TW468120B - Talk to learn system and method of foreign language - Google Patents

Abstract

The invention provides a talk to learn system of foreign language. Its advantages lay on: (1) integrate voice recognition technique, design various man-machine voice exchange environment, and create an excellent talk to learn environment, allowing user to talk to machine interactively without needing to face other people who speak, further breaking the mental barrier to speak foreign language, acquiring the talk to learn interest of foreign language and increasing the learning efficiency. (2) The system design emphasis targets at letting user speak, so the system design avoids to defy the difficult and weak points of voice recognition technique, and particularly focuses on the pronunciation uniqueness of Chinese to bring up the method in establishing the vocal template of voice recognition appropriate for Chinese pronunciation so that the voice recognition rate in this design is very high, making the man-machine interactive capability higher and the designed learning environment specifically suitable for Chinese. (3) The design opens the ""pronunciation garden"" for pronunciation error correction, providing the standard speaking pronunciation template. The talk to learn system of foreign language in the invention primarily includes five categories: (1) voice recognition module independent to speaker. (2) Talk to learn contents of foreign language. (3) Voice input/output module. (4) Teaching application software for speaking of foreign language and (5) display module.

Description

4 6 81 2 0 V. Description of the invention (l) The foreign language teaching of the present invention, from these products of sentence-making phonetic technology, people who learn foreign languages have better pronunciation. Psychological factors cannot be invented here: .Integrating the environment, creating a machine Interaction breaks the opening and improves learning. 2. The system is designed to avoid picking people's pronunciation characteristics and establish methods, machine interaction, and use. 3. Provided the standard department related to the development of electrical science to writing, which is the most important, not only the people, who dare not open the face to provide the voice of a system of spoken and spoken foreign language efficiency leaves the system voice, proposed to make This book is a foreign product type chapter. The major flaws in reading and speaking are those foreign languages that are good at foreign language recognition techniques that speak foreign languages and foreign languages with other people. The focus is on the identification technology, which is suitable for the Chinese design. The spoken and spoken language learning system and method, there are many 9 types. From vocabulary learning to grammar training to | J writing ’and listening. With the practice also gradually came to the market. But the overview is the environmental issue of oral practice. Many people who speak poorly, even those who have spoken in 0 language, also have many problems. Therefore, the current foreign language teaching products are a good learning method. Speaking learning system. Its advantage lies in the man-machine voice communication ring, which allows users to speak with people who are interested in speaking, so designing 'and especially targeting the constant high of the Chinese recognition acoustic template' increases the number of people suitable for Chinese speakers. Designed a variety of spoken language learning environments that need to face other obstacles. 'Learning environment that has difficulty in speaking to users and weak points. The pronunciation of non-speech recognition rate is abnormal. The pronunciation is wrong.' Foreign Spoken Pronunciation Templates

Page 5 4 6 81 2 0 V. Description of the invention (2) Friends. Faces: Application of language teaching questions. One set is suitable for: ,， 经 部 # Based on sexual or female voice environment 1% Five major achievements: music score. Garden The foreign language oral language learning system of the present invention mainly includes five remote- * speaker-independent speech recognition modules, two, foreign language alpha materials, three voice input / output modules, four foreign language spoken styles, and five display Module. We use the system described in the present invention to set up an environment for Chinese spoken English learning on a computer, named "Talk To Learn". After trials by users, we have achieved good results. First, let us not use M male voice ", or "Women" to select the corresponding male voice recognition acoustic template to improve the recognition rate of tolls. In addition to the two auxiliary functions of "Explanation" and "Exit", the system also includes the ability to: scene dialogue; grammar exercises ; Sentence practice; Words are connected. Hereinafter, embodiments of the present invention will be described with reference to the drawings. Brief Description of Drawings Fig. 1 is a structural diagram showing a foreign language oral language learning system of the present invention. Figure 2 shows a set of "Talk To Learn" masters of the spoken English environment. Figure 3 shows the control flow of "Talk To LearrT". Figure 4 shows the scene dialogue operation interface. Figure 5 is a flowchart showing a scenario dialogue. Fig. 6 is a control flowchart showing a scenario dialogue.

