JP6567372B2 - Editing support apparatus, editing support method, and program - Google Patents

Description

  Embodiments described herein relate generally to an editing support apparatus, an editing support method, and a program.

  In recent years, speech synthesis technology has been used in many places with improvements in sound quality and speaker characteristics. The speech synthesis technology is used in, for example, narrations of public facilities and public transportation, and interactive interfaces with entertainment and systems. In addition, attempts have been made to automate read-out speech of electronic books and the like using speech synthesis technology. In general, books contain unique phrases such as new words, unknown words and proper nouns, expressions and notations. In a speech synthesis system that synthesizes read-out speech, it is difficult to mechanically always correctly estimate the reading and accent of an arbitrary natural sentence (a kanji-kana mixed sentence). For this reason, a method of manually correcting a portion where the speech synthesis system cannot create a mechanically correct read-out speech is often used. For example, a method is used in which correct readings, accents, and the like are designated in advance by hand at locations where reading errors and accent errors occur.

JP2013-134396A

  However, there are cases in which it is difficult for a user who performs a reading voice editing operation to specify reading and accents intuitively.

The editing support apparatus according to the embodiment includes a storage unit, an analysis unit, a presentation unit, a storage control unit, and a search unit . The storage unit stores a character string and reading information including a reading of the character string and an accent type. The analysis unit analyzes the character string to be edited to obtain the number of mora of the character string to be edited. The presenting unit selects candidates for the reading information of the character string to be edited in a first presentation format for presenting a plurality of the reading information of the character string having the same number of mora as the number of mora of the character string to be edited. Present. The storage control unit includes the reading information including the same accent type as the accent type indicated by the reading information selected from the plurality of reading information candidates, and reading the character string to be edited, and the editing target. Are stored in the storage unit in association with each other. The search unit searches for a character string having the same number of mora as the character string to be edited and having a similarity with the character string to be edited higher than a threshold value. The presenting unit has, as a candidate for the reading information of the character string to be edited, a character having the same number of mora as the character string to be edited and having a similarity to the character string to be edited that is higher than a threshold value. Present a column. The similarity is weighted such that the position of a predetermined character and symbol matches the position of the predetermined character and symbol included in the character string to be edited, and the similarity is high. . The predetermined characters and symbols include at least one of a long sound, a prompt sound, a sound repelling sound, and a roaring sound.

The figure which shows the example of a function structure of the edit assistance apparatus 100 of embodiment. The figure which shows the example of case DB memorize | stored in the memory | storage part 101 of embodiment. The figure which shows the search example (in the case of a compound word) of the candidate of reading-out information of embodiment. The figure which shows the search example (in the case of the same kind of specific expression) of the candidate of the reading information of embodiment. The figure which shows the search example (when a long sound is included) of the candidate of reading-out information of embodiment. The figure which shows the example of a search of the candidate of the reading-out information of embodiment (when a prompt sound and a sound repellent are included). The figure which shows the search example (when it is not a common word) of the candidate of the reading-out information of embodiment. FIG. 5 is a diagram illustrating an example of a GUI that presents candidates for reading information according to the embodiment. 6 is a flowchart illustrating an example of an editing support method according to the embodiment. The figure which shows the example of the hardware constitutions of the edit assistance apparatus of embodiment.

  Hereinafter, embodiments of an editing support apparatus, an editing support method, and a program will be described in detail with reference to the accompanying drawings.

  FIG. 1 is a diagram illustrating an example of a functional configuration of an editing support apparatus 100 according to the embodiment. The editing support apparatus 100 according to the embodiment includes an input unit 1, an analysis unit 2, a generation unit 3, a search unit 4, a presentation unit 5, a storage control unit 6, a voice synthesis unit 7, a voice output unit 8, and a storage unit 101.

  The input unit 1 accepts an input indicating a character string (surface layer expression) to be edited in reading information through a character string designation (selection) operation by an editing operator (user). The reading information includes information indicating how to read the character string. The reading information includes, for example, information indicating an accent when a character string is read and a character string is pronounced. Note that the character string to be edited is an arbitrary character string designated as the object to be edited by the user, and is not limited to a unit having a linguistic meaning such as a word, a compound word, or a sentence. The character string to be edited in the reading information is, for example, a character string that is determined by the user to be inappropriate in reading or accent. When receiving the input indicating the character string to be edited, the input unit 1 inputs information indicating the character string to be edited to the analysis unit 2.

