JP3754750B2 - Accent setting method and apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an accent setting method and apparatus in which a user can set an accent of a registered word in registering a word in a speech synthesizer.
[0002]
[Prior art]
In general, in the standard accent, there are the following rules for the accent phrase of the N-Mora word.
(1) The accents of the first and second mora are inverted.
(2) An accent has one accent nucleus (from a high state to a low state).
(3) The accent of a word having no accent kernel is assumed to be 0 type.
(4) If an accent nucleus is present in the N-th order, it is assumed to be N-type.
[0003]
Specifically, an accent type as shown in FIG. 7 may be used for a 4-mora word “speech”. Here, FIGS. 7 (1) to (5) show the height of the accent of each mora in the accent 0 type to the 4 type, assuming that the hatched mora is low and the non-hatched mora is high. For example, FIG. 7A shows that the accent of mora “o” is low and the accents of mora “n”, “se”, “-”, and “ga” are high. The reason why the mora “ga” is added here is to distinguish the 0 type from the 4 type.
[0004]
A user interface in the conventional accent registration method will be described with reference to FIG. An input panel 30 is displayed on the display unit 12. The user inputs a notation into the notation input unit 31 using a keyboard or the like, and inputs a reading into the reading input unit 32. Here, the notation “speech” and reading “onse” are input. Then, the accent setting unit 33 divides the reading input to the reading input unit 32 into “m”, “m”, “n”, “se”, “−”, and adds a predetermined mora “ga” to the mora. A column is displayed. Here, type 0 (FIG. 5 (1)) is displayed. When the mouse is moved to a mora indicated by the user and clicked, the height of the mora's accent is reversed and the presence or absence of hatching is reversed. In order for the user to set a type 1 accent (Fig. 5 (2)) for this word, five operations to reverse the mora "o", "n", "se", "-" and "ga" Is required. When the user sets an accent from the first type (FIG. 5 (2)) to the second type (FIG. 5 (3)), two operations are required to reverse the mora “o” and “n”. When the user sets an accent from the 0 type (FIG. 5 (1)) to the 2 type (FIG. 5 (3)), the mora “se”, “−”, and “ga” are reversed. The accent set by the user in this way is newly registered in the dictionary 14 and used for voice synthesis and voice output of other application systems.
[0005]
[Problems to be solved by the invention]
With such a user interface, it takes time and effort when a user registers an accent without using a standard accent setting rule. Further, a user who does not know the standard accent rule sets an accent that does not follow the setting rule.
[0006]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, the present invention controls to display the input word reading divided into units of mora , instructs the standard accent setting mode, and sets the accent phrase delimiter for the word On the display, a designation for designating an arbitrary mora for the divided accent phrase unit is accepted, and when the standard accent setting mode is instructed, the specified mora and the height of the accent of each mora are In accordance with the accent setting rules determined according to the information, the accent of reading of each accent phrase is set, and when the standard accent setting mode is not instructed, the level state of only the instructed mora is changed to An accent setting method and apparatus for controlling to change the display of each changed mora are provided.
[0007]
In order to solve the above problems, the present invention preferably registers the set accent.
[0008]
In order to solve the above-mentioned problem, the present invention preferably defines an accent setting rule according to the indicated position of the mora and the height state.
[0009]
In order to solve the above-mentioned problem, the present invention preferably has a standard accent setting mode, and when the standard accent setting mode is instructed, an accent setting rule determined according to information of the instructed mora Then, the accent of the reading of the word is set, and when the standard accent setting mode is not instructed, the level of only the instructed mora is changed.
[0010]
In order to solve the above problems, the present invention preferably displays an attached mora that does not affect the accent of the word in addition to the input word reading divided into mora units.
[0011]
In order to solve the above-mentioned problems, the present invention preferably stores the accent setting rule in advance in correspondence with the accent type.
[0012]
In order to solve the above problems, in the present invention, the word preferably has a plurality of accent nuclei.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a functional block diagram of an apparatus for carrying out the present invention. The apparatus includes a control unit 11, a display unit 12, an operation unit 13, and a dictionary 14.
[0015]
FIG. 2 is a configuration diagram of an apparatus that implements the present invention. Reference numeral 1 denotes a CPU that implements the control unit 11, and processing according to the present invention including a flowchart described later according to a control program stored in the memory 3. To control. Reference numeral 2 denotes a display unit that realizes the display unit 12 and can display characters / images such as a CRT or a liquid crystal display. A memory 3 includes a dictionary 14, a control program, and a work work area. Reference numeral 4 denotes a keyboard and a mouse for executing the operation unit 13. Reference numeral 5 denotes a voice input / output unit, which can output a voice registered according to the present invention. Reference numeral 5 denotes a bus that enables data flow between the components.
