Speech synthesis support device, speech synthesis support method, and speech synthesis support program

Abstract

To enable text to be easily converted into accurate linguistic information.SOLUTION: A speech synthesis support device 10 comprises a conversion section 20A, a reception section 20C, and a registration section 20D. The conversion section 20A analyzes a text using a language dictionary 16A and converts the text into linguistic information for inputting to a synthetic speech dictionary 16B used for generating synthetic speech data. The reception section 20C accepts at least some of the modifications of the language information. The registration section 20D registers the text and the modified language information in the language dictionary 16A.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] An embodiment of the present invention relates to a speech synthesis support device, a speech synthesis support method, and a speech synthesis support program. [Background technology]

[0002] A technique is disclosed in which a text is analyzed using a language dictionary, linguistic information such as the phoneme sequence constituting a sentence and the position of an accent is generated, and synthetic speech data is generated from the linguistic information using a speech synthesis dictionary.

[0003] However, in conventional technology, if there is an error in the language information resulting from text analysis, the language information needs to be corrected each time analysis is performed through operation instructions from the user, making it difficult to easily convert text into accurate language information. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2022-081691 [Patent Document 2] JP 2023-038957 A Summary of the Invention [Problem to be solved by the invention]

[0005] An object of the present invention is to provide a speech synthesis support device, a speech synthesis support method, and a speech synthesis support program that can easily convert text into accurate linguistic information. [Means for solving the problem]

[0006] A speech synthesis support device according to an embodiment includes a conversion unit, a reception unit, and a registration unit. The conversion unit analyzes text using a language dictionary and converts the text into language information to be input to a synthetic speech dictionary used to generate synthetic speech data. The reception unit receives corrections to at least a portion of the language information. The registration unit registers the text and the corrected language information in the language dictionary in association with each other. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram illustrating an example of a speech synthesis support device. [Figure 2A] FIG. 2A is a schematic diagram of an example of a display screen. [Figure 2B] FIG. 2B is a schematic diagram of an example of a display screen. [Figure 2C] FIG. 2C is a schematic diagram of an example of a display screen. [Figure 3A] FIG. 3A is a schematic diagram of an example of a display screen. [Figure 3B] FIG. 3B is a schematic diagram of an example of a display screen. [Figure 3C] FIG. 3C is a schematic diagram of an example of a display screen. [Figure 4] FIG. 4 is a flowchart showing an example of the flow of information processing executed by the speech synthesis support device. [Diagram 5] FIG. 5 is a diagram showing the hardware configuration. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a speech synthesis support device, a speech synthesis support method, and a speech synthesis support program will be described in detail with reference to the accompanying drawings.

[0009] FIG. 1 is a diagram illustrating an example of a speech synthesis support device 10 according to the present embodiment.

[0010] The speech synthesis support device 10 is an information processing device that supports the conversion of text into accurate linguistic information.

[0011] The speech synthesis support device 10 includes a communication unit 12, a UI (user interface) unit 14, a storage unit 16, and a processing unit 20. The communication unit 12, the UI unit 14, the storage unit 16, and the processing unit 20 are connected to each other via a bus 18 so as to be able to communicate with each other.

[0012] The communication unit 12 communicates with other external information processing devices via a network, etc. The UI unit 14 includes a display unit 14A, an input unit 14B, and an audio output unit 14C.

[0013] The display unit 14A displays various types of information and is, for example, a display such as an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display, a projection device, or the like.

[0014] The input unit 14B receives user operations. The input unit 14B is, for example, a pointing device such as a digital pen, a mouse, or a trackball, or an input device such as a keyboard. The display unit 14A and at least a part of the input unit 14B may be integrally provided as a touch panel.

[0015] The audio output unit 14C is a speaker that outputs audio.

