An adaptive hearing aid remote fitting method, system and device

Abstract

The application discloses a kind of self-adapting hearing aid remote fitting method, system and equipment, by extracting hearing loss characteristics from the audiogram uploaded by user and acoustic condition information;Based on the maximum stable gain curve of hearing aid self-check, the feedback risk level of each frequency band is evaluated;The target gain corresponding to the preset prescription formula is calculated;Wherein, the prescription formula at least includes speech compensation equalizer enhancement formula and high-power fitting equalization formula;According to hearing loss characteristics, feedback risk level and target gain, candidate formula is screened from prescription formula, and formula parameter is generated, to be confirmed by fitter.It can greatly reduce the dependence on fitter experience by using the embodiment of the application, and provide strong algorithm support for non-prescription hearing aids in self-help fitting scenarios;At the same time, it can provide personalized gain scheme for moderate to severe hearing loss users, which fully compensates for loudness and actively suppresses whistling.

Description

Technical Field

[0001] This invention relates to the field of hearing aid fitting technology, and more particularly to an adaptive remote hearing aid fitting method, system, and device. Background Technology

[0002] Current remote fitting solutions typically involve voice communication, with the audiologist then issuing parameters based on a built-in fitting prescription formula.

[0003] However, existing remote fitting systems only serve as parameter distribution channels. Since fitters cannot verify the parameters on-site using real-ear measurement (REM) devices in remote fitting scenarios, manually selected parameters may not be suitable for actual needs. Summary of the Invention

[0004] The present invention aims to provide an adaptive remote hearing aid fitting method, system, and device, which can automatically recommend prescription formulas based on information submitted by the user, significantly reducing reliance on the experience of the audiologist and providing strong algorithmic support for non-prescription hearing aids in self-service fitting scenarios; at the same time, it can provide personalized gain solutions for users with moderate to severe hearing loss that fully compensate for loudness and actively suppress feedback.

[0005] In a first aspect, embodiments of the present invention provide an adaptive remote hearing aid fitting method, comprising: Hearing loss features are extracted from audiograms and acoustic condition information uploaded by users; Based on the maximum stable gain curve of the hearing aid self-test, assess the feedback risk level of each frequency band; The target gain corresponding to the preset prescription formula is calculated; wherein, the prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula; the voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in the moderate to severe hearing loss segment; the high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, and reduces the risk of howling through frequency band differentiated gain allocation and dynamic feedback safety margin; Based on hearing loss characteristics, feedback risk level, and target gain, candidate formulas are selected from the prescription formulas, and formula parameters are generated. The candidate formula and its parameters are sent to the audiologist, and in response to the audiologist's confirmation instruction, the formula parameters are sent to the user's hearing aid.

[0006] As an improvement to the above solution, the user's uploaded audiogram is obtained, and the user's hearing loss level and hearing loss configuration are extracted from the audiogram; Obtain acoustic condition information uploaded by the user, and extract the user's acoustic coupling method, user experience, and user age from the acoustic condition information; Hearing loss characteristics are obtained based on the degree of hearing loss, the hearing loss configuration, the acoustic coupling method, the user's experience, and the user's age.

[0007] As an improvement to the above scheme, when the prescription formula is a speech compensation equalizer enhancement formula, the target gain generated by calculating the preset prescription formula includes: The validity of the hearing thresholds at each frequency point in the audiogram is verified and missing values ​​are interpolated to obtain the verified hearing thresholds. The NAL-NL2 formula is used to calculate the first target gain at each frequency point under the preset decibel sound pressure level input. The second target gain at each frequency point is calculated using the DSL v5 formula under the preset decibel sound pressure level input. Based on predefined frequency-dependent weighting coefficients, the first target gain and the second target gain are weighted and summed to obtain the first weighted target gain; Calculate the loudness compensation term based on the verified hearing threshold; Based on the first weighted target gain, the loudness compensation term, and the compression ratio based on the degree of hearing loss, the target gain generated by the speech compensation equalizer enhancement formula at each decibel sound pressure level is obtained.

[0008] As an improvement to the above scheme, when the prescription formula is a high-power fitting equalization formula, the target gain generated by the preset prescription formula includes: Calculate the third target gain generated by the speech compensation equalizer enhancement formula at a preset decibel sound pressure level; Based on the maximum stable gain curve and the third target gain, calculate the feedback safety margin point by point; Based on the acoustic coupling method extracted from the acoustic condition information, the feedback safety margin at each frequency point is obtained; the feedback safety margin is an engineering empirical parameter. When the feedback safety margin is less than the feedback safety safety allowance, the feedback penalty term at each frequency point is calculated based on the feedback safety margin, the stable gain curve, and the feedback safety allowance. Based on the predefined frequency band compensation, generate frequency band differential compensation items for each frequency point; The first total loudness of the speech compensation equalizer enhancement formula and the second total loudness of the high-power fitting equalizer formula are calculated using the loudness model. If the first difference between the first loudness and the second loudness is greater than a preset difference threshold, then the cross-band migration compensation amount is calculated based on the first difference, the difference between the feedback safety margin and the feedback safety allowance; the migration compensation amount is used to compensate for the howling frequency band. Based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount, calculate the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level.

[0009] As an improvement to the above scheme, the step of calculating the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount includes: The third target gain, the frequency band differential compensation term, and the migration compensation amount are superimposed to obtain a first intermediate value; The feedback penalty term is used to inversely superimpose the first intermediate value to obtain the initial target gain; The initial target gain is back-substituted and verified based on the feedback safety margin, the first total loudness, and the second total loudness. If the back-substitution verification result shows that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the initial target gain is calculated iteratively. If the back-substitution verification result is that the initial target gain converges or the number of iterations is not less than the preset maximum number of iterations, then the initial target gain is used as the target gain generated by the high-power fitting equalization formula at the preset decibel sound pressure level. Based on the target gain generated by the high-power fitting equalization formula at a preset decibel sound pressure level, the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level is obtained.

[0010] As an improvement to the above scheme, if the back-substitution verification result shows that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the iterative calculation of the initial target gain includes: If the back-substitution verification result is that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the feedback safety margin and the feedback penalty term are sent to the feedback suppression module built into the hearing aid as a priority prompt for the initial notch frequency. The hearing aid feedback suppression module receives the equivalent maximum stable gain reduction based on the frequency band continuously triggered notch event. Based on the original maximum stable gain curve and the equivalent maximum stable gain reduction, it recalculates and updates the feedback safety margin, the feedback penalty term, and the migration compensation amount. The initial target gain is calculated iteratively based on the updated feedback safety margin, the feedback penalty term, and the migration compensation amount.

