Vocal music pronunciation training system and training auxiliary method thereof

Abstract

The invention provides a training system for vocal music sounding and a training auxiliary method thereof, and relates to the technical field of music education, and the training system comprises a data acquisition module used for data acquisition, a data processing module used for data processing, a data analysis module used for data analysis, and a data evaluation module used for data evaluation. According to the method, various data of trainees are acquired, so that the change of the trainees is judged according to comparison between training data and historical data, then environmental factors are used as variables to be input into the system, and an evaluation result is corrected, namely, the influence of the environmental factors is taken out to generate a more accurate evaluation report; and then advantages and disadvantages of the trainees are judged through evaluation results, auxiliary guidance is provided for training of the trainees, and the training effect of the trainees can be improved in a more targeted manner.

Description

Technical Field

[0001] The present invention relates to the technical field of music education, and in particular to a vocal phonation training system and a practice auxiliary method thereof. Background Art

[0002] The vocal training system and its auxiliary training methods belong to the field of music education, specifically a branch of vocal education, which involves the principles of human vocalization, breath control, scale practice, vowel practice, true and false voice practice, and clear pronunciation. Vocal education requires students to not only master singing skills, but also to continuously improve their singing skills through persistent practice. Effective vocal training is essential to improving students' vocal level and ability to express songs;

[0003] Generally speaking, students' voice training is mostly guided by a tutor one-on-one. The tutor relies on the tutor's subjective will to judge some problems that students may encounter in the voice production process, rather than comparing with the students' historical training results to find the points that need targeted training. Therefore, the training effect on students is not significantly improved. At the same time, it is impossible to provide targeted guidance to multiple students at the same time. Therefore, the present invention proposes a vocal training system and a practice auxiliary method thereof. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to provide a vocal training system and a training auxiliary method thereof.

[0005] In order to achieve the above-mentioned object, the present invention adopts the following technical scheme: a vocal voicing training system, comprising a data acquisition module for data acquisition, a data processing module for processing data, a data analysis module for analyzing data, and a data evaluation module for data evaluation;

[0006] The data acquisition module includes a time recording unit for recording training time, a voice acquisition unit for recording training sound, a heart rate detection unit for recording the heart rate of the trainee, an environmental data acquisition unit for collecting surrounding environment data, a data sorting unit for data sorting and numbering, and a data uploading unit for uploading data;

[0007] The data processing module includes a data receiving unit for data reception, a data classification unit for data classification, a speech recognition unit for converting speech data into text data, an AI learning unit for converting speech data into fundamental frequency data, and a data interaction unit for data interaction;

[0008] The data analysis module includes a data interaction unit for data interaction, a data and processing unit for data classification, a data matching unit for data matching, a matching result conversion unit for matching result conversion, a data analysis unit for data analysis, and an analysis result uploading unit for uploading analysis results;

[0009] The data evaluation module includes a data retrieval unit for data retrieval, a data receiving unit for data reception, an algorithm unit for adjusting evaluation results, a training evaluation production unit, and a storage unit for data storage.

[0010] Preferably, the time recording unit includes a timer, the voice collection unit is a recorder, the heart rate detection unit is a heart rate monitoring ring when worn, and the environmental data collection unit includes a temperature sensor, a humidity sensor and a dust particle detection sensor. The data collection steps are as follows: after the student starts voice training, the timer records the training time and the duration of high and low notes during the training process, the recorder is used to record the sound generated during the training process, the heart rate detection unit monitors the student's heart rate changes in real time and uploads them in real time, and the environmental data collection unit is used to collect surrounding temperature data, humidity data and dust content data in the air.

[0011] Preferably, the speech recognition unit is used to convert the sound data of the data acquisition unit module into text data. The converted text data does not need to be preprocessed and is directly uploaded to the data analysis module through the data interaction unit. The AI ​​learning unit is used to convert the speech data into fundamental frequency data. The specific steps of data processing are as follows: the data receiving unit receives the data collected by the data acquisition module, and then the data classification unit classifies the data. First, the data is classified according to the data format, and the data of different formats are classified to facilitate the processing of formats of the same format type. The speech recognition unit converts the text data, and the AI ​​learning unit converts the speech data into fundamental frequency data by the short-time average amplitude difference method. The processed data is uploaded to the data analysis module through the data interaction unit.

