CN107610691A - English vowel sounding error correction method and device - Google Patents

Abstract

The present invention relates to a kind of English vowel sounding error correction method, including：Step 1, the English vowel sounding acoustic model of pre-stored criteria；Step 2, the English Phonetics of typing measurand；Step 3, the vowel in the English Phonetics of measurand is identified；Step 4, typing measurand reads aloud the voice of the vowel identified；Step 5, the voice for reading aloud measurand the vowel identified carries out English vowel sounding acoustic analysis；Step 6, the English vowel sounding acoustic analyze data of measurand obtains first degree of deviation compared with the English vowel sounding acoustic model of standard；Step 7, error correction is carried out to the English vowel sounding of measurand according to first degree of deviation.By analyzing the English vowel sounding acoustic of measurand, and compared with the English vowel sounding acoustic model of the standard with prestoring, so as to be corrected to the English vowel sounding of measurand, to cause the English vowel sounding of measurand more accurate.

Description

English vowel pronunciation error correction method and device

technical field

The invention relates to the technical field of speech recognition, in particular to a method and device for correcting English vowel sounds.

Background technique

Language is a bridge for human communication, and language diversity is also an obstacle to communication. English is the dominant lingua franca with high frequency of use. "English variants" with different pronunciation characteristics in different countries in the world have become obstacles in communication. The study of English features in different native language backgrounds is the key to breaking through communication barriers and improving the speaker's English pronunciation. The recognition of the speaker's speech characteristics is an unavoidable process of recognizing the speaker's speech content. In order to improve the speaker's English pronunciation and improve communication barriers between people of different nationalities, it is particularly important to correct the speaker's English pronunciation when the speaker's voice deviates from the target language's phonetic norms within a certain range.

During pronunciation, the airflow exhaled from the lungs passes through the oral cavity, which acts as a resonator, and a voice with minimal resistance and no friction sound is produced. Although under normal circumstances the vocal cords vibrate when pronouncing vowels, it is also possible to make the vocal cords not vibrate and produce unvoiced or ear sounds.

From the perspective of pronunciation phonetics, vowels are usually classified according to the position of the tongue and the shape of the lips. When high vowels are pronounced, the tongue surface is arched and close to the palate. When low vowels are pronounced, the tongue is relatively flat. , the distance between the tongue surface and the upper palate is slightly larger, the position of the tongue is in the middle state when the middle vowel is pronounced, and the high, middle and low vowels are also classified according to the front and back columns. The position of the tongue surface and the shape of the lips are a criterion for classifying vowels.

According to pronunciation phonetics, vowels are a type of pronunciation formed by the airflow vibrating the vocal cords without being obstructed in the oral cavity. Different oral cavity shapes form different vowels; while consonants are formed when the airflow is obstructed in the oral cavity. Different consonants are formed by the place of pronunciation or the method of pronunciation.

Contents of the invention

The purpose of the present invention is to realize the error correction of the English vowel pronunciation of the measured object by comparing the measured object with the standard English vowel sounding model when English is the target language.

One aspect of the present invention provides a method for correcting English vowel pronunciation, including: step 1, pre-store a standard English vowel pronunciation acoustic model; step 2, input the English voice of the measured object; step 3, identify the measured object Vowels in the English voice; Step 4, input the voice of the measured object reading the recognized vowel; Step 5, the voice of the measured object reading the recognized vowel English Pronunciation acoustic analysis; step 6, the English vowel pronunciation acoustic analysis data of the measured object is compared with the English vowel pronunciation acoustic model of the standard to obtain a first degree of deviation; step 7, according to the first degree of deviation Error correction is performed on the English vowel pronunciation of the subject under test.

Described step 1 comprises: input the English speech of a plurality of standard English sample objects; Identify the vowel in the English speech of described a plurality of standard English sample objects; Carry out English vowel pronunciation acoustic analysis to the vowel of each sample object respectively ; Generating the standard English vowel pronunciation acoustic model according to the acoustic analysis results of English vowel pronunciation.

The step 2 includes: providing the voice material according to the nationality of the tested object, and recording the English voice of the tested object reading the voice material.

