JP6821728B2 - Text data voice playback device and text data voice playback program - Google Patents

Description

The present invention relates to a text data voice reproduction device and a text data voice reproduction program.

As a device that outputs digital contents stored in a storage unit in advance by voice, for example, a device having a configuration as disclosed in Patent Document 1 (Japanese Unexamined Patent Publication No. 2008-268648) is known.

Further, in recent years, a configuration as a device for supporting learning of children and students by using digital contents developed for learning has been proposed. As an example of such a text data voice reproduction device for learning support, for example, a configuration as disclosed in Patent Document 2 (Japanese Unexamined Patent Publication No. 2016-212168) is known.

Japanese Unexamined Patent Publication No. 2008-268648 Japanese Unexamined Patent Publication No. 2016-212168

By combining the configurations disclosed in Patent Document 1 and Patent Document 2, for example, text data of a textbook can be output as audio data. However, if the configurations of Patent Document 1 and Patent Document 2 are simply combined, the voice data of the conversational part and the non-conversational part have the same voice when the voice data output from the speaker includes a conversational sentence. If it is a kind of voice data, it becomes difficult to grasp the conversation content. This makes it difficult for the students who need learning support to fully understand the contents of the sentences output by the voice data, and there is a problem that it is not possible to provide sufficient learning support to the students.

Therefore, according to the present invention, when a child student who needs learning support learns a sentence in which normal sentences and conversational sentences are mixed, the voice type characteristic data used for generating voice data in the conversational part is appropriately changed. The main purpose is to provide a text data voice playback device and a text data voice playback program that can improve the comprehension of text contents by children and students who listen to voice data.

That is, the present invention stores at least a storage unit in which a voice type characteristic database including text data and a plurality of voice type characteristic data is stored, a text data acquisition unit that acquires the text data from the storage unit, and the text data. In the character discriminating unit that discriminates the type of the character data for each of the constituent character data, and in the character data constituting the text data, the character whose type of character data is a start-side quote and an end-side quote. A conversation part extraction unit that extracts a range separated by data as a conversation part, a voice data generation unit that generates voice data based on basic voice type characteristic data extracted from the voice type characteristic database from the text data, and the text data. In the conversation part included in the voice type characteristic data, the voice type characteristic data for the conversation part extracted from the voice type characteristic database is applied to the voice data generation unit in place of the basic voice type characteristic data. And an output unit that outputs the voice data generated by the voice data generation unit, and the conversation part extraction unit has another start between the start side quote and the end side quote. If there are side quotes and other ending quotes, the part separated by the other starting quotes and the other ending quotes is extracted as the conversational part in the conversation, and the conversation voice. The species change unit applies the voice type characteristic data for the conversation part extracted from the voice type characteristic database to the conversation part in the conversation to the voice data generation unit in place of the voice type characteristic data for the conversation part. It is a characteristic text data audio reproduction device.

As a result, when learning a sentence in which a normal sentence and a conversational sentence are mixed, the voice type characteristic data used for generating the voice data in the conversational part can be appropriately changed. That is, it is possible to improve the comprehension of the text content based on the voice data by the students by changing the voice type of the voice data in the basic part and the conversation part inserted in the middle position or the beginning position of the basic part. become. Further, even when the conversation part includes an in-conversation conversation part which is another conversation part, the voice type characteristic data of the voice data is different from the voice type characteristic data in the conversation part and the basic part. By applying the characteristic data, it is possible to facilitate the distinction of the text content by the students who have heard the voice data.

Further, when the person name and the voice type characteristic data corresponding to the person name are associated with the voice type characteristic database and the person name is included in the text data, the conversation voice type is changed. It is preferable that the unit applies the voice type characteristic data associated with the person's name to the voice data generation unit .

As a result, it is possible to automatically generate voice data using more natural voice type characteristic data for the content of the text data.

Also, the audio data generation unit, the generatable der Rukoto the audio data on the basis of the text data portion excluding a portion of the person's name is preferred.

Rukoto Accordingly, since the text data is not part important to understand the contents of the text from the voice data script format (a portion other than speech) is out of the product object of speech data, enhance the understanding of the contents of the text Can be done.

In addition, a storage unit in which at least a voice type characteristic database containing text data and a plurality of voice type characteristic data is stored, a text data acquisition unit that acquires the text data from the storage unit, and characters constituting the text data. In the character discriminating unit that discriminates the type of the character data for each of the data and the character data constituting the text data, the character data types are separated by the character data in which the type of the character data is a start side quote and an end side quote. The text data includes a conversation part extraction unit that extracts the specified range as a conversation part, a voice data generation unit that generates voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and the text data. The conversation part includes a conversation voice type changing unit that applies voice type characteristic data for the conversation part extracted from the voice type characteristic database to the voice data generation unit in place of the basic voice type characteristic data, and the conversation voice type changing unit. It includes an output unit that outputs the voice data generated by the voice data generation unit, and the voice type characteristic database is associated with the person name and the voice type characteristic data corresponding to the person name. When the text data includes the person's name, the conversation voice type changing unit applies the voice type characteristic data associated with the person's name to the voice data generation unit to generate the voice data. The unit may also be a text data audio reproduction device characterized in that the audio data can be generated based on the portion of the text data excluding the portion of the person's name .

As a result, when learning a sentence in which a normal sentence and a conversational sentence are mixed, the voice type characteristic data used for generating the voice data in the conversational part can be appropriately changed. That is, it is possible to improve the comprehension of the text content based on the voice data by the students by changing the voice type of the voice data in the basic part and the conversation part inserted in the middle position or the beginning position of the basic part. become. In addition, it is possible to automatically generate voice data using more natural voice type characteristic data for the content of the text data. Then, since the parts (parts other than the dialogue) that are not important for understanding the content of the text from the voice data of the script-format text data are excluded from the generation target of the voice data, the degree of understanding of the content of the text can be improved. ..

Also, the character discrimination unit Rukoto to determine the type of the character data by discriminating a character code that is given to the character data is preferable.

Thereby, it is possible to easily determine whether the character data constituting the text data is a start side quotation mark or an end side quotation mark (whether or not it is a quotation mark) .

