JP2011133658A - Device, method and program for synthesizing audio - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the reading correct answer rate even for unknown words not registered in one's own user dictionary by effectively using user dictionaries of other persons without preparing individual field dictionaries beforehand. <P>SOLUTION: An audio synthesizing device includes: a user registration word extracting part which extracts user registration words contained in a user dictionary from a synchronization object text whose input has been accepted by a text input part; a containing rate calculating part which calculates the containing rate of the extracted user registration words contained in the plurality of user dictionaries of other persons via an interface; a reference user dictionary selecting part which selects the user dictionaries of other persons used by the user dictionaries of other persons based on the containing rate calculated for each of the plurality of user dictionaries of other persons; and a reading determining part which determines readings of the unknown words not contained in a basic dictionary or the user dictionary using the selected user dictionaries of other persons to be used. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method for using a user dictionary used during speech synthesis.

  In the speech synthesis technology for inputting text and generating the reading speech, it is necessary to register many words in a language dictionary in which the reading for the word notation is registered in order to increase the correct reading rate of the text. However, as the number of registrations increases, the number of misreads with the same notation increases, causing a reading error.

  Therefore, a technique is disclosed that uses a user dictionary that is independently managed by a user as needed, based on a basic dictionary that collects words that frequently appear in Japanese. A method for selectively using specialized dictionaries that collect technical terms for each field is also disclosed. In addition, a method using another person's user dictionary is also disclosed (see, for example, Patent Document 1, Patent Document 2, Patent Document 3, Patent Document 4, and Patent Document 5).

Japanese Patent No. 3894479 JP 2007-206975 A Japanese Patent No. 3220096 JP-A-11-327871 Japanese Patent Laid-Open No. 2007-80019

  However, these methods have the following problems. First, when using a user dictionary, it is necessary to register a certain number of words by yourself, and the burden of registration work is large for the user.

  Also, when preparing specialized dictionaries by field in advance, it is practically impossible to prepare specialized dictionaries that cannot handle new words or that cover all the vast Japanese vocabulary. There is a problem. Furthermore, when a specialized dictionary is used, there is a problem that a specialized dictionary in an appropriate field cannot be selected unless the way of classifying the field is common between the user and the creator of the specialized dictionary.

  In the method of using another person's user dictionary, a plurality of other person's user dictionaries are examined for a certain notation unknown word, and a reading with a high registration rate is taken into the basic dictionary as a reading for the unknown word. However, in this method, since only one reading can be registered for one notation, it is not possible to cope with the same notation.

  Also, by comparing the degree of similarity between your user dictionary and another person's user dictionary, you can incorporate words that are included in another person's dictionary with a high degree of similarity but not in your own dictionary into your own dictionary. The user dictionaries are individually managed by individuals, and there is a problem that the probability that words similar to those are registered is small. For example, suppose that Mr. A is interested in politics, economy, and baseball, synthesizes those sentences well, and the words related to them are registered in the user. In addition, Mr. B is interested in performing arts, weather, and baseball, and registers words related to them as a user. Here, looking at the similarity between the user dictionary registered words of Mr. A and Mr. B, only the baseball part is similar, and 2/3 is not similar and cannot be shared. Also, if both A and B have registered only baseball-related words, even if the degree of similarity is high, the number of words that can be taken into A's dictionary because the degree of similarity is high, and the sharing effect Can not get much. Thus, the probability of sharing is small, and even if it can be shared, the effect is small.

  Therefore, the present invention aims to improve the reading accuracy rate even for unknown words that are not registered in one's user dictionary by effectively using another person's user dictionary. To do.

  In order to achieve the above object, a speech synthesizer disclosed below includes a text input unit that accepts an input of a text to be synthesized, and a user registration included in a user dictionary from a text to be synthesized that is accepted by the text input unit. User registration word extraction unit for extracting words, user registration word extracted by the user registration word extraction unit, content rate calculation unit for calculating the content rate included in the other person's user dictionary, and user calculated by the content rate calculation unit Based on the content rate of the registered word, a reference user dictionary selection unit that selects a user dictionary to be used from another user's dictionary and a user dictionary to be used selected by the reference user dictionary selection unit are used as a basic dictionary. And a reading determining unit that determines reading of unknown words not included in the user dictionary.

  According to the above configuration, it is possible to improve the correct reading rate even for unknown words that are not registered in the user dictionary by effectively using the user dictionary of another person.

1 is a block diagram showing the overall configuration of a speech synthesizer according to Embodiment 1 of the present invention. The flowchart which shows operation | movement of the speech synthesizer which concerns on Embodiment 1 of this invention. Diagram showing examples of various data The block diagram which shows the whole structure of the speech synthesizer which concerns on Embodiment 2 and 3 of this invention The block diagram which shows the whole structure of the speech synthesizer concerning Embodiment 4 of this invention.

