A vocal music practice assisting system based on vocal cord fatigue analysis

Abstract

This application provides a vocal practice assistance system based on vocal cord fatigue analysis, comprising: a surface-mounted electrode for attaching to a detection target to acquire electromyographic (EMG) signals; an EMG signal acquisition module for acquiring the EMG signals transmitted by the surface-mounted electrode and performing analog-to-digital conversion; a data transmission and reception module for sending the EMG signals acquired by the EMG signal acquisition module to a host computer and receiving relevant information; a preprocessing module for preprocessing the EMG signals transmitted by the data transmission and reception module to reduce noise; a signal separation module for separating the preprocessed EMG signals; a feature extraction module for acquiring the time-domain and frequency-domain features of the EMG signals; and a fatigue analysis module for linearly fitting the time-domain and frequency-domain feature curves and analyzing the vocal cord fatigue trend through the slope of the linear fitting equation. This system can monitor and analyze the degree of vocal cord fatigue in real time and provide a convenient and reliable vocal practice assistance device for vocal professionals.

Description

Technical Field

[0001] This application relates to the field of vocal cord monitoring technology, specifically to a vocal practice assistance system based on vocal cord fatigue analysis. Background Technology

[0002] In vocal music teaching and practice, music educators, practitioners, and vocal students often experience vocal cord fatigue, hoarseness, and sore throat due to excessive practice time or improper vocal techniques, which seriously affects their normal study, life, and work. Because different people have varying tolerances to vocal cord fatigue and can handle different amounts of practice, vocal cord fatigue is often difficult to detect unless discomfort is felt. This makes it hard to control the intensity and duration of training, easily leading to overuse of the vocal cords, which can be detrimental to vocal cord health in the long run.

[0003] The human vocal cords are formed by the vocal ligaments, located behind the anterior horn of the thyroid cartilage and between the arytenoid vocal cords and the vocal cord processes of the glottis, along with the vocal cord muscles and their surface mucosa. During phonation, the vocal cords tighten, the glottis narrows, and even almost closes. The airflow from the trachea and lungs continuously impacts the vocal cords, causing vibration and producing sound. Under the coordinated action of the laryngeal muscles, the glottis is controlled rhythmically. Electromyography (EMG) signals are the sum of electrical signals from individual muscle fibers. EMG signals can reflect the functional state of muscles and can effectively show changes in electrical displacement during muscle movement. Numerous studies have demonstrated a strong correlation between muscle fatigue and relevant EMG indicators. After acquisition, filtering, and separation processing, vocal cord EMG signals can provide a relatively intuitive reflection of the degree of vocal cord fatigue. However, weak vocal cord electromyography (EMG) signals are often accompanied by strong interference signals, including attenuation caused by impedance, interference from transmission line media, and random noise. While amplifying the vocal cord EMG signal, interference signals are inevitably amplified as well. Therefore, specialized equipment is needed for EMG signal acquisition and processing. There are generally two methods for EMG signal acquisition: needle electrodes and surface electrodes. The former is inserted into the muscle to acquire EMG signals, while the latter is placed on the skin to obtain surface EMG signals. Compared to needle electrodes, surface electrodes, due to their indirect and non-specific acquisition, have a larger detection range and lower spatial resolution. However, the acquired surface EMG signals are more susceptible to interference, resulting in a lower signal-to-noise ratio. Nevertheless, measuring EMG signals using surface electrodes is non-invasive, does not require physician intervention, and is the preferred method for acquiring EMG signals in non-clinical settings.

[0004] In summary, there is a great demand and practicality for portable, low-power, long-term wearable vocal practice aids that can monitor and personalize vocal cord fatigue levels in real time. It is necessary to apply the relevant technological research to practice and provide convenient and reliable vocal practice aids for the vocal cord health of music educators, practitioners, and vocal students. Summary of the Invention

[0005] To address the existing problems of real-time monitoring and personalized analysis of vocal cord fatigue, this application provides a vocal practice assistance system based on vocal cord fatigue analysis. This system can monitor and analyze the degree of vocal cord fatigue in real time and is easy to wear, providing a convenient and reliable vocal practice assistance device for the vocal cord health of vocal professionals.

