Hearing-aid sound transmission device system with intelligent frequency modulation and training method thereof

Abstract

A hearing-aid sound transmission device system with intelligent frequency modulation disclosed by the present invention comprises an electronic device or a sound transmission device, the electronic device is provided with a training APP, and the electronic device is configured with a data receiving end, a data transmitting end and a cloud storage interaction module. The sound transmission device comprises a microphone used for collecting sound, a sound transmission device used for playing audio, and a sound intensity sensor used for monitoring external sound intensity in real time. The training APP is internally provided with a preset voice sample library, a voice processing module, a frequency analysis module, an intelligent fine tuning module, a noise reduction filtering module and a hearing protection module, and the preset voice sample library comprises standard voice materials of at least seven specific frequency intervals; the sound transmission device is in wireless connection with the electronic equipment through a Bluetooth system, sound data collected by the microphone is transmitted to the training APP through the data receiving end, and the voice processing module, the frequency analysis module and the intelligent fine tuning module cooperatively process the data in sequence and generate optimized audio.

Description

Technical Field

[0001] This invention relates to the field of acoustic hearing aid systems, and in particular to a hearing aid transmission device system with intelligent frequency modulation and its training method. Background Technology

[0002] In today's society, there are a large number of people with hearing impairments such as deafness and hearing loss, which can be very inconvenient for them in social, entertainment, and learning activities. Most hearing aids on the market have limited functions and require special amplifiers, making them bulky and inconvenient.

[0003] Of course, there are some Bluetooth headphones available, but their structure is relatively complex and cannot be accepted or used by most users. Users find them too troublesome, and for people with speech impairments, there may be some unclear pronunciations or sentences when recognizing certain words. Therefore, sometimes it is necessary to listen to the passage several times to understand it. Summary of the Invention

[0004] The main objective of this invention is to propose a hearing aid transmission device system with intelligent frequency modulation and its training method. The system aims to control the transmission device with logical learning capabilities, thereby enabling repeated training of the system to achieve a matching relationship with the user, thus improving the user experience, making it more convenient and efficient.

[0005] To achieve the above objectives, the present invention proposes a hearing aid transmission device system with intelligent frequency modulation, comprising,

[0006] Electronic devices (such as watches, earphones, earphone cases, mobile phones, portable electronic screens, etc.), wherein the electronic devices are equipped with a training APP, and the electronic devices are configured with a data receiver, a data transmitter and a cloud storage interaction module;

[0007] An acoustic device, comprising a microphone for collecting sound, a sound transmission device for playing audio, and a sound intensity sensor for real-time monitoring of external sound intensity.

[0008] The training app includes a pre-set speech sample library, a speech processing module, a frequency analysis module, an intelligent fine-tuning module, a noise reduction and filtering module, and a hearing protection module.

[0009] The preset voice sample library contains standard voice materials in at least seven specific frequency ranges;

[0010] The sound transmission device and the electronic device are wirelessly connected via Bluetooth. The sound data collected by the microphone is transmitted to the training APP via the data receiving end. The voice processing module, frequency analysis module, and intelligent fine-tuning module process the data in sequence and generate optimized audio. The optimized audio is transmitted to the sound transmission device via the data transmitting end. The cloud storage interaction module realizes the cloud storage and retrieval of adjustment parameters, standard audio, and non-standard audio. The hearing protection module dynamically adjusts the audio gain threshold according to the sound intensity sensor data.

[0011] The speech processing module integrates an audio-to-text (ASR) model to convert standard speech materials and collected audio data into text and simultaneously compare semantic, speech, intonation, and speech rate differences.

[0012] The frequency analysis module integrates a Fourier Transform (FFT) unit and a Sound Intensity Response Function (RMS) algorithm unit, which is used to perform frequency spectrum conversion and sound intensity calculation on speech data with large semantic differences.