The step 3 includes: identifying the vowels in the English speech of the subject under test according to the formant peaks of the vowels.

The step 3 includes: identifying the vowel in the English speech of the subject under test according to the resonance peak of the vowel and the duration of the vowel.

The resonance peaks include a first resonance peak and a second resonance peak.

The step 7 includes: according to the data of the acoustic analysis of the English vowel pronunciation and the standard English vowel pronunciation acoustic model, adjusting the English vowel pronunciation of the subject under test with a visual image.

After the step 7, it also includes: re-entering the voice of the measured object reading the vowel again; performing an acoustic analysis of English vowel sounds on the voice of the measured object reading again; The English vowel phonation acoustic analysis data of the speech is compared with the English vowel phonation acoustic model of the standard to obtain a second deviation degree; English vowel pronunciation evaluation text.

The acoustic analysis of English vowel sounds includes: measuring the resonance peak of the input English vowel sounds; measuring the duration of the input English vowel sounds; according to the input resonance peak of the English vowel sounds and the The duration of generating the inputted English vowel phonation acoustic analysis data.

The resonance peaks include a first resonance peak and a second resonance peak.

The present invention also provides a storage device, wherein a plurality of instructions are stored, and the instructions are suitable for being loaded by a processor and executed as follows: Step 1, pre-store a standard English vowel sounding acoustic model; Step 2, input the measured object Step 3, identify the vowels in the English voice of the measured object; Step 4, input the voice of the measured object to read the recognized vowels; Step 5, to the measured object Reading the voice of the identified vowels aloud and performing an acoustic analysis of English vowel sounds; step 6, comparing the acoustic analysis data of English speech sounds of the measured object with the standard English vowel sound acoustic model to obtain a first deviation degree; step 7, correcting the English vowel pronunciation of the measured object according to the first degree of deviation.

Described step 1 comprises: input the English speech of a plurality of standard English sample objects; Identify the vowel in the English speech of described a plurality of standard English sample objects; Carry out English vowel pronunciation acoustic analysis to the vowel of each sample object respectively ; Generating the standard English vowel pronunciation acoustic model according to the acoustic analysis results of English vowel pronunciation.

The step 2 includes: providing voice material according to the nationality of the tested object, and recording the English voice of the tested object reading the voice material.

The step 3 includes: identifying the vowels in the English speech of the subject under test according to the formant peaks of the vowels.

The step 3 further includes: identifying the vowel in the English speech of the subject under test according to the resonance peak of the vowel and the duration of the vowel.

The resonance peaks include a first resonance peak and a second resonance peak.

The step 7 includes: according to the data of the acoustic analysis of the English vowel pronunciation and the standard English vowel pronunciation acoustic model, adjusting the English vowel pronunciation of the subject under test with a visual image.

After the step 7, it also includes: re-entering the voice of the tested object reading the recognized vowel again; performing an acoustic analysis of English vowel sounds on the voice of the tested object reading again; The English vowel phonation acoustic analysis data of the voice read aloud again by the subject is compared with the standard English vowel phonation acoustic model to obtain a second degree of deviation; The test object's English vowel pronunciation evaluation text.

The acoustic analysis of English vowel sounds includes: measuring the resonance peak of the input English vowel sounds; measuring the duration of the input vowel sounds; according to the input resonance peak of the English vowel sounds and the duration Acoustic analysis data of the entered English vowel phonation is generated.

The resonance peaks include a first resonance peak and a second resonance peak.

The present invention also provides an English vowel sound error correction device, comprising: a processor, adapted to implement instructions; and a storage device, adapted to store a plurality of instructions, the instructions are adapted to be loaded by the processor and executed as : Step 1, pre-store standard English vowel sound acoustic model; Step 2, input the English voice of the tested object; Step 3, identify the vowels in the English voice of the tested object; Step 4, input the tested object The object reads the voice of the recognized vowel; step 5, performs an acoustic analysis of English vowel sounds on the voice of the measured object reading the recognized vowel; step 6, the English element of the measured object The pronunciation acoustic analysis is compared with the standard English vowel pronunciation acoustic model to obtain a first deviation degree; step 7, correcting the English vowel pronunciation of the measured object according to the first deviation degree.