  When receiving the information indicating the character string to be edited from the input unit 1, the analysis unit 2 analyzes the character string to be edited.

  Specifically, the analysis unit 2 first estimates at least one reading of the character string to be edited in the reading information and provides estimated reading information indicating the reading of the character string to be edited in the reading information as a presentation unit described later. Enter 5. When receiving the estimated reading information from the analysis unit 2, the presenting unit 5 receives an input for confirming the reading of the character string to be edited via a GUI (see FIG. 7) described later.

  When the reading of the character string to be edited in the reading information is confirmed, the analysis unit 2 determines the segment of the character string to be edited by applying character type determination and morphological analysis. The character string delimiter is, for example, a boundary between consecutive character strings of the same type or a boundary of morpheme units. The analysis unit 2 divides the character string by dividing the character string. The analysis unit 2 calculates the number of mora for each divided character string and tentatively determines the most appropriate accent by the accent estimation process. Note that accent estimation processing can use conventionally known accent estimation processing.

  When the document including the character string to be edited has an explicit type (genre), the analysis unit 2 further includes the type (genre) in the analysis result of the character string to be edited. The type (genre) is, for example, a novel or an entertainment. Entertainment shows animation, a game, etc., for example.

  The analysis unit 2 inputs analysis result information indicating the analysis result of the character string to be edited to the generation unit 3.

  When receiving the analysis result information from the analysis unit 2, the generation unit 3 generates a candidate for the read-out information of the character string to be edited based on the analysis result information. Specifically, the generation unit 3 inputs information indicating a search condition based on the analysis result information to the search unit 4. Next, when the search unit 4 receives information indicating the search condition from the generation unit 3, the search unit 4 uses the search condition to acquire a candidate for reading information from the case DB stored in the storage unit 101. Then, the search unit 4 returns search result information including candidates for the read-out information to the generation unit 3. Since the number of reading information candidates included in the search result information may be enormous, the search unit 4 may narrow down the number of reading information candidates acquired from the case DB.

  FIG. 2 is a diagram illustrating an example of a case DB stored in the storage unit 101 according to the embodiment. The example DB of the embodiment includes an ID, a character string (surface expression), a part of speech, a reading / accent (intermediate language), a mora number, an accent type, a type (genre), and a registration number. ID is identification information for identifying a record in the case DB. The character string (surface representation) is one or more characters and symbols. The part of speech indicates the classification of the character string obtained by the morphological analysis and the structure (such as a compound word). The reading / accent (intermediate language) indicates how to read the character string and the position of the accent when the character string is pronounced. The intermediate language is a language that describes a reading control method such as reading and accent when synthesizing a character string.

  The number of mora indicates the number of segment units when a character string is pronounced. The accent type indicates the position of the accent when the character string is pronounced. The type (genre) is the type of the document from which the character string is acquired. The number of registrations indicates the number of registrations for which a combination of part of speech, reading / accent (intermediate language), number of mora, accent type, and type (genre) matches for a certain character string (surface expression). That is, the number of registrations indicates the number of times of registration (use) in past editing work.

  For example, a record with an ID of 3000 shows a case where the character string (surface expression) is “Tycopath”, the part of speech is a noun, and the reading / accent (intermediate language) is “Tycopass”. Here, “′” is an intermediate language symbol indicating the position of the accent. The presentation unit 5 described later presents a candidate for reading information using an intermediate language expressed as “Tycopath”.

  A record with an ID of 3000 shows a case where the number of mora is 5, the accent type is 4, and the type (genre) is entertainment. A record with ID 3000 indicates a case where the number of registrations is 18. Here, the registration number 18 is that the character string (surface expression) is “Tychopath”, the part of speech is a noun, the reading and accent (intermediate language) is “Tycopa's”, the number of mora is 5, The number of cases in which a combination is set in which the type is type 4 and the type (genre) is entertainment is shown.

  As shown in the case DB of FIG. 2, the registered content may differ depending on the type (genre) even for the same character string (surface expression).

  Hereinafter, a search process for candidates for reading information will be specifically described with reference to an example of a character string to be edited (surface expression).