[0013] The intelligent fine-tuning module has a built-in multi-task statistical analysis unit (T-Model) to statistically determine the frequency spectrum of hearing defects (e.g., sound-color spectrum, effectively analyzing existing background noise and high noise) by combining environmental factors such as dialect and speech distortion, and supports independent or synchronous loudness gain fixed-point adjustment for the left and right ears.

[0014] A training method for a hearing aid transmission device system with intelligent frequency modulation, based on the above-mentioned training method for an adjustable intelligent hearing aid transmission device system, includes the following steps:

[0015] S1: Construct a preset speech sample library, which contains standard speech materials in seven specific frequency ranges: 250Hz-300Hz, 500-600Hz, 1000-1100Hz, 1500Hz, 2000Hz, 4000Hz, and 8000Hz. At the same time, configure a natural dialogue speech acquisition task.

[0016] S2: Connect the sound transmission device to the electronic device via Bluetooth, start the training APP and complete the cloud storage account binding. The training APP will automatically load the preset voice sample library.

[0017] S3: Users repeat materials from a preset speech sample library through the microphone of the sound transmission device, and can also choose to participate in a natural dialogue task. The microphone simultaneously collects the repeated speech data and the natural dialogue speech data.

[0018] S4: The training APP optimizes the collected speech data by removing background noise and segmenting and classifying it. Through ASRModel, the optimized speech data and corresponding standard speech materials are converted into text and displayed on the electronic device interface for users to proofread the semantics and compare and analyze the differences in speech, intonation and speech rate.

[0019] S5: Extract speech segments with significant differences in semantic and speech features, convert them into frequency spectra using Fourier FFT transform, and calculate the sound intensity corresponding to different frequencies using the RMS algorithm to preliminarily identify frequencies of potential hearing defects.

[0020] S6: Using the T-Model multi-task statistical analysis unit, combined with the speech distortion factors caused by the user's dialect and education level, the initially identified defect frequencies are statistically analyzed in multiple dimensions to determine the frequency spectrum of the user's core hearing defects and the corresponding sound intensity requirements.

[0021] S7: The intelligent fine-tuning module adjusts the loudness gain of the corresponding frequency according to the frequency spectrum of the core defect. It supports users to manually select synchronous adjustment of the left and right ears or independent adjustment, and generates the initial optimized audio and plays it through the sound transmission device.

[0022] S8: Users provide feedback on audio clarity through silence, indistinct consonants, or loud repetition. The intelligent fine-tuning module iteratively adjusts the gain parameters according to the preset scale based on the feedback until a standard optimized audio acceptable to the user is generated. At the same time, the adjusted parameters, the standard optimized audio, and the non-standard audio during the adjustment process are uploaded and stored through the cloud storage interaction module.

[0023] S9: Continuously collect voice interaction data from daily use, repeat steps S4-S8, dynamically update the frequency spectrum of the user's hearing loss and adjustment parameters, and achieve dynamic matching between the system and the user's hearing condition.

[0024] Advantages of this invention:

[0025] 1. In actual design, the training software automatically communicates with hearing-impaired patients through dialogue and other interactive methods, thereby accurately locating the patient's insensitive points to audio frequencies, and fine-tuning the playback gain through the software to improve the patient's wearing experience and greatly reduce the patient's dependence on mobile phone operation.

[0026] In actual use, users can provide feedback on whether the audio is clear, thereby adjusting a certain segment of the output audio and saving it to the training software;

[0027] Furthermore, as users age, they may experience a decline in the functionality of certain audio functions. Therefore, by continuously learning and fine-tuning the device, continuous interaction with the user can be achieved.

[0028] 2. Breaking through the limitations of existing hearing aids' "single adjustment", it deeply integrates multiple technologies such as "specific frequency preset speech sample library + ASR semantic correction + FFT frequency spectrum conversion + RMS sound intensity calculation + T-Model multi-task statistics" to achieve intelligent processing of the entire process from speech acquisition to defect frequency localization, solving the technical pain point that existing devices cannot accurately identify the frequency of individual hearing defects.