[0011] As an improvement to the above scheme, the step of screening candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generating formula parameters, includes: If the hearing loss characteristics are a steep drop in high and medium frequencies, a closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and formula parameters are generated based on the target gain. If the hearing loss characteristics are severe or above, closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and the loudness compensation enhancement formula parameters are generated according to the target gain. If the hearing loss characteristics are moderate to severe, the acoustic coupling method is open eardrum or large vent, and the feedback risk level is high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters are generated according to the target gain. If the hearing loss characteristics are severe hearing loss, ear canal leakage, acoustic coupling method is semi-open tympanic membrane or poor coupling, and the feedback risk level is extremely high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters for strong feedback suppression are generated based on the target gain.

[0012] As an improvement to the above scheme, the prescription formula also includes NAL-NL2 and DSL v5. The step of selecting candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generating formula parameters, includes: If the hearing loss characteristics are mild to moderate hearing loss in new adult users, closed eardrums, and low feedback risk level, then the NAL-NL2 and the speech compensation equalizer enhancement formula are used as candidate formulas, and formula parameters are generated based on the target gain. If the hearing loss characteristic is that of a child user and the feedback risk level is low, then the children's version of DSL v5 is used as a candidate formula, and formula parameters are generated according to the target gain.

[0013] Secondly, embodiments of the present invention also provide an adaptive remote hearing aid fitting system, comprising: The hearing loss feature extraction module is used to extract hearing loss features from the audiograms and acoustic condition information uploaded by the user. The feedback risk level assessment module is used to assess the feedback risk level of each frequency band based on the maximum stable gain curve of the hearing aid self-test. The target gain calculation module is used to calculate the target gain corresponding to the preset prescription formula; wherein, the prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula; the voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in the moderate to severe hearing loss segment; the high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, and reduces the risk of howling through frequency band differentiated gain allocation and dynamic feedback safety margin; The candidate formula screening module is used to screen candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generate formula parameters. The parameter distribution module is used to send the candidate formula and its formula parameters to the audiologist, and in response to the audiologist's confirmation instruction, to distribute the formula parameters to the user's hearing aid.

[0014] Thirdly, embodiments of the present invention also provide an adaptive remote hearing aid fitting device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the adaptive remote hearing aid fitting method as described above.

[0015] Compared with existing technologies, this invention discloses an adaptive remote hearing aid fitting method, system, and device. It extracts hearing loss characteristics from user-uploaded audiograms and acoustic condition information; assesses the feedback risk level of each frequency band based on the maximum stable gain curve of the hearing aid's self-test; and calculates the target gain corresponding to a preset prescription formula. The prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula. The voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in moderate to severe hearing loss segments. The high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, reducing feedback risk through frequency band differentiated gain allocation and dynamic feedback safety margin. Candidate formulas are selected from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and formula parameters are generated. The candidate formulas and their parameters are sent to the audiologist, and in response to the audiologist's confirmation instruction, the formula parameters are sent to the user's hearing aid. By employing the embodiments of the present invention, the reliance on the experience of audiologists can be significantly reduced, providing strong algorithmic support for non-prescription hearing aids in self-service fitting scenarios; at the same time, it can provide personalized gain solutions for users with moderate to severe hearing loss that fully compensate for loudness while actively suppressing feedback. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of an adaptive remote hearing aid fitting method provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of an adaptive remote hearing aid fitting system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an adaptive remote hearing aid fitting device provided in an embodiment of the present invention. Detailed Implementation

[0017] 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 embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] In the description and claims, it should be understood that the terms "first," "second," etc., used in the description and claims are only for the purpose of distinguishing the description of the same technical features, and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated, nor necessarily the order of description or chronological order. The terms are interchangeable where appropriate. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature.

[0019] Traditional hearing aid fitting requires users to visit a physical fitting center for face-to-face hearing tests, parameter settings, and effect adjustments by an audiologist. However, the performance of hearing aids in a user's actual living environment (home, office, outdoors) differs significantly from that in a physical fitting center. A single fitting at a center cannot accurately reflect real-world usage scenarios, and users frequently require fine-tuning of parameters during use. Furthermore, for people with mobility impairments, users in remote areas, and the elderly, multiple trips to the fitting center are impractical.

[0020] This led to the development of remote fitting solutions based on voice communication. However, voice communication itself is an obstacle for hearing-impaired users, and most remote fitting systems only serve as parameter distribution channels. The built-in fitting formulas they use are still the traditional NAL-NL2 (National Acoustic Laboratories — Non-Linear version 2) or DSL v5 (Desired Sensation Level version 5), which have inherent limitations in different hearing configurations and usage scenarios.

[0021] The NAL series formula is the most widely used prescription formula in the global adult hearing aid fitting field. Almost all mainstream hearing aid brands include NAL-NL2 as a default or core option in their fitting software. Clinical guidelines, academic literature, and product manuals all list NAL-NL2 as the standard reference for adult fitting. It maximizes speech intelligibility while ensuring comfortable overall loudness.

[0022] The NAL-NL2 has a reasonable gain setting in the mid-frequency range (1-2 kHz), but its target gain in the high-frequency range (4-6 kHz) is relatively low, resulting in insufficient audibility of consonants. Some users complain that they can hear them but cannot understand them clearly. At the same time, for people with severe and profound hearing loss (hearing threshold greater than 80 dB HL), the NAL-NL2 actively reduces the target gain for safety reasons, resulting in insufficient loudness and users feeling that the sound is not loud enough.

[0023] DSL v5 is the de facto gold standard in the global field of pediatric hearing aid fitting, and it is recommended as a prescription in pediatric hearing rehabilitation guidelines in many countries. The fitting software of major hearing aid brands also generally includes DSL v5. It can restore amplified sound to near-normal loudness perception, ensuring that all speech components (especially consonants) are audible to the user.

[0024] The DSL v5 has sufficient high-frequency target gain, but its high-frequency target value is often too high, which can easily cause acoustic feedback howling. In addition, some adult new users may feel uncomfortable or even refuse to wear it when they first wear it due to the high total loudness.

[0025] Neither of the above two formulas provides a specific gain allocation strategy for users with ear canal structures that pose a high risk of feedback. This makes it easy for such users to experience feedback when pursuing sufficient loudness, while sacrificing loudness to avoid feedback severely affects the hearing aid effect.

[0026] Furthermore, in remote fitting scenarios, audiologists cannot verify the results on-site using real-ear measurement (REM) equipment, which places higher demands on the rationality and stability of the fitting formula output. There is an urgent need for an intelligent fitting formula system that can perform well under various hearing profiles and different otoacoustic conditions and is suitable for remote distribution.

