JP2580565B2 - Voice information dictionary creation device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] The present invention displays, on a display device, a character string indicating a syllable in voice data corresponding to a word or a phrase expressed in characters,
The utterance data such as accents, pitches, and pronunciation speeds are converted into a visual image and displayed, and the operator operates the screen to change the visual data to a visual image corresponding to the desired utterance data. The display state of the target image is converted into audio data and registered in the audio information dictionary.

[Industrial applications]

The present invention relates to a voice information dictionary creating apparatus that facilitates setting of voice data (especially utterance data) for performing voice synthesis.

[Conventional technology]

Recent advances in speech synthesis technology have made it possible to relatively easily perform speech synthesis in personal computers (hereinafter abbreviated as personal computers) and the like by using speech synthesis means such as speech synthesis cards.

However, in order to perform the above-mentioned speech synthesis, it is necessary to provide the speech synthesis unit with speech data including syllable data (including one or more syllables) and utterance data for determining an accent, a pitch, a pronunciation speed, and the like. .

In personal computers, for example, there is a trend to support TALK sentences in the BASIC language. Hereinafter, an example of specifying audio data in such a BASIC language will be described.

As an example, when "Yamagata" is uttered, it is necessary to specify the accent position and other utterance data following the character string indicating the syllable as the voice data to be specified.

For example, to make the above “Yamagata” uttered with an accent on the second note, it is described as TALK “Yamagata: A2”. In the above TALK sentence, “Yamagata” is a character string that specifies the syllable to be uttered, “A2” is utterance data that specifies that the second beat is to be accented, and “:” is a delimiter. .

In this way, in order to describe the specification of audio data in the TALK sentence supported by the BASIC language, various parameters such as accent positions and other parameters must be given as symbols and numerical values in addition to the character string representing the syllable . In this example, "A2" designates the position of the accent position of the word as an accent parameter. In addition to this, it is necessary to specify intonation, pitch, sounding speed, and the like in order to achieve natural sounding.

Specifying many of these parameters using symbols and numerical values is not only complicated in operation, but also difficult to grasp the relationship between the specified result and the problem. is there. In order to hear the actual utterance, it is necessary to execute only the TALK sentences included in the program one by one, which is a complicated operation.

Therefore, there is a method of providing a voice information dictionary storing voice data, searching the voice information dictionary, reading desired voice data, and substituting the read voice data in the TALK sentence in the case of the BASIC language.

[Problems to be solved by the invention]

However, as described above, in the related art, there is a problem that setting of audio data requires a complicated operation and cannot be confirmed on the spot. Therefore, in order to create a voice information dictionary, the complicated operation described above cannot be avoided.

The present invention operates on a screen using a visual image,
It is an object of the present invention to make it possible to easily set voice data by designating and to easily create a voice information dictionary.

[Means for solving the problem]

A voice information dictionary creation device according to the present invention is an information processing device including an input unit, a display device, a screen editing unit, an image conversion unit, a voice information dictionary, a registration unit, and a control unit.

In this device, under the control of the entire device by the control unit, the screen editing unit transmits the syllable data and the utterance data indicated by the character string specified through the input unit or the character string read from the voice information dictionary. The syllable data is displayed as it is on the display device as a character string, and the utterance data is converted into a visual image by the image conversion unit and displayed. The image converter corrects and displays the image, the image converter converts the corrected visual image into corresponding utterance data, and the registerer converts the converted utterance data into syllable data and a key for retrieval along with audio information. Register in the dictionary.

[Action]

Rather than specifying the utterance data (pitch, accent position, speed, etc.) of the voice data with simple symbols and numerical values, the sound is converted into a visual image and displayed on the screen,
Since the display is changed by the designation, it is easy to associate with the actual utterance state, and it is easy to generate and correct utterance data.