Described step 1 comprises: input the English speech of a plurality of standard English sample objects; Identify the vowel in the English speech of described a plurality of standard English sample objects; Carry out English vowel pronunciation acoustic analysis to the vowel of each sample object respectively ; Generating the standard English vowel pronunciation acoustic model according to the acoustic analysis results of English vowel pronunciation.

The step 2 includes: providing voice material according to the nationality of the sample object, and recording the English voice of the sample object reading the voice material.

The step 3 includes: identifying the vowels in the English speech of the subject under test according to the formant peaks of the vowels.

The step 3 includes: identifying the vowel in the English speech of the subject under test according to the resonance peak of the vowel and the duration of the vowel.

The resonance peaks include a first resonance peak and a second resonance peak.

The step 7 includes: according to the data of the acoustic analysis of the English vowel pronunciation and the standard English vowel pronunciation acoustic model, adjusting the English vowel pronunciation of the subject under test with a visual image.

After the step 7, it also includes: re-entering the voice of the tested object reading the recognized vowel again; performing an acoustic analysis of English vowel sounds on the voice of the tested object reading again; The English vowel pronunciation acoustic analysis data read aloud by the subject is compared with the standard English vowel pronunciation acoustic model to obtain a second degree of deviation; output the measured object according to the first degree of deviation and the second degree of deviation English vowel pronunciation evaluation text.

The acoustic analysis of English vowel sounds includes: measuring the resonance peak of the input English vowel sounds; measuring the duration of the input vowel sounds; according to the input resonance peak of the English vowel sounds and the duration Acoustic analysis data of the entered English vowel phonation is generated.

The resonance peaks include a first resonance peak and a second resonance peak.

The beneficial effects of the present invention are that, by analyzing the English vowel sounding acoustics of the measured object, and comparing with the pre-stored standard English vowel sounding acoustic model, the English vowel sounding of the measured object is corrected, so that Make the English vowel pronunciation of the tested object more accurate.

Description of drawings

The present invention will be further described below in conjunction with accompanying drawing and embodiment, in the accompanying drawing:

Fig. 1 is a flowchart of an English vowel pronunciation error correction method 100 according to an embodiment of the present invention;

2 is a flowchart of a method 200 for selecting sample objects in forming a standard English vowel pronunciation model according to an embodiment of the present invention;

FIG. 3 is a flow chart of a method 300 for forming a standard English vowel pronunciation model according to an embodiment of the present invention;

Fig. 4 is a visual English vowel pronunciation error correction diagram of the English vowel pronunciation error correction method according to an embodiment of the present invention;

FIG. 5 is a flow chart of an English vowel pronunciation error correction method 500 according to another embodiment of the present invention;

FIG. 6 is a flowchart of an English vowel pronunciation analysis method 600 according to an embodiment of the present invention;

FIG. 7 is an acoustic analysis diagram of English vowel pronunciation of men and women in different countries generated by the English vowel pronunciation acoustic analysis method 600 shown in FIG. 6 according to an embodiment of the present invention;

FIG. 8 is a diagram of a standard English vowel pronunciation model generated by using the English vowel pronunciation acoustic analysis method 600 shown in FIG. 6 according to an embodiment of the present invention.

detailed description

Now in conjunction with the accompanying drawings, the preferred embodiments of the present invention will be described in detail.

As shown in FIG. 1 , it is a flow chart of a method 100 for correcting English vowel pronunciation errors according to an embodiment of the present invention.

In step 101, a standard English vowel pronunciation model is pre-stored. For example, the vowel pronunciation model of the sample object of the English native language background may be saved as a standard English vowel pronunciation model. The sample object mentioned in the present invention refers to the speaker selected when forming a standard English vowel pronunciation model.

Step 103, inputting the English voice of the tested object. In a specific embodiment, audio materials can be provided for the subject to read aloud, and any other English words or sentences can also be read by the subject to be tested. When choosing to read aloud according to the phonetic material, the same phonetic material as that used when establishing a standard English vowel sounding model can be provided, or a different phonetic material than that used when building a standard English vowel sounding model.