[0027] Based on the above considerations, embodiments of the present invention provide an adaptive remote hearing aid fitting method. Please refer to... Figure 1 In this embodiment, the adaptive remote hearing aid fitting method is specifically executed through steps S1 to S5: S1. Extract hearing loss features from the audiograms and acoustic condition information uploaded by the user; S2. Based on the maximum stable gain curve of the hearing aid self-test, assess the feedback risk level of each frequency band; S3. Calculate the target gain corresponding to the preset prescription formula; wherein, the prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula; the voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in the moderate to severe hearing loss segment; the high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, and reduces the risk of howling through frequency band differentiated gain allocation and dynamic feedback safety margin; S4. Based on the hearing loss characteristics, feedback risk level, and target gain, candidate formulas are selected from the prescription formulas, and formula parameters are generated; S5. Send the candidate formula and its formula parameters to the audiologist. In response to the audiologist's confirmation instruction, send the formula parameters to the user's hearing aid.

[0028] Hearing loss is not uniformly distributed. Audiograms can accurately pinpoint the specific frequency bands and degree of a user's hearing loss and correlate it with the acoustic characteristics of their environment, providing a data foundation for subsequent personalized calculations.

[0029] The hearing aid has a built-in pure-tone audiometry function. Before initiating a remote fitting, users can self-measure their hearing thresholds at different frequencies in a quiet environment, following the instructions. This will ultimately generate a maximum stable gain curve. For example, the built-in pure-tone audiometry function of the hearing aid will measure hearing thresholds at 250, 500, 1000, 1500, 2000, 3000, 4000, 6000, and 8000 Hz. It is understood that this function is built into the hearing aid, and the specific selected test frequency and the principle of maximum stable gain curve generation do not affect the beneficial effects produced by the embodiments of the present invention.

[0030] Traditional remote fitting is based on NAL-NL2 or DSL v5 formulas, but these have inherent limitations in different hearing profiles and usage scenarios. For users with ear canal structures at high feedback risk, pursuing sufficient loudness can easily lead to feedback, while sacrificing loudness to avoid feedback severely affects the hearing aid effect. In this embodiment of the invention, the preset prescription formula includes at least the Voice Compensate Balancer Plus (VCB+) formula and the Power Fitting Balancer (PFB) formula.

[0031] It should be noted that the speech compensation equalizer enhancement formula is an adaptive fitting formula developed in this invention to address the common shortcomings of NAL-NL2 (insufficient high frequencies) and DSL v5 (excessive high frequencies), as well as their insufficient loudness for severe hearing loss. Its principle lies in using different reference formulas as benchmarks for different frequency bands and dynamically fusing them through an adaptive weighting function. This allows for the simultaneous acquisition of the mid-frequency intelligibility advantage of NAL-NL2 and the high-frequency audibility advantage of DSL v5, while also providing loudness compensation enhancement for moderate to severe hearing loss.

[0032] It should also be noted that the high-power fitting equalization formula is a dedicated fitting formula designed for user groups prone to acoustic feedback howling. Traditional formulas are highly likely to trigger howling when providing sufficient loudness for open-ear mold wearers, those with large vents, those with severe external auditory canal leakage, and those with insufficient coupling between the hearing aid and the ear canal. Passively reducing gain, on the other hand, leads to insufficient loudness, which is counterproductive. The high-power fitting equalization formula of this invention, through a dual strategy of frequency band differentiated gain allocation combined with dynamic feedback safety margin, significantly reduces the risk of howling while ensuring loudness.

[0033] When user characteristics differ, the applicable formulas also differ. This embodiment of the invention can select candidate formulas suitable for the user from a preset prescription formula library by matching the degree of matching of several dimensions such as hearing loss characteristics, feedback risk level and target gain, and obtain the formula parameters.

[0034] Remote fitting is designed for medical safety, so the audiologist must confirm its quality and safety before the parameters take effect. After the audiologist confirms, the parameters can be sent to the hearing aid in real time via the cloud + Bluetooth link and take effect immediately.

[0035] Preferably, the prescription formula in the embodiments of the present invention can be used as a built-in algorithm in the hearing aid firmware, or it can be issued as a parameter table after confirmation by the audiologist.

[0036] The above solution can automatically recommend prescription formulas based on information submitted by users, which greatly reduces the reliance on the experience of audiologists and provides strong algorithmic support for non-prescription hearing aids in self-service fitting scenarios. At the same time, by introducing speech compensation equalizer enhancement formulas and high-power fitting equalization formulas, it can provide personalized gain solutions for users with moderate to severe hearing loss that fully compensate for loudness and actively suppress feedback.

[0037] As a preferred implementation, step S1, extracting hearing loss features from the audiograms and acoustic condition information uploaded by the user, includes: Obtain the audiogram uploaded by the user, and extract the user's hearing loss level and hearing loss configuration from the audiogram; Obtain acoustic condition information uploaded by the user, and extract the user's acoustic coupling method, user experience, and user age from the acoustic condition information; Hearing loss characteristics are obtained based on the degree of hearing loss, the hearing loss configuration, the acoustic coupling method, the user's experience, and the user's age.

[0038] In some preferred embodiments, audiograms are used to calculate features such as PTA (mean hearing threshold), configuration classification, and symmetry to obtain the user's hearing loss level and hearing loss configuration.

[0039] Hearing loss severity ratings are used to classify and characterize the severity of a user's hearing loss. For example, hearing loss severity ratings include mild, moderate, moderate to severe, severe, and profound.

[0040] Hearing loss profiles are used to reflect the shape characteristics of the distribution of hearing loss at different frequencies. For example, hearing loss profiles include flat, sloping, steeply sloping, rising, valley, and noise notch profiles.

[0041] Acoustic coupling method is used to describe the acoustic connection or matching method between the hearing aid and the user's ear canal. Exemplary acoustic coupling methods include closed earmold, semi-open, open, and vent sizes.

[0042] By comprehensively considering the degree of hearing loss, the hearing loss configuration, the acoustic coupling method, the user's experience, and the user's age, a multi-dimensional comprehensive description set can be obtained to characterize the individual user's hearing loss status and usage background.

[0043] In the above scheme, based on the audiogram and acoustic condition information uploaded by the user, multi-dimensional individualized parameters are extracted to construct a multi-dimensional user hearing loss profile, ensuring that all subsequent formula calculations and parameter generation are based on fully individualized information.

