WO2010133072A1

WO2010133072A1 - Pronunciation evaluating device and method

Info

Publication number: WO2010133072A1
Application number: PCT/CN2009/075281
Authority: WO
Inventors: 陈淮琰; 张斌; 周骁
Original assignee: 无敌科技(西安)有限公司
Priority date: 2009-05-21
Filing date: 2009-12-03
Publication date: 2010-11-25
Also published as: CN101551952A

Abstract

A pronunciation evaluating device (1) and method are provided. The device (1) includes: a memorizing module (11) for memorizing a plurality of literal data(111), pronunciation voice information (112) corresponding to the literal data(111), pronunciation parameters (113) and phonetic symbol data (114); an inputting module (12) supplied for the user to select the literal data (111) by the user; a display module (13) for displaying the selected literal data (111); a pronunciation voice information output module (14) for playing the pronunciation voice information (112) corresponding to the selected literal data (111); a recording module (15) for recording the inputted voice information signal (151) by the user; an audio analyzing module (16) for analyzing the inputted voice information signal (151) and comparing the analyzed result with the pronunciation parameters (113) corresponding to the literal data (111) and the phonetic symbol data (114), generating the evaluating result based on the phonetic symbol data(114), and comparing the waveform of the inputted voice information signal (151) and the waveform of the pronunciation voice information (112) corresponding to the literal data (111); the display module (13) is controlled by the audio analyzing module (16) and displays the evaluated result (171) and the waveform compared result(172).

Description

Pronunciation evaluation device and method thereof

The invention relates to a sounding evaluation device and a method thereof, in particular to a sounding evaluation function.

At present, the electronic information industry is developing rapidly. Portable electronic consumer products, such as electronic dictionaries, mobile phones or personal digital assistants, are increasingly favored by many people, and the requirements for the functions of portable electronic consumer products are getting higher and higher. . Whether future portable electronic consumer products can better serve users has become the focus of high-tech product technology development. Among them, voice learning is one of the most frequently used and needed services for many users.

Although the English learning function in the portable electronic consumer products currently on the market has been relatively complete, the synchronization and targeted correction of English learning are not perfect, and the pronunciation problem of English learning has always been difficult to overcome. In the prior art, there have been various pronunciation learning devices and methods, such as a repeater, etc., to improve oral learning as soon as possible by allowing the user to follow up. However, for the user, the learning effect of reading only is limited. Therefore, how to let the user understand and learn the mistakes in the reading, so that mastering the correct pronunciation faster is an urgent problem to be solved.

Summary of the invention

The present invention has been made to solve the above-mentioned technical problems existing in the background art, and proposes a sounding evaluation apparatus and method thereof.

The technical solution of the present invention is: The present invention is a pronunciation evaluation device for evaluating user pronunciation, and the special feature is as follows: The device comprises: a storage module, storing a plurality of text data and a sound information corresponding to the text data, The reading parameter and the phonetic symbol data; the input module provides the user to select the text data; the display module displays the selected text data; the audio output module plays the pronunciation audio corresponding to the selected text data; and the recording module records the audio signal input by the user; The audio analysis module analyzes the input audio signal, compares the analysis result with the pronunciation parameter and the phonetic data of the corresponding text data, generates an evaluation result based on the phonetic data, and compares the waveform of the input audio signal with the sound waveform of the corresponding text data. The audio analysis module controls the display module to display the evaluation result and the waveform comparison result, and the input module, the recording module and the storage module are respectively connected to the audio analysis module, and the audio analysis module The audio output module and the display module are respectively connected.

The audio analysis module described above analyzes the input audio signal according to the LPC cepstrum technique.

The above phonetic data contains syllables, and the evaluation results include evaluation scores corresponding to each syllable. The above text data contains words, words or sentences.

A pronunciation evaluation method for evaluating user pronunciation, which is special in that: The method includes the following steps:

1) providing a plurality of text data and corresponding voice data, audio parameters, phonetic parameters and phonetic data;

2) selecting text data by the user;

3) displaying the selected text data and playing the pronunciation audio corresponding to the selected text data;

4) recording the audio signal input by the user;

5) analyzing the input audio signal, and comparing the analysis result with the pronunciation parameter and the phonetic data of the corresponding text data to generate an evaluation result based on the phonetic data;

6) comparing the waveform of the input audio signal with the waveform of the corresponding text data reading audio;

7) Display the evaluation results and waveform comparison results.