Step 105, identifying vowels in the recorded English speech of the subject to be tested.

Step 107, recording the voice of the tested subject reading the recognized vowels aloud.

Step 109 , performing an acoustic analysis of English vowel utterances on the recorded speech of the recognized vowels read aloud by the subject to obtain the acoustic analysis data of vowel utterances.

Step 111 , comparing the obtained English vowel pronunciation acoustic analysis data with a pre-stored standard English vowel pronunciation model to obtain a first degree of deviation.

Step 113 , correcting the English vowel pronunciation of the subject under test according to the first degree of deviation.

In a specific embodiment, after inputting the English voice read aloud by a certain measured object, first recognize the vowels in the English voice, and then input the voice of the measured object to read the recognized vowel, and to the measured object Read the voice of the recognized vowel aloud and perform acoustic analysis of English vowel pronunciation, and compare the obtained English vowel pronunciation acoustic analysis data with the pre-stored standard English vowel pronunciation model to obtain the English vowel pronunciation of the tested object If there is a difference from the standard English vowel pronunciation, the English vowel pronunciation of the subject under test can be corrected according to the difference.

In one embodiment, standard English sample objects need to be selected to establish a standard English vowel pronunciation acoustic model, which can be accomplished by using the method in FIG. 2 .

Step 201, select speech materials. The present invention is aimed at English as the target language, with vowels as the main object, and the corpus involves the presentation of all vowels in English in the structure of words and sentences. Sentence structure includes 5 kinds of simple sentence patterns in English. Semantics concerns predictable and unpredictable sentences. Predictable sentences include sentences with high predictability and sentences with low predictability. All words are high-frequency words, but include all English vowels. Voice materials can be designed by world-renowned phoneticians. Predictable and unpredictable sentences, which affect the perception of vowels, for example: the complete sentence is "spread some butter on the bread", even though the sentence is "spread some butter on the", although "bread" is not said, or there is no "butter" is recognized, but it is usually known as "butter/bread", which means that the sentence can be predicted, so it can also be recognized.

Step 203, select the standard English sample object to read the speech material aloud, and establish a speech library. In practical application, native English adults in California can be selected as standard English sample subjects. After selecting these people, they will read the selected speech materials aloud and record them to form a speech library.

Step 205, the selection of the standard English sample object representative, the listener with the same language background as the sample object can be selected in the United States, and the medium-term perception is carried out, and the male and female speakers who are in the middle of the perception are selected, and they are the most representative sample objects .

Step 207 , conduct overall acoustic analysis of English vowel pronunciation on the most representative sample objects perceived through the middle period, and form a standard acoustic model of English vowel pronunciation.

In one embodiment, after the most representative standard English pronunciation population is selected through the method shown in FIG. 2 , the standard English vowel pronunciation acoustic model can be generated using the method shown in FIG. 3 .

In step 301, the English voices of multiple standard English sample objects are input. After the voices are input, the voice material is formed. There is more than one vowel in the natural language flow, thus forming basic vowel data.

Step 303: Identify the vowels in the English speech of the plurality of standard English sample objects. In one embodiment, the vowel in the speech can be identified according to the resonance peak of the vowel, or in order to further improve the recognition accuracy, Vowels in speech are identified by their formants combined with their durations. In a specific embodiment, the resonant peak may be a first resonant peak and a second resonant peak of a vowel. Among them, the first resonance peak F1 represents the dimension of the lips, that is, the up-and-down dimension of pronunciation, and the second resonance peak F2 represents the dimension of the tongue, that is, the front-back dimension of pronunciation.

Step 305, performing an acoustic analysis of English vowel pronunciation on the vowels in the speech of each standard English sample object.

Step 307: Generate a standard English vowel pronunciation acoustic model according to the acoustic analysis data of English vowel pronunciation of a plurality of standard English sample objects obtained in step 305. In a specific implementation, the single vowel feature of each sample object will be different from other speakers, but a person's front, back, high and low vowels will be within a certain range, and the speaker will be classified according to this particularity. The overall measurement of the vowel sounds, and then determine the basic characteristics of the speaker according to the range of vowels.

