JPH096221A - Voice synthesizing apparatus - Google Patents

Abstract

PURPOSE: To make it possible to output voice with correct accent and high clearness by providing an accent information dictionary previously containing correlation of strings of letters and accent information to be given to the strings of letters. CONSTITUTION: Machine codes are attached to the right lower part of the surface of a character board 20. A character reading-out part 2 reads out a string of letters from the character board 20 set in a character board setting part 2b. A dictionary searching part 3 searches a dictionary 4 and obtains accent information, e.g. [1], written corresponding to the string of codes which shows the read-out string of letters, e.g. [UMI]. A voice rule synthesizing part 5 synthesizes respective sounds, [U] and [MI], from previously provided data of basic frequencies of respective sounds while giving an accent to the data where the [U] sound exists. Consequently, synthesized sound [U'MI(meaning sea)] having an accent at proper accent position can be sent out to a voice output terminal 7.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech synthesizer.

[0002]

2. Description of the Related Art As an educational toy that applies voice output,
Several things have already been put on the market, including the "Talking Aiueo" that is being sold by the A manufacturer.

FIG. 10 is a diagram showing the basic appearance, and the configuration will be described below.

In the figure, reference numeral 10a is a power switch for turning on / off the power. Reference numeral 10b is a syllable table panel, which is pressed by a finger to give an instruction. Reference numeral 10c is a control panel section for selecting a use mode of the toy such as a recording mode / playback mode / test mode. 10d is a speaker.

The operation of the above arrangement will be described below.

That is, the basic operation is that when a character on the syllable table panel is pressed with a finger, the sound is output as a voice. In record mode, press several syllables in succession to memorize. For example, if "Umi", then "U" and "Mi"
Press continuously. After that, when you switch to playback mode, "U
The voice is output. However, this voice is a so-called robot voice in which the speech units recorded in syllables are simply sequentially output, and the intonation and accent are not correctly controlled. In the test mode, the voice is shown first, and the content is answered on the panel to test the degree of agreement between the voice and the characters. As described above, by using this toy, it is possible to learn the correspondence between the voice and the character, and the fact that the characters are collected to form a word.

FIG. 11 is a block diagram for explaining the operation of the toy in the recording / playback mode.

Reference numeral 10 is an input terminal for inputting a syllable string using a syllable table panel. 2 interprets and holds the syllable string input by the character reading unit. Reference numeral 11 denotes a voice reading unit, which reads voice data of syllables corresponding to the syllable string held in 2 from the single syllable voice data 6a and sequentially outputs to the voice output terminal 7 when the reproduction mode is instructed. To do. In the fixed phrase voice data 6b, "Please press your favorite word."
The voice data of a fixed expression such as the above is stored and output according to the instruction.

[0009]

However, such a conventional voice education toy has the following problems. The first is that so-called robot voices, in which accents, intonations, rhythms, etc., are not added correctly because the syllable segment data recorded as vocalizations are simply connected and output. Secondly, the characteristics of speech are greatly different between the case where a syllable is uttered alone and the case where it is uttered as several consecutive syllables as a word. Since the syllable segment data is simply connected without considering the effect at all, the syllable connection is not smooth and the intelligibility is low. Third, in such educational toys, it is an important task to make them understand that each syllable constitutes a word, but just pressing the syllable table panel in order makes this problem understandable. Is difficult.

The present invention has been made in view of the above-mentioned problems of the prior art, and synthesizes a voice having a more accurate intonation and rhythm and a smoother connection between syllables than in the conventional art. It is intended to provide a speech synthesizer capable of outputting.

Further, in addition to the above objects, it is another object of the present invention to provide a speech synthesizer that makes it easier to understand that syllables form words, as compared with the conventional art.

[0012]

According to the present invention of claim 1,
A code reading unit for reading a code that is input in advance to a character or a word, a predetermined character string, and an accent information dictionary that has a correspondence relationship with accent information to be added to the character string in advance, and the reading A dictionary search unit that searches the accent information dictionary using a code string including one or more of the selected chords and extracts the accent information of the character string corresponding to the code string; and the extracted accent. A voice synthesizing device provided with voice rule synthesizing means for synthesizing voice based on a predetermined rule using information and the code string.