[Embodiment 1]
FIG. 1 is a block diagram showing the overall configuration of a speech synthesis apparatus 100 according to Embodiment 1 of the present invention. In FIG. 1, the speech synthesizer 100 includes a basic dictionary 110, a user dictionary 120, a user dictionary interface unit 130, a text input unit 140, a language processing unit 150, a waveform processing unit 160, and a speech output unit 170.

  The basic dictionary 110 is a dictionary that stores basic words necessary for speech synthesis. The user dictionary 120 is a user-specific dictionary in which a user registers words as needed. The other person's user dictionaries 121, 122, and 123 are dictionaries unique to other users in which other users individually register words at any time. The basic dictionary 110 stores, for example, word notation (= entry words), readings, accents, and the like that frequently appear in Japanese. The user dictionary 120 stores notation (= entry words), readings, accents, and the like of words registered by the user. The other person's user dictionaries 121, 122, 123 store notation (= entry words), readings, accents, and the like of words registered by other users. The user dictionary information can be exchanged between different users via the user dictionary interface unit 130 and the interface units 131 to 133 of another person's user dictionary.

  Text input unit 140 accepts input of text to be synthesized. For example, a configuration in which a user inputs text through a keyboard, a configuration in which electronic input is performed through a drive that reads a medium such as a CD or a flexible disk, a configuration in which text read by an OCR by a scanner is input, or wired or A configuration in which text is electronically received via a wireless network, or a combination thereof may be used.

  The input text is sent to the language processing unit 150. The language processing unit 150 analyzes and outputs the input text reading, accent, and the like. The language processing unit 150 includes a morphological analysis unit 151, an unknown word extraction unit 152, a user registered word extraction unit 153, a content rate calculation unit 154, a reference user dictionary selection unit 155, and a reading determination unit 156.

  The morpheme analysis unit 151 performs morpheme analysis using the basic dictionary 110 and the user dictionary 120 to determine reading. The unknown word extraction unit 152 extracts, as an unknown word, a word determined not to be registered in the basic dictionary 110 or the user dictionary 120 in the input text from the result of decomposing the text into words by morphological analysis. The user registration word extraction unit 153 extracts words registered in the user dictionary 120 from the input text from the result of morphological analysis. Note that the extraction processing in the unknown word extraction unit 152 and the user registered word extraction unit 153 may be extracted as a comparison result between a simple text and various dictionaries without performing morphological analysis.

  The content rate calculation unit 154 calculates how much the unknown word and the user registration word are included in the other person's user dictionaries 121 to 123. At this time, the user dictionary interface unit (UDIC-IF) 130 acquires information on the user dictionary of the other person via the line, and sends the information to the content rate calculation unit 154. Further, the user dictionary interface unit 130 provides information included in the user dictionary 120 of the user when there is a user dictionary information reference request from another person's speech synthesizer.

  The reference user dictionary selection unit 155 is based on the content ratio indicating how much the unknown word and the user registration word extracted by the unknown word extraction unit 152 and the user registration word extraction unit 153 are included in the other person's user dictionary. A user dictionary with a high content rate is determined as a user dictionary to be used from among a plurality of other user dictionaries. For example, it is determined as a user dictionary that should use another person's user dictionary whose content rate is a predetermined value or more. In addition, you may determine with the user dictionary which should utilize the user dictionary of others with the highest content rate. Alternatively, a predetermined number of other user dictionaries may be selected in descending order of content rate, and used as a user dictionary to be used.

  The reading determination unit 156 determines the reading of the input text, the accent, and the like using the word information included in the user dictionary of the other person to be used selected by the basic dictionary 110, the user dictionary 120, and the reference user dictionary selection unit 155. To do. Examples of the determination method include the following two methods. That is, (1) the word extracted as an unknown word by the unknown word extraction unit is compared with the notation of another person's user dictionary word, and the reading of a headword having a matching notation is set, or (2) By using words in the basic dictionary 110, the user dictionary 120, and another user's user dictionary to be used, the reading and accent can be determined by performing morphological analysis again. Note that the reading determination unit 156 may further determine accents and other necessary information in addition to reading.

  The waveform processing unit 160 generates synthesized speech data according to the reading and accent information output from the language processing unit 150. Although not shown, the waveform processing unit 160 may have a waveform dictionary for synthesizing speech. For example, the waveform processing unit 160 becomes a target accent for a speech unit in the waveform dictionary by digital signal processing using, for example, a PSOLA (Pitch Synchronous Overlap Add) method which is one of linear prediction analysis methods. Thus, it is possible to generate a synthesized speech by connecting while adjusting the pitch of the voice.

  The audio output unit 170 converts the audio data generated by the waveform processing unit 160 into a format corresponding to various audio formats and outputs the converted data.