[0044] In a preferred embodiment, when the prescription formula is a voice compensation equalizer enhancement formula, step S3, calculating the target gain corresponding to the preset prescription formula, includes: The validity of the hearing thresholds at each frequency point in the audiogram is verified and missing values ​​are interpolated to obtain the verified hearing thresholds. The NAL-NL2 formula is used to calculate the first target gain at each frequency point under the preset decibel sound pressure level input. The second target gain at each frequency point is calculated using the DSL v5 formula under the preset decibel sound pressure level input. Based on predefined frequency-dependent weighting coefficients, the first target gain and the second target gain are weighted and summed to obtain the first weighted target gain; Calculate the loudness compensation term based on the verified hearing threshold; Based on the first weighted target gain, the loudness compensation term, and the compression ratio based on the degree of hearing loss, the target gain generated by the speech compensation equalizer enhancement formula at each decibel sound pressure level is obtained.

[0045] In some preferred embodiments, the target gain of the speech compensation equalizer enhancement formula at a 65dB SPL input is... Represented as: ; in, For frequency; and These are frequency-dependent weighted coefficients. ; The first target gain of the NAL-NL2 formula at a 65dB SPL input; The second target gain for the DSL v5 formula at a 65dB SPL input; This is a loudness compensation item for moderate to severe hearing loss.

[0046] When operating in the low-frequency range, the target gain of the speech compensation equalizer enhancement formula should be biased towards NAL-NL2 to suppress uplink masking caused by excessive low-frequency amplification. When operating in the mid-frequency range, the target gain should be biased towards maximizing speech intelligibility, i.e., still primarily using NAL-NL2. When operating in the mid-to-high frequency range, the target gain should be balanced between NAL-NL2 and DSL v5. When operating in the high-frequency range, NAL-NL2 has insufficient high-frequency audibility; to compensate for this, the target gain of the speech compensation equalizer enhancement formula should be biased towards DSL v5.

[0047] For example, when At that time, take , ;when At that time, take , ;when At that time, take , ;when At that time, take , .

[0048] The loudness compensation term is designed for moderate to severe hearing loss and is used to compensate for the insufficient loudness caused by the active gain reduction of NAL-NL2 in the severe hearing loss range.

[0049] Preferably, the loudness compensation term is defined as a piecewise function, expressed as: ; in, and The slope coefficient is a frequency-dependent compensation coefficient. To ensure greater compensation for severe hearing loss, the relationship between the two must satisfy [condition missing]. Typical values ​​are , .

[0050] Based on this, the target gain of the speech compensation equalizer enhancement formula at each decibel sound pressure level is derived from the standard WDRC (Wide Dynamic Range Compression) relationship, and is expressed as: ; in, Sound pressure level in decibels; The target gain of the speech compensation equalizer enhancement formula at a 65dB SPL input; This is the compression ratio based on the degree of hearing loss.

[0051] In some preferred embodiments, the compression ratio is determined by the verified hearing threshold. It is derived adaptively through piecewise linear functions and is expressed as: .

[0052] In the above embodiments, when When the hearing loss is not severe, the compression ratio can be directly set to 1.0 to make the target gain perfectly linear; when When the compression ratio is controlled between 1.0 and 1.2, the target gain exhibits a slightly approximately linear relationship; when At that time, the hearing loss was moderate, and the compression ratio was controlled between 1.2 and 1.8; when At that time, the degree of hearing loss was in the moderate to severe range, and the compression ratio was controlled between 1.8 and 2.2; when At that time, the degree of hearing loss was in the severe range, and the compression ratio was controlled between 2.2 and 2.4; when At that time, the degree of hearing loss was profound. To prevent excessive compression from damaging the natural language quality, the compression ratio was set to the upper limit of 2.4.

[0053] This segmented derivation takes into account both the naturalness of compression and the effective use of the residual dynamic range. The more severe the hearing loss, the greater the compression ratio and the narrower the dynamic range, the more sufficient compression is needed. At the same time, an upper limit is set to protect the speech envelope.

[0054] In the above scheme, by using NAL-NL2 and DSL v5 as reference formulas, adaptive weighted fusion is performed based on different frequency bands, which can simultaneously obtain the mid-frequency intelligibility advantage of NAL-NL2 and the high-frequency audibility advantage of DSL v5; and the loudness compensation term based on the verified hearing threshold can also enhance the loudness compensation for moderate to severe hearing loss, preventing the impact on hearing aid effect due to sacrificing loudness to avoid feedback.

[0055] In a preferred embodiment, when the prescription formula is a high-power fitting equalization formula, step S3, calculating the target gain corresponding to the preset prescription formula, is performed through steps S321-S328: S321. Calculate the third target gain generated by the speech compensation equalizer enhancement formula at the preset decibel sound pressure level; S322. Calculate the feedback safety margin point by point based on the maximum stable gain curve and the third target gain. S323. Based on the acoustic coupling method extracted from the acoustic condition information, obtain the feedback safety margin at each frequency point; the feedback safety margin is an engineering empirical parameter. S324. When the feedback safety margin is less than the feedback safety safety allowance, calculate the feedback penalty term at each frequency point based on the feedback safety margin, the stable gain curve, and the feedback safety allowance. S325. Generate frequency band differential compensation items for each frequency point based on the predefined frequency band compensation. S326. Calculate the first total loudness of the speech compensation equalizer enhancement formula and the second total loudness of the high-power fitting equalizer formula using the loudness model. S327. If the first difference between the first loudness and the second loudness is greater than a preset difference threshold, then the cross-band migration compensation amount is calculated based on the first difference, the difference between the feedback safety margin and the feedback safety allowance; the migration compensation amount is used to compensate for the howling frequency band. S328. Calculate the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount.

[0056] Understandably, the calculation method for the target gain corresponding to the speech compensation equalizer enhancement formula at the preset decibel sound pressure level has been explained in the above scheme. The above target gain calculation can be directly used as the third target gain, and the relevant content will not be repeated here.

[0057] The maximum stable gain (MSG) curve can be obtained through the hearing aid's maximum stable gain (MSG) self-test procedure. Furthermore, the feedback safety margin can be calculated on a frequency-by-frequency basis.

[0058] In some preferred embodiments, the feedback safety margin is expressed as: ; in, The third target gain is obtained by taking the voice compensation equalizer enhancement formula at a preset decibel sound pressure level.

[0059] Feedback safety margin The actual safe distance from the howling threshold of this frequency band is the fundamental variable for all subsequent decisions of the voice compensation equalizer enhancement formula, and it runs through the three stages of feedback penalty calculation, cross-band loudness transfer and feedback suppression module linkage.