〔Example〕

FIG. 1 is a main block diagram showing a configuration of a speech information dictionary creating apparatus according to the present invention. Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

1 is an input unit, for example, a keyboard, 2 is a display device, 3 is a storage device, for example, a magnetic disk device or a floppy disk device, 4 is a screen editing unit, 5 is an image conversion unit, 6 is a registration unit, 7 is The control unit 8 is a voice information dictionary stored in the storage device 3, and 9 is a voice synthesis unit.

2 (a) to 2 (i) show one embodiment of the present invention.
FIG. 9 is a diagram for describing an example of a setting operation of audio data using a visual image.

In this embodiment pronunciation of "Hello" (the syllable),
An example of registration with desired utterance data will be described.

First, a character string “konnichiha” 10 indicating a syllable of a voice to be registered is designated from the input unit 1. This character string 10 is displayed on the screen of the display device 2 by the screen editing unit 4 as shown in FIG. In addition, a screen display (example: 11) corresponding to various utterance data is performed on the display device 2. If all the utterance data are determined, they are displayed on the screen at once, but an example will be described in which the utterance data is sequentially specified from a state where nothing is specified. Input designation 1 for display type of utterance data
Can be changed one after another. In the present embodiment, an example in which a keyboard is used as the input unit 1 will be described. However, the present invention is not limited to this. There is no such thing.

There are many types of utterance data, and an example will be described in which the accent position, the height of the ending and the beginning, the pitch of the voice, the pronunciation speed, and finally the distinction between male and female voices are determined in order. The order of determining these utterance data may be determined by a processing program, and may be specified from the input unit 1 (not shown). When specifying from the input unit 1, a type of utterance data is assigned to each of the function keys, and by selecting the function key, a method of selecting the type of utterance data to be determined, or an utterance on the screen is selected. The method can be implemented by various methods such as displaying the type of data and selecting a desired one from these.

In the present embodiment, designation is made from the input unit 1 as described above. For example, when the determination of the accent position is selected, as shown in FIG.
11 For example, '▼' is displayed at a predetermined position on the character string 10 (the top in the figure). The mark 11 can be moved to a desired position by operating a shift key of the input unit 1 as shown in FIG. 3B (third syllable in the figure). During this operation, it is also easy to actually utter the voice with the specified accent, and to be able to hear and confirm it.

If it is determined that the accent position is to be placed at the beginning of the character string 10 in this manner, the sign mark 11 is moved to the beginning of the character string 10 (see FIG. 3C). As a method of indicating the accent position, a method of changing the color of the character at the accent position or the like can be used in addition to the addition of the mark 11 such as “▼”.

Next, when the determination of the beginning and the end of the word is selected, the letter "ko" at the beginning and the letter "c" at the end can be moved up and down by the shift key as shown in FIG.
Then, the user again confirms the speech by actually uttering it, and selects a desired beginning and end height as shown in FIG.

When the determination of the next voice pitch is selected, a mark 12 indicating the voice pitch is displayed in front of the character string 10 as shown in FIG. By operating the shift key, the sign mark 12 is moved up and down, and actual utterance is made to determine a desired height (see FIG. 9G).

Next, when the user decides to determine the speaking speed, a sign 13 indicating the speaking speed is displayed on the screen as shown in FIG. The marking mark 13 in the figure is drawn by an image of a speedometer, and the pointer 14 thereof is moved by a shift key or a number key, and an appropriate speed is selected by actually speaking.

When the user decides whether to use a male voice or a female voice, a mark 15 of a male figure or a female figure is displayed as shown in FIG. Since this is alternately converted by pressing the shift key or the function key, either one is selected. The mark 15 may be displayed in both cases, and the selected one may be displayed with high brightness, or only the selected one may be displayed. Various methods may be used.

As described above, when desired utterance data is set, a key for search is input from the input unit 1 to specify registration. In addition,
A character expression may be used instead of a search key by using a correspondence between a character expression and syllable data designated as a reading as in a document processing device. The designation of registration can be executed by assigning “registration” to one of the function keys, or by displaying operation items on the screen and selecting one of them.