According to the present invention of claim 2, in addition to the above, one character or one word is displayed on the surface, and a plate-shaped member to which the assigned code is attached correspondingly, and the code reading means are incorporated. However, the speech synthesizing device is provided with an arranging means for arranging one or a plurality of the plate-shaped members side by side.

According to the present invention of claim 3, a code reading means for reading a code previously given to the plate-shaped member on which characters or words are displayed, and an arrangement means for arranging one or a plurality of the plate-shaped members side by side. And detecting means for detecting the arrangement state of the plate-like member in the arrangement means, and using the detection result, accent information corresponding to a code string including the read single or plural chords is generated. The speech synthesizing device is provided with: accent information generating means; and speech rule synthesizing means for synthesizing speech based on a predetermined rule using the accent information and the code string.

According to a fourth aspect of the present invention, a code reading means for reading a code previously given to a plate-shaped member on which characters or words are displayed, and an arrangement means for arranging one or a plurality of the plate-shaped members side by side. A detection means for detecting the arrangement state of the plate-like member in the arrangement means, a predetermined character string, and an accent information dictionary having a correspondence relationship with accent information to be given to the character string, and the reading. Dictionary search means for searching the accent information dictionary by using a code string including one or more of the selected codes, and extracting first accent information of the character string corresponding to the code string; and the detecting means. Accent information generating means for generating second accent information corresponding to the code string using the detection result by the first accent information or the previous accent information. The second accent information and the voice rule synthesizing unit for synthesizing the voice based on a predetermined rule using the code string, and the second accent when the detection result satisfies a predetermined criterion. This is a voice synthesizing device that uses information in preference to the first accent information.

According to the present invention of claim 5 to use the detection result by the detecting means, when a plurality of the plate-shaped members are arranged, the plurality of plate-shaped members arranged among the arranged plate-shaped members are used. A speech synthesizer which is to detect an arrangement position of a plate-shaped member pushed deeper than another plate-shaped member by the detecting means and utilize the detected position for position determination of an accent in the chord string. is there.

According to the sixth aspect of the present invention, the use of the detection result by the detection means means that the codes are attached to both front and back surfaces and that the surface can be identified as the front surface or the back surface. When the plate-shaped member to which the code is attached is arranged, the identification code is detected by the detection means, and the identification information obtained from the detected identification code is used as the position of the accent in the code string. It is a speech synthesizer that is to be used for decision.

According to a seventh aspect of the present invention, the detecting means detects that the plate-shaped member is pushed deeply, and utilizes the detection result to detect the voice synthesized by the voice rule synthesizing means. It is a speech synthesizer for outputting.

[0019]

According to the first aspect of the present invention, the code reading means reads a code which is input and which is given in advance to a character or a word, and the accent information dictionary stores a predetermined character string.
Preliminarily having a correspondence relationship with the accent information to be given to those character strings, the dictionary search means searches the accent information dictionary by using a code string including the read single or plural codes, and The accent information of the character string corresponding to the code string is extracted, and the voice rule synthesizing unit uses the extracted accent information and the code string to perform voice synthesis based on a predetermined rule.

According to the present invention of claim 2, in addition to the above, in the plate-shaped member, one character or one word is displayed on the surface, the assigned code is attached corresponding to it, and the arranging means is the code. A reading unit is built in, and one or a plurality of the plate-shaped members are arranged side by side.

Thus, the present invention can provide, for example, a voice educational toy having a high learning effect as an example of a voice synthesizing device having a voice output function with high clarity in which accent, intonation, and rhythm are correctly controlled. It will be possible.

In the present invention of claim 3, the code reading means reads the code previously given to the plate-shaped member on which characters or words are displayed, and the arrangement means arranges one or a plurality of the plate-shaped members side by side. Then, the detecting means detects the arrangement state of the plate-like member in the arranging means, and the accent information generating means uses the detection result to correspond to the code string including the read single or plural codes. Then, the voice rule synthesizing unit uses the accent information and the code string to perform voice synthesis based on a predetermined rule.