<When focusing on string delimiters>
FIG. 3 is a diagram illustrating a search example (in the case of a compound word) of candidates for reading information according to the embodiment. The example of FIG. 3 shows a case where the character string to be edited in the reading information is “speech synthesis”. “Speech synthesis” is a compound word of “speech” and “synthesis”. The analysis unit 2 determines the character string delimiter of “speech synthesis” as “| speech | synthesis |” based on the morpheme boundaries. The analysis unit 2 analyzes the “speech” to obtain analysis result information indicating that the number of mora is 4, the accent type is 1, and the part of speech is a noun. Further, the analysis unit 2 analyzes the “synthesis”, thereby obtaining analysis result information indicating that the number of mora is 4, the accent type is 0 type, and the part of speech is a subordinator. The search unit 4 uses the same attribute pattern as “speech synthesis” (“4 mora type 1: noun” and “4 mora type 0: sa variable noun”) according to the search conditions based on the analysis result information of “speech” and “synthesis”. To obtain a compound word having a combination of The search unit 4 acquires compound words such as “music analysis”, “voice recognition”, and “mineral synthesis” as compound words having the same attribute pattern as “speech synthesis”.

<When focusing on the types of specific expressions>
FIG. 4 is a diagram illustrating a search example of candidates for reading information according to the embodiment (in the case of the same kind of specific expression). The example of FIG. 4 shows a case where the character string to be edited in the reading information is “Magnacotra”. “Magnacotra” is an example of a proper expression. The analysis unit 2 determines the character string “Magnacotra” as “| Magnacotra |” based on the morpheme boundaries. The analysis unit 2 analyzes “Magnacotra”, so that the number of mora is 6, the accent type is 0 type, the part of speech is a noun, and the type (genre) of the source document is an animation game. The analysis result information indicating that is acquired. The search unit 4 obtains a character string having the same attribute pattern (combination of “6 mora”, “anime / game”, and “proprietary expression”) as “Magnacotra” according to the search condition based on the analysis result information of “Magnacotra”. To do. The search unit 4 acquires, as a character string having the same attribute pattern as “Magnacotra”, for example, a specific expression such as “Balmanion”, “Modokamogika”, and “Laranoazola”.

  As in the example of FIG. 4, the editing support apparatus 100 according to the embodiment has the same type (genre) when the document including the character string to be edited has a type (genre) indicating an explicit field and specialization. ) Can be enumerated as candidates for the reading information of the character string to be edited.

<When focusing on similarity>
When paying attention to the degree of similarity, the search unit 4 searches the character string to be edited for the reading information and the reading information for the character string having a high degree of similarity. A specific method of giving the similarity may be arbitrary.

  The similarity may be, for example, a character string distance in which weights are given to predetermined characters and symbols (long sound, prompting sound, repelling sound, roaring sound, etc.). The weighted character string distance is, for example, 0 when the long sound, prompt sound, repelling sound, and stuttering match, 1 when the character and symbol other than the long sound, sounding sound, repelling sound, and stuttering match, If the symbols do not match, 2 is assumed. In this case, for example, the similarity between “Masan” and “Osan” is 2 + 0 + 1 + 0 = 3. For example, the degree of similarity between “Masan” and “Masaru” is 1 + 2 + 1 + 2 = 6. Depending on the weighted character string distance, “Osan” and “Masaru” both match the character string “Masan” to be edited, but the sound and sound repellent match. The similarity of “Osan” is higher than the similarity of “Masaru”. The search unit 4 inputs information including a character string whose similarity with the character string to be edited is a threshold value or less to the generation unit 3.

  The search unit 4 may determine the degree of similarity between character strings without performing the threshold determination of the character string distance. For example, when a character string to be edited includes predetermined characters and symbols (long sound, prompting sound, repelling sound, roaring sound, etc.), the search unit 4 sets a character string having a high similarity to the character string to be edited as You may search for the character string in which the position of a predetermined character and symbol corresponds with the position of the predetermined character and symbol contained in the character string to be edited.