[0029] 3. The innovative integration of "sound intensity sensor + hearing protection module + low latency adjustment" constructs a secondary hearing damage protection mechanism, overcoming the shortcomings of existing equipment that lack adaptive protection in high-decibel environments;

[0030] 4. It achieves the linkage between "cloud storage + multi-device synchronization" and dynamic training, solving the industry problem of users having to repeat training when changing devices, and improving the system's practicality and convenience;

[0031] 5. As a better reference model, the key design point is not to increase the decibel level to improve the sound recognition, but to use multiple specific frequency ranges as specific models, and then adjust specific frequencies to the predetermined suitable hearing frequencies, thereby improving hearing aid efficiency and effectively avoiding secondary damage.

[0032] In other words, this model is a good and simple training model, which effectively reduces the complexity of training. Attached Figure Description

[0033] Figure 1 For training purposes Figure 1 ;

[0034] Figure 2 For training purposes Figure 2 . Detailed Implementation

[0035] The technical solutions in 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 a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0036] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, top, bottom, inside, outside, vertical, horizontal, longitudinal, counterclockwise, clockwise, circumferential, radial, axial, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0037] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0038] like Figures 1 to 2 As shown, including,

[0039] Electronic devices (such as watches, earphones, earphone cases, mobile phones, portable electronic screens, etc.), wherein the electronic devices are equipped with a training APP, and the electronic devices are configured with a data receiver, a data transmitter and a cloud storage interaction module;

[0040] An acoustic device, comprising a microphone for collecting sound, a sound transmission device for playing audio, and a sound intensity sensor for real-time monitoring of external sound intensity.

[0041] The training app includes a pre-set speech sample library, a speech processing module, a frequency analysis module, an intelligent fine-tuning module, a noise reduction and filtering module, and a hearing protection module.

[0042] The preset voice sample library contains standard voice materials in at least seven specific frequency ranges;

[0043] The sound transmission device and the electronic device are wirelessly connected via Bluetooth. The sound data collected by the microphone is transmitted to the training APP via the data receiving end. The voice processing module, frequency analysis module, and intelligent fine-tuning module process the data in sequence and generate optimized audio. The optimized audio is transmitted to the sound transmission device via the data transmitting end. The cloud storage interaction module realizes the cloud storage and retrieval of adjustment parameters, standard audio, and non-standard audio. The hearing protection module dynamically adjusts the audio gain threshold according to the sound intensity sensor data.

[0044] The speech processing module integrates an audio-to-text (ASR) model to convert standard speech materials and collected audio data into text and simultaneously compare semantic, speech, intonation, and speech rate differences.

[0045] The frequency analysis module integrates a Fourier Transform (FFT) unit and a Sound Intensity Response Function (RMS) algorithm unit, which is used to perform frequency spectrum conversion and sound intensity calculation on speech data with large semantic differences.

[0046] The Fourier transform is an important mathematical tool used to convert signals between the time and frequency domains. The following is an introduction to its core content: (i.e., using existing models in conjunction with audio frequency switching).

[0047] 1. Core Idea: The core of the Fourier transform is to decompose a complex signal into a superposition of sine waves of different frequencies. Any periodic signal can be represented as a combination of a series of sine waves, and aperiodic signals can also be transformed into a combination of sine waves of continuous frequencies through the concept of limits. This decomposition method helps to analyze the frequency components of a signal and reveal its essential characteristics.

[0048] 2. Mathematical Definition

[0049] Continuous Fourier Transform: For a time-domain function f(t), its Fourier transform is defined as:

[0050] F(ω)=∫−∞∞f(t)e−iωtdtF(ω)=∫−∞∞f(t)e−iωtdt

[0051] Where ω is the angular frequency, and F(ω) is the frequency domain representation. The inverse transform is:

[0052] f(t)=12π∫−∞∞F(ω)eiωtdωf(t)=2π1∫−∞∞F(ω)eiωtdω

[0053] Discrete Fourier Transform (DFT): Applicable to discrete signals, defined as:

[0054] X[k]=∑n=0N−1x[n]e−i2πkn / NX[k]=n=0∑N−1x[n]e−i2πkn / N

[0055] Where x[n] is the discrete time-domain signal, X[k] is the frequency-domain sequence, and NN is the signal length.