According to the present invention of claim 4, the code reading means reads the code previously given to the plate-shaped member on which characters or words are displayed, and the arranging means arranges one or a plurality of the plate-shaped members side by side. Then, the detecting means detects the arrangement state of the plate-like member in the arranging means, and the accent information dictionary has a correspondence relation between a predetermined character string and the accent information to be given to the character string in advance. The search means searches the accent information dictionary by using the code string including the read single or plural codes, and extracts the first accent information of the character string corresponding to the code string to obtain the accent information. The generation means generates the second accent information corresponding to the code string by using the detection result of the detection means, and the voice rule synthesizing means generates the first accent information. Information or the second accent information and the code string are used, voice synthesis is performed based on a predetermined rule, and when the detection result satisfies a predetermined criterion, the second accent information is It is used in preference to the first accent information.

Thus, for example, in the case of a voice toy as an example of a voice synthesizer, even for words that cannot be prepared in advance in the dictionary, such as the name of the user of the toy user, a dial plate as an example of a plate member is displayed. -By providing the correct accent position by using the back or the method of setting the dial, it is possible to provide a voice education toy with high clarity.

In the present invention of claim 7, the detecting means detects that the plate-shaped member is pushed deeply and outputs the voice synthesized by the voice rule synthesizing means using the detection result. To do.

[0026]

Embodiments of the present invention will be described below with reference to the drawings. (Embodiment 1) FIG. 1 is an external view of a voice educational toy according to one embodiment of the present invention, and the configuration of this embodiment will be described with reference to the drawing.

As shown in the figure, 2a is a power source SW for turning on / off the power source of the voice educational toy. The dial plate 20 as the plate member of the present invention is a square plate having a predetermined thickness, and one hiragana character is displayed on the surface thereof. A code (mechanical code) for giving syllable information to the voice education toy device of this embodiment is attached to the lower right portion of the surface of the dial plate 20. This code corresponds to the hiragana characters displayed on the surface of the dial 20. Here, FIG. 2 is a perspective view of the dial used in the present embodiment, and shows an example in which the hiragana character “A” is printed. Reference numeral 2b is a dial setting unit, which sequentially sets dials. Reference numeral 2c is a control panel unit, which sets a use mode of the toy such as a recording mode, a reproduction mode, and a test mode. 4d is a speaker. The code reading means of the present invention includes a dial setting portion 2b and the like. The fixed-sentence voice data 6b stores voice data of a fixed expression such as "please press a favorite word" and outputs it according to an instruction.

As shown in FIG. 1, the toy of this embodiment includes
There are a playback mode in which a plurality of dials 20 are set in the dial set part 2b when they are read out as words, and a test mode in which a correct dial is set according to the voice output from the speaker 4d. is there.

FIG. 3 is a block diagram showing the configuration of the toy in the recording / playback mode.

As shown in the figure, reference numeral 1 is a reproduction processing start instruction input terminal to which an instruction of "voice output" is inputted from a control panel or the like. A character reading unit 2 as a code reading unit of the present invention interprets and holds a character string from the dial set in the dial setting unit 2b. Reference numeral 4 is a dictionary as an accent information dictionary of the present invention, which holds information about notations and accents as described later.
Reference numeral 3 is a dictionary search unit that searches the input character string using the dictionary 4 to obtain information such as accent and reading. Reference numeral 5 is a rule synthesizing unit as a voice rule synthesizing means of the present invention, which synthesizes voices in accordance with reading and accent information and outputs the synthesized voices to the voice output terminal 7. The dictionary 4 is, for example, a dictionary in which correct accent information of words (character strings) generated by arranging hiragana characters is recorded, and they are recorded as follows. That is, the accent information about the character string "Umi" is the coat attached to the dial "U"
Accent information "1" is recorded in a form corresponding to a code string formed by sequentially arranging the codes attached to the dial plate. The "1" indicates that there is an accent position in the first syllable of "umi" represented as a chord string. As another example, in the case of "hashi (bridge)", the accent information "2" is recorded corresponding to "hashi (bridge)" represented as a code string different from the above example. There is. In this case, it is indicated that there is an accent at the second position of "hashi (bridge)" represented as a code string. In the case of an example in which no accent exists (so-called flat type accent) such as "Yes (ant)", the accent information is "0".

