JPS6125529A - Tongue position model forming apparatus - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] Industrial Application Field The present invention relates to a speech training device that extracts speech characteristic signals associated with speech and visually displays the feature signals on a display device for speech training. In particular, the present invention relates to a tongue position model creation device for measuring the tongue palate contact pattern during articulation and creating a tongue position model used for training.

Configuration of conventional example and its problems To detect the state of tongue contact during articulation, as shown in Fig. 1, a clock oscillator 2 supplies a minute voltage to a large number of transmitting electrodes (all 63 electrodes in this case) 1. The signal source 3 and distributor 4 sequentially apply a minute voltage to each transmitting electrode, and when the tongue contacts the transmitting electrode of the artificial palate, a minute current flows from the transmitting electrode to the human body and then to the receiving electrode 5. This current change is detected by the signal detector 6, taken into the controller 7, and stored in the memory 8. The stored tongue position data is displayed on the display 9 to monitor the contact state of the tongue.

In speech training, the tongue contact signal based on the tongue contact detection principle configured as described above is inputted on the same screen by the model pattern 1° and in real time by the speaker, as shown in the example display screen in Figure 2. Training is performed by comparing and displaying real-time patterns 11.

In general, model data for speech training may be a pattern that is abstracted and makes it easy to see the points of interest during training. In particular, the tongue contact pattern includes several types of typical putters Z and their respective transition patterns as shown in Figure 4. By using these as training models, sufficient training is possible. There is no need to use faithful pattern information as a model.

Here, since the tongue contact signal is composed of 63 points as shown in FIG. 1, one tongue contact pattern information requires 8 bytes as a memory container as shown in FIG. 3. Speech training requires a dynamic training model of tongue contact signals corresponding to words, and the problem is that the model pattern data of tongue contact at this time reaches several kilobytes, requiring a large memory capacity. .

Purpose of the Invention The present invention solves the above-mentioned conventional problems, and provides a tongue position determination method that classifies tongue contact signals into typical tongue contact patterns and outputs classified tongue contact pattern data as coded information. It is an object of the present invention to provide a tongue position model creation device that can create the coded information as tongue position model data during training.

Structure of the Invention The present invention provides a tongue position measuring device that defines the palate contact pattern of a speaker's tongue, classification of open, narrowed, and closed, and further classification of narrowed and closed into a plurality of categories, and classified tongue contact pattern data. A tongue position model creation device is provided with a tongue position determination unit that outputs the coded information as coded information, and a tongue position storage unit that stores the coded information as tongue position model data of measurement data of the tongue position measuring device, By classifying the contact pattern data and storing it as coded information, it is possible to create many tongue position models with a small storage capacity.In addition, since it is classified only into typical patterns,
When used as a training model, this model focuses on the feature points of the training words, so training can be performed effectively with an easy-to-read, abstracted model.

DESCRIPTION OF THE EMBODIMENT FIG. 6 shows a block diagram of a tongue position model creation device in an embodiment of the present invention. In Figure 6, 1
Reference numeral 2 refers to a tongue position sensor that is attached to the roof of the speaker's palate and detects the tongue contact pattern; 13 refers to a tongue position measuring device that scans the tongue position sensor 12 to measure the tongue contact pattern; and 14 refers to a tongue position measuring device 13. Tongue position classification unit 141 that performs classification of the contact patterns of open, narrow, and closed, with a tongue position determination unit that performs classification judgment on tongue contact data and outputs the classified tongue contact data as coded information; It further includes a narrow fit classification section 142 that classifies into a plurality of narrow fit patterns, and a closure classification section 143 that further classifies closures into a plurality of closure patterns. Reference numeral 15 denotes a tongue position storage unit that stores encoded information of the tongue position determination unit 14.

The operation of the tongue position model creation device of this embodiment configured as described above will be described below. The contact patterns of the tongue position can be classified as shown in FIG. That is, (10) an open state in which the tongue is completely separated from the roof of the mouth, consisting of the vowels Ca) and (o].