  In the above description, the speech synthesizer 100 includes the basic dictionary 110, the user dictionary 120, the user dictionary interface unit 130, the text input unit 140, the language processing unit 150, the waveform processing unit 160, and the speech output unit 170. However, the configuration of the speech synthesizer 100 is not limited to this. For example, the speech synthesizer 100 may be on a server connected to a network. In this case, for example, the text input unit 140 is configured to receive text input from a user terminal connected to the network, and the voice output unit 170 transmits the synthesized voice data to the user terminal via the network. It can be set as the structure to do. In addition, when the server stores the other person's user dictionary and accepts text input from a plurality of users, the user dictionary can be used for the plurality of users to synthesize speech for the input text. It is good also as a structure. Further, the functional units included in the speech synthesizer 100 may be distributed among a plurality of computers.

  Hereinafter, the operation of the speech synthesizer according to the first embodiment of the present invention will be described with reference to the flowchart in FIG. 2 and various data examples in FIG. 3. The user dictionaries of myself, Mr. A, Mr. B, and Mr. C are related to the texts that are usually read aloud as shown in FIGS. 3 (e), (a), (b), and (c). It is assumed that the word is registered as a user.

  First, it is assumed that the composition target text describing the IC card shown in FIG. 3D is input to the text input unit 140 (step S201).

  Next, the morpheme analysis unit 151 performs morpheme analysis on the input composition target text (step S202).

  The user registration word extraction unit 153 extracts, as a user registration word, a portion in which the word registered in the user dictionary is used from the result of decomposing the text into words by morphological analysis (step S203). Here, “RAM” and “ROM” registered in the user dictionary shown in FIG. 3E are extracted as user registration words.

  As a result of the morphological analysis, the unknown word extraction unit 152 extracts a portion that is not registered in the basic dictionary or the user dictionary as an unknown word (step S204). Here, it is assumed that “IC” and “EEPROM”, which are not registered in the basic dictionary or the user dictionary, are detected as unknown words. Incidentally, “EEPROM” is an abbreviation of “Electrically Erasable and Programmable Read Only Memory” and is a kind of ROM that can be electrically rewritten.

  Next, the content rate calculation unit 154 calculates a content rate in which the user registration word and the unknown word extracted by the user registration word extraction unit 153 and the unknown word extraction unit 152 are included in another person's dictionary (step S205). ). In comparison with Mr. A's user dictionary shown in FIG. 3A, a total of four words of user registration words (“RAM”, “ROM”) and unknown words (“IC”, “EEPROM”) are included. Among them, since only one of the first unknown words “IC” is included, the content rate is 25%. On the other hand, Mr. B's user dictionary shown in FIG. 3B includes all of the user registration words “RAM” and “ROM” and unknown words “IC” and “EEPROM”. Because the total of 4 words of user registration words and unknown words are all included, the content rate is 100%, and Mr. B can already read the text in the same field as the current synthesis target text It turns out that the nature is high. In addition, since Mr. C's user dictionary does not include any word that matches the total of four words of user registration words and unknown words, the content rate is 0%.

  The reference user dictionary selection unit 155 selects another person's user dictionary to be used based on the content rate (step S206). Here, it is determined that the user dictionary of Mr. B having the highest content rate is 100% and that the dictionaries of Mr. A and Mr. C having a low content rate are not used.

  In this description, all of the unknown words were included in Mr. B's dictionary with the highest content ratio. However, when all the unknown words are not included in the dictionary of the first other person with the highest content ratio, The second and third dictionaries that contain unknown words that are not included in the first other person's dictionary with a high content rate are selected from, for example, other people's user dictionaries whose content rate is equal to or greater than a certain threshold. A plurality of other user's dictionaries may be selected so that as many unknown words as possible are included.

  The reading determination unit 156 refers to the user dictionary of the other person selected in step S206 for the part extracted as the unknown word in the unknown word extraction result in step 204, and if there is a headword that can be used, the reading determination unit 156 Used to determine reading, accent, etc. Here, for the unknown words “IC” and “EEPROM”, reading “IC” and “EEPROM” in B's user dictionary and accent information are used to read the input text and determine the accent (step S207). ).

  With the above method, it is possible to correctly obtain a reading of a word that is not registered in the user dictionary. User-registered words and words included in unknown words are words that could not be covered by the basic word dictionary in which words that appear frequently in Japanese are registered, so they contain many proper nouns, technical terms, etc. It is a word that represents the characteristics of the text content. Therefore, by comparing these words with words registered in the other person's user dictionary, it is possible to estimate other users who have already read out the text having the same content. In addition, since the user dictionary is prepared for reading out text with similar contents, there is a high probability that the correct reading is inevitably registered, and it is possible to avoid the problem of misread words. Further, unlike the case where a conventional field-specific dictionary is used, it is not necessary to estimate a specific field from the text to be synthesized. Furthermore, since the latest user dictionary maintained by the user can be used for new words, the new words can be handled. In addition, although this embodiment demonstrated the structure which calculates two content rates of a user registration word and an unknown word, it is good also as a structure which calculates any one content rate.