Further, as another invention, a storage unit in which at least a voice type characteristic database including text data and a plurality of voice type characteristic data is stored, an output unit for outputting voice data based on the text data, and an operation control unit. a readable to the computer at least comprises, the operation control unit, and text data acquisition function to acquire the text data from the storage unit, the character data types for each of the character data constituting said text data In the character discrimination function for discriminating the text data and the character data constituting the text data, a range separated by the character data in which the type of the character data is the start side quote and the end side quote is extracted as a conversation part. The conversation part extraction function, the voice data generation function for generating voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and the conversation part included in the text data include the above. The conversation voice type change function that applies the voice type characteristic data for the conversation part extracted from the voice type characteristic database to the voice data generation function in place of the basic voice type characteristic data, and the voice data generation function generated by the voice data generation function. An output function for outputting voice data to the output unit and an output function for outputting audio data to the output unit are executed, and the conversation part extraction function performs the other start side quote and the other start side quote between the start side quote and the end side quote. If there is an end-side quote, the part separated by the other start-side quote and the other end-side quote is extracted as the conversation part in the conversation, and the conversation voice type change function is the conversation. text data voice and Rukoto characterized by applying to the audio data generating function instead a conversation in the conversation part for voice species characteristic data extracted from the voice species characteristic database in an inner talking part in voice type characteristics data for the talking part There is a playback program.

As a result, when learning a sentence in which a normal sentence and a conversational sentence are mixed, the voice type characteristic data used for generating the voice data in the conversational part can be appropriately changed. That is, it is possible to improve the comprehension of the text content based on the voice data by the students by changing the voice type of the voice data in the basic part and the conversation part inserted in the middle position or the beginning position of the basic part. become. Further, even when the conversation part includes an in-conversation conversation part which is another conversation part, the voice type characteristic data of the voice data is different from the voice type characteristic data in the conversation part and the basic part. By applying the characteristic data, it is possible to facilitate the distinction of the text content by the students who have heard the voice data.

Further, when the person name and the voice type characteristic data corresponding to the person name are associated with the voice type characteristic database and the person name is included in the text data, the conversation voice type is changed. feature Rukoto to apply the voice type characteristics data bundled with the person's name in the voice data generating function is preferable.

As a result, it is possible to automatically generate voice data using more natural voice type characteristic data for the content of the text data.

Furthermore, the to the sound data generating function, the can der Rukoto be based on the portion excluding the portion of the person's name in the text data to generate the audio data is preferred.

Rukoto Accordingly, since the text data is not part important to understand the contents of the text from the voice data script format (a portion other than speech) is out of the product object of speech data, enhance the understanding of the contents of the text Can be done.

Further, as another invention, a storage unit in which at least a voice type characteristic database including text data and a plurality of voice type characteristic data is stored, an output unit for outputting voice data based on the text data, and an operation control unit are provided. A text data acquisition function that is readable by at least a computer and causes the operation control unit to acquire the text data from the storage unit, and a type of the character data for each of the character data constituting the text data. In the character discrimination function for discriminating and the character data constituting the text data, a conversation in which the range separated by the character data whose type of character data is the start side quote and the end side quote is extracted as a conversation part. The partial extraction function, the voice data generation function for generating voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and the conversation portion included in the text data include the voice. The conversation voice type change function that applies the voice type characteristic data for the conversation part extracted from the species characteristic database to the voice data generation function instead of the basic voice type characteristic data, and the voice generated by the voice data generation function. An output function for outputting data to the output unit is executed, and the voice type characteristic database is associated with the person name and the voice type characteristic data corresponding to the person name, and is linked to the text data. When the person name is included, the conversation voice type change function applies the voice type characteristic data associated with the person name to the voice data generation function, and applies the text to the voice data generation function. There is a text data audio reproduction program characterized in that it is possible to generate the audio data based on a portion of the data excluding the portion of the person's name .

As a result, when learning a sentence in which a normal sentence and a conversational sentence are mixed, the voice type characteristic data used for generating the voice data in the conversational part can be appropriately changed. That is, it is possible to improve the comprehension of the text content based on the voice data by the students by changing the voice type of the voice data in the basic part and the conversation part inserted in the middle position or the beginning position of the basic part. become. In addition, it is important to understand the content of sentences from the voice data of script-format text data by automatically generating voice data using more natural voice type characteristic data for the content of text data. Since the non-speech parts (parts other than the dialogue) are excluded from the voice data generation target, it is possible to improve the understanding of the content of the text .

Also, the character determination function, Rukoto to determine the type of the character data by discriminating a character code that is given to the character data is preferable.

Thereby, it is possible to easily determine whether the character data constituting the text data is a start side quotation mark or an end side quotation mark (whether or not it is a quotation mark) .

According to the configuration of the text data voice reproduction device and the text data voice reproduction program in the present invention, when a child student who needs learning support learns a sentence in which ordinary sentences and conversation sentences are mixed, the voice in the conversation part The voice type characteristic data used for data generation can be changed as appropriate. That is, the difference between the basic part and the conversation part can be intuitively and clearly distinguished by appropriately changing the voice type of the voice data in the conversation part inserted in the middle position or the beginning position of the basic part and the basic part. Can be done. As a result, it becomes possible for students to improve their understanding of text content based on voice data. Further, even when the conversation part includes an in-conversation conversation part which is another conversation part, the voice type characteristic data of the voice data is different from the voice type characteristic data in the conversation part and the basic part. By applying the characteristic data, it is possible to facilitate the distinction of the text content by the students who have heard the voice data. Alternatively, it is important to understand the content of the text from the voice data of the script format text data by automatically generating the voice data using the more natural voice type characteristic data for the content of the text data. Since the non-speech parts (parts other than the dialogue) are excluded from the voice data generation target, it is possible to improve the understanding of the content of the text .

It is a schematic block diagram of the text data voice reproduction apparatus in this embodiment. It is explanatory drawing which shows the extraction example of the division text data when the text data of a sentence is plain text data. It is explanatory drawing which shows the extraction example of the division text data when the text data of a sentence is a tag delimited text data. It is explanatory drawing which shows an example of other text data. It is a schematic processing flow diagram of the text data voice reproduction program (text data voice reproduction method) in this embodiment. It is explanatory drawing which shows an example of the text data of a script format.

The text data voice reproduction device 100 in this embodiment will be described. As shown in FIG. 1, the text data voice reproduction device 100 in the present embodiment includes a storage unit 10, an operation control unit 20, and an output unit 30. Such a text data voice reproduction device 100 can be realized by a tablet computer or the like.