  FIG. 5A is a diagram illustrating a search example (when a long sound is included) of reading-out information candidates according to the embodiment. The example of FIG. 5A shows a case where the character string to be edited in the reading information is an “anarchist” including a long sound “−”. The analysis unit 2 determines the character string “anarchist” as “| anarchist |” based on the morpheme boundary. The analysis unit 2 analyzes the “anarchist” to obtain analysis result information indicating that the number of mora is 5, the accent type is 0 type, and the part of speech is a noun. The search unit 4 acquires a character string having the same attribute pattern (a combination of “5 mora” and “third character is long sound”) as “anarchist”, based on a search condition based on the analysis result information of “anarchist”. The search unit 4 acquires character strings such as “Sunny Rattle” and “Amilias” as character strings having the same attribute pattern as “Anarchist”.

  FIG. 5B is a diagram illustrating a search example of candidates for reading information according to the embodiment (in the case of including a prompt sound and a sound repellent). The example of FIG. 5B shows a case where the character string to be edited in the reading information is “masan” including the prompt sound “tsu” and the sound repellent “n”. The analysis unit 2 determines that the character string of “Masan” is “| Masan |” based on the morpheme boundary. The analysis unit 2 analyzes “Masan” to acquire analysis result information indicating that the number of mora is 3, the accent type is 0 type, and the part of speech is a noun. Based on the search condition based on the analysis result information of “Masan”, the search unit 4 has the same attribute pattern as “Masan” (a combination of “3 mora”, “second character is urging sound”, and “fourth character is sound repellent) ) Is obtained. The search unit 4 acquires, for example, character strings such as “Osan”, “Hatsusan”, and “Yoshisan” as character strings having the same attribute pattern as “Masan”.

<When focusing on the generality of specific expressions>
FIG. 6 is a diagram illustrating a search example (when not a general word) of reading information candidates according to the embodiment. The example of FIG. 6 shows a case where the character string to be edited in the reading information is “Saraden”. The analysis unit 2 determines the character string “区 切 り 嵯峨 嵯峨” as “| 嵯峨 嵯峨「 | ”based on the morpheme boundary. The analysis unit 2 analyzes the “嵯峨 嵯峨 鈿” to obtain analysis result information indicating that the mora number is 5, the accent type is 0 type, and the part of speech is an uncommon noun. The search unit 4 acquires, as an alternative word, a character string that has the same 5-mora as “嵯峨 嵯峨 嵯峨” and has a general noun attribute, based on the search condition based on the analysis result information of “嵯峨 嵯峨 鈿”. The search unit 4 uses, for example, a character string such as “Aiueo”, “Akusen to (Accent)”, “Akaishi (Red Stone)” and “Akaeboshi (Red Hat)” as an alternative word for “嵯峨 嵯峨 鈿”. To get.

  Returning to FIG. 1, upon receiving the search result information from the search unit 4, the generation unit 3 inputs the search result information to the presentation unit 5. Note that the generation unit 3 inputs a plurality of search result information to the presentation unit 5 when a candidate for the reading information can be searched from a plurality of viewpoints with respect to one editing target character string. For example, when the analysis unit 2 also analyzes that the above-described “嵯峨 嵯峨” may be a compound word of “嵯峨” and “screw”, the search unit 3 searches for search result information focusing on character string delimiters. Also get. In addition, since “Raku” is a decoration technique used in traditional crafts, if the type of the document from which “嵯峨 嵯峨” is acquired is specified by “Traditional craft” or the like, the search unit 3 focuses on the type of unique expression. Retrieve search result information. Furthermore, the search unit 3 can also acquire search result information focusing on the high degree of similarity to “嵯峨 嵯峨 嵯峨”.

  Next, the presentation unit 5 will be described. When the presentation unit 5 receives the estimated reading information from the analysis unit 2, the presentation unit 5 presents at least one reading of the character string to be edited in the reading information. When there are a plurality of reading candidates for the character string to be edited in the reading information, the presenting unit 5 presents a plurality of readings using, for example, a drop-down list. When the input unit 1 accepts an input indicating that the button 206 on the right side of the reading “Taikopas” is pressed, the speech synthesizer 7 performs speech synthesis, and the speech synthesized from the speech output unit 8 Is output. In addition, when there is no appropriate reading candidate of the character string to be edited in the reading information estimated by the analysis unit 2, the presentation unit 5 receives an input of reading the character string to be edited in the reading information.