In one embodiment, audio materials may be provided according to the nationality of the subject under test, and the English voice of the subject under test reading the audio material may be recorded. For example, for Chinese, the vowel with high error rate is: /i:～ /, /ε～ / and /u:~ /, so when providing speech materials, more speech materials containing the above-mentioned vowels can be provided, so as to correct the vowels that are often made wrong by the test object, so that the correction of English vowel sounds is more targeted.

In one embodiment, correcting the English vowel pronunciation of the subject under test according to the first degree of deviation may include: according to the data of the English vowel pronunciation acoustic analysis and the standard English vowel pronunciation acoustic model, using a visual image Adjust the pronunciation of English vowels of the test subject. As shown in Figure 4, it is a visual English vowel pronunciation error correction diagram of the English vowel pronunciation error correction method according to an embodiment of the present invention, for example, when a Chinese's English vowel pronunciation is corrected, the speech material provided is "a good looking woman", by adopting the English vowel pronunciation error correction method of the present invention to Pronunciation is corrected, first obtained is the coordinate position of the first formant and the second formant of the Chinese vowel sound represented by the dot in Fig. 4, and the triangle in Fig. 4 represents The coordinate positions of the first formant and the second formant of the standard vocalization, through this visual comparison, can let the speaker know how to adjust his English vowel pronunciation more intuitively.

Fig. 5 is a flow chart of an English vowel pronunciation error correction method 500 according to another embodiment of the present invention, wherein, steps 501 to 513 are the same as steps 101 to 113 in Fig. 1, and step 515 is to input the subject to read aloud and identify The voice of the vowel; Step 517, carry out the acoustic analysis of English vowel sound to the speech that the measured object reads aloud again; Step 519, the data of the English vowel sound acoustic analysis that step 517 obtains and standard English vowel sound acoustic model Comparing to obtain the second degree of deviation; step 521 , output the English vowel pronunciation evaluation text of the subject under test according to the first degree of deviation and the second deviation.

In a specific embodiment, the English vowel pronunciation evaluation text may include information such as the original English vowel pronunciation acoustic analysis diagram of the subject and the corrected English vowel pronunciation acoustic analysis diagram, so as to facilitate the subject to understand his own English vowel pronunciation. The pronunciation situation and the existing pronunciation problems that need to be corrected enable the test subjects to carry out purposeful English pronunciation exercises according to their own characteristics, so as to improve the test subjects' own English pronunciation.

When generating a standard English vowel pronunciation acoustic model or performing an English vowel pronunciation acoustic analysis on the measured object, the English vowel pronunciation acoustic analysis method shown in FIG. 6 can be used.

In step 601, English voice is firstly recorded. When a standard English vowel pronunciation acoustic model is generated, the input is the voice of the sample subject of standard English pronunciation; when the acoustic analysis of English vowel pronunciation is performed on the tested object, the input is the voice of the tested object.

Step 603 identifies vowels in English speech. In one embodiment, the vowel in the speech can be identified according to the resonance peak of the vowel, or the vowel in the speech can be identified by combining the resonance peak and the duration of the vowel in order to further improve the recognition accuracy. In a specific embodiment, the resonant peak may be a first resonant peak and a second resonant peak of a vowel.

Step 605 measures the first resonance peak F1 and the second resonance peak F2 of the vowel, and the F1 and F2 of the vowel resonance peak are not in a linear relationship. In a specific embodiment, the Hertz value of the formant can be converted into a Bark (Bark) value , the specific conversion formula is:

Bark = [(26.81 x F)/(1960+F)] – 0.53.

Step 607 measures the duration of vowels.

Step 609 generates English vowel pronunciation acoustic analysis data according to the data measured in steps 605 and 607 .

When used to generate a standard English vowel pronunciation acoustic model, the English vowel pronunciation acoustic analysis is performed on multiple standard English sample objects, and finally the standard English vowel pronunciation is generated based on the English vowel pronunciation acoustic analysis data of multiple sample objects Acoustic model of sound production.