[0012] Preferably, in the data analysis module, the processing of the data preprocessing unit includes data classification and cleaning of duplicate data, and the specific steps are as follows: the data interaction unit receives the processed data uploaded from the data processing module, and then deletes the duplicate data and classifies the data with the data preprocessing unit. The data preprocessing unit clears the duplicate data, and then classifies the data again. The classification is performed based on the format type of the data and the time of upload to ensure that the data uploaded by each student is aggregated together to facilitate processing of data for different students. Afterwards, the data matching unit matches the student historical data stored in the storage unit, and the matching result conversion unit converts the matching result into text. The data analysis unit is used to perform a preliminary analysis based on the converted matching result, and then upload the analysis result to the data evaluation for final evaluation.

[0013] Preferably, the storage unit in the data evaluation module is used to store the trainees' historical training data, and the algorithm processing unit is used to input variable factors into the preliminary analysis results uploaded by the data analysis module to improve the preliminary analysis results. The specific steps are as follows: the data receiving unit receives the preliminary analysis results from the data analysis module, and then the data retrieval unit retrieves the relevant environmental data, and inputs the environmental data as variables into the algorithm processing unit, and the algorithm processing unit corrects the analysis results, and completes the evaluation of this training through the corrected analysis results.

[0014] A method for assisting vocalization training, the specific steps of the method for assisting vocalization training are as follows:

[0015] S1. Data collection: The trainee puts on the heart rate monitoring ring and starts training. During the training, the timer records the training time and the duration of the high and low notes during the training. The recorder is used to record the sounds generated during the training. The heart rate detection unit monitors the changes in the trainee's heart rate in real time and uploads it in real time. The environmental data collection unit is used to collect the surrounding temperature data, humidity data and dust content data in the air, and then upload the data;

[0016] S2. Data processing: The data receiving unit receives the data collected by the data collection module, and then the data classification unit classifies the data. The classification standard is classified according to the format type of the data. After the classification is completed, the speech recognition unit converts the text data. The AI ​​learning unit converts the speech data into fundamental frequency data through the short-time average amplitude difference method. The processed data is uploaded to the data analysis module through the data interaction unit;

[0017] S3, data analysis: the data interaction unit receives the processed data uploaded from the data processing module, and then deletes duplicate data and classifies the data with the data preprocessing unit. The classification is based on the format type of the data and the upload time. Then, the data matching unit matches the student historical data stored in the storage unit. Specifically, the basic data of which student is uploaded in this time period is determined according to the upload time. Then, the historical data related to the student is retrieved from the storage unit according to the student information, and finally matched with the real-time data. The matching result conversion unit converts the matching result into text. The data analysis unit is used to perform preliminary analysis based on the converted matching result, and then upload the analysis result to the data evaluation for final evaluation.

[0018] S4, training evaluation: the data receiving unit receives the preliminary analysis results from the data analysis module, and then the data retrieval unit retrieves the relevant environmental data, and inputs the environmental data into the algorithm processing unit as a variable, and the algorithm processing unit corrects the analysis results, and completes the evaluation of this training through the corrected analysis results;

[0019] S5. Auxiliary guidance training: Based on the training evaluation data obtained in step S4, targeted training guidance is provided to the students according to the changes in the students' vocal training time, the changes in the vocalization time of the high and low notes, the changes in the heart rate, and the accuracy of the speech recognition text.

[0020] Preferably, in step S2, the specific process of using the AI ​​learning unit to convert data is as follows:

[0021] S2.1, select the voice data uploaded at a certain time, and convert the continuous signal into a discrete signal through A / D conversion;

[0022] S2.2. For the speech signal sample x(n+i) in each frame, the average value of the absolute value of the sample calculated by the short-time average amplitude function is as follows: N is the frame length, n is the start time index of the current frame, and within each frame, the amplitude difference of adjacent sample points is calculated. Assuming that the amplitudes of adjacent sample points are |x(n+i)| and |x(n+i+1)|, the amplitude difference is: D(n+i)=|x(n+i+1)|-|x(n+i)|. The short-time mean amplitude difference function (SMDF) is defined as the cumulative average of these differences, and the formula is as follows:

[0023] S2.3, estimate the fundamental frequency by finding the time interval corresponding to the valley value of the SMDF function;

[0024] S2.4. Perform outlier processing, remove excessively high and low fundamental frequency values, smooth the fundamental frequency sequence, and obtain the fundamental frequency value of the student's sound during the training process, so as to identify the frequency and changes of the student's high and low tones, thereby judging the student's training effect.