[Embodiment 2]
The speech synthesizer according to the present embodiment further includes a specialized dictionary 180 that is a dictionary in which specialized terms for each specialized field are registered, and the language processing unit 150 receives the input from the synthesis target text received by the text input unit 140. Further, a specialized dictionary registered word extraction unit 190 that extracts specialized dictionary registered words included in the specialized dictionary 180 is further provided. Furthermore, the content rate calculation unit 154 of the second embodiment includes a weighted content rate calculation unit 200, and the user registration word extracted by the user registration word extraction unit 153, the unknown word extracted by the unknown word extraction unit 152, and the specialized dictionary registration Weighting is performed according to the type of a word such as the specialized dictionary registered word extracted by the word extraction unit 190, and the content rate in which these words are included in the other person's user dictionaries 121, 122, 123 is calculated. The reference user dictionary selection unit 155 selects a user dictionary of another person whose weighted content rate is equal to or greater than a predetermined value based on the weighted content rate of the user registered words, unknown words, and specialized dictionary registered words calculated by the weighted content rate calculating unit. Select as user dictionary to be used.

  FIG. 4 is a block diagram showing the overall configuration of the speech synthesizer according to Embodiment 2 of the present invention. The configuration of the first embodiment further includes a specialized dictionary 180 and a specialized dictionary registered word extraction unit 190, and the content rate calculation unit 154 includes a weighted content rate calculation unit 200.

  When calculating the content rate, the weighted content rate calculation unit 200 increases the weight of the content rate of specific types of words such as user registration words, unknown words, and proper noun words, and specific parts of speech. For example, one user registered word, one unknown word, and one specialized dictionary registered word are extracted from the synthesis target text, and the user registered word and the unknown word are included in Mr. A's user dictionary. It is assumed that the dictionary includes unknown words and specialized dictionary registered words. At this time, if the content rate is calculated as it is, the content rate is 2 / 3≈66% for each of the user dictionaries of Mr. A and Mr. B, and there is no difference in the content rate. Here, it is assumed that the weighted content ratio calculation unit 200 sets the weight of the content ratio to be heavy for the specialized dictionary registered words. That is, if the number of specialized dictionary registration words actually contained is 1, but the weighting weight is set to twice, when calculating the content rate, the extracted user registration words, unknown words, The content rate is calculated by increasing the number of specialized dictionary registration words by weighting the total number of specialized dictionary registration words 3. That is, the weighted content of Mr. B is the extracted user registration word, unknown word, technical dictionary registration word for one word of the unknown word and two words of the professional dictionary registration word weighted and increased. The content rate is calculated by the total number of 3 and the content rate after weighting is (1 + 2) / 3≈100%. In addition, although the example applied to a user registration word, an unknown word, and a specialized dictionary registration word was shown as a word kind, the kind of word can be classified also by the kind of part of speech, such as a proper noun and a common noun. Further, the word type may be classified by combining the dictionary type and the part of speech, such as a common noun of the user registration word and a proper noun of the user registration word.

  The reference user dictionary selecting unit 155 is configured such that the weighted content rate calculating unit 200 weights the user registered words, unknown words, and specialized dictionary registered words according to the types of specific words, the user registered words, unknown words, and proper nouns. A user dictionary to be used is selected from another person's user dictionary based on the content rate of dictionary registered words. For example, a weighted content rate is selected as a user dictionary to be used by another person's user dictionary. In addition, you may determine with the user dictionary which should utilize the user dictionary of others with the highest content rate weighted.

  The reading determination unit 156 reads the input text using word information registered in the user dictionary of the other person to be used and selected by the basic dictionary 110, the user dictionary 120, the specialized dictionary 180, and the reference user dictionary selection unit 155. Determine accents.

  As a result, even if the text with the same content is read aloud, if the weight of specialized dictionary registered words is set to be heavy, the content rate of words with higher expertise is emphasized. It is possible to find a user dictionary of another person to be used while preventing an error with high reliability. In addition, when the weights for unknown words are set to be heavy, a user dictionary including more unknown words can be used, the reading grant rate for unknown words can be improved, and the correct reading rate according to the purpose can be improved. Improvements can be made.

[Embodiment 3]
The overall configuration of the speech synthesizer according to the third embodiment of the present invention is the block diagram shown in FIG.

  In the third embodiment, the weighted content rate calculation unit 200 performs different weighting according to the type of word and the type of part of speech when calculating the content rate of each of the user registered word, unknown word, and specialized dictionary registered word. . For example, weighting is performed so that the weights increase in the order of specialized dictionary registration word> user registration word of proper noun> unknown word> user registration word of common noun.