[0016] The storage unit 16 stores various data. The storage unit 16 is, for example, a semiconductor memory element such as a random access memory (RAM), a flash memory, a hard disk, an optical disk, or the like. The storage unit 16 may be a storage device provided outside the speech synthesis support device 10. The storage unit 16 may also be a storage medium. Specifically, the storage medium may store or temporarily store programs and various information downloaded via a local area network (LAN) or the Internet. The storage unit 16 may also be composed of multiple storage media.

[0017] In this embodiment, the storage unit 16 stores a language dictionary 16A and a synthetic speech dictionary 16B. The language dictionary 16A and the synthetic speech dictionary 16B will be described in detail later.

[0018] Next, the processing unit 20 will be described. The processing unit 20 executes various information processing. The processing unit 20 includes a conversion unit 20A, a display control unit 20B, a reception unit 20C, a registration unit 20D, a voice synthesis unit 20E, and a reproduction unit 20F. The conversion unit 20A, the display control unit 20B, the reception unit 20C, the registration unit 20D, the voice synthesis unit 20E, and the reproduction unit 20F are realized, for example, by one or more processors. For example, each of the above units may be realized by having a processor such as a CPU (Central Processing Unit) execute a program, that is, by software. Each of the above units may be realized by a processor such as a dedicated IC (Integrated Circuit), that is, by hardware. Each of the above units may be realized by using both software and hardware. When multiple processors are used, each processor may realize one of the units, or may realize two or more of the units.

[0019] At least one of the above units and at least a part of the information stored in the storage unit 16 may be mounted on a cloud server or the like that executes processing on the cloud.

[0020] The conversion unit 20A analyzes the text using the language dictionary 16A and converts the text into linguistic information.

[0021] Text is character data such as character strings and sentences handled by a computer. The language of the text is not limited. The language of the text is, for example, Japanese, English, etc. In this embodiment, a form in which the language of the text is Japanese will be described as an example. When the language of the text is Japanese, the text may be expressed in any of, for example, kanji, hiragana, katakana, alphanumeric characters, roman letters, etc.

[0022] The language dictionary 16A is a dictionary for analyzing text and outputting language information. For example, text and language information are registered in the language dictionary 16A in association with each other. The unit of text registered in the language dictionary 16A may be, for example, a unit including at least one morpheme.

[0023] Linguistic information is a feature of language extracted from text. Linguistic information is sometimes called linguistic feature. For example, linguistic information is information on the surrounding phonemes, information on pronunciation, accent position, reading of kanji, division of phrases (accent phrases), phrase-end position, sentence length, accent phrase length, mora length, mora position, accent type, part of speech, dependency, etc.

[0024] The conversion unit 20A acquires text designated by a user's operation instruction on the input unit 14B, etc. For example, the conversion unit 20A acquires the text by reading the text designated by a user's operation instruction on the input unit 14B, etc., from the storage unit 16 or an external information processing device, etc.

[0025] Then, the conversion unit 20A reads language information corresponding to the acquired text from the language dictionary 16A, thereby converting the text into language information.

[0026] The display control unit 20B displays various images on the display unit 14A.

[0027] FIG. 2A is a schematic diagram of an example of the display screen 30A. In this embodiment, a case where characters displayed on the display screen 30 are in Japanese will be described as an example. The display screen 30A is an example of the display screen 30 displayed on the display unit 14A. The display screen 30A and a display screen 30B described later are examples of a first display screen. The first display screen is a display screen 30 for displaying text and language information obtained by conversion of the text by the conversion unit 20A, and for accepting corrections to the language information.

[0028] The display screen 30A includes a text display field 40A, a language information display field 40B, a play button 40C, and a save / next button 40P.

[0029] The text display field 40A is a display field for the text to be converted by the conversion unit 20A. The language information display field 40B is a display field for language information corresponding to the text displayed in the text display field 40A and for accepting edits to the language information. The play button 40C is a button image area operated by the user when instructing the playback of synthetic speech data corresponding to the language information displayed in the language information display field 40B. The save / next button 40P is a button image area operated by the user when changing the text displayed in the text display field 40A to other text.