At least the text data TD, the classification condition KJ for classifying the text data TD, the character code information MCJ, and the voice type characteristic database KDB including a plurality of voice type characteristic data are stored in the storage unit 10. .. As such a storage unit 10, a non-volatile memory typified by a flash memory is preferably used. If the tablet computer is not a dedicated product of the text data audio reproduction device 100, it can be used as the text data audio reproduction device 100 by installing the text data audio reproduction program PGM in the storage unit 10 of the tablet computer.

The text data TD in this embodiment is so-called plain text data in which Japanese and non-Japanese (English) are mixed, as typified by English textbook data, but is not limited to this form. .. For example, plain text data such as Japanese and German may be used, or tag-delimited text data between Japanese and English may be adopted. That is, the text data TD in the present specification refers to a concept including any of Japanese and non-Japanese plain text data, Japanese and non-Japanese tag-delimited text data, and other text data. ..

The classification condition KJ in the present embodiment is for classifying the text data TD acquired from the storage unit 10 by the operation control unit 20 as the text data acquisition unit 21 under preset conditions. The text data TD in this embodiment is a plain text data format in Japanese and English. The division condition KJ for the text data TD in the plain text data format is set so that the range delimited by the positions such as punctuation marks, symbols, parentheses, line breaks, and spaces in the text data TD can be extracted as the division text data KTD. Has been done.

Specifically, as shown in FIG. 2A, the text data TD acquired from the storage unit 10 is the text data TD obtained by the operation control unit 20 as the division text data extraction unit 22 in accordance with the division condition KJ. Is separated by punctuation marks, symbols, parentheses, line breaks, spaces, etc. The divided text data KTD obtained in this way is in a state of being divided into a plurality of blocks as shown in FIG. 2 (B).

As shown in FIG. 3A, the division condition KJ when the text data TD is in the tag-delimited text data format can extract the range delimited by the tag as each division text data KTD. Is set to. Therefore, the text data TD shown in FIG. 3 (A) is converted into a plurality of divided text data KTDs as shown in FIG. 3 (B) by the operation control unit 20 as the divided text data extraction unit 22 in accordance with the division condition KJ. It will be divided.

For the divided text data KTD shown in FIGS. 2 and 3, the operation control unit 20 as the extraction number assigning unit 23 may execute a process of assigning serial number extraction numbers in the order of extraction. good. The divided text data KTD obtained as described above is stored in the storage unit 10 as the divided text data KTD (x: x is a natural number) in a state of being associated with the extraction number by the operation control unit 20. Hereinafter, in the present specification, the classified text data KTD to which the extraction number is assigned will be referred to as the classified text data KTD (x).

Further, in the present embodiment, as is clear from FIGS. 2 and 3, the number of the divided text data KTD (x) obtained by dividing the plain text data and the divided text data KTD (x) of the tag-delimited text data are obtained. ) Are in agreement. Even if the sentences have the same content, the content of the divided text data KTD (x) obtained by dividing the plain text data and the tag delimited text data may differ depending on the structure of the tag delimited text data. .. However, in the present embodiment, the difference in the division text data KTD (x) does not become a problem by performing the processing described later.

The division text data KTD (x) obtained by the division text data extraction unit 22 in this way is composed of one or more character data. Following this, the operation control unit 20 as the character discrimination unit 24 stores the character code assigned to the character data in the division text data KTD (x) for each division text data KTD (x). The character code information MCJ stored in the part 10 is referred to, and a process of determining whether the character data is a Japanese character code or a non-Japanese character code is executed.

Further, when the character discrimination unit 24 determines that the character code of a certain character data is a Japanese character code for each character data constituting each division text data KTD (x), the Japanese character discrimination counter NHC Add 1 to. Similarly, when it is determined that the character code of a certain character data is a non-Japanese character code, 1 is added to the non-Japanese discrimination counter HHC. The initial values of the Japanese discrimination counter NHC and the non-Japanese discrimination counter HHC are both 0. Then, the respective values of the Japanese discrimination counter NHC and the non-Japanese discrimination counter HHC in a certain division text data KTD (x) are associated with each division text data KTD (x) by the operation control unit 20. After being stored in the storage unit 10 in the state, a process of resetting each value to the initial value of 0 is executed.

This will be described based on the text data TD of FIG. In the divided text data KTD (1) shown in FIG. 2, the Japanese discrimination counter NHC = 6 and the non-Japanese discrimination counter HHC = 0, and these values are associated with the KTD (1) and stored in the storage unit 10. After that, the counter value is reset to 0. Further, in the divided text data KTD (2), the Japanese discrimination counter NHC = 14 and the non-Japanese discrimination counter HHC = 4, and these values are associated with the KTD (2) and stored in the storage unit 10. After that, the counter value is reset to the initial value of 0.

Then, the operation control unit 20 as the character data type ratio calculation unit 25 determines the ratio RT of the total value of the Japanese discrimination counter NHC to all the characters of the single or a plurality of division text data KTD (x) constituting the text data TD. Executes the process of calculating. The ratio RT in this embodiment is calculated by (total count value of Japanese discrimination counter NHC / (total count value of Japanese discrimination counter NHC + total count value of non-Japanese discrimination counter HHC)). The ratio RT of the count value of the Japanese discrimination counter NHC to the total of the count values of all the characters in the text data TD calculated in this way is stored in the state of being linked to the text data TD by the operation control unit 20. It is stored in 10.

This will be described based on the text data TD of FIG. Since the text data TD is composed of the divided text data KTD (1) and the divided text data KTD (2), the total number of character data constituting the text data TD is the total count value of the Japanese discrimination counter NHC + non-Japan. It is calculated by the total count value of the word discrimination counter HHC, and is 24. Of these, the total value of the Japanese discrimination counter NHC is 20, so the ratio RT of the total value of the Japanese discrimination counter to the total number of characters in the text data TD is (20/24) × 100, which is 83.33%. ..

Even if the number of divided text data KTD (x) differs between the plain text data format and the tag-delimited data format, the total value of the Japanese discrimination counter NHC in the text data TD. Since the total value of the non-Japanese discrimination counter HHC is the same as that of the non-Japanese discrimination counter HHC, the ratio RT of the total value of the Japanese discrimination counter to the total number of characters in the text data TD is calculated by (20/24) × 100 and is 83.33%. Therefore, it is understood that the difference in the data format of the text data TD does not matter in the present invention.

In addition, the storage unit 10 also has a predetermined ratio SRT that serves as a judgment criterion when deciding whether to apply the Japanese voice data generation engine NOE or the non-Japanese voice data generation engine HOE to the text data TD. It is remembered. It is preferable that the predetermined ratio SRT can be changed to a numerical value input by a data input means (not shown).