[0060] It should be noted that the feedback safety margin These are engineering experience parameters, and their values ​​are graded according to acoustic coupling conditions. For example, under closed earmold conditions, the feedback safety margin is 3dB; under semi-open conditions, the feedback safety margin is 5dB; under large vent conditions, the feedback safety margin is 6dB; under open fitting conditions, the feedback safety margin is 8dB; and under extreme sound leakage conditions, the feedback safety margin is 10dB.

[0061] By comparing the feedback safety margin and the feedback safety safety safety limit, it can be determined whether the current frequency band requires feedback penalties. Feedback penalties are reflected in the feedback penalty item; therefore, the feedback penalty item is only... It is activated at specific times to ensure that the target gain is always lower than the maximum stable gain, thus preventing howling from occurring at the source.

[0062] In some preferred embodiments, the feedback penalty term is represented as: .

[0063] It should also be noted that the frequency band differentiation compensation item acts independently of the aforementioned feedback penalty item, and the frequency band differentiation compensation item is a preset static frequency band compensation.

[0064] In some preferred embodiments, the frequency band differentiation compensation term is expressed as: ; Preferably, cubic spline interpolation is used for smoothing at the frequency band junctions, with a smoothing bandwidth of 1 / 6 octave.

[0065] In this embodiment of the invention, the total loudness calculation is based on the ANSI S3.4-2007 or ISO 532-2 loudness model, and the first total loudness of the speech compensation equalizer enhancement formula is calculated respectively. The second total loudness of the high-power fitting equalization formula (Unit: phon). In the calculation, the output of each frequency band is obtained from the frequency band input and the target gain, and then substituted into the loudness model for integration to obtain the total loudness.

[0066] To ensure that compensation is allocated to frequency bands further away from the howling line, a cross-frequency band migration compensation amount is also designed in this embodiment of the invention. This migration compensation amount... It is triggered only when. This is the preset difference threshold.

[0067] In some preferred embodiments, the migration compensation amount is expressed as: ; in, For assignable weights, ; ; This is the conversion factor from loudness level to decibel.

[0068] In this embodiment of the invention, the dynamic weighting ensures that more compensation is allocated to frequency bands farther away from the howling line, and for frequency band They will not participate in compensation.

[0069] Further, preferably, step S328, calculating the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount, includes: The third target gain, the frequency band differential compensation term, and the migration compensation amount are superimposed to obtain a first intermediate value; The feedback penalty term is used to inversely superimpose the first intermediate value to obtain the initial target gain; The initial target gain is back-substituted and verified based on the feedback safety margin, the first total loudness, and the second total loudness. If the back-substitution verification result shows that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the initial target gain is calculated iteratively. If the back-substitution verification result is that the initial target gain converges or the number of iterations is not less than the preset maximum number of iterations, then the initial target gain is used as the target gain generated by the high-power fitting equalization formula at the preset decibel sound pressure level. Based on the target gain generated by the high-power fitting equalization formula at a preset decibel sound pressure level, the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level is obtained.

[0070] In some preferred embodiments, the initial target gain generated by the high-power matching equalization formula at a preset sound pressure level (65 dB SPL) is expressed as follows: .

[0071] After calculating the initial target gain, back-substitute to verify the feedback safety margin. If the first and second total loudness values ​​do not converge, the initial target gain is iteratively calculated within a preset maximum number of iterations. For example, a typical value for the maximum number of iterations is 3.

[0072] Furthermore, after calculating the target gain corresponding to the high-power fitting equalization formula at the preset decibel sound pressure level, the target gain corresponding to each decibel sound pressure level can also be derived based on the standard WDRC (Wide Dynamic Range Compression) relationship, consistent with the calculation under the voice compensation equalizer enhancement formula. This will not be elaborated further here.

[0073] In the above scheme, the high-power fitting equalization formula can avoid acoustic feedback howling from the source while ensuring loudness through frequency band differential gain allocation and dynamic feedback safety margin strategy, providing an effective fitting algorithm for open fitting and severely hearing-impaired users.

[0074] Furthermore, preferably, if the back-substitution verification result indicates that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the iterative calculation of the initial target gain includes: If the back-substitution verification result is that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the feedback safety margin and the feedback penalty term are sent to the feedback suppression module built into the hearing aid as a priority prompt for the initial notch frequency. The hearing aid feedback suppression module receives the equivalent maximum stable gain reduction based on the frequency band continuously triggered notch event. Based on the original maximum stable gain curve and the equivalent maximum stable gain reduction, it recalculates and updates the feedback safety margin, the feedback penalty term, and the migration compensation amount. The initial target gain is calculated iteratively based on the updated feedback safety margin, the feedback penalty term, and the migration compensation amount.

[0075] In this embodiment of the invention, a high-power fitting equalization formula engine and a feedback suppression module built into the hearing aid exchange information bidirectionally. The feedback suppression module can be an adaptive notch filter, a phase inversion canceller, etc.

[0076] During the two-way information exchange, the feedback safety margin and feedback penalty term currently calculated by the high-power matching equalization formula are passed to the feedback suppression module as a priority indicator for the initial notch frequency. In actual use, if the feedback suppression module detects continuous notch triggering in a certain frequency band, it reports the equivalent maximum stable gain reduction via an event callback. This allows the high-power fitting equalization formula to dynamically update the maximum stable gain curve based on the decrease in equivalent maximum stable gain, and to recalculate the feedback safety margin, feedback penalty term, and migration compensation, issuing a new gain in the next parameter update cycle.

[0077] In some preferred embodiments, the dynamic update of the maximum stable gain curve is expressed as follows: ; in, This is the original maximum stable gain curve.

[0078] The preferred embodiment of the present invention does not provide users with a subjective scoring entry point, such as scoring loudness, clarity, comfort and howling sensation, and feeds the scores back to the audiologist as a basis for iteration.

[0079] In the above scheme, each update of the target gain incorporates the latest measured data on the actual feedback suppression capability of the hearing aid, thereby maximizing the use of the safe gain space of each frequency band while ensuring that no new feedback risk is added. At the same time, the forced sacrifice is transferred to other frequency bands through migration compensation, so as to maintain the overall hearing compensation effect without significant deterioration due to local suppression.

[0080] As a preferred implementation, step S4, screening candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generating formula parameters, includes: If the hearing loss characteristics are a steep drop in high and medium frequencies, a closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and formula parameters are generated based on the target gain. If the hearing loss characteristics are severe or above, closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and the loudness compensation enhancement formula parameters are generated according to the target gain. If the hearing loss characteristics are moderate to severe, the acoustic coupling method is open eardrum or large vent, and the feedback risk level is high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters are generated according to the target gain. If the hearing loss characteristics are severe hearing loss, ear canal leakage, acoustic coupling method is semi-open tympanic membrane or poor coupling, and the feedback risk level is extremely high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters for strong feedback suppression are generated based on the target gain.