  The presentation unit 5 accepts one or more pieces of search result information from the generation unit 3 after the reading of the character string to be edited in the reading information is confirmed by the confirm button 207, and candidates for the reading information included in the search result information Present.

  FIG. 7 is a diagram illustrating an example of a GUI (Graphical User Interface) that presents candidates for reading information according to the embodiment. The example of FIG. 7 shows a case where the character string to be edited in the reading information is “Tycopath”. When “Tychopath” is selected by the input unit 1 in a state where the document 201 including “Tychopath” is displayed, after reading “Tychopath” is confirmed, any of GUI202, GUI203, and GUI204 is determined. The candidate for the reading information of “Tycopath” is presented through the one. The GUI displayed first is determined according to the initial setting of the editing support apparatus 100 or the like. Further, the GUI 202, the GUI 203, and the GUI 203 can be switched by an operation such as cursor movement and selection with a mouse and a tap on a liquid crystal touch panel, for example.

  The GUI 202 presents candidates for the reading information of the character string to be edited in the first presentation format. The first presentation format is a presentation format in which a plurality of read-out information of a character string having the same number of mora as the number of mora of the character string to be edited is presented. In the first presentation format, candidates for reading information are presented using the search results obtained in FIGS.

  A symbol 205 is an accent symbol indicating the position of the accent. The method for indicating the position of the accent may be arbitrary. The position of the accent may be specified by, for example, a line provided above and below the character string, the height of the character position, and the size of the character size.

  In the example of FIG. 7, among the 5 mora character strings having the same number of mora as “Tycopath”, the accent type is “Rafresile”, the accent type is “1”, “Yuki Saiki”, and the accent type is “4”. A case is shown in which “Mentikatsu”, “Sakashion” with an accent type of 2, and “Fushigiso” with an accent type of 0 are presented.

  In addition, when the search part 4 has obtained a plurality of pieces of search result information acquired under different search conditions, any method for presenting candidates for read-out information presented at a time may be used. For example, the presentation unit 5 may obtain a representative example from each search result information to present a mixture of read-out information candidates obtained according to different search conditions, or separate categories (GUI 202a, GUI 202b, GUI 202c). ,...) Candidates for reading information may be presented for each search condition.

  For example, when the input unit 1 receives an input indicating that the button 206 on the right side of the “Rafflesille” notation is pressed, the speech synthesizer 7 performs speech synthesis of “Rafflesille” with an accent type of 3, and the speech output unit The synthesized voice is output from 8. The input indicating that the button 206 is pressed is, for example, pressing of the space key while “Raffles” is selected, pressing the button 206 using the cursor key, or the like. Since the user can refer to not only the visual information of the accent but also the voice of the reading information of other character strings, it is possible to determine whether the accent type is correct or not with a simple operation.

  The GUI 203 presents the candidates for the reading information of the character string to be edited in the second presentation format. The second presentation format presents a candidate for reading information based on the reading information history of the character string to be edited. Specifically, the presentation unit 5 acquires read-out information of the character string to be edited that has been stored in the case DB of the storage unit 101 in the past via the search unit 4 and presents it in the order of the number of registrations. In the example of FIG. 7, the number of registrations of “Tychopath” with an accent type of 4 is 18 cases, the number of “Tycopath” with an accent type of 3 is 4, and the registration of “Tycopath” with an accent type of 1 is registered. The case where the number is two is shown.

  The GUI 204 presents candidates for the reading information of the character string to be edited in the third presentation format. The third presentation format presents a plurality of reading information indicating all accent types that can be taken by the character string to be edited. In the example of FIG. 7, the accent type is “1 type” “Tycopath”, the accent type is “Type 2” “Tycopath”, the accent type is “Type 3” “Tycopass”, the accent type is “Type 4” Indicates a case where a “Tychopath” of type 0 is presented. The presentation unit 5 is visually decorated so that it can be easily distinguished from other accent-type “Tychopaths” by changing the character type and color of the accent-type “Tychopass” with the largest number of registrations. May be.

  Returning to FIG. 1, when the reading information is selected from a plurality of reading information candidates, the storage control unit 6 includes the same accent type as the accent type indicated by the reading information and reading of the character string to be edited. The read-out information and the character string to be edited are associated with each other and stored in the case DB of the storage unit 101. At this time, if the combination of reading / accent (intermediate language), number of mora, accent type, and type (genre) matches the existing combination stored in the past, the storage control unit 6 creates a new record. Without incrementing the registration number. If the combination of reading / accent (intermediate language), number of mora, accent type and type (genre) does not match the existing combination stored in the past, the storage control unit 6 creates a new record, Set the number of registered records to 1.