[0030] When the conversion unit 20A converts the text information into language information, the display control unit 20B displays the text information in a text display field 40A, and displays the language information in a language information display field 40B.

[0031] Returning to Figure 1, we will continue with the explanation.

[0032] Here, at least a part of the language information converted from the text using the language dictionary 16A may differ from the accurate language information of the text, such as phonemes and accents. The conversion from the text to the language information is performed using the language dictionary 16A. Therefore, if incorrect language information is registered in the language dictionary 16A, the conversion to the incorrect language information will always be performed when the same text is converted into language information.

[0033] Therefore, the receiving unit 20C receives a correction to at least a part of the language information.

[0034] For example, assume that the display screen 30A shown in Fig. 2A is displayed on the display unit 14A. Also assume that part of the language information displayed in the language information display field 40B is incorrect. In this case, the user operates the input unit 14B to change at least part of the language information displayed in the language information display field 40B to correct information.

[0035] FIG. 2B is a schematic diagram of an example of a display screen 30B.

[0036] The display screen 30B is an example of the display screen 30. The display screen 30B is also an example of a first display screen. The display screen 30B is an example of the display screen 30 displayed on the display unit 14A when a correction to language information is accepted. The display screen 30B includes a correction completion button 40D. The correction completion button 40D is a button image area operated by the user when issuing an instruction to save the language information in the synthetic speech dictionary 16B.

[0037] For example, the user operates input unit 14B to change at least a portion of the language information displayed in language information display field 40B of display screen 30A (see FIG. 2A) to correct information (see the underlined portion in language information display field 40B of display screen 30B in FIG. 2B).

[0038] Then, when the user changes at least a part of the language information displayed in the language information display field 40B to the correct information, the user operates the input unit 14B to display the correction complete button 40D.

[0039] Returning to Figure 1, we will continue with the explanation.

[0040] When the user operates the correction completion button 40D through an instruction from the input unit 14B, the reception unit 20C receives the correction completion instruction, and the display control unit 20B displays on the display unit 14A a second display screen showing the correspondence between the text and the corrected language information for which the correction has been accepted.

[0041] 2C is a schematic diagram of an example of a display screen 30C. The display screen 30C is an example of the display screen 30. Moreover, the display screen 30C is an example of a second display screen.

[0042] The display screen 30C includes a category display field 40E, a text display field 40F, a language information display field 40G, a registration button 40H, and a report button 40I.

[0043] The category display field 40E is a display field for the category to which the text belongs. The text display field 40F is a display field for the text. The language information display field 40G is a display field for corrected language information input for the text. The registration button 40H is a button image area operated by the user when registering the text displayed in the text display field 40F and the language information displayed in the language information display field 40G in the language dictionary 16A in association with each other. The report button 40I is a button image area operated by the user when outputting a pair of the text displayed in the text display field 40F of the display screen 30C and the language information displayed in the language information display field 40G to the conversion unit 20A as learning data.

[0044] The display control unit 20B may superimpose the display screen 30C, which is an example of the second display screen showing the correspondence between the text and the corrected language information, on the display screen 30B, which is an example of the first display screen, and display the superimposed display screen 30C as a pop-up display. The display control unit 20B may also display the display screen 30C, which is an example of the second display screen, at a position on the display surface of the display unit 14A that does not overlap with the display screen 30B, which is an example of the first display screen. The display control unit 20B may also display the display screen 30C on the display unit 14A instead of the display screen 30B.

[0045] The display control unit 20B may analyze the text using a known learning model or the like, and display the category to which the text belongs in the category display field 40E. The display control unit 20B may also display the category to which the text belongs, input by the user's operation instruction of the input unit 14B, in the category display field 40E. The display control unit 20B may also selectively extract a continuous character string including characters different from the pre-correction language information from the corrected language information displayed in the language information display field 40B of the display screen 30B, and display the extracted character string as the corrected language information in the language information display field 40G. In this case, the display control unit 20B may display text corresponding to the corrected language information in the text display field 40F.