[0016]
In the present invention, a standard user interface is used when setting accents for words, and a user interface is provided that does not require much trouble when setting accents.
[0017]
A user interface according to the present invention will be described with reference to FIG. An input panel 30 is displayed on the display unit 12. When setting the standard accent, the standard accent setting button 34 is clicked. By clicking this button 34, the setting is switched ON / OFF. If it is ON, the button 34 can be identified by marking it. Here, the “34” mark is displayed on the button 34, and it can be seen that the standard accent setting is ON. The user inputs a notation to the notation input unit 31 using the keyboard and inputs a reading to the reading input unit 32. Here, the notation “speech” and reading “onse” are input. When the focus is moved to the notation input unit 31, the input mode automatically enters the kanji kana mixed mode, and when the focus is moved to the reading input unit 32, the katakana input mode is automatically set. The accent setting section 33 displays a mora sequence in which the reading input to the reading input section 32 is divided into “o”, “on”, “se”, and “-” in units of mora, and a mora “ga” is added thereto. . The added mora “GA” is selected in advance as a mora that does not affect the accent of the word and is stored in the memory 3. In the accent setting section 33, type 0 (FIG. 3 (1)) is displayed as a default. When the mouse is moved to a mora designated by the user and clicked only once while the standard accent setting button 34 is ON, the height of the accent of each mora is inverted according to the standard accent rule. When the user sets a type 1 accent (FIG. 3 (2)) for this word, clicking on the mora “o” with the mouse will cause the mora “o”, “n”, “se”, “-”, and “ Is reversed. When the user sets an accent from type 1 (FIG. 3 (2)) to type 2 (FIG. 3 (3)), when mora “n” is clicked with the mouse, mora “o” and “n” are inverted. When the user sets an accent from the 0 type (FIG. 3 (1)) to the 2 type (FIG. 3 (3)), clicking on the mora “se” with the mouse, the mora “se”, “−” and “ "Is reversed. When the user sets an accent from type 2 (FIG. 3 (3)) to type 1 (FIG. 3 (2)), clicking on mora “o” or “n” with the mouse will result in mora “o” and “n”. "Is reversed. When the user sets an accent from type 2 (FIG. 3 (3)) to type 0 (FIG. 3 (1)), clicking on mora “ga” with the mouse will result in mora “se”, “-” and “g”. "Is reversed.
[0018]
A process from when a mora is clicked to the accent type is determined and set based on the clicked mora and the current situation and registered in the dictionary will be described using the flowchart shown in FIG.
[0019]
Here, a case will be described in which an accent is set for an N-mora word input to the reading input unit 32. At this time, the accent setting panel 33 displays N + 1 mora rows as described above. The displayed contents are all stored in the memory 3 such as notation, reading, accent height of each mora, ON / OFF state of the standard accent setting button, and updated whenever the state changes. In step 41, the user clicks the M-th mora with the mouse. In step 42, the height of the accent of the M mora eyes displayed is determined. If it is determined in step 42 that the accent of the M-mora is high, it is determined in step 43 whether M = 1. If it is determined in step 43 that M = 1, the accent to be set in step 44 is determined to be type 2, and the level of each mora is determined based on the type 2 rule, and the display is updated. If it is determined in step 43 that M = 1, it is determined in step 45 whether M = 2. If it is determined in step 45 that M = 2, the accent to be set is determined to be type 1 in step 46, the level of each mora is determined based on the type 1 setting rule, and the display is updated. If it is determined in step 45 that M = 2 is not satisfied, the accent to be set in step 47 is determined to be M-1 type, and the level of each mora is determined and displayed based on the M-1 type setting rule. Update. If it is determined in step 42 that the accent of the M mora eyes is low, it is determined in step 48 whether M = 1. If it is determined in step 48 that M = 1, the accent to be set in step 49 is determined as type 1, and the level of each mora is determined based on the type 1 setting rule, and the display is updated. If it is determined in step 48 that M = 1 is not satisfied, it is determined in step 50 whether M = N + 1. If it is determined in step 50 that M = N + 1, the accent to be set in step 51 is set to 0 type. Judgment is made, the level of each mora is determined based on the 0 type setting rule, and the display is updated. If it is determined in step 50 that M = N + 1, the accent to be set is determined to be M type in step 52, the level of each mora is determined based on the M type setting rule, and the display is updated. In step 51, the accent set by the user in this way is registered in the dictionary 14 and used for voice synthesis and voice output of other application systems. Each accent type setting rule is stored in the memory 3 in advance.