FIG. 7 is an acoustic analysis diagram of English vowel pronunciation of male and female test subjects of different nationalities generated by using the English vowel pronunciation acoustic analysis method 600 shown in FIG. Acoustic feature map of the test object, male on the left and female on the right, the abscissa is the F2 value of the second formant of the vowel, and the ordinate is the F1 value of the first formant of the vowel, and the formant F1 and F2 are divided by Hertz values are converted to Bark values.

The top layer in the picture shows the acoustic characteristics of the test subjects whose native language is Chinese. There is no obvious distinction between loose and vigorous vowels in the vowel pronunciation of the test subjects, which reflects the obvious interference of Chinese vowels and has obvious characteristics of Chinese accent. The middle layer in the figure shows the acoustic characteristics of the test subjects whose native language is Dutch. The vowel sounds of the test subjects are divided into tense and tense sounds, and some sounds obviously have negative transfer effect of Dutch phonetics. The bottom layer of the figure shows the acoustic characteristics of the test subjects whose native language is American English. The vowel pronunciation of the test subjects has a clear distinction between loose and tight vowels, which reflects the acoustic characteristics of the native English language and can be used as a standard acoustic model of English vowel pronunciation.

In one embodiment, as shown in FIG. 8 , it adopts the acoustic analysis method 600 of English vowel pronunciation shown in FIG. Diagram of a vowel pronunciation model.

It should be understood that the present invention does not limit the execution order of each step in the English vowel pronunciation error correction method, and the execution order of each step can be adjusted according to actual needs, as long as the technical solution of the present invention can be realized.

Those skilled in the technical field can understand that each step in the method for correcting English vowel pronunciation of the present invention can be implemented as a system, method or computer program product. Accordingly, various aspects of the present invention may be embodied in the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or a combination of hardware and software aspects.

It should be understood that each block of the flowchart above and combinations of blocks in the flowchart can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, the instructions produce Means for realizing the functions specified in one or more blocks in the flowchart. Computer program code for carrying out operations for the various aspects of the present invention can be written in any combination of one or more programming languages, including object-oriented programming languages—such as Java, C++, and C++ and similar programming languages.

These computer program instructions can also be stored in a computer-readable medium, and these instructions cause a computer, other programmable data processing apparatus, or other equipment to operate in a specific way, so that the instructions stored in the computer-readable medium generate information including Manufactures of instructions that implement the functions/actions specified in one or more blocks in flowcharts and/or block diagrams. The computer readable storage medium can be a portable computer disk, hard disk, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), portable compact disk read only memory (CD -ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. A computer readable storage medium may be any tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device.

It is also possible to load computer program instructions into a computer, other programmable data processing device, or other equipment, so that a series of operation steps are executed on the computer, other programmable device or other equipment, so as to generate a computer-implemented process, so that in Instructions executed on computers or other programmable devices provide processes for implementing the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams.

It should be understood that the above embodiments are only used to illustrate the technical solutions of the present invention, not to limit them. Those skilled in the art can modify the technical solutions described in the above embodiments, or modify some of the technical features. Perform equivalent replacements; and all these modifications and replacements should belong to the protection scope of the appended claims of the present invention.

Claims (30)