The Japanese voice data generation engine NOE is a division text data KTD obtained by dividing the text data TD acquired from the storage unit 10 by the text data acquisition unit 21 according to the division condition KJ and a voice type characteristic database KDB. This is for generating Japanese voice data NOD based on the voice type characteristic data (default setting is basic voice type characteristic data BKP) extracted from. As such a Japanese speech data generation engine NOE, a so-called Japanese TTS (Text To Speech) is preferably used. Similarly, the non-Japanese voice data generation engine HOE is non-Japanese based on the divided text data KTD and the voice type characteristic data extracted from the voice type characteristic database KDB (default setting is the basic voice type characteristic data BKP). It is for generating voice data HOD, and what is called non-Japanese TTS is preferably used.

Here, the voice type characteristic database KDB will be described. The multiple types of voice type characteristic data constituting the voice type characteristic database KDB are appropriately combined with elements for determining the characteristics of a plurality of voice data represented by a gender element, an age (generation) element, a speaking speed element, and the like. It will be. Specifically, the frequency, speaking speed, and the like of the voice data when being output to the output unit 30 are specified. More specifically, child boys, boy boys, adult boys, middle-aged boys, middle-aged boys, old boys, young girls, boys and girls, adult girls, middle-aged girls, middle-aged girls, old-aged girls and the like can be listed. It is also possible to set additional characteristic elements such as high voice, standard voice, low voice, clear voice, and husky voice for each of these voice type characteristic data.

Next, the operation control unit 20 as the voice data generation engine selection unit 26 compares the ratio RT associated with the text data TD stored in the storage unit 10 with the predetermined ratio SRT. Then, when the ratio RT associated with a certain text data TD is equal to or more than a predetermined ratio SRT, the Japanese voice data generation engine NOE and the voice type characteristic database KDB are added to all the character data in the certain text data TD. The process of applying the basic voice type characteristic data BKP extracted from is executed. Further, when the ratio RT in a certain text data TD is less than a predetermined ratio SRT, the Japanese voice data generation engine NOE is applied to the character data to which the Japanese character code is assigned in the certain text data TD. And the basic voice type characteristic data BKP extracted from the voice type characteristic database KDB is applied respectively, and the non-Japanese voice data generation engine HOE and the basic voice type characteristic data are applied to the character data to which the non-Japanese character code is assigned. The process of applying each BKP is executed.

That is, when the ratio of Japanese character data in the text data TD is high, the Japanese voice data generation engine NOE and the basic voice type characteristic data BKP are applied even for non-Japanese character data. , The entire text data TD is made possible to generate voice data of Japanese pronunciation by the basic voice type characteristic data BKP. The application of such a voice data generation engine is, for example, in an English textbook, when a Japanese translation of a sentence in which only Japanese names (proprietary nouns related to Japan) are written in English is output by voice, only Japanese names are pronounced in English. It is preferable in that it can be solved.

Further, when the ratio of non-Japanese (English) in the character data in the text data TD is high (when the ratio RT is less than the predetermined ratio SRT) as in the text data TD as shown in FIG. , Non-Japanese voice data generation engine HOE and basic voice type characteristic data BKP are applied to non-Japanese character data, respectively, and Japanese voice data generation engine NOE and basic voice are applied to Japanese character data. Species characteristic data BKP is applied respectively. In the text data TD shown in FIG. 4, only the first [Example 1] is generated by the Japanese voice data generation engine NOE and the basic voice type characteristic data BKP, and all the subsequent parts are non-Japanese voice data. Voice data will be generated by the generation engine HOE and the basic voice type characteristic data BKP.

In this way, it is possible to appropriately select a voice generation engine according to the ratio of Japanese in each character data constituting the text data TD. As a result, when generating the voice data of the text data TD, the application of the Japanese voice data generation engine NOE is prioritized, and it is easy for Japanese students to hear and the understanding of the content of the sentence can be improved. It is convenient.

Next, the operation control unit 20 as the conversation part extraction unit 27 sets the character data type to the start side quotation mark based on the character code information of all the character data constituting the text data TD determined by the character determination unit 24. Executes the process of detecting the character data that is and the character data that is the end quotation mark. Subsequently, the conversation part extraction unit 27 extracts the range (character data between the start side quotation mark and the end side quotation mark) separated by the extracted start side quotation mark and the end side quotation mark as the conversation part KB. To execute.

Next, the operation control unit 20 as the conversation voice type changing unit 28 basically uses the voice type characteristic data CKP for the conversation part extracted from the voice type characteristic database KDB of the storage unit 10 with respect to the portion extracted as the conversation part KB. The process to be applied instead of the voice type characteristic data BKP is executed. At this time, the change processing for the application state of the voice generation engine according to the ratio of Japanese in the character data constituting the text data TD is not performed. By doing so, when the text data TD includes a so-called conversation part, the voice type characteristic data CKP for the conversation part having a voice characteristic different from the basic voice type characteristic data BKP in the mainstream can be obtained. By applying it to the conversation part, the voice data of the mainstream and the conversation part KB can be clearly distinguished.

By the way, when the operation control unit 20 as the conversation part extraction unit 27 extracts that there is another start side quotation mark and another end side quotation mark between the start side quotation mark and the end side quotation mark, another The part separated by the start side quotation mark and the other end side quotation mark is extracted as the conversation part KKB in the conversation. When the conversation part KKB in the conversation is extracted, the operation control unit 20 as the conversation voice type change unit 28 replaces the voice type characteristic data CKP for the conversation part with respect to the conversation part KKB in the conversation for the conversation part. The process of applying the voice type characteristic data IKP is executed. The voice type characteristic data IKP for the conversation part in the conversation is different from the basic voice type characteristic data BKP and the voice type characteristic data CKP for the conversation part, and is the voice type characteristic data constituting the voice type characteristic database KDB. It is one of the data.

Further, since the conversation part KB returns to the conversation part KB again when the conversation part KKB in the conversation ends, the operation control unit 20 as the conversation voice type changing unit 28 is for the conversation part instead of applying the voice type characteristic data IKP for the conversation part in the conversation. The process of returning to the application of the voice type characteristic data CKP is executed. Then, when the conversation part KB ends, it returns to the mainstream part again, so that the operation control unit 20 as the conversation voice type changing unit 28 replaces the application of the conversation part voice type characteristic data CKP with the basic voice type characteristic data BKP. Execute the process to return to the application of.