I Figure 7 shows the interface for grammar practice. Figure 8 shows the operation interface of sentence making exercise.

Page 6 4 6 81 2 0

Figure 9 shows the single-word Guan I basket in® / Pingzi Guanyun operation interface: Figure shows the music field garden operation interface. Example H type = learning "ram one __ implementation example shown in Figure 1

The speech recognition module related to the speaker # is suitable for any Chinese users, and they do n’t need to perform complicated a H, go to training beforehand. You can perform man-machine interaction with the computer platform by practicing 'talk and speak'. A speaker independent Voice Recognition Engine is used. You can choose the speech recognition engine of the world's top speech recognition companies such as Dragon system, IBM and other companies. — 1 In order to specifically satisfy the pronunciation characteristics of Chinese users, we must improve the speech acoustic template that is only suitable for European and American languages while selecting the above speech recognition engine, that is, adding Chinese pronunciation characteristics to establish a suitable speech acoustic template. The method is as follows: Step 1: Pick the match! The phoneme provider (ie the recorder) 's selection requirements for recorders are as follows: (1) Different regions: The following can be selected. Ten categories of people: noble and excellent Taiwanese people, Shuangbei, Central China, South China, North China, Northeast Northwest, Southwest, Hong Kong articles, thin Chinese .. people. (2) Different ages:

4 6 812 V. Description of the invention (4) ------- Various types of adults at different ages of 18:60 can be selected. (3) Male-female ratio: Due to the large differences in the pronunciation of male and female voices, in order to avoid the difficulties and weaknesses of sound recognition technology, we must create acoustic templates for male and female ^ respectively, so male and female voice recorders each account for half. (4) Cultural level: It is necessary to select a recording person with a considerable foreign language level, who can read the required recording subject matter correctly and fluently. (5) Number of people: 50 or more male and female are preferred. Step 2: Extract the characteristic values of the sound from these recorded data and establish an acoustic template for speech recognition. 2. The subject matter of oral English teaching in foreign languages. It includes the following aspects: situational dialogues; grammar exercises; sentence making exercises; vocabulary relations and musical notation. Among them, the music score field provides standard spoken pronunciation of words and sentences in all teaching topics. Two • Voice input / ^ output module

Microphone input provides users with voice input. It inputs the user's voice into the computer, and the (speaker-independent) speech recognition engine matches it with a specific template in the acoustic template, that is, speech recognition. ·

Speaker is to provide users with standard spoken foreign language pronunciation. The sound card is the main. To complete the A / D conversion.

Page 8 46 8 彳 20 V. Description of the invention (5) IV. Foreign language spoken language teaching application software It packs and integrates the above functions well, and provides users / users with a convenient, lively and relaxed learning environment. V. Display module The display device displays the subject matter of foreign language oral English teaching and various auxiliary information to remind users to successfully complete various exercises. The process of oral foreign language learning can be described as follows, please refer to Figure 1 (the numbers in the figure are the control steps of oral foreign language teaching): 1-— Step 1. Input the voice ① The user inputs the voice signal through the Microphone. Step 2 · Analog-to-digital conversion ② The sound card converts the voice signal from an analog signal into a digital medicament (input) that can be processed by a computer, and converts the digital speech and sound signals stored in the computer into an analog one-sign (sign). Step 3: Speech recognition (1) ③ The speech recognition module (referred to as the speech recognition module) that is related to the speaker recognizes the input voice signal and compares it with the hypothetical "template" of the "objects". Yes, if it matches the "template" of a certain "target object η11" as r-one one-----'one, the speech recognition result judges that the input is the "object η". ④ Foreign language ρ language teaching The application software (referred to as the application software) judges the voice input as the " target rT. ⑤ The judgment result is displayed on the display device. Step three. Voice recognition (two) ③ The voice recognition module cannot be found _ 佳 .. . "What" object η " " template "