  The reference user dictionary selection unit 155 performs user registration after the weighted content rate calculation unit 200 performs weighting so that the weights increase in the order of specialized dictionary registration word> user registration word of proper noun> unknown word> user registration word of common noun. Based on the content of words, unknown words, and specialized dictionary registered words, a user dictionary of another person that can be used is selected from among a plurality of other person's user dictionaries. For example, the user dictionary of another person whose content rate is equal to or higher than a predetermined value is selected as a user dictionary of another person to be used. The other person's user dictionary with the highest content rate may be determined as the other person's user dictionary. In addition, as described above, not only weighting is performed according to the type of word or part of speech of each of the user registered word, unknown word, and specialized dictionary registered word, but also the appearance frequency (number of appearances) of the word itself. It is good also as a structure which weights. For example, if the word “timely hit” is extracted as an unknown word in the synthesis target text related to baseball, and “timely hit” appears multiple times in the text to be synthesized, it is called “timely hit”. You may perform the weighting which makes a weight heavy according to the appearance frequency of a word.

  The reading determination unit 156 determines the reading of the input text, the accent, and the like using the user dictionary words of other users to be used selected by the basic dictionary 110, the user dictionary 120, the specialized dictionary 180, and the reference user dictionary selection unit 155. .

  This makes it possible to emphasize the content rate of specialized dictionary registered words and the content rate of user dictionary registered words that are likely to represent the content of the text to be synthesized, and to read out text with similar content However, it is possible to find a user dictionary of another person to be used while avoiding the problem of the same notation word with a high probability. In the second and third embodiments, the three content rates of the user dictionary registered word, unknown word, and specialized dictionary registered word are calculated, but the content rate is calculated for any one of the three. Also good.

[Embodiment 4]
The speech synthesizer according to the present embodiment selects a plurality of user dictionaries to be used by the reference user dictionary selection unit 155, and when the unknown word reading differs between the plurality of user dictionaries to be used, Of the plurality of user dictionaries to be used, an unknown word reading and accent are determined using the user dictionary having the highest content rate.

  FIG. 5 is a diagram illustrating a configuration of the reference user dictionary selection unit 155 and the reading determination unit 156 in the speech synthesis apparatus 100 according to Embodiment 4 of the present invention. In contrast to the second and third embodiments, the reference user dictionary selection unit 155 further includes a content rate comparison unit 157, the same notation misreading word detection unit 158, and a priority determination unit 159.

  The content rate comparison unit 157 compares the content rate of each other person's dictionary calculated by the content rate calculation unit 154, and selects the other person's user dictionary having a certain content rate or higher as a user dictionary to be used. .

  The same notation different word detection unit 158 is unknown between another user's dictionary that is equal to or higher than the certain content rate selected by the content rate comparison unit, or between the user's user dictionary and the selected user's user dictionary. It is detected whether there is a user dictionary in which different readings are registered for the word and the entry word of the user dictionary registered word.

  When there are a plurality of user dictionaries in which different readings are registered with respect to an unknown word entry word, the priority determination unit 159 determines the content rate of the user registered word or the unknown word alone or the unknown in the other person's user dictionary. Prioritize which user dictionary reading is used according to the number of registered words. For example, you may set a high priority to the user dictionary of others with a high content rate of a user registration word. Also, a high priority may be set for the user dictionary of another person who has a large number of registered headwords for unknown words.

  If priorities are determined in this way, in the former case, it is possible to focus on reducing reading errors of misread words, and in the latter case, it is possible to focus on increasing the coverage of unknown words so that unknown words can be read. become.

  The reading determination unit 156 includes a temporary dictionary creation unit 161, a morpheme analysis unit 162, and an accent determination unit 163, and uses words in the basic dictionary 110, the user dictionary 120, and other users' user dictionary to be used, Perform morphological analysis to determine reading and accents.

  When the other person's user dictionary to be used is determined, the temporary dictionary creating unit 161 includes the determined other person's user dictionary in his basic dictionary and user dictionary, and creates a temporary dictionary for the text to be synthesized.

  The morpheme analysis unit 162 performs morpheme analysis on the synthesis target text again using a temporary dictionary, and gives a reading.

  The accent determination unit 163 performs an accent combination process while referring to an accent determination rule and the like, and determines a final accent.

  Unlike Embodiment 1, morphological analysis is performed again on the first language dictionary that does not contain unknown words by performing morphological analysis again on the synthesis target text using a language dictionary that enables reading of unknown words. Compared to the case, it is possible to cope with phenomena such as compound word reading change and accent coupling, and the accuracy of morphological analysis is increased. If morphological analysis is performed with a language dictionary that does not contain unknown words, it is possible to identify the unknown word part, but it cannot cope with accent concatenation and reading changes with words before and after the unknown word. The accuracy of the portion is improved by the configuration.

  As described above, according to the present embodiment, an unknown word that is not registered in one's own user dictionary can be obtained by effectively using another user's dictionary without preparing a field-specific dictionary in advance. Also, the correct reading rate can be improved. In addition, it is possible to deal with the problem of differently read words and to correctly analyze readings for unknown words that are not registered in the user dictionary. Further, it is possible to reduce the user dictionary registration effort and increase the probability of reading correctly. In addition, it can handle new words.