In this way, after the voice data generation engine to be used and the voice type characteristic data are associated with the text data TD, the operation control unit 20 as the voice data generation unit 29 is finally applied. The generation of voice data is executed based on the voice type characteristic data and the text data TD. The voice data generated in this way is stored in the storage unit 10 by the voice data generation unit 29.

In the above-described embodiment, the ratio RT is calculated for the entire text data TD, but the ratio RT may be calculated for each division text data KTD (x). .. In this case, the ratio RT is the percentage of the number of Japanese character data to the number of all character data constituting the divided text data KTD (x) (the total count value of the Japanese discrimination counter NHC in the divided text data KTD (x)). / (Total count value of the Japanese discrimination counter NHC in the classified text data KTD (x) + total count value of the non-Japanese discrimination counter HHC in the classified text data KTD (x))) × 100 can be calculated. Since the data processing after calculating the ratio RT can be performed in the same manner as in the present embodiment described above, detailed description here will be omitted.

By doing so, a voice generation engine corresponding to the ratio of Japanese in the character data constituting the divided text data KTD (x) is appropriately selected, and the voice data corresponding to the divided text data KTD (x) is generated. Can be executed. Then, the voice data (at least one of Japanese voice data NOD and non-Japanese voice data HOD) generated in correspondence with the divided text data KTD (x) is stored by the operation control unit 20 by the operation control unit 20 as in the above-described embodiment. It is stored in 10.

When the voice data generation unit 29 completes the generation of voice data corresponding to the text data TD, the operation control unit 20 as the voice data reproduction control unit transfers the voice data stored in the storage unit 10 to the output unit 30 such as a speaker. Execute the output process. In this way, by outputting voice data from the speaker that clearly distinguishes between the text text and the conversation part in the text data TD and the distinction between the Japanese part and the English part, even children and students who need learning support It is convenient in that the learning efficiency of foreign language learning can be greatly improved.

Next, the text data voice reproduction program PGM and the text data voice reproduction method in the present embodiment will be described with reference to FIG. As shown in FIG. 1, the text data voice reproduction program PGM in the present embodiment is stored in the storage unit 10 in advance in a state in which it can be read by a computer. When the user taps a predetermined icon displayed on the screen of the tablet computer, the text data voice reproduction program PGM is started (processing starts). When the text data voice reproduction program PGM is started, the titles of a plurality of text data TDs stored in the storage unit 10 are displayed on the screen, and the operation control unit 20 displays the text data to be processed for the user. Requests TD selection. It should be noted that the beginning portion of the text data TD and the like may be provided with a description portion of the title name separately from the text of the text data TD. When the user selects the title of the target for reproducing the audio data from the text data TD displayed on the screen, the text data audio reproduction program PGM operates the operation control unit 20 to exert the text data acquisition function. , The text data TD of the specified title is acquired from the storage unit 10 (text data acquisition step: S1).

Next, the text data voice reproduction program PGM operates the operation control unit 20 so as to exert the division text data extraction function, and the text data acquisition step: the text data TD acquired in S1 is classified text based on the division condition KJ. The data KTD is extracted (division text data extraction step: S2). At this time, the text data voice reproduction program PGM operates the operation control unit 20 to exert the extraction number assigning function, and assigns an extraction number to each of the segmented text data KTDs extracted in the segmented text data extraction step: S2. (Extraction number assignment step: S3) may be performed.

Next, the text data voice reproduction program PGM operates the operation control unit 20 to exert the character discrimination function, and stores the character code assigned to the characters constituting each division text data KTD (x) in the storage unit 10. The character code information MCJ is referred to to determine whether each character code is a Japanese character code or a non-Japanese character code (character determination step: S4). At this time, the operation control unit 20 as a character discrimination function counts 1 in the Japanese discrimination counter NHC when the character code is a Japanese character code, and when the character code is a non-Japanese character code. 1 is counted in the non-Japanese discrimination counter HHC. The initial values of the Japanese discrimination counter NHC and the non-Japanese discrimination counter HHC are 0.

When the determination of the character code (type of character data) for all the character data of the divided text data KTD (x) is completed, the operation control unit 20 as the character discrimination function receives the divided text data KTD (x). The count values of the Japanese discrimination counter NHC and the non-Japanese discrimination counter HHC are linked and stored in the storage unit 10. Further, the operation control unit 20 as a character discrimination function stores the Japanese discrimination counter NHC and the non-Japanese discrimination counter HHC in the storage unit 10, and then executes a process of resetting each numerical value to the initial value of 0. ..

Next, the text data voice reproduction program PGM operates the operation control unit 20 to exert the character data type ratio calculation function to calculate the ratio RT of the Japanese character data type in the text data TD (character data type ratio calculation). Step: S5). The ratio RT of the Japanese character data type in the present embodiment is the total number of characters of all the division text data KTD (x) constituting the text data TD (total value of Japanese discrimination counter NHC + total value of non-Japanese discrimination counter HHC). ) Is the percentage of the number of characters to which the Japanese character code is assigned (total value of the Japanese discrimination counter NHC).

Next, the text data voice reproduction program PGM operates the operation control unit 20 to exert the voice data generation engine selection function, and applies the text data TD characters (text data) to which voice data generation engine to generate voice data. A judgment is made about the generation (voice data generation engine selection step: S6).

Specifically, the motion control unit 20 compares the ratio RT of the Japanese character data type calculated in the character data type ratio calculation step: S5 with the predetermined ratio SRT as a determination material stored in advance in the storage unit 10. Then, it is determined whether or not the ratio RT of the Japanese character data type is equal to or higher than the predetermined ratio SRT. More specifically, when the ratio RT of the Japanese character data type of the text data TD is equal to or more than the predetermined ratio SRT, the operation control unit 20 as the voice data generation engine selection function is used with respect to the text data TD. The basic voice type characteristic data BKP extracted from the Japanese voice data generation engine NOE and the voice type characteristic database KDB are applied respectively. On the contrary, when the ratio RT of the Japanese character data type of a certain text data TD is less than the predetermined ratio SRT, the non-Japanese voice data generation engine HOE and the voice type characteristic are given to the text data TD. The basic voice type characteristic data BKP extracted from the database KDB is applied respectively.

Further, in the character data type ratio calculation step: S5 and the voice data generation engine selection step: S6, the number of characters to which the Japanese character code in the division text data KTD (x) is assigned to the number of characters in each division text data KTD (x). The process of determining the application of the voice data generation engine to each section text data KTD (x) may be executed based on the percentage of.