Next, a more specific example will be described.

When the dials 20 on which the letters "u" and "mi" are individually printed on the surface are set in the dial set portions 2b, respectively, the character reading portion 2 is a code representing the syllable string "umi". Read the column. This code string is "U"
The code on the dial of No. and the code on the dial of "Mi" are arranged in order. Next, the dictionary search unit 3 searches the contents of the dictionary 4 for the same code string as the read code string. Then, the written accent information "1" is obtained corresponding to the code string representing the character string "Umi".

The voice rule synthesizing unit 5 further uses the accent information "1" and the code string corresponding to the accent information "1" and the "u" and "mi" sounds, respectively, to further add "U" to the basic frequency data. The composition is done with an accent at the position. As a result, it is possible to generate the synthesized sound "Uumi (sea)" with the correct accent position. Here, "'(called a literal)" is attached to the right shoulder of a character to indicate that the character (syllable) has an accent.

By the way, in the above example, one hiragana character (that is, one syllable) is printed on one dial plate, but in addition to this, in addition to the dial plate showing one syllable, the word ( That is, a dial showing a plurality of syllables can also be used. Furthermore, the homophone meaning words whose accent cannot be determined only by the hiragana character string (for example, "hashi '(bridge)",
It is also possible to print "Hashi (chopsticks)" and the like separately on the front and back of the same dial and attach different codes corresponding to each.

Further, the start of voice output may be designated from the control panel, or the voice output may be automatically started when a predetermined time has elapsed after the setting of the dial.
Both are acceptable. It is more preferable to add information such as intonation and rhythm to the dictionary 4 in addition to the accent information.

As described above, the present embodiment comprises a dial which represents one character or one word on one sheet and a device for setting the dial, which device represents the characters represented by the set dial. Character reading means for reading, a character string holding means for holding the read character string, a dictionary means for holding notation and accent information, a dictionary searching means for searching the dictionary means, and a voice rule synthesizing means. It is a voice toy.

As described above, according to the present embodiment, it is possible to solve the problems of the prior art and to output a natural and highly distinct voice with correct accent, intonation and rhythm. In addition, since the operation of setting the dial is involved, it is possible to objectively, physically, and visually understand that the syllable constitutes a word. As a result, the user is able to more clearly understand the relationship between the syllable and the word. (Embodiment 2) FIG. 4 is a perspective view showing a front surface (see the left side in the drawing) and a back surface (see the right side in the drawing) of a dial 20 used in another embodiment according to the present invention. Dial 2
0 is provided with a code for giving syllable information to the toy device and a front / back identification code for identifying the front side or the back side, and the color is arranged so that the front side and the back side have different colors. Has been done. Although FIG. 4 shows the case of the dial having the character "A" printed thereon, the same applies to the dial showing other characters.

FIG. 5 is a view showing the external appearance of this embodiment.
The configuration and operation are almost the same as those in the above-described first embodiment except that a detecting means described later is provided.

FIG. 6 is a block diagram of the toy in the recording / playback mode. In addition to the block diagram of the first embodiment, there is a dial face front / back determination unit 8 as the detection means of the present invention. The dial front / back determination unit 8 recognizes the front / back of the dial 20 from the front / back identification code, and determines the position where the front / back of the dial 20 is first inverted as a position with an accent.

As shown in the figure, in the case of "ko (front) ne (back) ko (front)", it is judged that "ne" has an accent. When the position of the accent can be set from the front and back of the dial, the dictionary search processing is not performed. When the accent positions cannot be determined for the front and back of all the dials, a dictionary search is performed.

As described above, according to the present embodiment, the dials which represent one character or one word by one sheet and have different colors on the front and back sides,
It comprises a device for setting the dial, the device comprising:
A character reading means for reading a character represented by the set dial, a character string holding means for holding the read character string, an accent reading means for reading a different color position of the dial, and a voice rule synthesizing means. It is a voice toy.

As described above, according to this embodiment, even if a word that cannot be prepared in advance in the dictionary 4, such as the name of a child who uses this toy, is correctly and easily accented,
It is possible to output a natural and highly distinct voice with intonation and rhythm.