[0006] The technical solution adopted by this application to solve its technical problem is: a vocal practice auxiliary system based on vocal cord fatigue analysis, including surface-adhesive electrodes, an electromyography signal acquisition module, a data transmission and reception module, a preprocessing module, a signal separation module, a feature extraction module, and a fatigue analysis module.

[0007] The surface-adhesive electrode is used to attach to the detection target to acquire electromyographic signals.

[0008] The electromyography (EMG) signal acquisition module is used to acquire EMG signals transmitted by surface-attached electrodes and perform analog-to-digital conversion.

[0009] The data sending and receiving module sends the electromyographic signals acquired by the electromyographic signal acquisition module to the host computer and receives relevant information.

[0010] The preprocessing module preprocesses the electromyographic signals transmitted by the data sending and receiving modules to reduce noise.

[0011] The signal separation module separates the preprocessed electromyographic signals.

[0012] The feature extraction module is used to obtain the time-domain and frequency-domain features of electromyographic signals.

[0013] The fatigue analysis module performs linear fitting on the time-domain and frequency-domain characteristic curves and analyzes the fatigue change trend of the vocal cords by using the slope of the linear fitting equation.

[0014] In one specific implementation, the electromyography (EMG) signal acquisition module includes an analog front-end for acquiring EMG signals transmitted from surface-attached electrodes and performing analog-to-digital conversion, a microcontroller for realizing related data interaction and control, and a power supply module.

[0015] In one specific implementation, the data sending and receiving module, under the control of the microcontroller, sends the electromyographic signals acquired by the electromyographic signal acquisition module to the host computer, and parses the commands received from the host computer through the microcontroller.

[0016] In one specific implementation, the preprocessing module visualizes the electromyographic signals and performs bandpass filtering, notch filtering, and wavelet denoising.

[0017] In one specific implementation, the signal separation module uses independent component analysis technology to de-aliased separate electromyographic signals.

[0018] In one specific implementation, the time-domain features include the root mean square value and integral electromyographic value of the electromyographic signal in the time domain, and the frequency-domain features include the average power frequency and median frequency of the electromyographic signal in the frequency domain.

[0019] In one specific implementation, the fatigue analysis module performs windowing calculations on the time-domain and frequency-domain characteristics of the electromyographic signal to meet signal processing requirements.

[0020] In one specific implementation, before the electromyographic signal transmitted by the surface-attached electrode is input to the analog front end, an electrostatic protection element is used to protect the input transient, prevent electrostatic breakdown and interference with the electromyographic signal acquisition module, and prevent leakage current from causing injury.

[0021] In one specific implementation, the vocal practice assistance system based on vocal cord fatigue analysis also includes a wearable neck collar with a through hole, in which the surface-adhesive electrode is disposed and forms a protrusion to fit closely to the monitoring target.

[0022] In one specific implementation, the surface-mount electrode is made of a flexible conductive material.

[0023] The advantages of this application are:

[0024] 1. The vocal practice assistance system based on vocal cord fatigue analysis provides a complete vocal cord protection assistance solution. By quantitatively analyzing the degree of vocal cord fatigue, it can provide timely warnings before the user's vocal cords become uncomfortable due to overuse caused by difficulty in controlling the intensity and duration of training, so as to achieve the effect of scientifically and effectively protecting the vocal cords.

[0025] 2. The vocal practice assistance system based on vocal cord fatigue analysis uses flexible material surface-adhesive electrodes in conjunction with a lightweight and wearable neck collar to non-invasively collect electromyographic signals. The wearable components are lightweight and convenient, and the electromyographic signal acquisition module and data transmission and reception module are small in size and weight, making them portable. It can communicate wirelessly with a host computer, and movement within a certain range does not affect data transmission. At the same time, the system has low power consumption, high information transmission efficiency and stability, is easy to operate, and has multiple functions to meet user needs.

[0026] 3. The vocal practice assistance system based on vocal cord fatigue analysis collects, filters, and separates vocal cord electromyography signals, and then performs fatigue analysis through feature extraction. This can reflect the degree of vocal cord fatigue in a relatively intuitive and real-time manner. It analyzes multiple features in the time and frequency domains to comprehensively evaluate the degree of fatigue with high accuracy.