Specifically, the character data type ratio calculation step: S5 is (total value of the Japanese discrimination counter NHC in the classified text data KTD (x)) / (total of the Japanese discrimination counter NHC in the classified text data KTD (x)). A process of calculating the ratio RT is performed by multiplying the value + the total value of the non-Japanese discrimination counter HHC in the division text data KTD (x) × 100 (%). Next, the voice data generation engine selection step: S6 compares the ratio RT of each division text data KTD (x) with the predetermined ratio SRT, and with respect to each division text data KTD (x). The process of determining the application of the Japanese voice data generation engine NOE or the non-Japanese voice data generation engine HOE may be performed.

Next, the text data voice reproduction program PGM operates the motion control unit 20 to exert the conversation part extraction function, and detects the start side quotation mark and the end side quotation mark from the character data constituting the text data TD. , The range separated by the character data of the start side quotation mark and the end side quotation mark is extracted as the conversation part KB (conversation part extraction step: S7).

Next, the text data voice reproduction program PGM operates the motion control unit 20 to exert the conversation voice type changing function, and the voice type characteristic database of the storage unit 10 is used for the character data of the portion extracted as the conversation part KB. A process of applying the voice type characteristic data CKP for the conversation part extracted from the KDB in place of the basic voice type characteristic data BKP is executed (conversation voice type change step: S8). At this time, the change processing for the application state of the voice generation engine according to the ratio of Japanese in the character data constituting the text data TD is not performed. By doing so, when the text data TD includes a so-called conversation part, the voice type characteristic data CKP for the conversation part having a voice characteristic different from the basic voice type characteristic data BKP in the mainstream can be obtained. By applying it to the conversation part, it is possible to clearly distinguish the voice data of the mainstream and the conversation part.

By the way, in the text data voice reproduction program PGM, the operation control unit 20 as a conversation part extraction function has another start side quotation mark and another end side quotation mark between the start side quotation mark and the end side quotation mark. When extracted, the part separated by other start side quotation marks and other end side quotation marks is extracted as the conversation part KKB in the conversation. Then, when the conversation part KKB in the conversation is extracted, the text data voice reproduction program PGM operates the motion control unit 20 to exert the conversation voice type change function, and causes the conversation part KKB to have the voice type characteristics for the conversation part. Instead of the data CKP, a process of applying the voice type characteristic data IKP for the conversation part in the conversation is executed (step for changing the conversation voice type in the conversation: S9). As described above, the voice type characteristic data IKP for the conversation part in the conversation is different from the basic voice type characteristic data BKP and the voice type characteristic data CKP for the conversation part.

When the conversation part KKB in the conversation ends, the text data voice reproduction program PGM returns to the conversation part KB again. Therefore, the text data voice reproduction program PGM has the voice type characteristics for the conversation part in the conversation so that the motion control unit 20 can exert the function as the conversation voice type change function. Instead of applying the data IKP, the process of returning to the application of the voice type characteristic data CKP for the conversation part is executed. Then, when the conversation part KB ends, the text data voice reproduction program PGM returns to the mainstream part again, so that the motion control unit 20 exerts the function as the conversation voice type change function, and the voice type characteristic data for the conversation part. Instead of applying the CKP, the process of returning to the application of the basic voice type characteristic data BKP is executed.

In this way, after associating the text data TD with the voice data generation engine to be used and the voice type characteristic data, the text data voice reproduction program PGM provides the motion control unit 20 with a voice data generation function. The voice data is generated based on the text data TD by using the voice type characteristic data and the voice data generation engine finally applied (voice data generation step: S10). The voice data generated in this way is stored in the storage unit 10 by the motion control unit 20.

When the generation of the voice data corresponding to the text data TD is completed, the text data voice reproduction program PGM operates the operation control unit 20 so as to exert the voice data output function, and stores the voice data stored in the storage unit 10. A process of outputting to the output unit 30 of a speaker or the like is executed (audio data output step: S11). At this time, the text data voice reproduction program PGM performs a process of displaying the characters of the text data TD stored in the storage unit 10 in the operation control unit 20 on the screen in a state synchronized with the output timing of the voice data from the speaker. You can also run it. In this way, by outputting voice data from the speaker that clearly distinguishes between the text text and the conversation part in the text data TD and the distinction between the Japanese part and the English part, the learning efficiency of foreign language learning for children and students is greatly improved. It is convenient in that it can be improved to.

When the output processing of the voice data in the text data TD of the title selected at the start of processing to the speaker is completed, the operation control unit 20 displays the processing end on the screen and displays the screen requesting the selection of the next title (processing end). ). Here, the case where one text data TD is selected at the start of processing is described, but if a plurality of text data TDs are selected at the start of processing, the output processing of the voice data in the plurality of text data TDs to the speaker is performed. Can also be done.

The text data voice reproduction device 100, the text data voice reproduction program PGM, and the text data voice reproduction method in the present embodiment have been described above, but the text data voice reproduction device 100 and the text data voice reproduction program PGM in the present invention have been described above. It is not limited to the form.

For example, in the present embodiment, when the division text data KTD is extracted from the text data TD, an extraction number assignment process is performed in which the extraction numbers are assigned to each division text data KTD in the order of extraction from the text data TD. However, the assignment of the extraction number to the division text data KTD is not essential and can be omitted.

Further, the text data TD is constructed by referring to the character code assigned to each of the character data constituting the text data TD to the character code information MCJ without dividing the text data TD into the classified text data KTD (x). It is also possible to adopt a form in which the count value of the Japanese discrimination counter NHC and the count value of the non-Japanese discrimination counter HHC are directly calculated for all the character data to be performed to calculate the ratio RT.

Further, the character discrimination unit 24 in the present embodiment discriminates the type of character data by referring to the character code assigned to the character data constituting the text data TD to the character code information MCJ. It may be possible to determine only whether the data type is start quotes or end quotes.

Further, in the above embodiment, the operation control unit 20 as the character data type ratio calculation unit 25 calculates the ratio RT of the Japanese character data type in the character data constituting the text data TD or the divided text data KTD. Then, the operation control unit 20 as the voice data generation engine selection unit 26 compares the calculated ratio RT with the predetermined ratio SRT stored in the storage unit 10 in advance, and generates Japanese voice data according to the comparison result. The process of determining the application of the engine NOE or the non-Japanese voice data generation engine HOE is adopted, but the present invention is not limited to this form.