  When the speech synthesizer 7 receives an input indicating the output of the speech of a candidate for reading information via the GUIs 202 to 204, the speech synthesizing unit 7 performs speech synthesis based on the reading information. The voice output unit 8 outputs the voice synthesized by the voice synthesis unit 7. In the conventional method of manually correcting the reading information, it is necessary to learn a special notation method for designating reading and accents. For compound words, phrases, etc., knowledge about the division and combination of accent phrases is also required. The editing support apparatus 100 according to the embodiment includes the voice synthesis unit 7 and the voice output unit 8, so that even a user who has not mastered a special notation or the like can edit the reading information while listening to the voice. be able to.

  Next, an example of the editing support method of the embodiment will be described.

  FIG. 8 is a flowchart illustrating an example of the editing support method according to the embodiment. First, the input unit 1 receives an input indicating selection of a character string (surface expression) to be edited for reading information (reading and accent) (step S1). Next, the presentation unit 5 presents candidates for reading the character string to be edited estimated by the analysis unit 2 (step S2). When there are a plurality of reading candidates, the presenting unit 5 presents a plurality of reading candidates using a drop-down list or the like.

  Next, it is determined whether or not the presentation unit 5 has received an input indicating that the button 206 for outputting the reading candidate voice has been pressed (step S3). When an input indicating that the button 206 for outputting the reading candidate voice is pressed is received (step S3, Yes), the voice output unit 8 outputs the voice synthesized by the voice synthesis unit 7 (step S4). If an input indicating that the button 206 for outputting the reading candidate sound is not received (No in step S3), the process proceeds to step S5.

  Next, it is determined whether or not the presentation unit 5 has received an input indicating that the button 207 for confirming a reading candidate has been pressed (step S5). When an input indicating pressing of the button 207 for confirming a reading candidate is received (step S5, Yes), the analysis unit 2 analyzes the character string to be edited in the reading information (step S6). When an input indicating that the button 207 for confirming a reading candidate is not received is received (No in step S5), the process returns to step S2.

  Next, based on the analysis result information obtained by the analysis unit 2, the generation unit 3 generates a candidate for reading information of the character string to be edited (step S7). Next, the presentation unit 5 presents the candidates for the reading information by using the GUI determined by the initial setting or the like among the above-described GUIs 202 to 204 (see FIG. 7) (step S8).

  Next, it is determined whether or not the presentation unit 5 has received an input indicating that the button 206 for outputting the speech of the reading information candidate is pressed (step S9). When an input indicating that the button 206 for outputting the voice of the reading information candidate is pressed is received (step S9, Yes), the voice output unit 8 outputs the voice synthesized by the voice synthesis unit 7 (step S10). When an input indicating that the button 206 for outputting the voice of the reading-out information candidate is not received is received (No in step S9), the process proceeds to step S11.

  Next, the presentation unit 5 determines whether or not the accent has been determined based on whether or not an operation for selecting one piece of reading information has been received from a plurality of reading information candidates (step S11).

  When the confirmation of the accent has not been performed (No at Step S11), the presentation unit 5 determines whether or not an input indicating switching of a GUI that presents a candidate for the reading information has been received (Step S12). If an input indicating GUI switching is not received (step S12, No), the process returns to step S8. When an input indicating GUI switching is received (step S12, Yes), the presentation unit 5 switches the reading information candidate to be presented to the user to the reading information candidate acquired by another search condition (step S13). The process returns to step S8.

  When the accent is confirmed (Yes in step S11), the storage control unit 6 confirms the same accent type as the accent type indicated by the reading information selected from the plurality of reading information candidates in step S11, and the setting in step S5. The read-out information including the reading of the edited character string to be edited and the character string to be edited selected in step S1 are associated and stored in the case DB of the storage unit 101 (step S14).

  Finally, an example of a hardware configuration of the editing support apparatus 100 according to the embodiment will be described.