[0027] 4. The vocal practice assistance system based on vocal cord fatigue analysis can personalize the fatigue benchmark according to the different tolerance levels of different people to vocal cord fatigue. In addition, the hardware system is composed of a low-cost analog front-end and microcontroller with high integration and excellent performance, which significantly reduces the overall cost of the system while ensuring high-precision and low-noise acquisition of vocal cord electromyography signals. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of a vocal practice assistance system based on vocal cord fatigue analysis according to this application.

[0029] Figure 2 This is a schematic diagram of a wearable neck collar and surface-fitted electrodes for a vocal practice assist system based on vocal cord fatigue analysis, as described in this application.

[0030] Figure 3 This is a schematic diagram of the electromyography signal acquisition module and data transmission and reception module of a vocal practice auxiliary system based on vocal cord fatigue analysis according to this application.

[0031] Explanation of key figure labels:

[0032] 1-Surface-mounted electrode; 2-Through hole; 3-Wearable neck collar. Detailed Implementation

[0033] This application provides a vocal practice assistance system based on vocal cord fatigue analysis, which solves the existing problems of real-time monitoring and personalized analysis of vocal cord fatigue. The overall idea is as follows:

[0034] Please see Figure 1This application provides a vocal practice assistance system based on vocal cord fatigue analysis, including a surface-adhesive electrode 1, an electromyography (EMG) signal acquisition module, a data transmission and reception module, a preprocessing module, a signal separation module, a feature extraction module, and a fatigue analysis module. The surface-adhesive electrode 1 is used to attach to a detection target to acquire EMG signals; the EMG signal acquisition module is used to acquire the EMG signals transmitted by the surface-adhesive electrode 1 and perform analog-to-digital conversion; the data transmission and reception module sends the EMG signals acquired by the EMG signal acquisition module to a host computer and receives relevant information; the preprocessing module preprocesses the EMG signals transmitted by the data transmission and reception module to reduce noise; the signal separation module separates the preprocessed EMG signals; the feature extraction module acquires the time-domain and frequency-domain features of the EMG signals; and the fatigue analysis module performs linear fitting on the time-domain and frequency-domain feature curves and analyzes the vocal cord fatigue change trend by using the slope of the linear fitting equation. This system provides a complete vocal cord protection assistance program, quantitatively analyzes the degree of vocal cord fatigue, and provides timely warnings before users experience discomfort due to overuse of their vocal cords caused by difficulty in controlling the intensity and duration of training, thus achieving a scientific and effective effect in protecting the vocal cords.

[0035] Preferably, in this example, the vocal practice assistance system based on vocal cord fatigue analysis also includes a wearable neck collar 3, which has a through hole 2. A surface-adhesive electrode 1 is disposed in the through hole 2 and forms a protrusion to ensure close adhesion to the monitoring target. The surface-adhesive electrode 1 is made of a flexible conductive material. The lightweight wearable neck collar 3 uses a ring-shaped wearing method, which can fix the surface-adhesive electrode 1 inside the lightweight wearable neck collar 3, making the electrode fixation more secure and comfortable, and ensuring effective data acquisition time. Please refer to... Figure 2 The surface-mount electrode 1, made of 8-channel flexible conductive material, can have multiple through holes 2 for electrode positions at the throat in the support layer of the lightweight wearable neck collar 3. The flexible conductive material can be filled into the through holes 2 by in-situ casting, molding, or bonding to form protrusions. This allows the surface-mount electrode 1 made of flexible conductive material to not only form a tight attachment with the detection target, but also to have good adaptability to the detection target due to its flexibility. At the same time, it avoids problems such as electrode corrosion, oxidation, and signal deviation caused by electrode deterioration due to direct contact between the detection target and the metal electrode and the skin, which are common with metal electrodes. This helps to extend the life of the device.