For example, instead of the character data type ratio calculation unit 25, the operation control unit 20 as a Japanese character code presence / absence determination unit assigns a Japanese character code to the character data constituting the division text data KTD (x) or the text data TD. If it is determined whether or not the data is generated and the operation control unit 20 as the voice data generation engine selection unit 26 has at least one Japanese character code, the operation control unit 20 as the voice data generation engine selection unit 26 determines the division text data KTD (x) or the text data. Japanese voice data generation engine NOE is applied to all character data of TD, and non-Japanese voice data generation engine HOE is applied to all character data in divided text data KTD (x) or text data TD in other cases. It is also possible to execute the processing to be performed.

In addition, the above modification is also applied to the character code ratio calculation function and the voice data generation engine selection function in the text data voice reproduction program PGM, and the character code ratio calculation step and the voice data generation engine selection step in the text data voice reproduction method. It can be applied in the same way. By adopting such a modification, Japanese voice data NOD is generated when even one Japanese character is included, and non-Japanese only when no Japanese character is included. The voice data HOD can be generated.

Further, in the present embodiment, a process of generating Japanese voice data based on the text data TD and generating non-Japanese voice data and then outputting the data to the output unit 30 is performed. The voice data generation engine The process of sequentially outputting the voice data generated by any of the voice data generation engines selected by the selection unit 26 to the output unit 30 may be performed. According to this configuration, it is possible to omit the process of storing the audio data in the storage unit 10 and the configuration of the audio data reproduction control unit.

In the present embodiment, the text data TD describes a form in which the Japanese part and the English part as the non-Japanese part are mixed, but the text data TD composed of only sentences in a single language is used. In that case, the determination and processing as to whether it is Japanese or non-Japanese can be omitted.

Further, in the present embodiment, a process of synchronizing the character data of the text data TD, which is the original data of the voice data, with the output of the voice data to the output unit 30 and outputting it to the display is performed, but this process is omitted. You may.

Further, in the above embodiments, some modifications are listed in the detailed description, but the text data voice reproduction device 100 or the text data voice reproduction program PGM is used for the user at the time of activation. The screen may be displayed so that the user can select what kind of processing is to be performed. Providing such a processing menu selection function is convenient in that learning support can be appropriately provided according to the situation of the child and student.

When the text data TD to be used is in the so-called script format, the voice type characteristic database KDB is associated with the voice type characteristic data JKP corresponding to the person name, which is the voice type characteristic data corresponding to the role name (person name) of the script in advance. You can also let it. At this time, when the motion control unit 20 acts as the conversation voice type change unit 28 (as the conversation voice type change function) and detects a person name from the text data TD, the next person name or the start side quote and the end side quote For the character data up to the detection of, the Japanese voice data NOD is applied by the voice data generation unit 29 (voice data generation function) by applying the voice type characteristic data JKP corresponding to the person name associated with the extracted person name. Alternatively, non-Japanese voice data HOD may be generated.

This will be described using the script-format text data TD shown in FIG. By performing the same processing as in the present embodiment, the text data TD shown in FIG. 6 (A) is divided into the divided text data KTD (x) shown in FIG. 6 (B), and each divided text data KTD (x). The application of the voice generation engine is determined according to the proportion of Japanese in the character data that composes. Then, the operation control unit 20 of the conversation voice type change unit 28 (as a conversation voice type change function) indicates that the person name is included in the division text data KTD (x) (according to the instruction of the text data voice reproduction program PGM). When it is detected, the voice type characteristic data JKP corresponding to the person name associated with the person name detected from the voice type characteristic database KDB is extracted and applied to the voice data generation unit 29 (voice data generation function). As an example of the method of detecting a person's name in the script-type divided text data KTD (x), it is possible to adopt a method of detecting the character string immediately before the colon as the person's name with reference to the position of the colon (:). it can.

For example, since the person name detected in KTD (1) is Taro, the voice type characteristic data of "male part 1" as the person name corresponding voice type characteristic data JKP is applied. Since the person name detected in the next division text data KTD (2) is Jack, the voice type characteristic data of "male part 2" as the person name corresponding voice type characteristic data JKP is applied. Since the person name detected in the next division text data KTD (3) is Jill, the voice type characteristic data of "Woman No. 1" as the person name corresponding voice type characteristic data JKP is applied. In the same manner thereafter, the voice data characteristic data JKP corresponding to the person name associated with the person name detected from each of the divided text data KTD (x) is extracted from the voice type characteristic database KDB, and the voice data generation unit 29 (voice data). The process applied to the generation function) is repeatedly executed. In the script-format text data TD as shown in FIG. 6, the user can also select whether or not to generate or output voice data for the person name portion. As a result, in the script-format text data TD, the voice data of only the so-called dialogue portion can be output from the output unit 30, which is convenient in that the content of the text can be easily understood.

Here, the voice type characteristic database KDB of the storage unit 10 is in a state in which the person name and the voice type characteristic data JKP corresponding to the person name are linked in advance, but the present invention is not limited to this form. A plurality of types of male or female voice type characteristic data are stored in the storage unit 10 as personal name corresponding voice type characteristic data JKP, and the motion control unit 20 of the conversation voice type change unit 28 (as a conversation voice type change function) (According to the instruction of the text data voice reproduction program PGM) In the order of appearance in the divided text data KTD (x), the gender data of the person name is stored in the storage unit 10 in advance according to the gender of the person name, and is referred to in the gender data. (Judgment by doing) Person name correspondence Voice type characteristic data JKP sequentially selects male voice type characteristic data or female voice type characteristic data and applies it to the voice data generation unit 29 (voice data generation function). It is also possible to adopt the form of

Further, in the present embodiment, in the script-format text data TD, a form in which a proper noun is used as a person's name is described, but the person's name is unique such as man 1, woman 1, child 1, etc. It is also possible to adopt a form that uses a person's name that is not a noun.

In addition to the above, various modifications described in the specification for this embodiment and other known configurations can be appropriately combined.