[0081] In this embodiment of the invention, the speech compensation equalizer enhancement formula overcomes the inherent defects of NAL-NL2's low high frequencies and DSL v5's excessive high frequencies, ensuring clear and audible high-frequency consonants while preserving mid-frequency intelligibility. Therefore, for users with a moderate to high high-frequency drop-off, closed eardrums, and a moderate feedback risk level, as well as users with severe or higher hearing loss, closed eardrums, and a moderate feedback risk level, the speech compensation equalizer enhancement formula can be adaptively recommended.

[0082] It is worth noting that for users with severe or higher hearing loss, closed eardrums, and a medium level of feedback risk, the recommended voice compensation equalizer enhancement formula further provides graded compensation for the hearing loss segment through a loudness compensation term, so as to significantly improve the loudness deficiency of NAL-NL2 in the severe hearing loss segment.

[0083] The high-power fitting equalization formula, through a combination of frequency band differentiated gain allocation, dynamic feedback penalty, and cross-band loudness conservation, avoids acoustic feedback howling at its source while ensuring loudness. It is particularly suitable for open-fit and severely hearing-impaired users. Therefore, the high-power fitting equalization formula is recommended for users with moderate to severe hearing loss, acoustic coupling methods of open diaphragm or large vents, and a high feedback risk level; and for users with severe hearing loss, ear canal leakage, acoustic coupling methods of semi-open diaphragm or poor coupling, and an extremely high feedback risk level. For the latter, a formula parameter for strong feedback suppression is further generated based on the target gain.

[0084] It should be noted that, in this embodiment of the invention, strong feedback suppression is achieved by linking with the built-in feedback suppression module of the hearing aid, and the specific implementation method can be found in the above-described two-way information exchange process.

[0085] In the above scheme, the adaptive formula recommendation mechanism can automatically select prescription formulas based on multi-dimensional features, which greatly reduces the reliance on the experience of the optician. Even junior opticians can give qualified prescriptions based on the recommendation results.

[0086] In a preferred embodiment, if the prescription formula further includes NAL-NL2 and DSL v5, then step S4, screening candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generating formula parameters, further includes: If the hearing loss characteristics are mild to moderate hearing loss in new adult users, closed eardrums, and low feedback risk level, then the NAL-NL2 and the speech compensation equalizer enhancement formula are used as candidate formulas, and formula parameters are generated based on the target gain. If the hearing loss characteristic is that of a child user and the feedback risk level is low, then the children's version of DSL v5 is used as a candidate formula, and formula parameters are generated according to the target gain.

[0087] In some preferred embodiments, the candidate formula selection strategy is shown in Table 1.

[0088] Table 1 In some preferred embodiments, the adaptive remote hearing aid fitting method of this invention is implemented through an application. Users initiate remote audio and video calls, upload personal information and acoustic condition information, and connect to the hearing aid via Bluetooth through their mobile application. The audiologist receives user calls, manages user files, invokes the adaptive fitting formula engine, adjusts hearing aid parameters in real time, and sends the adjustments to the user's terminal through their application. The adaptive fitting formula engine incorporates four prescription formulas: NAL-NL2, DSL v5, VCB+ (Voice Compensate Balancer Plus), and PFB (Power Fitting Balancer), and has the ability to automatically recommend the optimal formula based on the user's audiogram, hearing loss configuration, and feedback risk level.

[0089] More preferably, based on real-time communication protocols such as WebRTC, it supports low-latency, high-fidelity two-way audio and video transmission, while performing real-time voice recognition on the voices of the audiologist and the user, and displaying subtitles on both parties' interfaces in real time.

[0090] The adaptive remote hearing aid fitting method provided by this invention can automatically recommend prescription formulas based on information submitted by the user, greatly reducing the reliance on the experience of the audiologist and providing strong algorithmic support for non-prescription hearing aids in self-service fitting scenarios. At the same time, by introducing a speech compensation equalizer enhancement formula and a high-power fitting equalization formula, it can provide personalized gain solutions for users with moderate to severe hearing loss that fully compensate for loudness and actively suppress feedback.

[0091] This invention provides an adaptive remote hearing aid fitting system. Please refer to [link / reference]. Figure 2The adaptive remote hearing aid fitting system includes a hearing loss feature extraction module 11, a feedback risk level assessment module 12, a target gain calculation module 13, a candidate formula screening module 14, and a parameter distribution module 15, wherein: The hearing loss feature extraction module 11 is used to extract hearing loss features from the audiograms and acoustic condition information uploaded by the user. Feedback risk level assessment module 12 is used to assess the feedback risk level of each frequency band based on the maximum stable gain curve of the hearing aid self-test. The target gain calculation module 13 is used to calculate the target gain corresponding to the preset prescription formula; wherein, the prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula; the voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in the moderate to severe hearing loss segment; the high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, and reduces the risk of howling through frequency band differentiated gain allocation and dynamic feedback safety margin; The candidate formula screening module 14 is used to screen candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level and target gain, and generate formula parameters. The parameter sending module 15 is used to send the candidate formula and its formula parameters to the audiologist, and in response to the audiologist's confirmation instruction, send the formula parameters to the user's hearing aid.

[0092] In a preferred embodiment, the hearing loss feature extraction module 11 is specifically used for: Obtain the audiogram uploaded by the user, and extract the user's hearing loss level and hearing loss configuration from the audiogram; Obtain acoustic condition information uploaded by the user, and extract the user's acoustic coupling method, user experience, and user age from the acoustic condition information; Hearing loss characteristics are obtained based on the degree of hearing loss, the hearing loss configuration, the acoustic coupling method, the user's experience, and the user's age.

[0093] In a preferred embodiment, when the prescription formula is a speech compensation equalizer enhancement formula, the target gain calculation module 13 is specifically used for: The validity of the hearing thresholds at each frequency point in the audiogram is verified and missing values ​​are interpolated to obtain the verified hearing thresholds. The NAL-NL2 formula is used to calculate the first target gain at each frequency point under the preset decibel sound pressure level input. The second target gain at each frequency point is calculated using the DSL v5 formula under the preset decibel sound pressure level input. Based on predefined frequency-dependent weighting coefficients, the first target gain and the second target gain are weighted and summed to obtain the first weighted target gain; Calculate the loudness compensation term based on the verified hearing threshold; Based on the first weighted target gain, the loudness compensation term, and the compression ratio based on the degree of hearing loss, the target gain generated by the speech compensation equalizer enhancement formula at each decibel sound pressure level is obtained.