  The configuration described in the above embodiment merely shows a specific example, and does not limit the technical scope of the present invention. Any configuration can be employed within the scope of the effects of the present invention.

  In the embodiment of the present invention, a software program for realizing the above-described embodiment (in the embodiment, a program corresponding to the flowchart shown in FIG. 2) is supplied to the apparatus, and a computer of the apparatus is supplied. This includes the case where it is also achieved by reading and executing the program. Therefore, in order to realize the functional processing described in this embodiment by a computer, the program itself installed in the computer is also an embodiment of the present invention. That is, a program for realizing the functional processing of the present invention is also included in one aspect of the embodiment.

  The following additional notes are further disclosed with respect to the above-described first to fourth embodiments.

(Appendix 1)
A speech synthesizer having a basic dictionary and a user-specific user dictionary,
A text input part that accepts input of text to be synthesized;
An interface unit that can refer to a plurality of other user's dictionaries;
A user-registered word extraction unit that extracts a user-registered word included in the user dictionary from the text to be synthesized that is accepted by the text input unit;
About a user registration word extracted by the user registration word extraction unit, a content rate calculation unit that calculates a content rate included for each of the plurality of other user's dictionaries via the interface unit,
Based on the content rate calculated for each of the plurality of other person's user dictionaries, a reference user dictionary selecting unit that selects the other person's user dictionary to be used from the other person's user dictionary;
A speech synthesizer comprising: a reading determination unit that determines reading of an unknown word that is not included in either the basic dictionary or the user dictionary using the user dictionary of another person to be used selected by the reference user dictionary selection unit.

(Appendix 2)
Furthermore, an unknown word extraction unit that extracts an unknown word that is not included in the basic dictionary or the user dictionary of the user from the synthesis target text,
The speech synthesizer according to appendix 1, wherein the content rate calculation unit calculates the content rate included in each of the plurality of other user's dictionaries via the interface unit for the unknown word extracted by the unknown word extraction unit.

(Appendix 3)
The content rate calculation unit calculates the content rate that the user registration word extracted by the user registration word extraction unit and the unknown word extracted by the unknown word extraction unit are included in another user's dictionary,
The reference user dictionary selection unit calculates a user dictionary of another person to be used from the plurality of other person's user dictionaries based on a user registration word calculated for each of the plurality of other person's user dictionaries and an unknown word content rate. The speech synthesizer according to appendix 2, which is selected.

(Appendix 4)
In addition, a specialized dictionary in which specialized words for each specialized field are registered,
A specialized dictionary registered word extraction unit that extracts a specialized dictionary registered word included in the specialized dictionary from the synthesis target text accepted by the text input unit;
The content rate calculation unit calculates the content rate that the user registration word extracted by the user registration word extraction unit and the professional dictionary registration word extracted by the specialized dictionary registration word extraction unit are included in the other person's user dictionary,
The reference user dictionary selection unit selects a user dictionary of another person to be used from the plurality of other person's user dictionaries based on the user registration word calculated by the content rate calculation unit and the content rate of the specialized dictionary registration word. The speech synthesizer according to appendix 1.

(Appendix 5)
An unknown word extraction unit that extracts unknown words that are not included in the basic dictionary or the user dictionary of the user from the text to be synthesized;
A specialized dictionary in which specialized words for each specialized field are registered,
A specialized dictionary registration word extraction unit that extracts a specialized dictionary registration word included in the specialized dictionary from the synthesis target text accepted by the text input unit;
The content rate calculation unit may include at least one of a user registration word extracted by the user registration word extraction unit, an unknown word extracted by the unknown word extraction unit, and a dictionary registration word extracted by a specialized dictionary registration word extraction unit. Based on the content rate included in the other person's user dictionary,
The speech synthesis device according to attachment 1, wherein the reference user dictionary selection unit selects a user dictionary of another person to be used from the plurality of other person's user dictionaries based on the content rate calculated by the content rate calculation unit.

(Appendix 6)
The content rate calculation unit calculates the content rate by weighting according to the type of word,
The reference user dictionary selection unit selects any other user's user dictionary to be used from the plurality of other person's user dictionaries based on the content rate after the content rate calculation unit is weighted. A speech synthesizer according to claim 1.

(Appendix 7)
A plurality of other user dictionaries to be used by the reference user dictionary selection unit;
When the reading of unknown words is different among the plurality of other user dictionary to be used, the reading determining unit uses the other user dictionary having the highest content rate among the selected user dictionary to be used. The speech synthesizer according to any one of appendices 1 to 6, which determines reading of an unknown word.