In the above step 5), the input audio signal is analyzed according to the LPC cepstrum technique.

The above phonetic data includes syllables, and the evaluation results include evaluation scores corresponding to each syllable.

The above text data contains words, words or sentences.

The sounding evaluation device and the method thereof provided by the invention first display words to the user, analyze the user's pronunciation after recording the user's pronunciation, and compare the analysis result with the word pronunciation parameters and the phonetic data, and generate the basis The evaluation results of the phonetic data, and comparing the waveform of the user's pronunciation and the waveform of the pronunciation message of the corresponding word, finally display the evaluation result and the waveform comparison result. Thereby, the user is more clearly aware of the problem of his own pronunciation, and improves the efficiency of the user's correction of the pronunciation.

DRAWINGS

Figure 1 is a block diagram of an embodiment of the apparatus of the present invention;

2 is a schematic diagram 1 of a display interface of an embodiment of the device of the present invention;

3 is a schematic diagram 2 of a display interface of an embodiment of the device of the present invention;

Figure 4 is a flow chart of the method of the present invention.

Among them, 1- pronunciation evaluation device, 11-storage module, 111-character data, 112-audio audio, 113- pronunciation parameters, 114-phonetic data, 12-input module, 13 display module, 14 audio output module, 1-5 Recording module, 151-input audio signal, 16-audio analysis module, 171-test result, 172-waveform alignment result.

detailed description

Referring to FIG. 1 , the sounding evaluation device 1 includes a storage module 11 , an input module 12 , a display module 13 , an audio output module 14 , a recording module 15 , and an audio analysis module 16 . The storage module 11 stores a plurality of text data 111 and a read audio signal 112 corresponding to the text data 111 . , the pronunciation parameter 113 and the phonetic data 114. Text data 111 contains words, words, or statements. The pronunciation audio 112 is the correct pronunciation of the corresponding text data 111, and can be used as a reference for the user to learn the pronunciation. The phonetic symbol data 114 contains syllables. For example, the phonetic data of the word "abbreviation" includes five syllables such as "money", violent, ",", "mu," and "Sh ll". The pronunciation parameter 113 may include a pronunciation audio. The LPC cepstrum parameter of 112 or other audio analysis parameters. The storage module 11 can be a built-in memory, a memory card or an optical storage medium.

The input module 12 provides the user with the choice of text data. The input module 12 can be a keyboard, a button group, a cursor controller or a touch module. After the user selects, the display module 13 displays the text data 111 selected by the user, and the audio output module 14 plays the pronunciation audio 112 corresponding to the selected text data 111, so that the user can first listen to the correct pronunciation of the selected text data. The display module 13 displays the phonetic symbol data 114 and other related data of the selected text data.

After the read audio 112 of the selected text data 111 is played, the recording module 15 is activated to record the audio signal 151 input by the user. At this time, the display module 13 can display a prompt message to remind the user to start reading the selected text data 111. Next, the audio analysis module 16 analyzes the input audio signal 151, and compares the analysis result with the pronunciation parameter 113 and the phonetic data 114 of the corresponding character data 111 to generate a result 171 based on the phonetic data 114. The audio analysis module 16 compares the waveform of the input audio signal 151 with the waveform of the read audio 112 of the corresponding text data 111, and controls the display module 13 to display the evaluation result 171 and the waveform comparison result 172. The evaluation result 171 includes the evaluation score of each syllable corresponding to the phonetic symbol data 114. The user views the evaluation result 171 to know which syllable pronunciation has a problem, and further corrects the pronunciation. For example, when practicing the pronunciation of the word "abbreviation", if the evaluation score of ", '§" is lower, the user should pay more attention to the pronunciation and practice more, and observe the waveform comparison result to understand their pronunciation and Where the difference in correct pronunciation is, the effect of effectively correcting the pronunciation has been achieved. The audio analysis module 16 is implemented in a software manner in which the processor executes the related audio analysis program.