[0094] In a preferred embodiment, when the prescription formula is a high-power fitting equalization formula, the target gain calculation module 13 is specifically used for: Calculate the third target gain generated by the speech compensation equalizer enhancement formula at a preset decibel sound pressure level; Based on the maximum stable gain curve and the third target gain, calculate the feedback safety margin point by point; Based on the acoustic coupling method extracted from the acoustic condition information, the feedback safety margin at each frequency point is obtained; the feedback safety margin is an engineering empirical parameter. When the feedback safety margin is less than the feedback safety safety allowance, the feedback penalty term at each frequency point is calculated based on the feedback safety margin, the stable gain curve, and the feedback safety allowance. Based on the predefined frequency band compensation, generate frequency band differential compensation items for each frequency point; The first total loudness of the speech compensation equalizer enhancement formula and the second total loudness of the high-power fitting equalizer formula are calculated using the loudness model. If the first difference between the first loudness and the second loudness is greater than a preset difference threshold, then the cross-band migration compensation amount is calculated based on the first difference, the difference between the feedback safety margin and the feedback safety allowance; the migration compensation amount is used to compensate for the howling frequency band. Based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount, calculate the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level.

[0095] Further, preferably, the step of calculating the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount includes: The third target gain, the frequency band differential compensation term, and the migration compensation amount are superimposed to obtain a first intermediate value; The feedback penalty term is used to inversely superimpose the first intermediate value to obtain the initial target gain; The initial target gain is back-substituted and verified based on the feedback safety margin, the first total loudness, and the second total loudness. If the back-substitution verification result shows that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the initial target gain is calculated iteratively. If the back-substitution verification result is that the initial target gain converges or the number of iterations is not less than the preset maximum number of iterations, then the initial target gain is used as the target gain generated by the high-power fitting equalization formula at the preset decibel sound pressure level. Based on the target gain generated by the high-power fitting equalization formula at a preset decibel sound pressure level, the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level is obtained.

[0096] Furthermore, preferably, if the back-substitution verification result indicates that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the iterative calculation of the initial target gain includes: If the back-substitution verification result is that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the feedback safety margin and the feedback penalty term are sent to the feedback suppression module built into the hearing aid as a priority prompt for the initial notch frequency. The hearing aid feedback suppression module receives the equivalent maximum stable gain reduction based on the frequency band continuously triggered notch event. Based on the original maximum stable gain curve and the equivalent maximum stable gain reduction, it recalculates and updates the feedback safety margin, the feedback penalty term, and the migration compensation amount. The initial target gain is calculated iteratively based on the updated feedback safety margin, the feedback penalty term, and the migration compensation amount.

[0097] In a preferred embodiment, the candidate formula filtering module 14 is specifically used for: If the hearing loss characteristics are a steep drop in high and medium frequencies, a closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and formula parameters are generated based on the target gain. If the hearing loss characteristics are severe or above, closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and the loudness compensation enhancement formula parameters are generated according to the target gain. If the hearing loss characteristics are moderate to severe, the acoustic coupling method is open eardrum or large vent, and the feedback risk level is high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters are generated according to the target gain. If the hearing loss characteristics are severe hearing loss, ear canal leakage, acoustic coupling method is semi-open tympanic membrane or poor coupling, and the feedback risk level is extremely high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters for strong feedback suppression are generated based on the target gain.

[0098] In a preferred embodiment, if the prescription formula further includes NAL-NL2 and DSL v5, then the candidate formula screening module 14 is further configured to: If the hearing loss characteristics are mild to moderate hearing loss in new adult users, closed eardrums, and low feedback risk level, then the NAL-NL2 and the speech compensation equalizer enhancement formula are used as candidate formulas, and formula parameters are generated based on the target gain. If the hearing loss characteristic is that of a child user and the feedback risk level is low, then the children's version of DSL v5 is used as a candidate formula, and formula parameters are generated according to the target gain.

[0099] The adaptive remote hearing aid fitting system provided by this invention can automatically recommend prescription formulas based on information submitted by the user, greatly reducing the reliance on the experience of the audiologist and providing strong algorithmic support for non-prescription hearing aids in self-service fitting scenarios. At the same time, by introducing a speech compensation equalizer enhancement formula and a high-power fitting equalization formula, it can provide personalized gain solutions for users with moderate to severe hearing loss that fully compensate for loudness and actively suppress feedback.

[0100] Please see Figure 3 The present invention also provides an adaptive hearing aid remote fitting device, including a processor 31, a memory 32, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements an adaptive hearing aid remote fitting method as described above. The working principles and beneficial effects of the two are one-to-one, so they will not be described in detail here.

[0101] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0102] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. An adaptive remote hearing aid fitting method, characterized in that, include: Hearing loss features are extracted from audiograms and acoustic condition information uploaded by users; Based on the maximum stable gain curve of the hearing aid self-test, assess the feedback risk level of each frequency band; The target gain corresponding to the preset prescription formula is calculated; wherein, the prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula; the voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in the moderate to severe hearing loss segment; the high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, and reduces the risk of howling through frequency band differentiated gain allocation and dynamic feedback safety margin; Based on hearing loss characteristics, feedback risk level, and target gain, candidate formulas are selected from the prescription formulas, and formula parameters are generated. The candidate formula and its parameters are sent to the audiologist, and in response to the audiologist's confirmation instruction, the formula parameters are sent to the user's hearing aid.

2. The adaptive remote hearing aid fitting method as described in claim 1, characterized in that, The extraction of hearing loss features from user-uploaded audiograms and acoustic condition information includes: Obtain the audiogram uploaded by the user, and extract the user's hearing loss level and hearing loss configuration from the audiogram; Obtain acoustic condition information uploaded by the user, and extract the user's acoustic coupling method, user experience, and user age from the acoustic condition information; Hearing loss characteristics are obtained based on the degree of hearing loss, the hearing loss configuration, the acoustic coupling method, the user's experience, and the user's age.