  FIG. 9 is a diagram illustrating an example of a hardware configuration of the editing support apparatus 100 according to the embodiment. The editing support apparatus 100 according to the embodiment includes a control device 301, a main storage device 302, an auxiliary storage device 303, a display device 304, an input device 305, a communication device 306, and a speaker 307. The control device 301, main storage device 302, auxiliary storage device 303, display device 304, input device 305, communication device 306, and speaker 307 are connected via a bus 310.

  The control device 301 executes the program read from the auxiliary storage device 303 to the main storage device 302. The main storage device 302 is a memory such as a ROM and a RAM. The auxiliary storage device 303 is a memory card, an SSD (Solid State Drive), or the like.

  The display device 304 displays information. The display device 304 is a liquid crystal display, for example. The input device 305 receives input of information. The input device 305 is, for example, a keyboard and a mouse. Note that the display device 304 and the input device 305 may be a liquid crystal touch panel having both a display function and an input function. The communication device 306 communicates with other devices. The speaker 307 outputs sound.

  A program executed by the editing support apparatus 100 according to the embodiment is an installable or executable file and can be read by a computer such as a CD-ROM, a memory card, a CD-R, and a DVD (Digital Versatile Disk). It is stored in a storage medium and provided as a computer program product.

  In addition, the program executed by the editing support apparatus 100 according to the embodiment may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Further, the program executed by the editing support apparatus 100 according to the embodiment may be configured to be provided via a network such as the Internet without being downloaded.

  Further, the program executed by the editing support apparatus 100 of the embodiment may be configured to be provided by being incorporated in advance in a ROM or the like.

  The program executed by the editing support apparatus 100 of the embodiment has a module configuration including functions that can be realized by the program among the functional configurations of the editing support apparatus 100 of the above-described embodiment.

  Functions realized by the program are loaded into the main storage device 302 by the control device 301 reading the program from a storage medium such as the auxiliary storage device 303 and executing the program. That is, the function realized by the program is generated on the main storage device 302.

  Note that part or all of the functions of the editing support apparatus 100 according to the embodiment may be realized by hardware such as an IC (Integrated Circuit).

  As described above, in the editing support apparatus 100 according to the embodiment, when the user determines the reading information (reading and accent type) of the character string to be edited, the user reads out another character string related to the character string. Candidate information can be switched across, and voice can be heard as needed. Further, the determined reading information can be accumulated in the storage unit 101. Thus, the user can more easily obtain reading information indicating a desired voice according to the user's skill.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 Input part 2 Analysis part 3 Generation part 4 Search part 5 Presentation part 6 Storage control part 7 Speech synthesis part 8 Speech output part 100 Editing support apparatus 101 Storage part 301 Control apparatus 302 Main storage apparatus 303 Auxiliary storage apparatus 304 Display apparatus 305 Input Device 306 Communication device 307 Speaker 310 Bus

Claims (12)