When all the dials 20 set in the dial set portion 2b are the front surface or the back surface, a flat type accent is added and synthesized to remove the dictionary search portion 3 and the dictionary 4 portion. It is also possible to have a different system. (Embodiment 3) FIG. 7 is a perspective view showing the outer appearance of another embodiment of the present invention, and is substantially the same as the embodiment 1 except that the configuration and operation are provided with a detecting means described later. Is.

In the present embodiment, the dial set portion 2b is constructed so that the dial set can be slightly pushed in.

FIG. 8 shows a block diagram of the toy of this embodiment in the reproduction mode.

In the case of this embodiment, instead of the dial face front / back determination unit 8 of the block diagram of Embodiment 2, there is a dial position determination unit 9 as the detection means of the present invention. The dial position determination unit 9 detects the pressed position of the dial when one dial is pushed deeper into the dial setting unit 2b than the other dials. , The position where the dial is first pushed is determined as the position with the accent. The three dials 20 that have been set
When the dial 20 of "Ne" is pushed in, it is determined that "Ne" has an accent. If the position of the accent can be set from the position where the dial 20 is pressed, the dictionary search process is not performed. However, if there is no dial that has been pushed deeper than the other dials and the accent position cannot be determined by the dial position determination unit 9, the dictionary 4
The above-mentioned dictionary search referring to is performed.

As described above, this embodiment comprises the dial which represents one character by one sheet and the device for setting the dial, and the device reads the character represented by the set dial. The voice toy includes a reading unit, a character string holding unit that holds the read character string, a unit that detects the pushing position of the set dial, and a voice rule synthesizing unit.

As described above, according to the present embodiment, even words that cannot be prepared in advance in the dictionary such as the names of children who use this toy can easily be given natural accents, intonations, rhythms, and natural intelligibility. It is possible to output high sound.

When there is no pressed dial, it is possible to make a simple system by deleting the parts of the dictionary retrieval unit 3 and the dictionary 4 by adding a flat accent and combining them. (Embodiment 4) FIG. 9 is a block diagram showing a configuration of a voice toy according to another embodiment of the present invention in a reproduction mode.

The appearance of this embodiment is the same as that of the third embodiment. However, unlike the above-described embodiment, there is no reproduction process start input terminal 1 and the dial position determination unit 9 detects that the dial 20 is pushed in to start the voice creation process.

Thus, it is possible to easily operate the dial 20 without pressing the dial 20 and operating the reproduction control panel.

As described above, the present embodiment comprises the dial which represents one character by one sheet and the device which sets the dial, and the device reads the character which the dial represents. It has a reading means, a character string holding means for holding the read character string, a means for detecting the indented position of the set dial, and a voice rule synthesizing means. It is a voice toy having a function of detecting and activating a voice synthesizer.

As described above, according to this embodiment, it is possible to solve the problems of the prior art and to output a natural and highly distinct voice with correct accent, intonation and rhythm. Further, since the operation of setting the dial 20 is involved, it is possible to more clearly understand the relationship between syllables and words.

Further, according to the above-described embodiment, it is possible to output a voice with high clarity while associating a kana with a voice and to provide an effective educational toy.

The respective means and the like of the present invention may be realized by software using a computer, or may be realized by using a dedicated hardware circuit having these respective functions.

In the above embodiment, the case where Japanese hiragana is used has been described, but the present invention is not limited to this.
It is possible to realize a voice toy having the same function and effect even if it includes kanji, romaji or English.

Further, in the above embodiment, the case of being configured as an educational toy has been described, but not limited to this, for example,
It can also be realized as an educational tool that can be used for learning English.

Further, in the above embodiment, the case where the dial plate is used as the plate member of the present invention has been described, but the present invention is not limited to this, and may of course be configured as a card-shaped thin plate.

[0059]

As is clear from the above description, the present invention has an advantage that it is possible to output a voice with higher clarity than the conventional one.

[Brief description of drawings]

FIG. 1 is an external view of a voice toy according to a first embodiment of the present invention.

FIG. 2 is a perspective view of a dial according to the first embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a voice toy according to a first embodiment of the present invention.