[0025] Preferably, in step S3, the data matching mainly includes training duration data matching, speech-to-text data matching, high-note and low-note utterance duration data matching and heart rate change data matching. The training duration data matching mainly determines the change in the trainees' vital capacity, the speech-to-text data matching mainly determines the clarity of the trainees' pronunciation, the high-note and low-note utterance duration matching mainly determines the high-low-note conversion site of the trainees, and the heart rate change data matching determines the stage effect of the trainees.

[0026] Preferably, in step S4, the correction of the analysis result specifically includes the following steps:

[0027] S4.1. Obtaining preliminary analysis results, i.e., evaluation results of preliminary analysis completed by the data analysis module through the data analysis unit;

[0028] S4.2, the data retrieval module retrieves the environmental data collected in step S1 and inputs the environmental data into the system. The environmental data includes the temperature, humidity and air index data during the training process:

[0029] S4.3. Use machine learning algorithms to compare environmental data with historical data to identify data outliers. Improve the evaluation results by replacing or deleting outliers. Then fit the evaluation result data to generate new analysis result data. Repeat the above steps based on the outliers to generate a variety of different analysis results. Divide the analysis result data into multiple subsets to repeat training and testing. Finally, obtain new evaluation results combined with abnormal environmental data.

[0030] Compared with the prior art, the advantages and positive effects of the present invention are:

[0031] 1. In the present invention, by collecting the training data of the students, and then classifying and analyzing the collected data, the instructor only needs to evaluate the results of the students' initial training and propose targeted training guidance. After that, after each training, the students only need to collect training data to analyze and calculate the students' various data, namely, vital capacity data, the ability to convert high and low tones, and the clarity of pronunciation, so as to judge which aspects of the students need to be improved, and then propose targeted training, which more effectively improves the training efficiency of the students. This design only needs to collect data at regular intervals to complete the training guidance of the students, which is more comparative and targeted.

[0032] 2. In the present invention, environmental data is also collected. In certain special environments, such as high temperature, dryness or a large amount of dust particles in the air, the students' voice will be affected. For example, students in a high temperature environment will be more impetuous, and dry or humid conditions will affect the students' breathing. Poor air quality is more likely to affect the students' physical and mental health and training results. Therefore, in the present invention, in the system that uses environmental factors as variable input values, the algorithm is used to extract the influence of abnormal values ​​in the environmental data on the evaluation results, so as to obtain more effective evaluation result data, avoid errors in the instructor's analysis and judgment of the students due to environmental factors, and further improve the students' training effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A system diagram of a vocal training system of the present invention;

[0034] Figure 2 The present invention is a flow chart of a method for assisting vocalization practice. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Example 1

[0037] like Figure 1 As shown, the present invention provides a technical solution: a vocal training system, comprising a data acquisition module for data acquisition, a data processing module for processing data, a data analysis module for analyzing data, and a data evaluation module for data evaluation;

[0038] The data acquisition module includes a time recording unit for recording training time, a voice acquisition unit for recording training sound, a heart rate detection unit for recording the heart rate of the trainee, an environmental data acquisition unit for collecting surrounding environment data, a data sorting unit for data sorting and numbering, and a data uploading unit for uploading data;

[0039] The data processing module includes a data receiving unit for data reception, a data classification unit for data classification, a speech recognition unit for converting speech data into text data, an AI learning unit for converting speech data into fundamental frequency data, and a data interaction unit for data interaction;

[0040] The data analysis module includes a data interaction unit for data interaction, a data and processing unit for data classification, a data matching unit for data matching, a matching result conversion unit for matching result conversion, a data analysis unit for data analysis, and an analysis result uploading unit for uploading analysis results;

[0041] The data evaluation module includes a data retrieval unit for data retrieval, a data receiving unit for data reception, an algorithm unit for adjusting evaluation results, a training evaluation production unit, and a storage unit for data storage.