[0046] When the user operates the report button 40I by operating the input unit 14B to receive a report instruction, the display control unit 20B outputs a pair of the text displayed in the text display field 40F included in the display screen 30C and the language information displayed in the language information display field 40G as learning data to the conversion unit 20A. The display control unit 20B may also output learning data further including a category displayed in the category display field 40E included in the display screen 30C to the conversion unit 20A.

[0047] The conversion unit 20A, which receives the learning data, may use multiple learning data, including previously received learning data, to learn a learning model using a known method that takes text as input and outputs corrected language information.

[0048] When the user operates the correction completion button 40D or the registration button 40H by operating the input unit 14B as an instruction to the reception unit 20C to receive the correction completion instruction, the registration unit 20D associates the text with the corrected language information and registers it in the language dictionary 16A. In this embodiment, when the user operates the registration button 40H displayed on the display screen 30C by operating the input unit 14B as an instruction to the reception unit 20C to receive the correction completion instruction, the display screen 30D associates the text with the corrected language information and registers it in the language dictionary 16A.

[0049] At this time, the registration unit 20D registers the corrected language information in the language dictionary 16A by overwriting the corrected language information with the language information registered in the storage unit 16. Through these processes, the incorrect language information associated with the text registered in the language dictionary 16A is updated to the corrected language information.

[0050] The registration unit 20D may further associate a category to which the text received via the display screen 30C belongs, and register the associated text in the language dictionary 16A. The storage unit 16 may store a language dictionary 16A for each category. In this case, the registration unit 20D may register the text and the corrected language information corresponding to the text in the language dictionary 16A corresponding to the category to which the text received via the display screen 30C belongs, in association with each other.

[0051] In the above description, it is assumed that text and language information are registered in one-to-one correspondence in the language dictionary 16A. However, a plurality of different pieces of language information may be registered in the language dictionary 16A in correspondence with one text.

[0052] In this case, when the conversion unit 20A acquires text specified by a user's operation instruction on the input unit 14B, the conversion unit 20A converts one piece of text into multiple pieces of language information by reading multiple pieces of language information corresponding to the acquired text from the language dictionary 16A.

[0053] Then, the display control unit 20B displays the text on the display screen 30. Furthermore, the display control unit 20B displays a plurality of pieces of language information corresponding to the text on the display screen 30 in a selectable manner.

[0054] 3A and 3B are schematic diagrams of an example of the display screen 30. Fig. 3A shows a display screen 30E as an example of the display screen 30. Fig. 3B shows a display screen 30F as an example of the display screen 30. The display screens 30E and 30F are examples of a first display screen.

[0055] The display screen 30E includes a text display field 40A, a language information display field 40B, a play button 40C, a correction complete button 40D, and a next candidate button 40J.

[0056] The display control unit 20B displays the acquired text in the text display field 40A. The display control unit 20B also displays one of the multiple pieces of language information corresponding to the text in the language information display field 40B. Then, every time a next candidate display instruction is input by operating the next candidate button 40J through an operation instruction from the input unit 14B by the user, the display control unit 20B displays another piece of language information corresponding to the text in the language information display field 40B. For example, when the next candidate button 40J is operated through an operation instruction from the input unit 14B by the user while the display control unit 20B displays the display screen 30E on the display unit 14A, the display control unit 20B displays, for example, a display screen 30F on the display unit 14A in which language information corresponding to the text different from the currently displayed language information is displayed in the language information display field 40B (see FIG. 3A and FIG. 3B). This allows the user to sequentially check multiple different language information for one text.