Page 9 4 6 812 0 V. Description of the invention (6)-Matching with the input speech ’, the speech recognition result judges that the input is, wrong input. ④ Application shadow: The body judges the voice input as I, and makes a mistake. ⑤ The judgment result is displayed. Because the focus of this design system is to let the user speak, the design system avoids the challenges and weaknesses in speech recognition. In particular, the foreign language oral teaching theme is designed to have as much pronunciation (acoustic template) as possible: Object " as a hypothetical object. As long as the user's pronunciation is basically close to the candidate " I 象 " template 11, it will be easy to match it, and it will not conceal with other "templates" to the emperor, making the user confident The mind has been greatly improved, and the motivation of the foreign language has been greatly improved. Step 4 • Standard g language output According to the result of step 3 · speech recognition (1), the computer platform will give different results ⑥0⑧, and output different standard spoken and used The talker talks. Ren's content. Outgoing call MW loses W. No. out of the East letter to the phonetic transcription. Oral change Electron tune transfer Ask for a sample number Transcribe the word when reading a pragmatic letter. With the quasi-voices, the weekly card team will be able to release the words, but the answer is simply to use / Φ © ⑧ to make clauses. We use the system described in the present invention to establish a suitable match on the computer. The environment for Chinese spoken English learning, named "Talk To Learn", has been tested by some users and has achieved good results. The details are as follows:

ηίίΐί m

ill

Page 10 4 6 812 0 ----------- V. Description of the invention ^ " " ------- First let the user choose according to sexuality " 男 查 ", Or, female voice ", use the corresponding male-lead or female voice recognition acoustic farming competition, so ^ high cognition [? 1. Yu and then the main face, as shown in the second group _, except" quotation " ; And „Retreat, M. In addition to the two auxiliary functions, the system also contains the five Taiyi one function ... Ability: Psalm pair, Xiangxiang practice; Sentence practice; Single-word connection; Music score garden. The flow of the control system running as shown in Figure 3 ^ ', the following five functions are described separately: 1. Scenario dialogues Scenario dialogues are simulations of various scenarios in real life, allowing the user and Deng to conduct Interactive communication. As shown in Figure 4. Situational dialogue operation interface, after the computer asks a question in the "question area", the user can choose one of the answers borrowed from the system to answer. Because these answers are not correct, It just simulates various possible situations, so that the person can answer according to their own feelings; and the system ... will also judge by speech and speech recognition, and conduct the corresponding dialogue based on your answer, so it can be achieved Interactive dialogue between users and the system. Let ’s take a look at the design of the topic of conversation with “Introduction” as an example. I topic: Introduce yourself

Introduction:

Bobby has found that Robot Computer is friendly. So Bobby wants to make friends with him. Robot is also interested in Bobby. So he is going

4 6 81 2 0 V. Description of Invention (8) to ask some quest ions about Bobby and Bobby ’s family.

Dialog:

Robot: 氺 1: Can you tell roe something about yourself? * 2: First, what's your name? ^ 3: Then how old are you? 氺 4: But I really want to know. * 5: I guess you're 6 ， Right?: Then you are 12 now, right?