A storage unit for storing a character string and reading information including a reading and an accent type of the character string;
An analysis unit that obtains the number of mora of the edit target character string by analyzing the edit target character string;
A presentation unit for presenting candidates for the reading information of the character string to be edited in a first presentation format for presenting a plurality of the reading information of the character string having the same number of mora as the number of mora of the character string to be edited When,
Associating the reading information including the same accent type as the accent type indicated by the reading information selected from a plurality of reading information candidates and the reading of the character string to be edited, and the character string to be edited A storage control unit for storing in the storage unit;
A search unit that searches for a character string that has the same number of mora as the edit target character string and that has a similarity to the edit target character string that is higher than a threshold;
The presenting unit has, as a candidate for the reading information of the character string to be edited, a character having the same number of mora as the character string to be edited and having a similarity to the character string to be edited that is higher than a threshold value. Present a column,
The similarity is weighted such that the position of a predetermined character and symbol matches the position of a predetermined character and symbol included in the character string to be edited. ,
The predetermined characters and symbols include at least one of a long sound, a prompt sound, a repelling sound, and a roaring sound.
Editing support device. A speech synthesizer that performs speech synthesis based on the reading information selected from the reading information candidates;
A voice output unit that outputs the synthesized voice;
The editing support apparatus according to claim 1, further comprising: The candidates for the plurality of reading information include reading information of a plurality of character strings having the same number of mora as the character string to be edited and having different accent types.
The editing support apparatus according to claim 1. The storage unit further stores the character string break position in association with each other,
The analysis unit further obtains a break position of the character string to be edited,
The plurality of read-out information candidates include read-out information having the same number of mora as the edit target character string and having the same delimiter position as the edit target character string delimiter,
The editing support apparatus according to claim 1. The storage unit further stores the type of document from which the character string is acquired in association with each other,
The analysis unit further acquires the type of document from which the character string to be edited is acquired,
The presenting unit, as a candidate for the reading information of the character string to be edited, includes the number of mora of the character string to be edited among the character strings of the same type as the type of the document from which the character string to be edited is acquired. Presenting a plurality of the reading information of the character string having the same number of mora as
The editing support apparatus according to claim 1. The presenting unit further includes a second presentation format that presents candidates for the reading information based on the reading information history of the character string to be edited, and when the second presentation format is selected, Presenting the read-out information of the character string to be edited stored in the storage unit in descending order;
The editing support apparatus according to claim 1. The presenting unit further includes a third presenting format that presents a plurality of the reading information indicating all accent types that can be taken by the character string to be edited, and when the third presenting format is selected, Presenting a plurality of the reading information indicating all accent types that can be taken by the character string to be edited as candidates,
The editing support apparatus according to claim 1. The presentation unit includes a second presentation format of the reading information based on a reading information history of the character string to be edited, and a plurality of reading information indicating all accent types that the character string to be edited can take. A third presentation format, and by simultaneously applying two or more presentation formats of the first presentation format, the second presentation format, and the third presentation format, the editing Present candidate reading information for the target string,
The editing support apparatus according to claim 1. The presenting unit selects candidates for the reading information having the same accent type as the accent type included in the reading information having the largest number among the reading information of the character string to be edited previously stored in the storage unit. To emphasize,
The editing support apparatus according to claim 1. The presenting unit switches and presents the read-out information candidates searched by the different search conditions from the storage unit according to a user operation.
The editing support apparatus according to claim 1. An editing support method of an editing support apparatus comprising a storage unit for storing a character string and reading information including a reading of the character string and an accent type,
The editing support device analyzing the character string to be edited to obtain the number of mora of the character string to be edited;
In the first presentation format in which the editing support apparatus presents a plurality of the reading information of the character string having the same number of mora as the number of mora of the character string to be edited, the candidate of the reading information of the character string to be edited Presenting steps,
The editing support apparatus includes the reading information including the same accent type as the accent type indicated by the reading information selected from the plurality of reading information candidates, and reading the character string to be edited, and the editing target. The character string having the same number of mora as the edit target character string and the similarity to the edit target character string is higher than a threshold is searched. And including steps,
The presenting step has the same number of mora as that of the character string to be edited as a candidate for the reading information of the character string to be edited, and the similarity to the character string to be edited is higher than a threshold value. Presents a string,
The similarity is weighted such that the position of a predetermined character and symbol matches the position of a predetermined character and symbol included in the character string to be edited. ,
The predetermined characters and symbols include at least one of a long sound, a prompt sound, a repelling sound, and a roaring sound.
Editing support method. A computer comprising a storage unit for storing a character string and reading information including a reading of the character string and an accent type,
An analysis unit that obtains the number of mora of the edit target character string by analyzing the edit target character string;
A presentation unit for presenting candidates for the reading information of the character string to be edited in a first presentation format for presenting a plurality of the reading information of the character string having the same number of mora as the number of mora of the character string to be edited When,
Associating the reading information including the same accent type as the accent type indicated by the reading information selected from a plurality of reading information candidates and the reading of the character string to be edited, and the character string to be edited A storage control unit for storing in the storage unit,
Having the same number of mora as the character string to be edited, and functioning as a search unit for searching for a character string having a similarity to the character string to be edited higher than a threshold,
The presenting unit has, as a candidate for the reading information of the character string to be edited, a character having the same number of mora as the character string to be edited and having a similarity to the character string to be edited that is higher than a threshold value. Present a column,
The similarity is weighted such that the position of a predetermined character and symbol matches the position of a predetermined character and symbol included in the character string to be edited. ,
The predetermined characters and symbols include at least one of a long sound, a prompt sound, a repelling sound, and a roaring sound.
program.

Images