1. A method for correcting errors in pronunciation of English vowels, is characterized in that, comprising: Step 1, pre-store the standard English vowel pronunciation acoustic model; Step 2, input the English voice of the tested object; Step 3, identifying vowels in the English speech of the tested object; Step 4, recording the voice of the subject under test reading the recognized vowel aloud; Step 5, performing an acoustic analysis of English vowel pronunciation on the voice of the measured subject reading the recognized vowel aloud; Step 6, comparing the English vowel pronunciation acoustic analysis data of the measured object with the standard English vowel pronunciation acoustic model to obtain a first degree of deviation; Step 7 Correcting the English vowel pronunciation of the subject under test according to the first degree of deviation. 2. English vowel pronunciation error correction method as claimed in claim 1, is characterized in that, described step 1 comprises: Input the English voice of multiple standard English sample objects; identifying vowels in English speech of the plurality of standard English sample objects; Acoustic analysis of English vowel pronunciation for each sample object's vowel separately; The standard English vowel pronunciation acoustic model is generated according to the English vowel pronunciation acoustic analysis result. 3. The English vowel pronunciation error correction method as claimed in claim 1, characterized in that, said step 2 comprises: providing voice material according to the nationality of said tested object, and typing said measured object to read said voice aloud English phonetics of the material. 4. The English vowel pronunciation error correction method according to claim 1, characterized in that, said step 3 comprises: identifying the vowel in the English speech of the subject under test according to the resonance peak of the vowel. 5. The English vowel pronunciation error correction method as claimed in claim 1, characterized in that, said step 3 comprises: identifying the English language of the measured object according to the resonance peak value of the vowel and the duration of the vowel. Vowels in speech. 6. The method for correcting errors in English vowel pronunciation according to claim 4 or 5, wherein the resonance peak comprises a first resonance peak and a second resonance peak. 7. English vowel pronunciation error correction method as claimed in claim 1, is characterized in that, described step 7 comprises: according to the data of described English vowel pronunciation acoustic analysis and the English vowel pronunciation acoustic model of described standard, The vocalization of English vowels of the subject under test is corrected with a visual image. 8. English vowel pronunciation error correction method as claimed in claim 1, is characterized in that, after described step 7, also comprises: Recording again the voice of the subject under test reading the recognized vowel sound aloud again; Carrying out acoustic analysis of English vowel pronunciation to the voice read aloud again by the subject under test; The English vowel sounding acoustic analysis data of the speech read aloud again by the subject under test is compared with the standard English vowel sounding acoustic model to obtain a second degree of deviation; Outputting the English vowel pronunciation evaluation text of the measured object according to the first degree of deviation and the second degree of deviation. 9. English vowel pronunciation error correction method as claimed in claim 1,2 or 8, is characterized in that, described English vowel pronunciation acoustic analysis comprises: Measuring the resonance peak of the recorded English vowel sounds; Measuring the duration of the recorded English vowel sounds; Acoustic analysis data of the subject's English vowel pronunciation is generated according to the recorded resonance peak of the English vowel pronunciation and the duration. 10. The method for correcting errors in English vowel pronunciation according to claim 9, wherein the resonance peak comprises a first resonance peak and a second resonance peak. 11. A memory device in which a plurality of instructions are stored, said instructions being adapted to be loaded by a processor and executed as: Step 1, pre-store the standard English vowel pronunciation acoustic model; Step 2, input the English voice of the tested object; Step 3, identifying vowels in the English speech of the tested object; Step 4, recording the voice of the subject under test reading the recognized vowel aloud; Step 5, performing an acoustic analysis of English vowel pronunciation on the voice of the measured subject reading the recognized vowel aloud; Step 6, comparing the English vowel pronunciation acoustic analysis data of the measured object with the standard English vowel pronunciation acoustic model to obtain a first degree of deviation; Step 7, correcting the English vowel pronunciation of the subject under test according to the first degree of deviation. 12. The storage device according to claim 11, wherein the step 1 comprises: Input the English voice of multiple standard English sample objects; identifying vowels in English speech of the plurality of standard English sample objects; Acoustic analysis of English vowel pronunciation for each sample object's vowel separately; The standard English vowel pronunciation acoustic model is generated according to the English vowel pronunciation acoustic analysis result. 13. The storage device according to claim 11, wherein the step 2 comprises: providing voice materials according to the nationality of the subject under test, and recording the English voice of the subject under test reading the voice material aloud. 14. The storage device according to claim 11, wherein the step 3 comprises: identifying the vowel in the English speech of the object under test according to the resonance peak of the vowel. 