[0042] Among them, the time recording unit includes a timer, the voice collection unit is a recorder, the heart rate detection unit is a heart rate monitoring ring when worn, and the environmental data collection unit includes a temperature sensor, a humidity sensor and a dust particle detection sensor. The data collection steps are as follows: after the student starts voice training, the timer records the training time and the duration of the high and low notes during the training process, the recorder is used to record the sound generated during the training process, the heart rate detection unit monitors the student's heart rate changes in real time and uploads it in real time, and the environmental data collection unit is used to collect surrounding temperature data, humidity data and dust content data in the air. The student's training data is obtained through the time recording unit, the voice collection unit and the heart rate detection unit. The environmental data collection unit obtains variable data, which can realize rapid data collection, so that the instructor can compare according to the collected data, and then judge the student's training effect, rather than relying on subjective will to provide training guidance, making the training guidance more effective.

[0043] Among them, the speech recognition unit is used to convert the sound data of the data acquisition unit module into text data. The converted text data does not need to be preprocessed and is directly uploaded to the data analysis module through the data interaction unit. The AI ​​learning unit is used to convert the speech data into fundamental frequency data. The speech data is converted into fundamental frequency data, which can obtain the duration and conversion of the student's high and low tones, so as to analyze the student's skills in some special vocal skills. The specific steps of data processing are as follows: the data receiving unit receives the data collected by the data acquisition module, and then the data classification unit classifies the data. After the classification is completed, the speech recognition unit converts the text data. By converting the speech recognition into text, the clarity of the student's pronunciation can be judged. If the text converted from the student's speech data is closer to the lyrics of the training song, it means that the pronunciation is clear, and vice versa. The AI ​​learning unit converts the speech data into fundamental frequency data through the short-time average amplitude difference method, and the processed data is uploaded to the data analysis module through the data interaction unit.

[0044] Among them, in the data analysis module, the processing of the data preprocessing unit includes data classification and cleaning of duplicate data. The specific steps are as follows: the data interaction unit receives the processed data uploaded from the data processing module, and then deletes the duplicate data and classifies the data with the data preprocessing unit. The classification is based on the format type of the data. Then, the data matching unit matches the student historical data stored in the storage unit, and the matching result conversion unit converts the matching result into text. The data analysis unit is used to perform a preliminary analysis based on the converted matching result, and then upload the analysis result to the data evaluation for final evaluation. Through the preliminary analysis of the student's training data, it can be judged whether the student's certain data is enhanced or reduced compared with the historical training, thereby judging which aspects of the student need to be strengthened.

[0045] Among them, the storage unit in the data evaluation module is used to store the historical training data of the students, and the algorithm processing unit is used to input the variable factors into the preliminary analysis results uploaded by the data analysis module to improve the preliminary analysis results. The specific steps are as follows: the data receiving unit receives the preliminary analysis results from the data analysis module, and then the data retrieval unit retrieves the relevant environmental data, and inputs the environmental data as a variable into the algorithm processing unit, and the algorithm processing unit corrects the analysis results. The evaluation of this training is completed through the corrected analysis results. The environmental data is collected. In some special environments, such as high temperature, dryness or a large number of dust particles in the air, it will affect the students' voice. For example, in a high temperature environment, the students' mentality will be more impetuous, and dry or humid conditions will affect the students' breathing. In the case of poor air quality, it is more likely to affect the students' physical and mental health and training results. Therefore, in the system of taking environmental factors as variable input values ​​in the present invention, the influence of abnormal values ​​in the environmental data on the evaluation results is taken out through the algorithm, so as to obtain more effective evaluation result data, avoid errors in the instructor's analysis and judgment of the students due to environmental factors, and further improve the training effect of the students.

[0046] In this embodiment, by collecting the training data of the students, and then classifying and analyzing the collected data, the instructor only needs to evaluate the results of the students' initial training and propose targeted guidance training. After that, after each training, the students only need to collect training data to analyze and calculate the students' various data, namely vital capacity data, the ability to convert high and low notes, and the clarity of pronunciation, so as to determine which aspects of the students need to be improved, and then propose targeted training, so as to more effectively improve the students' training efficiency. This design only needs to collect data at regular intervals to complete the training guidance of the students, which is more comparative and targeted.