[0056] 3. Physical meaning

[0057] Time-domain perspective: describes how signals change over time, such as sound waveforms and stock price trends.

[0058] Frequency domain perspective: Displays the frequency components of a signal and the amplitude and phase information of each frequency. Through Fourier transform, time-domain signals can be converted into frequency-domain signals, facilitating the analysis of the signal's frequency characteristics, such as filtering, noise reduction, and feature extraction.

[0059] 4. Main application areas

[0060] Signal processing: used for audio noise reduction, image filtering, modulation and demodulation in communication systems, etc.

[0061] Image processing: Implementing image compression (such as JPEG), edge detection, image enhancement, etc.

[0062] Vibration analysis: In mechanical engineering, the location and extent of faults are diagnosed by analyzing the frequency components of equipment vibration signals.

[0063] The sound intensity response function H(I) represents the output response of the system under the action of sound intensity II. The output can be physical quantities such as sound pressure, particle velocity, and energy absorption rate. It reflects the system's sensitivity and response characteristics to sound waves of different intensities. For example:

[0064] In noise control, it is used to evaluate the attenuation effect of sound insulation materials on noise of different intensities, that is, to convert existing noise filtering into sound intensity switching.

[0065] The intelligent fine-tuning module has a built-in multi-task statistical analysis unit (T-Model) to statistically determine the frequency spectrum of hearing defects (e.g., sound-color spectrum, effectively analyzing existing background noise and high noise) by combining environmental factors such as dialect and speech distortion, and supports independent or synchronous loudness gain fixed-point adjustment for the left and right ears.

[0066] 1. In actual design, the training software automatically communicates with hearing-impaired patients through dialogue and other interactive methods, thereby accurately locating the patient's insensitive points to audio frequencies, and fine-tuning the playback gain through the software to improve the patient's wearing experience and greatly reduce the patient's dependence on mobile phone operation.

[0067] In actual use, users can provide feedback on whether the audio is clear, thereby adjusting a certain segment of the output audio and saving it to the training software;

[0068] Furthermore, as users age, they may experience a decline in the functionality of certain audio functions. Therefore, by continuously learning and fine-tuning the device, continuous interaction with the user can be achieved.

[0069] 2. Breaking through the limitations of existing hearing aids' "single adjustment", it deeply integrates multiple technologies such as "specific frequency preset speech sample library + ASR semantic correction + FFT frequency spectrum conversion + RMS sound intensity calculation + T-Model multi-task statistics" to achieve intelligent processing of the entire process from speech acquisition to defect frequency localization, solving the technical pain point that existing devices cannot accurately identify the frequency of individual hearing defects.

[0070] 3. The innovative integration of "sound intensity sensor + hearing protection module + low latency adjustment" constructs a secondary hearing damage protection mechanism, overcoming the shortcomings of existing equipment that lack adaptive protection in high-decibel environments;

[0071] 4. It achieves the linkage between "cloud storage + multi-device synchronization" and dynamic training, solving the industry problem of users having to repeat training when changing devices, and improving the system's practicality and convenience.

[0072] Specifically, the standard voice materials in the preset voice sample library include standard voice materials in specific frequency ranges of 250Hz-300Hz, 500-600Hz, 1000-1100Hz, 1500Hz, 2000Hz, 4000Hz, and 8000Hz.

[0073] Specifically, it can be the following everyday expressions, such as...

[0074] "Thank you" at 250Hz-300Hz

[0075] "It's nice to meet you" at 500-600Hz

[0076] "No problem, I think it's fine" for 1000-1100Hz.

[0077] "Is it okay?" at 1500Hz

[0078] 2000Hz sounds pretty good!

[0079] Please provide a detailed explanation of 4000Hz!

[0080] The 8000Hz frequency is a constant chorus of "Come quick, help me out!"