10 Memory unit,
20 Motion control unit,
21 text data acquisition unit, 22 division text data extraction unit,
23 Extraction number assignment unit, 24 character discrimination unit, 25 character data type ratio calculation unit,
26 Voice data generation engine selection part, 27 Conversation part extraction part, 28 Conversation voice type change part,
29 Voice data generator,
30 Output section,
100 Text data audio playback device,
TD text data,
KJ classification conditions,
KTD division text data,
KTD (x) Category text data,
MCJ character code information,
NHC Japanese discrimination counter,
HHC non-Japanese discrimination counter,
RT rate,
SRT predetermined ratio,
NOE Japanese voice data generation engine,
HOE non-Japanese voice data generation engine,
KDB Voice Species Database,
BKP basic voice characteristic data,
CKP Voice type characteristic data for conversation part,
IKP Voice type characteristic data for conversation part in conversation,
KB conversation part,
KKB In-conversation conversation part,
NOD Japanese voice data,
HOD non-Japanese voice data,
JKP person name correspondence voice type characteristic data,
PGM text data audio playback program

Claims (10)

A storage unit that at least stores a voice type characteristic database containing text data and multiple voice type characteristic data,
A text data acquisition unit that acquires the text data from the storage unit,
A character discriminating unit that discriminates the type of the character data for each of the character data constituting the text data,
In the character data constituting the text data, a conversation part extraction unit that extracts a range separated by the character data whose types of character data are start-side quotation marks and end-side quotation marks as conversation parts, and
A voice data generation unit that generates voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and a voice data generation unit.
For the conversation part included in the text data, a conversation voice in which the voice type characteristic data for the conversation part extracted from the voice type characteristic database is applied to the voice data generation unit in place of the basic voice type characteristic data. Seed change department and
An output unit for outputting the voice data generated by the voice data generation unit is provided .
The conversational part extractor, if there is another starter quote and another end quote between the starter quote and the end quote, and the other starter quote. The part separated by the other end quotation marks is extracted as the conversation part in the conversation.
The conversation voice type changing unit substitutes the voice type characteristic data for the conversation part extracted from the voice type characteristic database in the conversation part in the conversation into the voice data generation unit in place of the voice type characteristic data for the conversation part. A text data audio player that is characterized by its application . In the voice type characteristic database, the person name and the voice type characteristic data corresponding to the person name are associated with each other.
When the text data includes the person's name, the conversation voice type changing unit applies the voice type characteristic data associated with the person's name to the voice data generation unit. Item 1. The text data voice reproduction device according to Item 1. The text data voice reproduction device according to claim 2, wherein the voice data generation unit can generate the voice data based on a portion of the text data excluding the part of the person's name . A storage unit that at least stores a voice type characteristic database containing text data and multiple voice type characteristic data,
A text data acquisition unit that acquires the text data from the storage unit,
A character discriminating unit that discriminates the type of the character data for each of the character data constituting the text data,
In the character data constituting the text data, a conversation part extraction unit that extracts a range separated by the character data whose types of character data are start-side quotation marks and end-side quotation marks as conversation parts, and
A voice data generation unit that generates voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and a voice data generation unit.
For the conversation part included in the text data, a conversation voice in which the voice type characteristic data for the conversation part extracted from the voice type characteristic database is applied to the voice data generation unit in place of the basic voice type characteristic data. Seed change department and
An output unit for outputting the voice data generated by the voice data generation unit is provided.
In the voice type characteristic database, the person name and the voice type characteristic data corresponding to the person name are associated with each other.
When the text data includes the person name, the conversation voice type changing unit applies the voice type characteristic data associated with the person name to the voice data generation unit.
The voice data generation unit is a text data voice reproduction device capable of generating the voice data based on a portion of the text data excluding the portion of the person's name . The text data according to any one of claims 1 to 4, wherein the character discriminating unit determines the type of the character data by discriminating the character code assigned to the character data. Voice playback device. Read by a computer that has at least a storage unit that stores at least a voice type characteristic database containing text data and a plurality of voice type characteristic data, an output unit that outputs voice data based on the text data, and an operation control unit. It's possible,
The operation control unit,
A text data acquisition function for acquiring the text data from the storage unit,
A character discrimination function that discriminates the type of the character data for each of the character data constituting the text data, and
In the character data constituting the text data, a conversation part extraction function for extracting a range separated by the character data whose type of character data is a start side quotation mark and an end side quotation mark as a conversation part, and
A voice data generation function that generates voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and
For the conversation part included in the text data, a conversation voice in which the voice type characteristic data for the conversation part extracted from the voice type characteristic database is applied to the voice data generation function in place of the basic voice type characteristic data. Seed change function and
An output function for outputting the voice data generated by the voice data generation function to the output unit is executed.
The conversation part extraction function may be used with other start-side quotes, if there are other start-side quotes and other end-side quotes between the start-side quotes and the end-side quotes. The part separated by the other end quotation marks is extracted as the conversation part in the conversation.
The conversation voice type change function replaces the voice type characteristic data for the conversation part extracted from the voice type characteristic database in the conversation part in the conversation with the voice type characteristic data for the conversation part to the voice data generation function. It applied to text data audio playback program to Rukoto features. In the voice type characteristic database, the person name and the voice type characteristic data corresponding to the person name are associated with each other.
If it contains the person's name into the text data, the conversation voice species changing function, characterized Rukoto to apply the voice type characteristics data bundled with the person's name in the voice data generation function The text data voice reproduction program according to claim 6. The text data voice reproduction program according to claim 7, wherein the voice data generation function can generate the voice data based on a portion of the text data excluding the part of the person's name . Read by a computer that has at least a storage unit that stores at least a voice type characteristic database containing text data and a plurality of voice type characteristic data, an output unit that outputs voice data based on the text data, and an operation control unit. It's possible,
The operation control unit,
A text data acquisition function for acquiring the text data from the storage unit,
A character discrimination function that discriminates the type of the character data for each of the character data constituting the text data, and
In the character data constituting the text data, a conversation part extraction function for extracting a range separated by the character data whose type of character data is a start side quotation mark and an end side quotation mark as a conversation part, and
A voice data generation function that generates voice data based on the basic voice type characteristic data extracted from the voice type characteristic database from the text data, and
For the conversation part included in the text data, a conversation voice in which the voice type characteristic data for the conversation part extracted from the voice type characteristic database is applied to the voice data generation function in place of the basic voice type characteristic data. Seed change function and
An output function for outputting the voice data generated by the voice data generation function to the output unit is executed.
In the voice type characteristic database, the person name and the voice type characteristic data corresponding to the person name are associated with each other.
When the text data includes the person's name, the conversation voice type changing function applies the voice type characteristic data associated with the person's name to the voice data generation function.
A text data voice reproduction program characterized in that the voice data generation function can generate the voice data based on a portion of the text data excluding the part of the person's name . The text data according to any one of claims 6 to 9, wherein the character discrimination function determines the type of the character data by discriminating the character code assigned to the character data. Voice playback program.

Images