of long-term wearing comfort.

[0109] From the moment the system is powered on, each module enters its working state. The ear canal micro-pressure sensing module continuously collects data, the auditory perception state inference module calculates the perception state index in real time, the multimodal adjustment decision module makes decisions periodically, the collaborative execution drive module drives the actuator according to instructions, and the user adaptive learning unit accumulates data and optimizes thresholds in the background. Together, they form an efficient, intelligent, and adaptive dynamic ear pressure balance adjustment system.

[0110] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0111] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic ear pressure balance adaptive adjustment system, characterized in that, include: The ear canal micro-pressure sensing module is used to collect the absolute pressure value and pressure change rate of the sealed air cavity in the ear canal in real time. The auditory perception state inference module is connected to the ear canal micro-pressure sensing module. It is used to infer the current tympanic membrane deformation state of the user and its influence on sound perception characteristics based on the received pressure data and the preset tympanic membrane deformation-auditory perception mapping model, and output a perception state index that characterizes the degree of auditory perception deviation. The multimodal adjustment decision module is connected to both the ear canal micro-pressure sensing module and the auditory perception state inference module. It is used to receive pressure data and perception state index, and generate specific adjustment instructions based on preset multi-level decision logic. The collaborative execution drive module is connected to the multimodal adjustment decision module and is used to receive adjustment commands and drive the corresponding physical actuators, which include a micro-electrically actuated pressure relief valve and a digital audio compensation processor.

2. The dynamic ear pressure balance adaptive adjustment system according to claim 1, characterized in that, The ear canal micro-pressure sensing module includes a high-precision microelectromechanical system pressure sensor array, which simultaneously monitors the static pressure at multiple locations within the ear canal. The ear canal micro-pressure sensing module integrates a signal conditioning circuit and an analog-to-digital converter to filter, reduce noise, and digitize the original pressure signal, outputting a smooth absolute pressure value and the pressure change rate obtained through differential calculation.

3. The dynamic ear pressure balance adaptive adjustment system according to claim 1, characterized in that, The inference process of the auditory perception state inference module is as follows: the received real-time pressure data is input into the tympanic membrane deformation-auditory perception mapping model, and a set of estimated tympanic membrane deformation parameters are calculated. Based on this set of deformation parameters, the sensing gain or attenuation of several key frequency points under the current state is calculated; by combining the sensing changes of these frequency points, a dimensionless sensing state index is calculated and output through a weighted summation algorithm.

4. The dynamic ear pressure balance adaptive adjustment system according to claim 1, characterized in that, The multimodal regulation decision module has three preset decision thresholds: pressure balance threshold, perception compensation threshold, and emergency pressure relief threshold. The decision logic of the multimodal adjustment decision module is as follows: compare the continuously monitored absolute pressure value of the ear canal with the pressure balance threshold. If the absolute pressure value exceeds the pressure balance threshold, it is further determined whether the sensing state index also exceeds the sensing compensation threshold; if the sensing state index does not exceed the sensing compensation threshold, a first type of adjustment command is generated. If the perception state index also exceeds the perception compensation threshold, a second type of adjustment instruction is generated. If the absolute pressure value exceeds the emergency pressure relief threshold, a third type of regulation command is generated. If the absolute pressure value is within the pressure balance threshold but the sensing state index exceeds the sensing compensation threshold, a fourth type of adjustment command is generated.

5. The dynamic ear pressure balance adaptive adjustment system according to claim 4, characterized in that, The first type of adjustment command only includes opening and duration parameters for the micro electrically actuated pressure relief valve; The second type of adjustment command is a composite command, which includes control parameters for the micro-electrically actuated pressure relief valve and audio signal correction parameters sent to the digital audio compensation processor; The third type of adjustment command forces the micro-electrically actuated pressure relief valve to perform a rapid pressure relief operation at its maximum opening. The fourth type of adjustment instruction only includes audio signal correction parameters sent to the digital audio compensation processor.

6. The dynamic ear pressure balance adaptive adjustment system according to claim 1, characterized in that, The collaborative execution drive module drives the corresponding physical execution mechanism according to the different types of adjustment instructions received; When the command involves a miniature electrically actuated pressure relief valve, the opening degree and duration parameters in the command are converted into precise pulse width modulation signals to control the piezoelectric ceramic actuator to drive the valve core to perform micron-level displacement. When the instruction involves a digital audio compensation processor, the audio signal correction parameters are loaded into the processor's digital filter core to perform real-time inverse compensation of the amplitude and phase frequency responses of the audio signal to be output.

7. The dynamic ear pressure balance adaptive adjustment system according to claim 1, characterized in that, The audio signal correction process of the digital audio compensation processor runs on an independent digital signal processing kernel; The kernel receives correction parameters from the multimodal adjustment decision module in real time and applies a dynamic equalization algorithm based on an infinite impulse response filter structure. Based on changes in the perception state index, the algorithm dynamically adjusts the gain and quality factor of up to 10 independent parametric equalization bands within milliseconds.

8. The dynamic ear pressure balance adaptive adjustment system according to claim 1, characterized in that, It also includes a user adaptive learning unit; This unit continuously records stress data, the types of regulation instructions executed, and subsequent effect feedback data in historical regulation events. Based on this historical data, the gradient descent algorithm is periodically used to fine-tune the specific values ​​of the stress balance threshold and the perception compensation threshold in the multimodal regulation decision module.

9. A dynamic ear pressure balance adaptive adjustment method, characterized in that, Performed by the system according to any one of claims 1 to 8, the method includes the following steps: The ear canal micro-pressure sensing module continuously monitors the absolute pressure value and the rate of pressure change in the ear canal. Using the auditory perception state inference module, based on the obtained pressure data and the preset tympanic membrane deformation-auditory perception mapping model, the current perception state index is calculated and output. In the multimodal regulation decision module, real-time pressure data and perceived state index are compared and analyzed with preset multi-level decision thresholds, and corresponding regulation instructions are generated based on the comparison results. The collaborative execution drive module parses and executes the generated adjustment instructions to drive the micro electro-actuated pressure relief valve and / or digital audio compensation processor to perform corresponding physical adjustment or electroacoustic compensation operations.

10. An in-ear headphone with dynamic ear pressure balance adaptive adjustment, characterized in that, Includes the dynamic ear pressure balance adaptive adjustment system as described in any one of claims 1 to 8.

Images