(Appendix 8)
A speech synthesis method that uses a basic dictionary and a user-specific user dictionary,
A text input step in which the computer accepts input of the text to be synthesized;
A user-registered word extracting step in which the computer extracts a user-registered word included in a user dictionary from the synthesis target text received in the text input step;
About the user registration word extracted in the user registration word extraction step, the computer calculates a content ratio included in each of a plurality of other user's dictionaries via the interface unit,
A reference user dictionary selection step in which the computer selects a user dictionary of another person to use from the user dictionary of the other person based on the content rate calculated for each of the plurality of other person user dictionaries;
Using a user dictionary of another person to be used selected by the reference user dictionary selecting unit, and a reading determining step in which the computer determines a reading of an unknown word that is not included in the basic dictionary or the user dictionary. Synthesis method.

(Appendix 9)
Furthermore, the computer includes an unknown word extraction step of extracting an unknown word that is not included in the basic dictionary or the user's user dictionary from the composition target text,
9. The speech synthesis method according to appendix 8, wherein in the content rate calculation step, the content rate included in each of the plurality of other user's dictionaries is calculated via the interface unit for the unknown word extracted in the unknown word extraction step.

(Appendix 10)
A speech synthesis program that uses a basic dictionary and a user-specific user dictionary,
On the computer,
A text input step that accepts input of text to be synthesized;
A user registration word extraction step for extracting a user registration word included in the user dictionary from the synthesis target text accepted in the text input step;
About the user registration word extracted in the user registration word extraction step, a content rate calculation step for calculating a content rate included for each of a plurality of other user's dictionaries via the interface unit;
A reference user dictionary selection step of selecting a user dictionary of another person to be used from the user dictionary of the other person based on the content rate calculated for each of the plurality of other person user dictionaries;
A speech synthesis program for executing a reading determination step for determining reading of an unknown word that is not included in either the basic dictionary or the user dictionary using the user dictionary of another person to be used selected by the reference user dictionary selection unit .

(Appendix 11)
Furthermore, the computer is caused to execute an unknown word extraction step for extracting an unknown word that is not included in the basic dictionary or the user dictionary of the user from the text to be synthesized.
The speech synthesis program according to appendix 10, wherein in the content rate calculation step, the content rate included in each of the plurality of other user's dictionaries is calculated via the interface unit for the unknown word extracted in the unknown word extraction step.

(Appendix 12)
In the content rate calculation step, the user registration word extracted in the user registration word extraction step, and the content rate that the unknown word extracted in the unknown word extraction step is included in another user's dictionary,
In the reference user dictionary selecting step, based on the user registration word calculated for each of the plurality of other person's user dictionaries and the content rate of unknown words, the other person's user dictionary to be used from the plurality of other person's user dictionaries is obtained. The speech synthesis program according to appendix 11, which is selected.

(Appendix 13)
In the computer,
Furthermore, from the synthesis target text accepted by the text input unit, to execute a specialized dictionary registration word extraction step for extracting specialized dictionary registration words included in a specialized dictionary in which specialized words for each specialized field are registered,
In the content rate calculation step, the user registration word extracted in the user registration word extraction step, and the content rate in which the specialized dictionary registration word extracted in the specialized dictionary registration word extraction step is included in another user's dictionary,
In the reference user dictionary selection step, a user dictionary of another person to be used is selected from the plurality of other person's user dictionaries based on the user registration word calculated in the content ratio calculation step and the content ratio of the specialized dictionary registration word. The speech synthesis program according to attachment 11.

100 speech synthesizer 110 basic dictionary 120 user dictionary 130 user dictionary interface unit 140 text input unit 150 language processing unit 151 morpheme analysis unit 152 unknown word extraction unit 153 user registered word extraction unit 154 content rate calculation unit 155 reference user dictionary selection unit 156 Reading decision unit 157 Content rate comparison unit 158 Same notation different word detection unit 159 Priority decision unit 160 Waveform processing unit 161 Temporary dictionary creation unit 162 Morphological analysis unit 163 Accent decision unit 170 Speech output unit 180 Special dictionary 190 Special dictionary registration word Extraction unit 200 Weighted content rate calculation unit

Claims (9)