[0020]
In the flowchart of FIG. 4, the processing based on the assumption that the standard accent setting is ON has been described. However, after the mora is clicked in S41, it is determined whether the standard accent setting is ON or OFF. The processing after S42 may be continued, and when OFF, only the mora clicked at S41 may be reversed.
[0021]
An example in which the present invention is applied to a case where a word having a plurality of accent nuclei is input will be described with reference to FIG. An input panel 30 is displayed on the display unit 12. The user inputs a notation to the notation input unit 31 and inputs a reading to the reading input unit 32. Here, the notation “total profit” and the reading “Zen | Rieki” are input from the keyboard. Here, the symbol “|” indicates an accent phrase delimiter. By reading this symbol and inputting it from the keyboard, the user determines and specifies the accent phrase delimiter. The accent setting section 33 displays a mora sequence in which readings are divided into “m”, “m”, “n”, “|”, “li”, “d”, and “ki”, and a mora “ga” is added thereto. Here, type 0 (FIG. 6 (1)) designated as the default for both the accent phrases “Zen” and “Rieki” is displayed. When the user moves the mouse to the mora indicated in each accent phrase and clicks once, the height of the accent of each mora is inverted in accordance with the standard accent rule in each accent phrase. When the user sets an accent (“(2))” of type 1 for “Zen” and type 1 for “Rieki”, when the user clicks mora “Ze” with the mouse, the mora “Ze” “ When “M” is reversed and the mora “Li” is clicked with the mouse, the mora “Li” “D” “K” “G” is reversed (FIG. 6 (2)). Accents set by the user in this way are registered in the dictionary 14 and used for voice synthesis and voice output of other application systems.
[0022]
【The invention's effect】
As described above, according to the present invention, word accents can be registered with a small number of operations.
[Brief description of the drawings]
FIG. 1 is a functional configuration diagram of an apparatus that implements the present invention. FIG. 2 is a configuration diagram of an apparatus that implements the present invention. FIG. 3 is a diagram showing a display unit. Fig. 6 shows the display unit. Fig. 7 shows the standard accent.

Claims (12)

Display control means for controlling the input word reading to be displayed divided into mora units;
Holding means for holding height information of the accent of each mora;
An instruction means for instructing a standard accent setting mode;
Delimiter setting means for setting an accent phrase delimiter for the word;
On the display, designation means for designating an arbitrary mora for the accent phrase unit delimited by the delimiter setting means ,
When a standard accent setting mode is instructed by the mode instructing means, the accent type of each accent phrase is read according to an accent setting rule determined according to the designated mora and the height information of the accent of each mora. And when the standard accent setting mode is not instructed by the instruction means, the setting means for changing the height state of only the designated mora,
The display control means changes the display of each mora that has been changed to a height based on the height information of the accent of each mora determined based on the accent set by the setting means. Accent setting device.   The accent setting apparatus according to claim 1, wherein the accent set by the setting unit is registered.   2. The accent setting apparatus according to claim 1, wherein the setting unit determines an accent setting rule according to the position of the mora and the height information instructed by the instruction unit.   2. The accent setting device according to claim 1, wherein the display control means displays an attached mora that does not affect the accent of the word in addition to the input word divided into mora units. .   2. The accent setting apparatus according to claim 1, wherein the accent setting rule is stored in advance corresponding to an accent type.   The accent setting apparatus according to claim 1, wherein the receiving unit receives an instruction from a pointing device. A display control step for controlling the reading of the input word to be divided into mora units and displayed;
A delimitation step for setting an accent phrase delimiter for the word;
An accepting step of accepting designation of an arbitrary mora in the divided accent phrase unit designated on the display;
A determination step of determining whether or not the standard accent setting mode is instructed;
When said standard accented setting mode is instructed in accordance with an accent configuration rules determined in accordance with the specified Mora and the elevation information for each mora accent, set accents readings of each accent phrase, the If the standard accent setting mode is not instructed, a setting step for changing the height state of only the designated mora;
A second display control step of changing the display of each mora that has been changed to a height based on the height information of the accent of each mora determined based on the accent set in the setting step. Characteristic accent setting method. The accent setting method according to claim 7 , further comprising a registration step of registering the set accent. The accent setting method according to claim 7 , further comprising a determination step of determining an accent setting rule according to the instructed mora position and height state. 8. The accent setting method according to claim 7 , wherein an attached mora that does not affect the accent of the word is displayed in addition to the input word reading divided into mora units. The accent setting method according to claim 7 , wherein the accent setting rule is stored in advance corresponding to an accent type. The accent setting method according to claim 7 , wherein the receiving step receives an instruction from a pointing device.

Images