[0047] Example 2

[0048] like Figure 2 As shown, a vocal training auxiliary method, the vocal training auxiliary method has the following specific steps:

[0049] S1. Data collection: The trainee puts on the heart rate monitoring ring and starts training. During the training, the timer records the training time and the duration of the high and low notes during the training. The recorder is used to record the sounds generated during the training. The heart rate detection unit monitors the changes in the trainee's heart rate in real time and uploads it in real time. The environmental data collection unit is used to collect the surrounding temperature data, humidity data and dust content data in the air, and then upload the data;

[0050] S2. Data processing: The data receiving unit receives the data collected by the data collection module, and then the data classification unit classifies the data. After the classification is completed, the speech recognition unit converts the text data. The AI ​​learning unit converts the speech data into fundamental frequency data through the short-time average amplitude difference method. The processed data is uploaded to the data analysis module through the data interaction unit;

[0051] S3, data analysis: the data interaction unit receives the processed data uploaded from the data processing module, and then deletes duplicate data and classifies the data with the data preprocessing unit, and the classification is performed according to the format type of the data. Then, the data matching unit matches the student historical data stored in the storage unit, and the matching result conversion unit converts the matching result into text. The data analysis unit is used to perform preliminary analysis based on the converted matching result, and then uploads the analysis result to the data evaluation for final evaluation;

[0052] S4, training evaluation: the data receiving unit receives the preliminary analysis results from the data analysis module, and then the data retrieval unit retrieves the relevant environmental data, and inputs the environmental data into the algorithm processing unit as a variable, and the algorithm processing unit corrects the analysis results, and completes the evaluation of this training through the corrected analysis results;

[0053] S5. Auxiliary guidance training: Based on the training evaluation data obtained in step S4, targeted training guidance is provided to the students according to the changes in the students' vocal training time, the changes in the vocalization time of the high and low notes, the changes in the heart rate, and the accuracy of the speech recognition text.

[0054] Among them, in step S2, the specific process of using the AI ​​learning unit to convert data is as follows:

[0055] S2.1. Select the voice data uploaded in a certain period of time, first determine which student's training data the voice data belongs to based on the time between uploads, and after determination, convert the continuous signal into a discrete signal through A / D conversion;

[0056] S2.2. For the speech signal sample x(n+i) in each frame, the average value of the absolute value of the sample calculated by the short-time average amplitude function is as follows: N is the frame length, n is the start time index of the current frame, and within each frame, the amplitude difference of adjacent sample points is calculated. Assuming that the amplitudes of adjacent sample points are |x(n+i)| and |x(n+i+1)|, the amplitude difference is: D(n+i)=|x(n+i+1)|-|x(n+i)|. The short-time mean amplitude difference function (SMDF) is defined as the cumulative average of these differences, and the formula is as follows:

[0057] S2.3, estimate the fundamental frequency by finding the time interval corresponding to the valley value of the SMDF function;

[0058] S2.4. Perform outlier processing, remove excessively high and low fundamental frequency values, smooth the fundamental frequency sequence, and then obtain the student's speech fundamental frequency value. Determine the frequency and change time of the student's high and low tones from the fundamental frequency value.

[0059] Among them, in step S3, the data matching mainly includes training duration data matching, speech-to-text data matching, high-note and low-note pronunciation duration data matching and heart rate change data matching. The training duration data matching is mainly used to judge the changes in the students' vital capacity, the speech-to-text data matching is mainly used to judge the clarity of the students' pronunciation, the high-note and low-note pronunciation duration matching is mainly used to judge the students' high-low-tone conversion site, and the heart rate change data matching is used to judge the students' stage effect.

[0060] Among them, in step S4, the correction of the analysis result specifically includes the following steps:

[0061] S4.1. Obtain preliminary analysis results and retrieve corresponding analysis results according to the time when the voice data was uploaded;

[0062] S4.2, the data retrieval module retrieves the environmental data collected in step S1 and inputs the environmental data into the system. The environmental data includes temperature, humidity and air index data. When the environmental data is input, the time of uploading the environmental data corresponds to the time of the voice data uploaded by the trainee, ensuring that the real-time uploaded environmental data corresponds to each trainee to avoid errors:

[0063] S4.3. Use machine learning algorithms to compare environmental data with historical data to identify data outliers. Improve the evaluation results by replacing or deleting outliers. Then fit the evaluation result data to generate new analysis result data. Repeat the above steps based on the outliers to generate a variety of different analysis results. Divide the analysis result data into multiple subsets to repeat training and testing. Finally, obtain new evaluation results combined with abnormal environmental data.

[0064] In this embodiment, various data of the students are collected, and the real-time collected data is compared with the historical data to determine the changes in the data, thereby determining the training results of the students, and then being able to give more targeted training suggestions to improve their vocal level. Secondly, the method of using data collection to judge the training effect can truly achieve efficient vocal training. Compared with traditional methods, vocal training is more efficient and the training guidance for students is more effective, avoiding the low efficiency and low accuracy of traditional one-on-one training by instructors.