[0081] Of course, the vocabulary database is not limited to the words mentioned above; it can also filter pre-defined words based on existing dialogues.

[0082] In a preferred embodiment, the vocabulary is used only when the user's trained vocabulary reaches a predetermined range value. That is, a predetermined vocabulary and seven specific frequency bands are trained first to improve the recognition accuracy.

[0083] The specific frequency speech sample library covers the core frequency range of human daily communication (250Hz-8000Hz), and each frequency corresponds to high-frequency usage sentences, which is more targeted than general speech materials and significantly improves the efficiency of defective frequency recognition.

[0084] It effectively changes the existing training system without addressing the issue of standardized design.

[0085] Specifically, the multi-task statistical analysis unit (T-Model) achieves multi-task collaborative switching and efficient analysis by embedding specific expressions of hearing aid frequency recognition, sound intensity matching, and semantic correction tasks into a pre-trained model. That is, it introduces existing big data models to reduce the cost of purchasing databases, such as using large models like BERT, GPT, or Deepseek.

[0086] Of course, whether or not to apply a model can be set according to actual needs. That is, the large model can aggregate big data based on existing stored data and frequency modulation rules to achieve analysis of more data, thereby improving further training refinement, improving existing shortcomings, and achieving automatic recognition.

[0087] The task embedding method of the multi-task statistical analysis unit is clearly defined, and the specific tasks of the hearing aid scenario are deeply bound to the pre-trained model, which solves the problem of low efficiency of multi-task analysis and ensures the accuracy and real-time performance of frequency recognition.

[0088] That is, deep training is carried out using a combination of stand-alone and network-based modes.

[0089] Specifically, the intelligent fine-tuning module supports users to trigger parameter adjustments through three feedback methods: silence, indistinct consonants, and loud repetition. The gain adjustment is gradually increased according to a preset scale, and the single adjustment range does not exceed 5%.

[0090] The feedback triggering mechanism and calibrated gain adjustment take into account both user ease of operation and hearing safety, avoiding damage caused by excessive gain adjustment;

[0091] In other words, frequency modulation (FM) involves not only adjusting the decibel level but also adjusting the audio frequency.

[0092] Specifically, the hearing protection module has a preset high decibel safety threshold. When the sound intensity sensor detects that the external sound intensity exceeds 85dB, it reduces the audio gain to below the threshold in real time, and the gain adjustment delay does not exceed 10ms.

[0093] The 10ms low latency design solves the problem of communication stuttering in existing Bluetooth hearing aids, ensuring smooth normal communication;

[0094] Specifically, the noise reduction filtering module supports an environment adaptive mode and a manual selection mode. The environment adaptive mode can automatically identify three scenarios: sleep, communication, and outdoor, and switch the filtering parameters accordingly.

[0095] The manual selection mode supports users to customize the noise filtering intensity and provides scenario-based noise reduction modes, breaking through the limitations of traditional single noise reduction, adapting to different usage scenarios, and improving the user experience.

[0096] Specifically, the sound transmission device is a loudspeaker, a cochlear implant adapter, or bone conduction headphones.

[0097] The output audio generated by the training app can be switched between sound and vibration modes, with the vibration mode frequency matching the corresponding sound frequency.

[0098] The adaptability of the sound transmission device and the switching of audio modes have expanded the system's applicable population (people with mild to severe hearing impairment) and enhanced product compatibility.

[0099] Specifically, the cloud storage interaction module supports multi-device synchronization. When a user changes electronic devices or sound transmission devices, they can log in to their account to access historical adjustment parameters and hearing defect frequency spectrum data without having to repeat the training.

[0100] The multi-device synchronization function enables a "one-time training, lifelong reuse" usage model, reducing user costs and enhancing product competitiveness.

[0101] This means that the device stores the information in the phone's database to reduce internet access. When frequency modulation fails to achieve the desired hearing aid effect, the decibel level is adjusted for bidirectional control. If none of the above methods can improve recognition accuracy, the sound transmission device needs to be changed.