Page 12

Bobby: 1 a: Yes, certainly. Lb: All right. 1 c: What do you want to know? 2a: My name is Bobby. 2b: I'm Bobby. 2c: Bobby, BOBBY, Bobby. 3a: I ' m 12 years old. 3b: It's a secret. 3c: Just gU £ ^ JS. 4a: I'm 12. ^ 6 81 2 0 Five 'Description of invention (9) 4b: Six years ago I was 6. 4c: Don 't ask people about their age! 5a: Yes, you're right. 5b: Yeah! You've guessed it! 5c: Right, I'm 12. 6a: Yes, i * m 12 years old. 6b: You 're good at math. 6c: How clever you are! And for each _machine_device pronunciation-one possible dialogue is the language, where the number is the pronunciation of the machine, the sentence, there are three possible dialogues. This sound identifies the object to be identified. For example, the machine! * 1 sentence, the possible dialogue is 1 & or ^ bite, and the device then speaks different sentences according to-different answers', so as to achieve the function of love it. The data on the live route is shown in Figure 5, and the rectification m n k is as follows (refer to the map and Figure 6): a control process of the live process Step 1 * Turn in.. ① The user inputs the voice message of a certain object through Microphone. 2. Analog-to-digital conversion. ② The sound card converts the voice signal from analog signal to digital signal (round rotation) and the computer storage and juice processing machine is replaced by Analog signal (output). Secret Digital Voice Signal Transfer

4 6 81 2 0 V. Description of the invention (ίο) ③ The speaker-independent speech module (referred to as speech recognition module) recognizes the input of the speech signal 'and it is related to the hypothetical " object " " Template M is compared. If the match with a " template of the residual image " is successful, the speech recognition result determines that the input is a V object n ". ④ The foreign language oral teaching software. The software (referred to as the application software) judges the speech input as "the target object". ⑤ Display the judgment result on the display device. Step 3. Speech recognition (2) ⑥ The speech recognition module cannot find any of the "templates" of the two objects Π "that match the input speech, then the speech recognition result judges that the input is 11 wrong input ". '..' ， © Application software. The voice input is judged as “wrong input”. ⑧The judgment result is displayed. Step 4 · Standard spoken language output. According to the third step. Speech recognition (a) of the “Ziguo” computer half .. will give different The result is ©© ⑥, and different standard spoken language is output to talk with the messenger .. The user can also request the computer platform to give the standard alpha pronunciation of any sentence / word in the conversation content at any time. ⑥ The application software calls the standard spoken language壹 一 库. @ Is a sound card that converts digital signals to analog signals and outputs them. ⑧ It is a system for playback. — As shown above-just ~ is the content of self-introduction in the scene dialogue. Other parts such as' , Xingsheng Hobby " etc., no more detailed list here

Page 14

β 81 P V. Description of the Invention (li)-Examples. In this kind of situational dialogue that simulates real life, I believe that it will allow users to speak up, and improve their speaking skills in life. 2. Grammar exercises Grammar exercises allow users to practice oral English while practicing English grammar. The specific operation interface is shown in Figure 7. The grammar practice operation interface ’Among them, the topics listed in the " topic area " are as follows:

He is surprised the Robot can speak

English. (A) which (b) so (c) because (d) by Four of the answers are to be identified-objects, and the control process is similar to the contextual dialogue. It can be seen that the topics listed in the topic area allow the two to practice grammar week. The usual practice is to choose the number of the correct answer. But in this environment, it is the answer chosen by the user. In this way, the system can perform speech recognition on this, so as to determine whether the rabbit used the test to answer correctly. In this situation, the ‘use test has both practiced grammar’ and spoken language, and thus has obtained a comprehensive practice of grammar and spoken language. 3. Sentence-making exercises Sentence-building ..._. Xi is to provide a set of words or phrases, and use the test to combine them into a complete sentence. Its interface is shown in the second figure, and a set of (four) words or phrases are listed in the subject area, so that users (they test them into sentences. It is different from traditional interactive methods (such as keyboard or mouse input) Is using

See page 15 46 81 2 0 V. Description of the invention (12) The person can read the corresponding word or phrase _ on the Internet to make a sentence. As in the above example, the four words or phrases are: (a) he is surprised (b) can speak English (c) the Robot (d) because when the user reads any of the Tan phrases, the system will perform speech recognition 'and The recognition rabbit_question is displayed in the " meal question area. &Quot; In this way, after reading aloud four groups of words or phrases, a complete sentence will be displayed in the " answer area " and can tell the user whether the answer is correct. For example, if the sentence after making a sentence is surprisedhe is surprised because the Robot can speak -—— .. Eng 1 i sh "', then the result of this sentence is correct. Therefore, when doing the sentence making exercise', I also got oral practice. Four * words related