[0065] In this embodiment, the upload time will be displayed during the data upload process, and the identity information of the corresponding student will be judged based on the time when the data was uploaded. The data uploaded in the same time period will be reviewed offline by the instructor. After the review is completed, the instructor will provide guidance on the student's training based on the relevant data.

[0066] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A vocal training system, characterized in that: It includes a data acquisition module for data acquisition, a data processing module for processing data, a data analysis module for analyzing data, and a data evaluation module for data evaluation; The data acquisition module includes a time recording unit for recording training time, a voice acquisition unit for recording training sound, a heart rate detection unit for recording the heart rate of the trainee, an environmental data acquisition unit for collecting surrounding environment data, a data sorting unit for data sorting and numbering, and a data uploading unit for uploading data; The data processing module includes a data receiving unit for data reception, a data classification unit for data classification, a speech recognition unit for converting speech data into text data, an AI learning unit for converting speech data into fundamental frequency data, and a data interaction unit for data interaction; The data analysis module includes a data interaction unit for data interaction, a data and processing unit for data classification, a data matching unit for data matching, a matching result conversion unit for matching result conversion, a data analysis unit for data analysis, and an analysis result uploading unit for uploading analysis results; The data evaluation module includes a data retrieval unit for data retrieval, a data receiving unit for data reception, an algorithm unit for adjusting evaluation results, a training evaluation production unit, and a storage unit for data storage.

2. A vocal training system according to claim 1, characterized in that: The time recording unit includes a timer, the voice collection unit is a recorder, the heart rate detection unit is a heart rate monitoring ring when worn, and the environmental data collection unit includes a temperature sensor, a humidity sensor and a dust particle detection sensor. The data collection steps are as follows: after the student starts voice training, the timer records the training time and the duration of high and low notes during the training process, the recorder is used to record the sound generated during the training process, the heart rate detection unit monitors the student's heart rate changes in real time and uploads them in real time, and the environmental data collection unit is used to collect surrounding temperature data, humidity data and dust content data in the air.

3. A vocal training system according to claim 2, characterized in that: The speech recognition unit is used to convert the sound data of the data acquisition unit module into text data. The converted text data does not need to be preprocessed and is directly uploaded to the data analysis module through the data interaction unit. The AI ​​learning unit is used to convert the speech data into fundamental frequency data. The specific steps of data processing are as follows: the data receiving unit receives the data collected by the data acquisition module, and then the data classification unit classifies the data. First, the data is classified according to the data format, and the data of different formats are classified to facilitate the processing of formats of the same format type. The speech recognition unit converts the text data, and the AI ​​learning unit converts the speech data into fundamental frequency data through the short-time average amplitude difference method. The processed data is uploaded to the data analysis module through the data interaction unit.

4. A vocal training system according to claim 3, characterized in that: In the data analysis module, the processing of the data preprocessing unit includes data classification and cleaning of duplicate data, and the specific steps are as follows: the data interaction unit receives the processed data uploaded from the data processing module, and then deletes the duplicate data and classifies the data with the data preprocessing unit. The data preprocessing unit clears the duplicate data, and then classifies the data again. The classification is performed based on the format type of the data and the time of upload to ensure that the data uploaded by each student is aggregated together to facilitate processing of data for different students. Afterwards, the data matching unit matches the student historical data stored in the storage unit, and the matching result conversion unit converts the matching result into text. The data analysis unit is used to perform a preliminary analysis based on the converted matching result, and then upload the analysis result to the data evaluation for final evaluation.

5. A vocal training system according to claim 4, characterized in that: The storage unit in the data evaluation module is used to store the students' historical training data, and the algorithm processing unit is used to input variable factors into the preliminary analysis results uploaded by the data analysis module to improve the preliminary analysis results. The specific steps are as follows: the data receiving unit receives the preliminary analysis results from the data analysis module, and then the data retrieval unit retrieves the relevant environmental data, and inputs the environmental data as variables into the algorithm processing unit. The algorithm processing unit corrects the analysis results and completes the evaluation of this training through the corrected analysis results.