When the registration is designated, the image conversion unit 5 reads the above-mentioned screen state, that is, the position of the marking mark 11 of the accent position, the position of the character at the beginning and end of the word, and converts it into the corresponding utterance data 22. The registration unit 6 writes the converted utterance data 22 into the audio information dictionary 8 provided in the storage device 3 together with the designated key.

FIG. 3 is a conceptual diagram showing the structure of the voice information dictionary 8,
This shows a state in which the character string 10 determined as described above and various utterance data 22 are stored in association with the designated key 21.

As shown in the figure, the utterance data 22 are all recorded by a combination of a symbol indicating the item of the utterance and a numerical value indicating the degree thereof in accordance with a predetermined description format. Therefore, the character string 10 and the utterance data 22 are read out by designating the key 21 corresponding to the desired voice data, and are given to the voice synthesizer 9 to synthesize the desired voice. In the present embodiment, the keys 21 are configured by combining alphanumeric characters and numbers, but this is an example, and the configuration is arbitrary. It may be a key expressed in kanji and kana.

As described above, in the present embodiment, the voice is converted into a visual image on the display device 2 and displayed, the voice data is set by operating the screen in an interactive manner, and the candidate data is actually converted using a known technique. Can be uttered. Therefore, the setting of the candidate data is not only easy, but also can be actually heard and confirmed by ear, so that the desired voice data can be easily and accurately set and registered. When the registered voice data is corrected, the voice data is read from the dictionary and processed in the same manner.

The ability to easily and accurately set and register desired voice data in this manner not only facilitates the creation of a general-purpose voice information dictionary, but also allows the use of special voice data synthesis systems. It is possible to arbitrarily create a dictionary or a dictionary storing only necessary vocabulary.

Further, the present embodiment is not only convenient for a system developer, but also can be used to enable a user of the system to register necessary voice data as necessary during operation of the system.

In the above embodiment, the case where the character string 10 is composed of five characters (five syllables) of “Konichiha” has been described, but this is an example, and the number of characters of the character example 10 does not need to be particularly limited.
Therefore, in FIG. 3, the column of the character string 10 is set to five characters, but this is also for the sake of illustration, and the required number of digits can be taken. In addition, it goes without saying that the column of the utterance data 22 may be set to the number of digits as required.

〔The invention's effect〕

As described above, according to the present invention, voice data (pitch, accent position, speed, etc.) of voice data is not specified by simple symbols and numerical values, but is converted into a visual image. Display on the screen and change the display according to designation, it is easy to associate with the actual utterance state, and it is easy to generate and correct utterance data. Therefore, it is easy and accurate to set the audio data, so that a desired audio information dictionary can be easily created. As a result, the configuration of a system using speech synthesis and the development of software are facilitated.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of the configuration of the present invention, FIG. 2 is an explanatory diagram of one embodiment of the present invention, and FIG. 3 is a diagram showing the configuration of the voice information dictionary created in the above-described embodiment. In the figure, 1 is an input unit, 2 is a display device, 3 is a storage device, 4 is a screen editing unit, 5 is an image conversion unit, 6 is a registration unit, 7 is a control unit, 8 is a voice information dictionary, 10 is Character strings representing syllables, 11, 12, 13, 14, and 15 are indication marks, 21 is a search key, and 22 is utterance data.

Claims (1)

(57) [Claims] An information processing apparatus comprising: an input unit, a display device, a screen editing unit, an image conversion unit, a voice information dictionary, a registration unit, and a control unit, wherein the control unit controls the entire device. The screen editing unit displays the syllable data as a character string as it is, of the voice data composed of syllable data and utterance data indicated by a character string specified via the input unit or a character string read from the voice information dictionary In addition to displaying on the device, the utterance data is converted into a visual image by the image conversion unit and displayed, and if there is an instruction from the input unit, the visual image is corrected and displayed, and the image conversion unit is corrected. The registration unit converts the visual image into corresponding utterance data, and registers the converted utterance data in a speech information dictionary together with syllable data and a key for retrieval. Audio information dictionary creating apparatus characterized by was Unishi.