Referring to FIG. 2, the pronunciation evaluation device displays the English word "abbreviation" selected by the user to practice the pronunciation, and the pronunciation evaluation device first plays the built-in pronunciation of the English word "abbreviation", so that the user can listen to the pronunciation first and learn against the phonetic data. . After that, the pronunciation evaluation device displays a prompt message to remind the user to read the word, and the pronunciation evaluation device records the user's voice. Referring to Figure 3, after confirming that the user has finished typing, the pronunciation evaluation device is turned on for audio analysis and waveform comparison, and the evaluation scores and waveform comparison results of each syllable are displayed. For example, the user's pronunciation has lower scores in the three syllables such as ",i,,, "i" and "sh weaving". After observing the waveform, it can be found that the users in the three syllables seem to be over-sounding, so Users can understand the shortcomings of their pronunciation.

Referring to Fig. 4, the method can be applied to an electronic device having a signal processing function, such as a computer, a portable electronic dictionary, a mobile phone, or a personal digital assistant (PDA). This method contains the following steps. In step S1, a plurality of character data and pronunciation audio, pronunciation parameters and phonetic data corresponding to the character data are provided, for example, the data is stored in advance in a storage module of the electronic device. The phonetic data contains syllables, and the pronunciation parameters include LPC cepstral parameters of the audio tones or other audio analysis parameters. The storage module is built-in memory, memory card or optical storage media. Next, in step S2, the user selects the text data of the pronunciation to be practiced, displays the selected text data in step S3 and plays the pronunciation audio corresponding to the selected text data, and displays the phonetic symbol of the selected text data on the screen of the electronic device. Data or other relevant data.

After the preset time allows the user to read the text data, the prompt information may be displayed on the screen of the electronic device to remind the user to read the selected text data, and the audio signal input by the user is recorded in step S4, and then the input is analyzed in step S5. The audio signal is compared with the pronunciation parameter and the phonetic data of the corresponding text data to generate an evaluation result based on the phonetic data. The evaluation result includes the evaluation score of each syllable corresponding to the phonetic data, and the evaluation score may be an absolute score, a relative score, or a weight score. In step S6, the waveform of the input audio signal and the waveform of the sound of the corresponding text data are compared. Finally, the evaluation result and the waveform comparison result are displayed in step S7. Among them, the user views the evaluation results to know which syllable pronunciation has a problem, and further corrects the pronunciation. The waveform comparison results give the user a clearer picture of why the pronunciation is poor.

Claims

Claim

A sounding evaluation device, comprising: a storage module, storing a plurality of text data and corresponding sound data, sound reading parameters and phonetic symbol data; an input module, providing a user to select text data; a display module , displaying the selected text data; an audio output module, playing the sound information corresponding to the selected text data; a recording module recording the audio signal input by the user; an audio analysis module, analyzing the input audio signal, and analyzing the result with the corresponding text data Comparing the pronunciation parameters and the phonetic data, generating the evaluation result based on the phonetic data, and comparing the waveform of the input audio signal with the sound waveform of the corresponding text data, the audio analysis module controls the display module to display the evaluation result and the waveform comparison result, The input module, the recording module and the storage module are respectively connected to the audio analysis module, and the audio analysis module is respectively connected to the audio output module and the display module.

2. The sounding evaluation apparatus according to claim 1, wherein: said audio analysis module analyzes an input audio signal according to an LPC cepstrum technique.

3. The sounding evaluation apparatus according to claim 1, wherein: said phonetic symbol data includes a syllable, and the evaluation result includes an evaluation score corresponding to each syllable.

4. The sounding evaluation apparatus according to claim 1, wherein the character data includes a word, a word or a sentence.

5. A sounding evaluation method, characterized in that: the method comprises the following steps:

2) selecting text data by the user;

4) recording the audio signal input by the user;

7) Display the evaluation results and waveform comparison results.

6. The sounding evaluation method according to claim 5, wherein: in the step 5), the input audio signal is analyzed according to the LPC cepstrum technique.

7. The sounding evaluation method according to claim 5, wherein: the phonetic symbol data includes a syllable, and the evaluation result includes an evaluation score corresponding to each syllable.

8. The sounding evaluation method according to claim 5, wherein the character data includes a word, a word or a sentence.