(b) Constriction in which both sides of the tongue are in contact with the roof of the mouth, and vowels (i), Ce) and consonants (s)
, (sj), (hi).

(c) Closure with the tongue and roof of the mouth closed, consonant Ct: 1. (d), (n), [r), Cki: 1.

These classifications are performed by the tongue position determining section 14. The tongue position determination process of the tongue position determination section 14 will be explained with reference to the flowchart of FIG.

16...The number n of contact points of the 63 points of tongue position data measured by the tongue position measuring device 13 is calculated.

17... Determine whether or not there is contact at the tongue position.

18...If the number of contact points at the tongue position is n-0, output the tongue position release code CD=o to the tongue position storage unit 160 19...If the number of contact points at the tongue position is 0, Figure 7 23
The presence or absence of contact on the central axis as shown in is determined.

20...Perform the contact determination branching process in 19 above.

21... When there is contact with any electrode on the central axis of the tongue position, the tongue position is further subjected to closure classification processing and the resulting code data is output to the tongue position storage section 16.

22 - If there is no contact on the tongue position center axis, the tongue position is further subjected to a classification process, and the resulting code data is output to the tongue position storage unit 16.

The closure classification processing 21 and the narrowing classification processing 22 can be realized by performing more detailed judgment processing on the tongue position contact data shown in FIG. 7, and therefore their explanation will be omitted here.

As described above, according to the present embodiment, by providing a tongue position determination unit that classifies measured tongue position data into open, multiple indented, and multiple closed, and outputs the results as coded information, it is possible to reduce It is possible to create a large number of tongue position models based on the storage capacity.

For example, when there are 16 types of tongue position classification, for 8 bytes of tongue position measurement data at 63 points, the encoded tongue position model data becomes -bytes, and the data compression performance becomes ``■''.

When using a tongue position model created by the tongue position model creation device of the present invention in a training device, a typical pattern of the tongue position is stored in the training device in advance, and the tongue position model can be trained only by specifying the above code.

Although nine types of typical tongue position patterns are shown in FIG. 4, more detailed patterns may be created by adding transition patterns of each mortar-shaped pattern.

Effects of the Invention The tongue position model creation device of the present invention has a tongue position determination unit that opens tongue position data, classifies it into multiple narrowings, and multiple closures, and outputs the results as coded information. It is possible to create a large number of tongue position models based on the storage capacity.

In addition, by using the typical pattern of tongue position as a training model, it is abstracted and easy to see as a training model.
There is a practical effect on the lateral line.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing the principle of tongue position measurement, Fig. 2 is a diagram showing an example of tongue position display, Fig. 3 is a memory configuration diagram of tongue position data, and Fig. 4 is a diagram showing an example of tongue position display.
The figure shows an example of the classification of typical tongue position patterns, FIG. 6 is a block diagram of a tongue position model creation device according to an embodiment of the present invention, and FIG. 6 is a tongue position classification of the tongue position model creation device of the present invention. FIG. 7, which is a flowchart of the determination process, is a pattern diagram showing the electrode arrangement of the tongue position sensor. 13...Tongue position measuring device, 14...Tongue position determination section, 16...Tongue position storage section. Name of agent: Patent attorney Toshio Nakao and 1 other person 2nd
Figure 3 Figure 4 Cp20C9=j CD=2 Cρゞ3CO=l
CO□7 ””θ (71(d)(n) (and)
(Ka) Figure 5 Figure 6 Figure 7

Claims (1)

[Claims] A tongue position measuring device that measures the fornix contact pattern of the tongue of a speaker, and a tongue position measuring device that classifies the palate contact pattern of the tongue into open, constricted, and closed states, and further classifies the constricted and closed states into multiple types.
A tongue position model consisting of a tongue position determination unit that outputs classified tongue position pattern data as coded information, and a tongue position data storage unit that stores the coded information as a tongue position model of measurement data of the tongue position measuring device. Creation device.