The connection of the word f is to make the person who finds it agree with the known word or righteousness, =, so as to grasp more words and lay the foundation for the spoken dialogue. The single-word related operation interface is not as good as the j9 picture. The single-word related operation interface is not as good as the first picture. The left, ,, and the question! Area "lists a group of single words,-and the 1 answer compliance area" on the right has a sequence Messy antonyms. Ask the person who writes the correct antonym for each word. It looks like this: 1 η foolish

friend clean c1 ever dirty- light enemy

out V. Description of the invention (13) During heavy practice, when a word of 矣 j® is reversed (as shown in the figure, the user must find the antonym of the word 铉 on the right. Make ^ = will choose the f force antonym to pass Read into the system in turn, let the system enter the language recognition, and then a line between the two words. 胄 Complete part :: This completes the ^ character connection this time. At the same time, the system can show whether the answer and = True. In this way, the user can also exercise ^ when doing single-word running. The control process is similar to scenario H e】】-Five. Music spectrum field $ An exercise on the soil will give users a chance to exercise Spoken language. In order to allow users to master the correct reading of the words or sentences in each exercise, the practice of designing the musical notation field f is designed. In any of the above four exercises, the eighth must let the king know how to pronounce a word or sentence. He can train for these waivers or phrases to improve the level of the heart a. The operation interface of the music score field is shown in Figure 10 in the music score field. For example, when the user is in situation In the dialogue, when we want to say "we lj", if he ca n’t read or thinks that the reading is not accurate, he can enter the music field for training. As shown in section ι〇_ =, in the "standard pronunciation waveform area" , Will display the standard pronunciation waveform, and there are M standard pronunciation playback keys ", so that the user can listen to the standard pronunciation. However," users can press the "Record button" to record after learning. After recording Will display the waveforms of the two households' tones in the "User's hair, tone waveform area", and let the households compare with the standard .pronounced waveforms to know which parts are read correctly and which two parts

Page 17 Fifth, the invention description (14) You may also read and listen to the two households accurately. After this, you can master the word or the functions described above, and take a comprehensive look with the present invention. What kind of exercises in the function, he is to perform foreign language speaking in the context of the present invention. Although the present invention limits the present invention, within the scope and scope, when the range is good enough, it can be attached after viewing. What is the difference in pronunciation? After repeated comparisons / training, the correct phrase is pronounced.的 The five major functions of the learning environment are not relevant ', so the major functions are not described in detail here. We found that (except that the environment allows the messenger and the machine to be integrated into oral practice), both must be answered in spoken language. &Quot; A foreign language "Spoken language learning system" will be very effective in learning the language. It has been revealed in a preferred embodiment how those skilled in the art can make changes and retouching without making any changes. Because the scope defined in the scope of the patent application is compared, it is not their own. Whether the pronunciation is believed or not, the user must have another pair of contextual dialogue interactions, regardless of the use of such one. Believe in the result. However, the benefit of the present invention is that the present invention can provide the outside story, or the environment king. Used spirit

Claims (1)