[0102] A training method for a hearing aid transmission device system with intelligent frequency modulation, based on the above-mentioned training method for an adjustable intelligent hearing aid transmission device system, includes the following steps:

[0103] S1: Construct a preset speech sample library, which contains standard speech materials in seven specific frequency ranges: 250Hz-300Hz, 500-600Hz, 1000-1100Hz, 1500Hz, 2000Hz, 4000Hz, and 8000Hz. At the same time, configure a natural dialogue speech acquisition task.

[0104] S2: Connect the sound transmission device to the electronic device via Bluetooth, start the training APP and complete the cloud storage account binding. The training APP will automatically load the preset voice sample library.

[0105] S3: Users repeat materials from a preset speech sample library through the microphone of the sound transmission device, and can also choose to participate in a natural dialogue task. The microphone simultaneously collects the repeated speech data and the natural dialogue speech data.

[0106] S4: The training APP optimizes the collected speech data by removing background noise and segmenting and classifying it. Through ASRModel, the optimized speech data and corresponding standard speech materials are converted into text and displayed on the electronic device interface for users to proofread the semantics and compare and analyze the differences in speech, intonation and speech rate.

[0107] S5: Extract speech segments with significant differences in semantic and speech features, convert them into frequency spectra using Fourier FFT transform, and calculate the sound intensity corresponding to different frequencies using the RMS algorithm to preliminarily identify frequencies of potential hearing defects.

[0108] S6: Using the T-Model multi-task statistical analysis unit, combined with the speech distortion factors caused by the user's dialect and education level, the initially identified defect frequencies are statistically analyzed in multiple dimensions to determine the frequency spectrum of the user's core hearing defects and the corresponding sound intensity requirements.

[0109] S7: The intelligent fine-tuning module adjusts the loudness gain of the corresponding frequency according to the frequency spectrum of the core defect. It supports users to manually select synchronous adjustment of the left and right ears or independent adjustment, and generates the initial optimized audio and plays it through the sound transmission device.

[0110] S8: Users provide feedback on audio clarity through silence, indistinct consonants, or loud repetition. The intelligent fine-tuning module iteratively adjusts the gain parameters according to the preset scale based on the feedback until a standard optimized audio acceptable to the user is generated. At the same time, the adjusted parameters, the standard optimized audio, and the non-standard audio during the adjustment process are uploaded and stored through the cloud storage interaction module.

[0111] S9: Continuously collect voice interaction data from daily use, repeat steps S4-S8, dynamically update the frequency spectrum of the user's hearing loss and adjustment parameters, and achieve dynamic matching between the system and the user's hearing condition.

[0112] Specifically, in step S3, the collection time for natural dialogue voice data is no less than 10 minutes. During the collection process, the training APP automatically records the dialogue scenario (home, outdoor, indoor communication) and stores the scenario information in association with the voice data. The collection time of natural dialogue and scenario labeling enrich the diversity and scenario adaptability of the training data and ensure the accuracy of dynamic updates.

[0113] In step S5, the frequency resolution accuracy of the Fourier FFT transform is no less than 1Hz, and the sound intensity calculation error of the RMS algorithm is no more than 3%. High-precision frequency resolution and sound intensity calculation improve the accuracy of defect frequency identification and provide reliable data support for subsequent adjustments.

[0114] In step S6, the statistical sample size of the T-Model multi-task statistical analysis unit is no less than 1,000 sets, including voice interaction data of different dialects and age groups, to ensure that the defect frequency recognition accuracy is no less than 95%. The large sample size statistics and multi-factor consideration avoid recognition bias caused by factors such as dialect and education level, and ensure that the system is adapted to different user groups.

[0115] In step S8, the preset scale of the gain parameter is 0.5dB / level, and the maximum gain adjustment range does not exceed 30dB to avoid secondary hearing damage to the user; the scaled gain limitation ensures hearing protection throughout the entire training process, solving the problem of lack of safety control in existing training methods.