[0036] Specifically, please refer to Figure 3The electromyography (EMG) signal acquisition module includes an analog front-end for acquiring EMG signals transmitted from surface-attached electrodes 1 and performing analog-to-digital conversion, a microcontroller for related data interaction and control, and a power supply module. Under the control of the microcontroller, the data transmission and reception module sends the EMG signals acquired by the EMG signal acquisition module to the host computer and parses the commands received from the host computer via the microcontroller. The EMG signal acquisition module aims to realize a simple, efficient, and versatile wireless non-invasive EMG signal acquisition device, improve information transmission efficiency, and achieve high system stability. The analog front-end can specifically use an ADS1299, and the microcontroller can specifically use a 32-bit PIC microcontroller. Before the EMG signals transmitted from surface-attached electrodes 1 are input to the analog front-end, electrostatic discharge (ESD) protection components are used for input transient protection. Specifically, before the EMG signals are input to the ADS1299 analog front-end, a TPD4E1B06DCKR high-speed ESD protection device can be used for input transient protection to prevent electrostatic discharge from the human body from interfering with the normal operation of the EMG signal acquisition module and to prevent system leakage from harming the human body.

[0037] In this example, the ADS1299 analog front-end, with its high level of integration and excellent performance, enables the creation of scalable medical instrument systems with significantly reduced size, power consumption, and overall cost, reducing the overall volume and weight of the EMG signal data acquisition section. The ADS1299 analog front-end features high performance with 24-bit precision, low power consumption, and low noise, along with a sampling rate of 250 SPS to 16 kSPS and a 24x amplification gain. It also boasts extremely high input impedance (1 TΩ), extremely low input bias current (300 pA), extremely low input reference noise (1 μVpp), a high common-mode rejection ratio (-110 dB), high resolution (0.1 μV), and extremely low operating power consumption (39 mW). Furthermore, it incorporates a built-in filtering algorithm to effectively remove environmental noise interference and filter out mains interference. Its eight signal input channels are differential, providing 24 bits of data in two's complement format. In this example, the microcontroller for the control section uses a 32-bit PIC microcontroller to perform analog-to-digital conversion on the ADS1299 analog front-end and control data interaction with related communication modules. The PIC32 microcontroller can initialize and configure the ADS1299 analog front-end, including configuring the sampling rate and amplifier gain, and selecting the data reception mode and reference potential. The PIC32 microcontroller firmware development includes data definition, definition and declaration of various function functions, etc. The main function includes writing the function call interface, parsing SDK commands, and defining the EMG signal data packet format. The loop function includes registering the various registers of the ADS1299 analog front-end, checking serial port data, detecting EMG signal data, and controlling related communication modules. The default data transmission rate is 250Hz, and the total system input voltage is less than 6V. An MCP1754T-3302E / OT (3.3V regulator) can be used to provide operating voltage for each module. A TLV70025QDDCRQ1 (+2.5V regulator) and TPS72325DBVR (-2.5V regulator) provide reference voltage for the ADS1299 analog front-end. An LM2664M6 / NOPB voltage converter provides the required reverse voltage for the system. The electromyography (EMG) signal acquisition module low-pass filters and amplifies the acquired analog EMG signals, then converts them into digital signals via analog-to-digital conversion.

[0038] In this example, the data transmission and reception module uses Wi-Fi technology to achieve wireless data transmission. Specifically, it uses the ESP8266 module to implement Wi-Fi communication. The ESP8266 module requires very little external circuitry to achieve the corresponding functions, reducing the space occupied on the PCB board. Designed specifically for mobile devices, wearable electronics, and IoT applications, the ESP8266 module employs several proprietary technologies to achieve low power consumption. Its power supply method aligns with the design principles of wearable devices and wireless communication, enabling a data acquisition duration of at least 3 hours. The ESP8266 module can perform data transmission using TCP, UDP, or UDPx3 protocols. Simultaneously, the Wi-Fi-enabled data transmission and reception module incorporates an MCP1754T-3302E / OT (3.3V regulator) to provide the required 3.3V operating voltage for the ESP8266 module. The data transmission and reception module receives SDK commands from the host computer and parses them using a PIC32 microcontroller. Simultaneously, it packages and sends electromyographic signals (e.g., digital signals) acquired by the electromyographic signal acquisition module to a PC or mobile device. The PIC32 microcontroller controls the Wi-Fi transmission and reception of the data. By employing SPI full-duplex communication and robust connector integration, it utilizes two PCBs smaller than 3cm × 3cm × 3cm. The PCB layout uses a star grounding configuration, minimizing PCB area while improving noise immunity. The main chips on the first PCB may include an ADS1299 analog front-end, a TLV70025QDDCRQ1 (+2.5V regulator), and a TPS72325DBVR (-2.5V regulator) to provide reference voltage for the ADS1299 analog front-end. An LM2664M6 / NOPB voltage converter provides the required reverse voltage for the system, and six TPD4E1B06DCKR high-speed ESD protection devices for input transient protection. The main chips on the second PCB may include a PIC32MX250F128B-I / SS microcontroller for control functions, an MCP1754T-3302E / OT (3.3V regulator) to provide operating voltage for related modules, and an ESP-12E module with an ESP8266 to implement Wi-Fi transmission. This allows the data transmission and reception modules to transmit data via Wi-Fi to a host computer program on a mobile phone or PC for signal processing.