B312J1_New 8_6? 〇 6 years of supplementary __ Sixth, the scope of patent application 1. A foreign language spoken language learning system, including: a voice input / output module, the voice input / output module can input users Speech signal and output standard spoken pronunciation; a speaker-independent speech recognition module with a plurality of hypothetical object templates to recognize the input speech signal and a plurality of hypothetical Object template comparison; a foreign language oral teaching application module, by which the speech signal received by the speech input / output module and the speaker-independent speech recognition module are compared as a specific object ; And a display module 'displays the determination result of the foreign language spoken language teaching application module. 2 The foreign language spoken language learning system described in item 1 of the scope of patent application, wherein the input / output module further includes a sound card, which can convert a voice signal from an analog signal into a digital signal that can be processed by a computer and store the computer Digital voice signals are converted into analog signals. 3. The foreign language spoken language learning system according to item 1 of the scope of patent application, wherein the speaker-independent speech recognition module further includes a speaker-independent speech recognition engine and a Chinese speech acoustic template. 4. The foreign language spoken language learning system as described in item 3 of the scope of patent application, wherein the speaker-independent speech recognition engine is a speech recognition engine of IBM Corporation. 5. The foreign language spoken language learning system according to item 3 of the scope of patent application, wherein the speaker-independent speech recognition engine is a speech recognition engine of Dragon System. 6. Spoken foreign language learning system as described in item 3 of the scope of patent application 0549-5270TiVFl.pt c page 19 46 8120 t still amended / corrected / added __ case number 8910767] amended month and day, six, the scope of the patent application one of the 'the Chinese voice acoustic template, its establishment method , Including: selecting people from different regions, different ages, and foreign language abilities as phoneme providers; and ~ extracting the sound feature values in the recording data of these phoneme providers, and establishing an acoustic template for speech recognition. 7. The foreign language spoken language learning system as described in item 1 of the scope of the patent application, which further includes a foreign language spoken language teaching subject matter, providing standard spoken pronunciation in all teaching subjects. 8. The foreign language spoken language learning system described in item 1 of the scope of patent application, which includes a method of contextual dialogue. This method is to ask a question from the question area. The user can choose one from the answers provided by the system. Answer, because these answers are not right or wrong, they just simulate various possible situations, so users can answer based on their own feelings, and the system will judge by voice recognition and conduct corresponding dialogue based on the answers. , And implement interactive dialogue exercises between users and the system. 9_ Foreign language spoken language learning system as described in item 丨 of the patent application scope, which includes 'the method of grammar practice, this method is to ask a question for the user to practice grammar from the subject area'. The user can provide it from the system. Pick one of the answers to answer, so that the system can perform speech recognition on it, so as to determine whether the user answered correctly. 10. The foreign language spoken language learning system as described in item 1 of the scope of the patent application, where 'also' includes a method of making sentences. This method is to provide a set of words or phrases in the subject area, and let users combine them into a complete Sentence 'When the user reads any one of the phrases aloud, the system will perform speech recognition and display the recognition results in the answer area, after reading all the words or words aloud 0549-5270TOl.ptc Page 20 4 6 81 2 0 6. After applying for a patent, a complete sentence will show whether the answer is correct. It is not displayed in the problem area and can be used. 11 · If the scope of the patent application is applied, it also includes the party related to the word: So; m language learning system '-group of words, and there is a method in the answer area. Listed in the problem area For the word, the user must find the antonym on the right or the synonym user will choose the antonym with the same name; the son ’s antonym or synonyms' make the system perform speech recognition, and then after the j is finished, the system will display whether correct. Day-to-day line between lists, and made by 12 'foreign language spoken language learning system described in item 1 of the patent scope, where' more includes music scores, gardens, and squares, and the β area will significantly dry up. Method 疋 In the standard pronunciation waveform, the dagger will show the waveform of the standard car pronunciation, and the user can listen to the standard pronunciation. &Amp; User :: Put the key to make the recording record * 德 # 合 ym Users can also press The record button performs recording, and the & user pronunciation waveform area displays the waveform of the user's pronunciation. 13.