15. The storage device according to claim 11, wherein the step 3 comprises: recognizing the vowel in the English speech of the object under test according to the resonance peak of the vowel and the duration of the vowel . 16. The storage device according to claim 14 or 15, wherein the resonance peak comprises a first resonance peak and a second resonance peak. 17. The storage device according to claim 11, characterized in that, said step 7 comprises: according to the data of the English vowel pronunciation acoustic analysis and the standard English vowel pronunciation acoustic model, the visual image Adjust the pronunciation of English vowels of the subject under test. 18. The storage device according to claim 11, further comprising: Recording again the voice of the subject under test reading the recognized vowel sound aloud again; Carrying out acoustic analysis of English vowel pronunciation to the voice read aloud again by the subject under test; The English vowel sounding acoustic analysis data of the speech read aloud again by the subject under test is compared with the standard English vowel sounding acoustic model to obtain a second deviation; Outputting an English vowel pronunciation evaluation text of the subject under test according to the first degree of deviation and the second deviation. 19. The storage device according to claim 11, 12 or 18, wherein the acoustic analysis of English vowel sounds comprises: Measuring the resonance peak of the recorded English vowel sounds; Measuring the vocalization duration of the recorded vowels; The recorded English vowel sound acoustic analysis data is generated according to the recorded resonance peak of English vowel sound and the duration. 20. The storage device according to claim 19, wherein the resonance peak comprises a first resonance peak and a second resonance peak. 21. An English vowel pronunciation error correction device, comprising: a processor adapted to implement the instructions; and a storage device adapted to store a plurality of instructions adapted to be loaded and executed by the processor as: Step 1, pre-store the standard English vowel pronunciation acoustic model; Step 2, input the English voice of the tested object; Step 3, identifying vowels in the English speech of the tested object; Step 4, recording the voice of the subject under test reading the recognized vowel aloud; Step 5, performing an acoustic analysis of English vowel pronunciation on the voice of the measured subject reading the recognized vowel aloud; Step 6, comparing the English vowel pronunciation acoustic analysis data of the measured object with the standard English vowel pronunciation acoustic model to obtain a first degree of deviation; Step 7, correcting the English vowel pronunciation of the subject under test according to the first degree of deviation. 22. English vowel pronunciation error correction device as claimed in claim 21, is characterized in that, described step 1 comprises: Input the English voice of multiple standard English sample objects; identifying vowels in English speech of the plurality of standard English sample objects; Acoustic analysis of English vowel pronunciation for each sample object's vowel separately; The standard English vowel pronunciation acoustic model is generated according to the English vowel pronunciation acoustic analysis result. 23. The English vowel pronunciation error correction device as claimed in claim 21, characterized in that, said step 2 comprises: providing voice materials according to the nationality of the tested object, and inputting the tested object to read the voice English phonetics of the material. 24. The device for correcting errors in English vowel pronunciation according to claim 21, characterized in that, said step 3 comprises: identifying the vowel in the English speech of the subject under test according to the resonance peak value of the vowel. 25. The English vowel pronunciation error correction device as claimed in claim 21, wherein said step 3 comprises: identifying the English language of the measured object according to the resonance peak of the vowel and the duration of the vowel. Vowels in speech. 26. The English vowel pronunciation error correction device according to claim 24 or 25, wherein the resonance peak comprises a first resonance peak and a second resonance peak. 27. The English vowel pronunciation error correction device as claimed in claim 21, wherein said step 7 comprises: according to the data of the English vowel pronunciation acoustic analysis and the standard English vowel pronunciation acoustic model, The pronunciation of English vowels of the subject under test is adjusted with a visual image. 28. English vowel pronunciation error correction device as claimed in claim 21, is characterized in that, after described step 7, also comprises: Recording again the voice of the subject under test reading the recognized vowel sound aloud again; Carrying out acoustic analysis of English vowel pronunciation to the voice read aloud again by the subject under test; The English vowel sounding acoustic analysis data of the speech read aloud again by the subject under test is compared with the standard English vowel sounding acoustic model to obtain a second deviation; Outputting the speaker's English vowel pronunciation evaluation text according to the first deviation degree and the second deviation. 29. The English vowel utterance error correction device as claimed in claim 21, 22 or 28, wherein said English vowel utterance acoustic analysis comprises: Measuring the resonance peak of the recorded English vowel sounds; Measuring the vocalization duration of the recorded vowels; The recorded English vowel sound acoustic analysis data is generated according to the recorded resonance peak of English vowel sound and the duration. 30. The English vowel pronunciation error correction device according to claim 29, wherein the resonance peak comprises a first resonance peak and a second resonance peak.