FIG. 4 is a perspective view of a dial plate according to a second embodiment of the present invention.

FIG. 5 is an external view of a voice toy according to a second embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a voice toy according to a second embodiment of the present invention.

FIG. 7 is an external view of a voice toy according to a third embodiment of the present invention.

FIG. 8 is a block diagram showing a configuration of a voice toy according to a third embodiment of the present invention.

FIG. 9 is a block diagram showing a configuration of a voice toy according to a fourth embodiment of the present invention.

FIG. 10 is an external view of a conventional audio toy.

FIG. 11 is a block diagram showing a configuration of a conventional audio toy.

[Explanation of symbols]

 1 Playback start instruction input end 2 Character reading part 3 Dictionary search part 4 Dictionary 5 Voice rule synthesis part 6a Single syllable voice data 6b Fixed sentence voice data 7 Voice output end 8 Dial face front and back determination part 9 Dial plate position determination part 10 Syllable string Input terminal 2a, 10a Power switch 2b Dial set part 10b 50 syllabary panel (syllable table panel part) 2c, 10c Control panel 2d, 10d Speaker 11 Voice data reading part

Claims (7)

[Claims] 1. Accent information having a correspondence relationship between a code reading unit that reads a code that has been previously given to a character or a word that is input, a predetermined character string, and accent information that should be given to these character strings. A dictionary, and a dictionary search unit that searches the accent information dictionary by using a code string including the read single or plural codes, and extracts the accent information of the character string corresponding to the code string, A voice synthesizing device, comprising: a voice rule synthesizing means for synthesizing a voice based on a predetermined rule using the extracted accent information and the code string. 2. A plate-shaped member having one character or one word displayed on the surface thereof, the plate-shaped member corresponding to the attached code, and the code reading means built-in, and one or a plurality of the plate-shaped members. 2. The speech synthesizing apparatus according to claim 1, further comprising an arranging unit for arranging the two. 3. A code reading means for reading a code previously given to a plate-shaped member on which letters or words are displayed, an arrangement means for arranging one or a plurality of the plate-shaped members side by side, and the arrangement means. A detection unit that detects the arrangement state of the plate-shaped member; and an accent information generation unit that uses the detection result to generate accent information that corresponds to the code string that includes the read single or plural codes. A voice synthesizing device comprising: a voice rule synthesizing unit for synthesizing a voice based on a predetermined rule using the accent information and the code string. 4. A code reading means for reading a code previously given to a plate-shaped member on which letters or words are displayed, an arrangement means for arranging one or a plurality of the plate-shaped members side by side, and the arrangement means. Detecting means for detecting an arrangement state of the plate-like member, a predetermined character string, an accent information dictionary having a correspondence relationship with the accent information to be given to the character string in advance, the read singular or plural A dictionary search unit that searches the accent information dictionary by using a code string including each code and extracts the first accent information of the character string corresponding to the code string, and a detection result by the detection unit are used. Accent information generating means for generating second accent information corresponding to the code string, and the first accent information or the second accent information. Voice information and voice rule synthesizing means for performing voice synthesizing based on a predetermined rule using the code string, and if the detection result satisfies a predetermined criterion, the second accent information is used. Is used with priority over the first accent information. 5. Utilizing the detection result of the detecting means means that a plurality of the plate-shaped members are arranged,
Among the plurality of arranged plate-like members, the arrangement position of the plate-like member pushed deeper than the other plate-like member is detected by the detecting means, and the detected position is the position of the accent in the chord string. The speech synthesizer according to claim 3 or 4, wherein the speech synthesizer is used for determination. 6. The use of the detection result of the detecting means means that the code is attached to both front and back surfaces, and an identification code capable of identifying whether the surface is a front surface or a back surface is attached. When the plate-shaped member is arranged, the identification code is detected by the detection means, and the identification information obtained from the detected identification code is used to determine the position of the accent in the code string. The speech synthesizer according to claim 3 or 4, characterized in that. 7. The detecting means detects that the plate-shaped member is pushed deeply and outputs the voice synthesized by the voice rule synthesizing means using the detection result. The speech synthesizer according to claim 5.