A speech synthesizer having a basic dictionary and a user-specific user dictionary,
A text input part that accepts input of text to be synthesized;
An interface unit that can refer to a plurality of other user's dictionaries;
A user-registered word extraction unit that extracts a user-registered word included in the user dictionary from the text to be synthesized that is accepted by the text input unit;
About a user registration word extracted by the user registration word extraction unit, a content rate calculation unit that calculates a content rate included for each of the plurality of other user's dictionaries via the interface unit,
Based on the content rate calculated for each of the plurality of other person's user dictionaries, a reference user dictionary selecting unit that selects the other person's user dictionary to be used from the other person's user dictionary;
A speech synthesizer comprising: a reading determination unit that determines reading of an unknown word that is not included in either the basic dictionary or the user dictionary using the user dictionary of another person to be used selected by the reference user dictionary selection unit. Furthermore, an unknown word extraction unit that extracts an unknown word that is not included in the basic dictionary or the user dictionary of the user from the synthesis target text,
The speech synthesizer according to claim 1, wherein the content rate calculation unit calculates a content rate included in each of the plurality of other user's dictionaries via the interface unit for the unknown word extracted by the unknown word extraction unit. . The content rate calculation unit calculates the content rate that the user registration word extracted by the user registration word extraction unit and the unknown word extracted by the unknown word extraction unit are included in another user's dictionary,
The reference user dictionary selection unit calculates a user dictionary of another person to be used from the plurality of other person's user dictionaries based on a user registration word calculated for each of the plurality of other person's user dictionaries and an unknown word content rate. The speech synthesizer according to claim 2, which is selected. In addition, a specialized dictionary in which specialized words for each specialized field are registered,
A specialized dictionary registered word extraction unit that extracts a specialized dictionary registered word included in the specialized dictionary from the synthesis target text accepted by the text input unit;
The content rate calculation unit calculates the content rate that the user registration word extracted by the user registration word extraction unit and the professional dictionary registration word extracted by the specialized dictionary registration word extraction unit are included in the other person's user dictionary,
The reference user dictionary selection unit selects a user dictionary of another person to be used from the plurality of other person's user dictionaries based on the user registration word calculated by the content rate calculation unit and the content rate of the specialized dictionary registration word. The speech synthesizer according to claim 1. An unknown word extraction unit that extracts unknown words that are not included in the basic dictionary or the user dictionary of the user from the text to be synthesized;
A specialized dictionary in which specialized words for each specialized field are registered,
A specialized dictionary registration word extraction unit that extracts a specialized dictionary registration word included in the specialized dictionary from the synthesis target text accepted by the text input unit;
The content rate calculation unit may include at least one of a user registration word extracted by the user registration word extraction unit, an unknown word extracted by the unknown word extraction unit, and a dictionary registration word extracted by a specialized dictionary registration word extraction unit. Based on the content rate included in the other person's user dictionary,
The speech synthesizer according to claim 1, wherein the reference user dictionary selection unit selects a user dictionary of another person to be used from the plurality of other person user dictionaries based on the content rate calculated by the content rate calculation unit. . A speech synthesis method that uses a basic dictionary and a user-specific user dictionary,
A text input step in which the computer accepts input of the text to be synthesized;
A user-registered word extracting step in which the computer extracts a user-registered word included in a user dictionary from the synthesis target text received in the text input step;
About the user registration word extracted in the user registration word extraction step, the computer calculates a content ratio included in each of a plurality of other user's dictionaries via the interface unit,
A reference user dictionary selection step in which the computer selects a user dictionary of another person to use from the user dictionary of the other person based on the content rate calculated for each of the plurality of other person user dictionaries;
Using a user dictionary of another person to be used selected by the reference user dictionary selecting unit, and a reading determining step in which the computer determines a reading of an unknown word that is not included in the basic dictionary or the user dictionary. Synthesis method. A speech synthesis program that uses a basic dictionary and a user-specific user dictionary,
On the computer,
A text input step that accepts input of text to be synthesized;
A user registration word extraction step for extracting a user registration word included in the user dictionary from the synthesis target text accepted in the text input step;
About the user registration word extracted in the user registration word extraction step, a content rate calculation step for calculating a content rate included for each of a plurality of other user's dictionaries via the interface unit;
A reference user dictionary selection step of selecting a user dictionary of another person to be used from the user dictionary of the other person based on the content rate calculated for each of the plurality of other person user dictionaries;
A speech synthesis program for executing a reading determination step for determining reading of an unknown word that is not included in either the basic dictionary or the user dictionary using the user dictionary of another person to be used selected by the reference user dictionary selection unit . In the computer,
Further, an unknown word extraction step for extracting unknown words that are not included in the basic dictionary or the user dictionary of the user from the text to be synthesized is performed,
In the content rate calculation step, the user registration word extracted in the user registration word extraction step, and the content rate that the unknown word extracted in the unknown word extraction step is included in another user's dictionary,
In the reference user dictionary selecting step, based on the user registration word calculated for each of the plurality of other person's user dictionaries and the content rate of unknown words, the other person's user dictionary to be used from the plurality of other person's user dictionaries is obtained. The speech synthesis program according to claim 7 to be selected. In the computer,
Furthermore, from the synthesis target text accepted by the text input unit, to execute a specialized dictionary registration word extraction step for extracting specialized dictionary registration words included in a specialized dictionary in which specialized words for each specialized field are registered,
In the content rate calculation step, the user registration word extracted in the user registration word extraction step, and the content rate in which the specialized dictionary registration word extracted in the specialized dictionary registration word extraction step is included in another user's dictionary,
In the reference user dictionary selection step, a user dictionary of another person to be used is selected from the plurality of other person's user dictionaries based on the user registration word calculated in the content ratio calculation step and the content ratio of the specialized dictionary registration word. The speech synthesis program according to claim 7.

Images