JPH02706B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a foreign language pronunciation training device.

Structure of conventional examples and their problems In recent years, various teaching materials for learning conversation in foreign languages, especially English and American, have been sold in Japan, and language broadcasts on television and radio have become popular. However, learning the correct pronunciation of a foreign language is quite difficult, and even if you think you are pronouncing it correctly, it is often difficult for a native speaker of the foreign language to tell that it is not the correct pronunciation with your own ears. For this reason, objective evaluation becomes possible by extracting the features of the voice, displaying it visually, and comparing it with the model display to determine whether it is correct or not. In practice, in speech training for speech-impaired people, formants are extracted from speech for vowel training, and the first formant F is plotted on the horizontal axis.
1 and the second formant F2 on the vertical axis,
A method of classifying and displaying vowels using combinations of F1 and F2 has been proposed, and for this purpose studies have been conducted on how to easily extract formants.

On the other hand, the progress of semiconductor technology in recent years has been remarkable.
As personal computers have become cheaper and relatively easy for individuals to obtain, it is possible to create a personal computer that incorporates the above-mentioned formant extraction and display functions and has a training function for foreign languages, especially foreign vowels. It will be done.

However, as described above, the configuration of the device that displays only two types of audio information: the model of the target foreign language and the audio features extracted from the practice utterances has the following serious drawbacks as a training device. .

In other words, when the target model and the practice utterance do not match, the problem is that it is difficult to know how to change the pronunciation method to get closer to the target model, and the user has to do it by trial and error. It had

Purpose of the Invention The present invention is intended to solve the above-mentioned problems, and is to provide a foreign language in which it is possible to understand how to get closer to the target model when the practice utterance shows a mismatch with the target model. The purpose is to provide practice equipment.

Structure of the Invention The present invention includes a feature extraction section that extracts features of input speech, a visual display section that displays the features of the speech, and features of the native language speech uttered by the practitioner himself/herself extracted by the feature extraction section. a foreign language feature storage section that stores features of the foreign language speech to be practiced; a foreign language feature storage section that stores the features of the foreign language speech to be practiced; the features of the speech extracted by the feature extraction section; The foreign language practice device is equipped with a control unit that displays on the visual display unit the characteristics of the voice read from the storage unit, and the device displays the characteristics of the voice registered by the practitioner, the characteristics of the foreign language voice, and the voice of the practice utterance. Three types of voice characteristic information are displayed on the same screen, so that the voice characteristics registered by the practitioner can be used as guidelines for changing the speaking method.

DESCRIPTION OF EMBODIMENTS FIG. 1 is a block diagram showing the configuration of a foreign language training device according to an embodiment of the present invention.

In the figure, 1 is a feature extraction unit that extracts features of the voice uttered by the practitioner, for example, formants;
Reference numeral 2 denotes a practitioner characteristic storage unit that stores characteristics of the native language sounds uttered by the practitioner himself/herself, such as the native language sounds uttered by the practitioner prior to practice, such as "a",
Feature extraction unit 1 for “i”, “u”, “e”, and “o”
Memorize features such as formants extracted by
Reference numeral 3 denotes a foreign language feature storage section that stores the features of foreign language speech to be practiced; 4 a visual display section that displays the features of the speech; when the speech feature is a formant, the formant plane is made to correspond to the display screen; . Reference numeral 5 denotes a control unit for displaying on the visual display unit 4 the voice features extracted by the feature extraction unit 1 and the voice features read out from the practitioner characteristic storage unit 2 and the foreign language characteristic storage unit 3; Consists of a microcomputer. 6 is a character pattern storage section that stores character patterns to be displayed, and 7 is a keyboard.

The operation of the foreign language training device configured as described above will be explained in accordance with the practice procedure using FIGS. 1 to 3.

In addition, the five Japanese vowels of the native language voice are assumed to be the characteristics of the practitioner, and the American vowel |ε| is assumed to be the foreign language to be practiced, and the voice characteristics are explained using formants as an example.

FIG. 2 shows an example of a screen displayed on the visual display unit 4 in the embodiment. In FIG. 2, the horizontal axis is the first formant F ₁ and the vertical axis is the second formant F2. In the same figure, "a", "i",
"U", "E", and "O" indicate the positions on the F1 and F2 planes of Japanese vowels uttered by the practitioner himself.
The hatched region of ε indicates the correct pronunciation region of American vowels to be practiced. Further, X indicates the position on the F1 and F2 planes of the features extracted from the practice utterances.