—A foreign language spoken learning method, including: output input / output module 'by performing the voice input / input of the user's voice signal and rotating out of the standard spoken pronunciation; track one — speaker-independent speech recognition The speech recognition mold tantalum with right pass 轱 桐 昱 #consider, and the speaker who should run into the speaker > Beer. * Object template with the last few false di, to perform more hypothetical object templates Compare; The plural letters in the two quote groups subscribe to a foreign language spoken language teaching application module, and also stand up to the twisted f received by the output module. The 5 " input / module is compared to the speaker. The result of the speech recognition comparison is determined to be-a specific object; and a display with extremely fine -_ JJ ---_, and Wei II 7f ^ ^ ^ Spoken teaching application module determination 4 6 812 0 Public \ Amendment / Correction / Addition of "佟 正" __ Case No. 89107671__ £ _ -m- VI. Results of Patent Application. 14. The method of learning foreign languages as described in item 13 of the scope of patent application, wherein the input / output module further includes a sound card, which can convert a voice signal from an analog signal into a digital signal that can be processed by a computer and a computer The stored digital voice signals are converted into analog signals. 1 5. The spoken foreign language learning method according to item 13 of the scope of the patent application, wherein the speaker-independent speech recognition module further includes a speaker-independent speech recognition engine and a Chinese speech acoustic template. 16 · The spoken language learning method for foreign languages as described in Item 15 of the scope of the patent application, wherein the speaker-independent speech recognition engine is a speech recognition engine of 1BM. 1 7 ‘A foreign language spoken language learning method as described in item 5 of the patent application’, wherein the speaker-independent speech recognition engine is a speech recognition engine of Dragon System. 18. The method for learning spoken foreign languages as described in Item 5 of the scope of the patent application, wherein the method for establishing the Chinese speech acoustics template includes: selecting people from different regions, different ages, and foreign language abilities as phoneme providers; and The sound feature values in the recorded data provided by the phonemes are extracted to establish an acoustic template for speech recognition. 1 9. The foreign language spoken language learning method as described in item 3 of the patent application scope, which further includes performing a foreign language spoken language teaching subject matter 'to provide standard spoken pronunciation in all teaching topics. 2 0. The method for learning spoken foreign languages as described in item 丨 3 of the patent application park, which further includes a method of contextual dialogue, which is based on the problem area. 0549-527 (OTFl.ptc Page 22 46 812 u Amend the year A's external bill. 〖Modify / Correct / Fill the month__ Case No. 89107671 VI. After a patent application park asks a question, the user can download it from the system Choose one of the answers provided to answer. Because these answers are not right or wrong, they just simulate various possible situations, so users can answer according to their own feelings. The system will also judge by voice recognition, and According to the answer, the corresponding dialogue is carried out, and the interactive dialogue practice between the user and the system is realized. 21. As described in item 13 of the scope of the patent application, a foreign language learning method 'where' includes a method of grammar practice, which The method is to ask a question for the user to practice grammar from the question area. The user can choose one of the answers provided by the system to answer, so that the system can perform speech recognition on it to determine whether the user answered correctly. 2 2. As for the foreign language spoken language learning method described in item 丨 3 of the scope of application for patents, among which, the method of sentence making exercises is also included, In this method, a set of words or phrases is provided in the question area, so that the user can combine them into a complete sentence. When the user reads any of these words aloud, the system will perform a voice $ Don 'and display the recognition result in the answer area. After reading all the words or groups aloud, a complete sentence will be displayed in the answer area, and the user can be told whether the answer is correct. 23. A foreign language spoken language learner as described in item 13 of the scope of patent application = The method of associating with the early bracketing 'This method is to list the list of benefit groups in the title area' and the answer area has a set of antonyms or two, and the user must find the antonym or synonym of the word on the right : ， The person who has input the selected antonym or synonym into the system by reading aloud for system speech recognition, and then a line between the two words, the system shows whether the answer is correct after doing yuan. 0549-5270TOFl.ptc Page 23 Apply for a foreign language spoken language as described in 3 items of the patent 46 8120 ~ M ^ 671_4- Month Amendment VI. Patent Application Scope Law, which also includes the method of sound spectrum garden, this method is standard The waveform of the standard pronunciation is displayed in the pronunciation waveform area, and the standard pronunciation playback key allows the user to listen to the standard pronunciation. The user can also press the record button to record. After the recording, the user pronunciation is displayed in the user pronunciation waveform area. Waveform. 0549-5270TWFl.ptc Page 24