[0116] In step S9, the dynamic update cycle can be set by the user, with a default cycle of 30 days. When the frequency change of the user's hearing loss is detected to exceed 10%, the instant update process is automatically triggered. The dynamic update cycle and instant triggering mechanism ensure the timeliness of parameter adjustment and avoid the cumbersome operation caused by frequent updates.

[0117] It also includes a hearing status assessment step: the training app generates a quarterly hearing status report based on historical defect frequency spectra, gain adjustment parameters, and user feedback data, highlighting hearing change trends and precautions for users. This hearing status assessment function overcomes the limitations of existing training methods that only focus on adjustment effects while neglecting hearing monitoring, providing users with full-cycle hearing management services and enhancing product added value.

[0118] The training software is equipped with a filtering device, which can directly filter out unnecessary audio, such as reducing external noise interference when sleeping. Of course, the filtering device can also be understood as filtering parameters, thereby achieving noise reduction.

[0119] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A hearing aid transmission device system with intelligent frequency modulation, characterized in that, include, An electronic device, which is equipped with a training APP and is configured with a data receiver, a data transmitter and a cloud storage interaction module; An acoustic device, comprising a microphone for collecting sound, a sound transmission device for playing audio, and a sound intensity sensor for real-time monitoring of external sound intensity. The training app includes a pre-set speech sample library, a speech processing module, a frequency analysis module, an intelligent fine-tuning module, a noise reduction and filtering module, and a hearing protection module. The preset voice sample library contains standard voice materials in at least seven specific frequency ranges; The sound transmission device and the electronic device are wirelessly connected via Bluetooth. The sound data collected by the microphone is transmitted to the training APP via the data receiving end. The voice processing module, frequency analysis module, and intelligent fine-tuning module process the data in sequence and generate optimized audio. The optimized audio is transmitted to the sound transmission device via the data transmitting end. The cloud storage interaction module realizes the cloud storage and retrieval of adjustment parameters, standard audio, and non-standard audio. The hearing protection module dynamically adjusts the audio gain threshold according to the sound intensity sensor data. The speech processing module integrates a speech-to-text module to convert standard speech materials and collected sound data into text and simultaneously compare semantic, speech, intonation, and speech rate differences. The frequency analysis module integrates a Fourier transform unit and a sound intensity response function algorithm unit, which is used to perform frequency spectrum conversion and sound intensity calculation on speech data with large semantic differences. The intelligent fine-tuning module has a built-in multi-task statistical analysis unit, which is used to statistically determine the frequency spectrum of hearing loss by combining environmental factors such as dialect and speech distortion, and supports independent or synchronous loudness gain fixed-point adjustment for the left and right ears.

2. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The preset voice sample library includes standard voice materials in specific frequency ranges: 250Hz-300Hz, 500-600Hz, 1000-1100Hz, 1500Hz, 2000Hz, 4000Hz, and 8000Hz.

3. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The multi-task statistical analysis unit achieves multi-task collaborative switching and efficient analysis by embedding specific expressions of hearing aid frequency recognition, sound intensity matching, and semantic correction tasks into a pre-trained model.

4. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The intelligent fine-tuning module supports users to trigger parameter adjustments through three feedback methods: silence, indistinct consonants, and loud repetition. The gain adjustment increases gradually according to a preset scale, with a single adjustment range not exceeding 5%.

5. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The hearing protection module is preset with a high decibel safety threshold. When the sound intensity sensor detects that the external sound intensity exceeds 85dB, it reduces the audio gain to below the threshold in real time, and the gain adjustment delay does not exceed 10ms.

6. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The noise reduction filtering module supports both an environment adaptive mode and a manual selection mode. The environment adaptive mode can automatically identify three scenarios: sleep, conversation, and outdoor, and switch filtering parameters accordingly. The manual selection mode allows users to customize the noise filtering intensity.

7. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The sound transmission device is a loudspeaker, a cochlear implant adapter, or bone conduction headphones. The output audio generated by the training app can be switched between sound and vibration modes, with the vibration mode frequency matching the corresponding sound frequency.