6. A vocal training auxiliary method, according to a vocal training system according to claims 1-5, characterized in that: The specific steps of the vocal training auxiliary method are as follows: S1. Data collection: The trainee puts on the heart rate monitoring ring and starts training. During the training, the timer records the training time and the duration of the high and low notes during the training. The recorder is used to record the sounds generated during the training. The heart rate detection unit monitors the changes in the trainee's heart rate in real time and uploads it in real time. The environmental data collection unit is used to collect the surrounding temperature data, humidity data and dust content data in the air, and then upload the data; S2. Data processing: The data receiving unit receives the data collected by the data collection module, and then the data classification unit classifies the data. The classification standard is classified according to the format type of the data. After the classification is completed, the speech recognition unit converts the text data. The AI ​​learning unit converts the speech data into fundamental frequency data through the short-time average amplitude difference method. The processed data is uploaded to the data analysis module through the data interaction unit; S3, data analysis: the data interaction unit receives the processed data uploaded from the data processing module, and then deletes duplicate data and classifies the data with the data preprocessing unit. The classification is based on the format type of the data and the upload time. Then, the data matching unit matches the student historical data stored in the storage unit. Specifically, the basic data of which student is uploaded in this time period is determined according to the upload time. Then, the historical data related to the student is retrieved from the storage unit according to the student information, and finally matched with the real-time data. The matching result conversion unit converts the matching result into text. The data analysis unit is used to perform preliminary analysis based on the converted matching result, and then upload the analysis result to the data evaluation for final evaluation. S4, training evaluation: the data receiving unit receives the preliminary analysis results from the data analysis module, and then the data retrieval unit retrieves the relevant environmental data, and inputs the environmental data into the algorithm processing unit as a variable, and the algorithm processing unit corrects the analysis results, and completes the evaluation of this training through the corrected analysis results; S5. Auxiliary guidance training: Based on the training evaluation data obtained in step S4, targeted training guidance is provided to the students according to the changes in the students' vocal training time, the changes in the vocalization time of the high and low notes, the changes in the heart rate, and the accuracy of the speech recognition text.

7. A vocal training auxiliary method according to claim 6, characterized in that: In step S2, the specific process of using the AI ​​learning unit to convert data is as follows: S2.1, select the voice data uploaded at a certain time, and convert the continuous signal into a discrete signal through A / D conversion; S2.

2. For the speech signal sample x(n+i) in each frame, the average value of the absolute value of the sample calculated by the short-time average amplitude function is as follows: N is the frame length, n is the start time index of the current frame, and within each frame, the amplitude difference of adjacent sample points is calculated. Assuming that the amplitudes of adjacent sample points are |x(n+i)| and |x(n+i+1)|, the amplitude difference is: D(n+i)=|x(n+i+1)|-|x(n+i)|. The short-time mean amplitude difference function (SMDF) is defined as the cumulative average of these differences, and the formula is as follows: S2.3, estimate the fundamental frequency by finding the time interval corresponding to the valley value of the SMDF function; S2.

4. Perform outlier processing, remove excessively high and low fundamental frequency values, smooth the fundamental frequency sequence, and obtain the fundamental frequency value of the student's sound during the training process, so as to identify the frequency and changes of the student's high and low tones, thereby judging the student's training effect.

8. A vocal training auxiliary method according to claim 7, characterized in that: In step S3, the data matching mainly includes training duration data matching, speech-to-text data matching, high-note and low-note pronunciation duration data matching, and heart rate change data matching. The training duration data matching is mainly used to determine the changes in the students' vital capacity, the speech-to-text data matching is mainly used to determine the clarity of the students' pronunciation, the high-note and low-note pronunciation duration matching is mainly used to determine the students' high-low-tone conversion site, and the heart rate change data matching is used to determine the students' stage effect.

9. A vocal training auxiliary method according to claim 8, characterized in that: In step S4, the correction of the analysis result specifically includes the following steps: S4.

1. Obtaining preliminary analysis results, i.e., evaluation results of preliminary analysis completed by the data analysis module through the data analysis unit; S4.2, the data retrieval module retrieves the environmental data collected in step S1 and inputs the environmental data into the system. The environmental data includes the temperature, humidity and air index data during the training process: S4.

3. Use machine learning algorithms to compare environmental data with historical data to identify data outliers. Improve the evaluation results by replacing or deleting outliers. Then fit the evaluation result data to generate new analysis result data. Repeat the above steps based on the outliers to generate a variety of different analysis results. Divide the analysis result data into multiple subsets to repeat training and testing. Finally, obtain new evaluation results combined with abnormal environmental data.

Images