[0039] In this example, the preprocessing module visualizes the electromyography (EMG) signal data. For instance, it plots corresponding curves or waveforms based on the EMG signal data transmitted from the lower-level computer, such as the EMG signal acquisition module, and performs bandpass filtering, notch filtering, and wavelet denoising. Specifically, the preprocessing module visualizes the EMG signal data using the upper-level computer, then performs bandpass filtering, notch filtering, and wavelet denoising. Wavelet spatial domain filtering is used to decompose the surface EMG signal using wavelets to obtain high-frequency and low-frequency coefficients. The correlation of wavelet coefficients at corresponding points on different scales is then used to achieve filtering. Finally, the original signal is reconstructed using a wavelet transform reconstruction algorithm to obtain the denoised signal.

[0040] The signal separation module separates the preprocessed electromyographic (EMG) signals. Due to the complexity of EMG signal aliasing methods, independent component analysis (ICA) is used in blind signal separation correlation techniques to de-alias the EMG signals. ICA has wide applications in fields such as communication, array signal processing, biomedical signal processing, speech signal processing, signal analysis, and process control, including image denoising and feature extraction.

[0041] After separating the vocal cord electromyography (EMG) signal, it proceeds to the feature extraction module. Time-domain features include the root mean square (RMS) and integrated EMG values ​​of the EMG signal in the time domain, while frequency-domain features include the average power frequency and median frequency of the EMG signal in the frequency domain. The feature extraction module extracts the RMS and integrated EMG values ​​of the vocal cord EMG signal in the time domain, as well as the average power frequency and median frequency in the frequency domain. The RMS value of the vocal cord EMG signal represents the overall signal energy level, the integrated EMG value reflects the change in signal intensity over time, and the average power frequency and median frequency reflect the frequency domain changes of the signal.

[0042] The fatigue analysis module performs windowing calculations on the extracted time-domain and frequency-domain features to meet signal processing requirements and performs linear fitting on the obtained feature curves. For example, the least squares method can be used for linear fitting, and the slope of the linear fitting equation is used to analyze fatigue trend changes. The windowing method is used to calculate the features before and after signal separation, and linear fitting is performed on the feature curves before and after separation to obtain the linear fitting equations for each feature curve. The slopes of the linear fitting equations for each feature curve of the vocal cord electromyography (EMG) signal are compared, and the vocal cord fatigue state of the test subject is analyzed based on the trend changes of the time-domain and frequency-domain indicators. The root mean square value and integral EMG value of the surface EMG signal in the time domain increase with the increase of fatigue level, while the median frequency and average power frequency in the frequency domain decrease with the increase of fatigue level. The degree of vocal cord fatigue can be monitored in real time. Simultaneously, through a certain number of experiments, a general fatigue benchmark can be obtained. By comparing the obtained analysis values ​​with the general fatigue benchmark values, it can be determined whether the fatigue level has exceeded the threshold. If it exceeds the threshold, suggestions such as stopping practice can be given. In addition, fatigue benchmarks can be customized according to individual user differences; rest should be initiated when the fatigue level exceeds this value.

[0043] In summary, this application provides a vocal practice assistance system based on vocal cord fatigue analysis, which can monitor and analyze the degree of vocal cord fatigue in real time and is easy to wear, providing a convenient and reliable vocal practice assistance device for the vocal cord health of vocal professionals.

[0044] Finally, it should be noted that the above embodiments are merely examples for clearly illustrating this application and are not intended to limit the implementation. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this application.