8. The adjustable intelligent hearing aid sound transmission device system according to claim 1, characterized in that, The cloud storage interaction module supports multi-device synchronization. When a user changes electronic devices or sound transmission devices, they can log in to their account to access historical adjustment parameters and hearing defect frequency spectrum data without having to repeat the training.

9. A training method for a hearing aid transmission device system with intelligent frequency modulation, characterized in that, The training method for the adjustable intelligent hearing aid transmission device system based on any one of claims 1-8 includes the following steps: S1: Construct a preset speech sample library, which contains standard speech materials in seven specific frequency ranges: 250Hz-300Hz, 500-600Hz, 1000-1100Hz, 1500Hz, 2000Hz, 4000Hz, and 8000Hz. At the same time, configure a natural dialogue speech acquisition task. S2: Connect the sound transmission device to the electronic device via Bluetooth, start the training APP and complete the cloud storage account binding. The training APP will automatically load the preset voice sample library. S3: Users repeat materials from a preset speech sample library through the microphone of the sound transmission device, and can also choose to participate in a natural dialogue task. The microphone simultaneously collects the repeated speech data and the natural dialogue speech data. S4: The training APP optimizes the collected speech data by removing background noise and segmenting and classifying it. Through ASRModel, the optimized speech data and corresponding standard speech materials are converted into text and displayed on the electronic device interface for users to proofread the semantics and compare and analyze the differences in speech, intonation and speech rate. S5: Extract speech segments with significant differences in semantic and speech features, convert them into frequency spectra using Fourier FFT transform, and calculate the sound intensity corresponding to different frequencies using the RMS algorithm to preliminarily identify frequencies of potential hearing defects. S6: Using the T-Model multi-task statistical analysis unit, combined with the speech distortion factors caused by the user's dialect and education level, the initially identified defect frequencies are statistically analyzed in multiple dimensions to determine the frequency spectrum of the user's core hearing defects and the corresponding sound intensity requirements. S7: The intelligent fine-tuning module adjusts the loudness gain of the corresponding frequency according to the frequency spectrum of the core defect. It supports users to manually select synchronous adjustment of the left and right ears or independent adjustment, and generates the initial optimized audio and plays it through the sound transmission device. S8: Users provide feedback on audio clarity through silence, indistinct consonants, or loud repetition. The intelligent fine-tuning module iteratively adjusts the gain parameters according to the preset scale based on the feedback until a standard optimized audio acceptable to the user is generated. At the same time, the adjusted parameters, the standard optimized audio, and the non-standard audio during the adjustment process are uploaded and stored through the cloud storage interaction module. S9: Continuously collect voice interaction data from daily use, repeat steps S4-S8, dynamically update the frequency spectrum of the user's hearing loss and adjustment parameters, and achieve dynamic matching between the system and the user's hearing condition.

10. The training method for the hearing aid transmission device system with intelligent frequency modulation according to claim 1, characterized in that, In step S3, the collection time for natural dialogue voice data is no less than 10 minutes. During the collection process, the APP is trained to automatically record the dialogue scene and associate the scene information with the voice data for storage. In step S5, the frequency resolution of the Fourier FFT transform is no less than 1Hz, and the sound intensity calculation error of the RMS algorithm does not exceed 3%. In step S6, the statistical sample size of the T-Model multi-task statistical analysis unit is no less than 1000 sets, including voice interaction data of different dialects and age groups, to ensure that the defect frequency recognition accuracy is no less than 95%; In step S8, the preset scale of the gain parameter is 0.5dB / level, and the maximum gain adjustment range does not exceed 30dB to avoid secondary hearing damage to the user. In step S9, the dynamic update cycle can be set by the user, with a default cycle of 30 days. When the frequency change of the user's hearing loss is detected to exceed 10%, the instant update process is automatically triggered. It also includes a hearing status assessment step: the training app generates a quarterly hearing status report based on historical defect frequency spectrum, gain adjustment parameters and user feedback data, prompting users to note hearing change trends and precautions.

Images