3. The adaptive remote hearing aid fitting method as described in claim 1, characterized in that, When the prescription formula is a speech compensation equalizer enhancement formula, the target gain generated by the preset prescription formula includes: The validity of the hearing thresholds at each frequency point in the audiogram is verified and missing values ​​are interpolated to obtain the verified hearing thresholds. The NAL-NL2 formula is used to calculate the first target gain at each frequency point under the preset decibel sound pressure level input. The second target gain at each frequency point is calculated using the DSL v5 formula under the preset decibel sound pressure level input. Based on predefined frequency-dependent weighting coefficients, the first target gain and the second target gain are weighted and summed to obtain the first weighted target gain; Calculate the loudness compensation term based on the verified hearing threshold; Based on the first weighted target gain, the loudness compensation term, and the compression ratio based on the degree of hearing loss, the target gain generated by the speech compensation equalizer enhancement formula at each decibel sound pressure level is obtained.

4. The adaptive remote hearing aid fitting method as described in claim 1, characterized in that, When the prescription formula is a high-power fitting equalization formula, the target gain generated by the preset prescription formula includes: Calculate the third target gain generated by the speech compensation equalizer enhancement formula at a preset decibel sound pressure level; Based on the maximum stable gain curve and the third target gain, calculate the feedback safety margin point by point; Based on the acoustic coupling method extracted from the acoustic condition information, the feedback safety margin at each frequency point is obtained; the feedback safety margin is an engineering empirical parameter. When the feedback safety margin is less than the feedback safety safety allowance, the feedback penalty term at each frequency point is calculated based on the feedback safety margin, the stable gain curve, and the feedback safety allowance. Based on the predefined frequency band compensation, generate frequency band differential compensation items for each frequency point; The first total loudness of the speech compensation equalizer enhancement formula and the second total loudness of the high-power fitting equalizer formula are calculated using the loudness model. If the first difference between the first loudness and the second loudness is greater than a preset difference threshold, then the cross-band migration compensation amount is calculated based on the first difference, the difference between the feedback safety margin and the feedback safety allowance; the migration compensation amount is used to compensate for the howling frequency band. Based on the third target gain, the frequency band differentiation compensation term, the feedback penalty term, and the migration compensation amount, calculate the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level.

5. The adaptive remote hearing aid fitting method as described in claim 4, characterized in that, The step of calculating the target gain corresponding to each decibel sound pressure level produced by the high-power fitting equalization formula based on the third target gain, the frequency band differential compensation term, the feedback penalty term, and the migration compensation amount includes: The third target gain, the frequency band differential compensation term, and the migration compensation amount are superimposed to obtain a first intermediate value; The feedback penalty term is used to inversely superimpose the first intermediate value to obtain the initial target gain; The initial target gain is back-substituted and verified based on the feedback safety margin, the first total loudness, and the second total loudness. If the back-substitution verification result shows that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the initial target gain is calculated iteratively. If the back-substitution verification result is that the initial target gain converges or the number of iterations is not less than the preset maximum number of iterations, then the initial target gain is used as the target gain generated by the high-power fitting equalization formula at the preset decibel sound pressure level. Based on the target gain generated by the high-power fitting equalization formula at a preset decibel sound pressure level, the target gain generated by the high-power fitting equalization formula at each decibel sound pressure level is obtained.

6. The adaptive remote hearing aid fitting method as described in claim 5, characterized in that, If the back-substitution verification result indicates that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the initial target gain is iteratively calculated, including: If the back-substitution verification result is that the initial target gain has not converged and the number of iterations is less than the preset maximum number of iterations, then the feedback safety margin and the feedback penalty term are sent to the feedback suppression module built into the hearing aid as a priority prompt for the initial notch frequency. The hearing aid feedback suppression module receives the equivalent maximum stable gain reduction based on the frequency band continuously triggered notch event. Based on the original maximum stable gain curve and the equivalent maximum stable gain reduction, it recalculates and updates the feedback safety margin, the feedback penalty term, and the migration compensation amount. The initial target gain is calculated iteratively based on the updated feedback safety margin, the feedback penalty term, and the migration compensation amount.

7. The adaptive remote hearing aid fitting method as described in claim 1, characterized in that, The step of selecting candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generating formula parameters, includes: If the hearing loss characteristics are a steep drop in high and medium frequencies, a closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and formula parameters are generated based on the target gain. If the hearing loss characteristics are severe or above, closed eardrum, and the feedback risk level is medium, then the speech compensation equalizer enhancement formula is used as a candidate formula, and the loudness compensation enhancement formula parameters are generated according to the target gain. If the hearing loss characteristics are moderate to severe, the acoustic coupling method is open eardrum or large vent, and the feedback risk level is high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters are generated according to the target gain. If the hearing loss characteristics are severe hearing loss, ear canal leakage, acoustic coupling method is semi-open tympanic membrane or poor coupling, and the feedback risk level is extremely high, then the high-power fitting equalization formula is used as a candidate formula, and formula parameters for strong feedback suppression are generated based on the target gain.

8. The adaptive remote hearing aid fitting method as described in claim 1, characterized in that, The prescription formula also includes NAL-NL2 and DSL v5. The step of selecting candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generating formula parameters, includes: If the hearing loss characteristics are mild to moderate hearing loss in new adult users, closed eardrums, and low feedback risk level, then the NAL-NL2 and the speech compensation equalizer enhancement formula are used as candidate formulas, and formula parameters are generated based on the target gain. If the hearing loss characteristic is that of a child user and the feedback risk level is low, then the children's version of DSL v5 is used as a candidate formula, and formula parameters are generated according to the target gain.

9. An adaptive remote hearing aid fitting system, characterized in that, include: The hearing loss feature extraction module is used to extract hearing loss features from the audiograms and acoustic condition information uploaded by the user. The feedback risk level assessment module is used to assess the feedback risk level of each frequency band based on the maximum stable gain curve of the hearing aid self-test. The target gain calculation module is used to calculate the target gain corresponding to the preset prescription formula; wherein, the prescription formula includes at least a voice compensation equalizer enhancement formula and a high-power fitting equalizer formula; the voice compensation equalizer enhancement formula is based on frequency band adaptive fusion of NAL-NL2 and DSL v5 to enhance loudness compensation in the moderate to severe hearing loss segment; the high-power fitting equalizer formula is constructed based on the voice compensation equalizer enhancement formula, and reduces the risk of howling through frequency band differentiated gain allocation and dynamic feedback safety margin; The candidate formula screening module is used to screen candidate formulas from the prescription formulas based on hearing loss characteristics, feedback risk level, and target gain, and generate formula parameters. The parameter distribution module is used to send the candidate formula and its formula parameters to the audiologist, and in response to the audiologist's confirmation instruction, to distribute the formula parameters to the user's hearing aid.

10. An adaptive remote hearing aid fitting device, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the adaptive remote hearing aid fitting method as described in any one of claims 1 to 8.

Images