Claims

1. A vocal practice assistance system based on vocal cord fatigue analysis, used to monitor and analyze the vocal cord fatigue state of a user in real time during vocal practice, characterized in that, include: A wearable neck collar with at least one pair of electrode holes on its inner surface corresponding to the body surface projection position of the user's vocal cords. The surface-adhesive electrode, made of a flexible conductive material, is disposed in the electrode through-hole and protrudes toward the skin contact side, so as to maintain close contact with the skin of the throat when the user wears the collar to make a sound, so as to collect mixed electromyographic signals including vocal cord electromyographic signals and interference from other neck muscles. An electromyography (EMG) signal acquisition module, which is electrically connected to the surface-attached electrodes, is used to acquire the mixed EMG signals and perform analog-to-digital conversion. The EMG signal acquisition module includes an analog front-end, a microcontroller, and a power supply module. The analog front-end adopts a high-precision, low-noise biopotential measurement analog front-end. The data transmission and reception module is used to transmit the digital electromyography signals acquired by the electromyography signal acquisition module to the host computer and receive information from the host computer. The preprocessing module, located in the host computer, is used to perform bandpass filtering, notch filtering, and wavelet denoising preprocessing on the received digital electromyography signals. The signal separation module, located in the host computer, is pre-configured with a hybrid model based on the anatomical location and electrophysiological characteristics of the vocal cords, sternocleidomastoid muscle, and trapezius muscle. Using independent component analysis technology, the module performs directional blind source separation on the pre-processed hybrid electromyographic signals according to the hybrid model, so as to separate the specific vocal cord electromyographic signals corresponding to vocal cord vibration from the hybrid electromyographic signals and suppress interference signals from the sternocleidomastoid muscle and trapezius muscle. The feature extraction module, located in the host computer, is used to extract time-domain features and frequency-domain features from the separated specific vocal cord electromyography signal. The time-domain features include root mean square value and integrated electromyography value, and the frequency-domain features include average power frequency and median frequency. The fatigue analysis module, located in the host computer, is used to perform windowing calculations and linear fitting on the extracted time-domain and frequency-domain features in the vocal practice time series. It analyzes the fatigue change trend of the vocal cord muscles under continuous vocalization by analyzing the slope of the linear fitting equation, and generates warning or suggestion information for vocal practice when the trend indicates that the fatigue level exceeds a preset threshold. A personalized calibration module, located in the host computer, is used to guide the user through a series of standard vocal exercises during initial use. Based on the specific vocal cord electromyography signals collected during this period and the user's subjective fatigue feedback, a personalized vocal cord fatigue baseline threshold is established for the user. The preset threshold on which the warning or suggestion information generated by the fatigue analysis module is based is the personalized vocal cord fatigue baseline threshold.

2. The vocal practice assistance system based on vocal cord fatigue analysis as described in claim 1, characterized in that, Under the control of the microcontroller, the data sending and receiving module sends the electromyographic signals acquired by the electromyographic signal acquisition module to the host computer, and parses the commands received from the host computer through the microcontroller.

3. The vocal practice assistance system based on vocal cord fatigue analysis as described in claim 2, characterized in that, The preprocessing module visualizes the electromyographic signals and performs bandpass filtering, notch filtering, and wavelet denoising.

4. A vocal practice assistance system based on vocal cord fatigue analysis as described in any one of claims 1-3, characterized in that, The signal separation module uses independent component analysis technology to de-alias and separate electromyographic signals.

5. The vocal practice assistance system based on vocal cord fatigue analysis as described in claim 4, characterized in that, The time-domain features include the root mean square value and integral electromyographic value in the time domain of the electromyographic signal, and the frequency-domain features include the average power frequency and median frequency in the frequency domain of the electromyographic signal.

6. The vocal practice assistance system based on vocal cord fatigue analysis as described in claim 5, characterized in that, The fatigue analysis module performs windowing calculations on the time-domain and frequency-domain features of electromyographic signals to meet signal processing requirements.

7. The vocal practice assistance system based on vocal cord fatigue analysis as described in claim 6, characterized in that, Before the electromyographic signals transmitted by the surface-attached electrodes are input to the analog front end, an electrostatic protection element is used to protect the input transients, prevent electrostatic breakdown and interference with the electromyographic signal acquisition module, and prevent leakage current from causing injury.

Images