FIG. 3 is a flowchart showing the operation of this embodiment.

(b) Native language utterance registration mode This mode is a mode in which the five vowels of the mother tongue are uttered and registered so that they can be reliably pronounced correctly. The practitioner selects the native language utterance registration mode using the keyboard 7.

(a) When this mode is selected, the control unit 5 instructs the utterance of Japanese vowels to be registered in a predetermined order. This instruction is given, for example, by blinking a lamp.

(b) The practitioner utters "a" according to the instructions of this lamp.

(c) The feature extraction unit 1 extracts formant features when the above-mentioned "a" is uttered.

(d) The extracted voice features are stored in the practitioner feature storage section 2.

(e) Repeat the same procedure for the vowels ``i'', ``u'', ``e'', and ``o''.

(b) Practice mode (f) Select foreign language sounds to practice.

(g) Since the selected foreign language sound is “ε”, the control unit 5 reads out the sound feature of “ε” from the foreign language feature storage unit 3 and also reads the sound feature of “ε” from the character pattern storage unit 6.
The character "ε" is read out from the character "ε", and the correct answer area (the hatched part) is displayed on the screen of the visual display section 4 together with the character "ε" as shown in FIG. At the same time, the control section 5 reads out the voice characteristics of "A" . . . "O" stored in the practitioner characteristic storage section 2, and displays the corresponding characters on the screen.

(h) Next, the practitioner practices saying “ε”.

(i) The feature extraction unit 1 extracts the features of this practice voice uttered as "ε".

(j) Erase the display from the previous practice utterance.

(k) The control unit 5 controls the extracted formant F1,
"X" (i) at the display position on the screen corresponding to F2
Display.

(l) If the result of the practice utterance is “X” (i), the next practice utterance should be “U” or “E”.
Practice by imitating the articulation method used to pronounce.
“U” and “e” are the sounds you uttered,
Since it is something that can be reliably pronounced, we have already learned how to change the way we pronounce it to get it closer to the model "ε".
Similarly, if the position is "X" (ii), practice utterances assuming "a" or "o", and if the position is "X" (iii), "i" or "e", etc. Therefore, it is possible to perform an articulation method that brings the pitch closer to the target "ε" region.

(m) If you want to practice other foreign language sounds, return to step (f).

As described above, according to this embodiment, it is possible to know how to change the vocalization method to get closer to the target model when the practice vocalization and the target model do not match.

Note that in the above embodiment, the first feature of the voice is
Although the second formant is used, other audio features may be used, such as local peaks in frequency. Furthermore, although the embodiment has been described for practicing vowels, it is also possible to practice consonants by registering and displaying native language consonants that can be reliably pronounced. Furthermore, although the embodiment shows an example in which a Japanese person practices American English, the invention can also be applied to foreign language practice between other languages.

Effects of the Invention The foreign language training device of the present invention visually displays three types of audio information: a target model, a sound that the practitioner can reliably pronounce in addition to the practice vocalization sounds, and the target model. This will have a great practical effect on foreign language practice, as you will be able to learn how to change your vocalization method to get closer to the target model when your practice utterances do not match.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a foreign language practice device according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of a visual display screen in the same embodiment, and FIG. 3 is a foreign language practice device in the same embodiment. This is a flowchart showing the steps. 1...Feature extraction unit, 2...Practice feature storage unit,
3...Foreign language feature storage unit, 4...Visual display unit, 5
...control section.

Claims (1)

[Claims] 1. A feature extraction unit that extracts the features of the voice; a practitioner feature storage unit that stores the features of the native language voice uttered by the practitioner himself/herself extracted by the feature extraction unit; and the characteristics of the foreign language voice to be practiced. a visual display section for displaying voice characteristics; a voice feature extracted by the feature extraction section; and a voice read from the practitioner characteristic storage section and the foreign language characteristic storage section. and a control unit that displays the characteristics of the voice on the visual display unit, and displays three types of voice characteristic information on the same screen: the voice characteristics registered by the practitioner, the foreign language voice characteristics, and the voice characteristics of the practice utterance. A foreign language practice device characterized by being configured to display.