[0057] Then, when the correct language information for the text is displayed in the language information display field 40B among the multiple pieces of language information, the user operates the correction complete button 40D by operating the input unit 14B. The registration unit 20D accepts the language information displayed in the language information display field 40B when the correction complete button 40D was operated as correct language information after correction. The registration unit 20D then assigns a flag indicating that the language information accepted as the corrected language information among the multiple pieces of language information corresponding to the text in the language dictionary 16A is correct language information.

[0058] Furthermore, if the flag has already been added to one of the pieces of language information corresponding to the text in the language dictionary 16A, the registration unit 20D may replace the flag with the language information received as the corrected language information. In this case, a flag indicating that the language information is correct is added to one of the pieces of language information corresponding to one text registered in the language dictionary 16A.

[0059] Furthermore, if the flag has already been added to at least one of the multiple pieces of language information corresponding to the text in the language dictionary 16A, the registration unit 20D may further add a flag to the language information received as the corrected language information. In this case, 0, 1, or multiple flags are added to each of the multiple pieces of language information corresponding to one text registered in the language dictionary 16A.

[0060] The display control unit 20B may control the display of the display screen 30E and the display screen 30F next time so that, among the multiple pieces of language information corresponding to one text in the language dictionary 16A, the language information to which the greater number of flags are assigned is displayed in the language information display field 40B in descending order. Then, each time the user operates the next candidate button 40J by operating the input unit 14B, the display control unit 20B may control the display of the language information to which the fewer number of flags are assigned in sequence in the language information display field 40B.

[0061] Furthermore, when the correction completion button 40D on the display screen 30E and the display screen 30F is operated, the display control unit 20B may display the display screen 30C on the display unit 14A (see FIG. 2C) in the same manner as described above, and execute the same process as described above.

[0062] Furthermore, the display control unit 20B may display on the display screen 30 a selectable list of a plurality of pieces of language information in the language dictionary 16A that correspond to the text acquired by the conversion unit 20A.

[0063] 3C is a schematic diagram of an example of a display screen 30G. The display screen 30G is an example of the display screen 30. Moreover, the display screen 30G is an example of the first display screen and the second display screen.

[0064] Display screen 30G includes a text display field 40A, a plurality of language information display fields 40B (language information display field 40B1 to language information display field 40B4), a plurality of play buttons 40C (play button 40C1 to play button 40C4) arranged within each display frame of the plurality of language information display fields 40B, a play button 40M, and a correction complete button 40D.

[0065] The display control unit 20B displays the acquired text in the text display field 40A. The display control unit 20B also displays a plurality of pieces of language information corresponding to the text in a plurality of language information display fields 40B (language information display fields 40B1 to 40B4), respectively. FIG. 3C shows an example of a state in which four pieces of language information are displayed in the language information display fields 40B1 to 40B4, respectively. Within the display frames of the language information display fields 40B1 to 40B4, play buttons 40C1 to 40C4 are respectively arranged. The play buttons 40C1 to 40C4 are examples of the play button 40C.

[0066] The user then operates the input unit 14B to select a display field for the correct language information for the text from among the multiple pieces of language information displayed on the display screen 30G, and then operates the correction complete button 40D. When the correction complete button 40D is operated, the registration unit 20D accepts the selected language information from among the multiple pieces of language information displayed on the display screen 30G as corrected language information. Then, a flag indicating that the language information accepted as corrected language information from among the multiple pieces of language information corresponding to the text in the language dictionary 16A is correct language information may be added.

[0067] Similarly to the above, if the flag has already been added to one of the pieces of language information corresponding to the text in the language dictionary 16A, the registration unit 20D may replace the flag with the language information received as the corrected language information. In this case, a flag indicating that the language information is correct is added to one of the pieces of language information corresponding to one text registered in the language dictionary 16A.

[0068] Furthermore, if the flag has already been added to at least one of the multiple pieces of language information corresponding to the text in the language dictionary 16A, the registration unit 20D may further add a flag to the language information received as the corrected language information. In this case, 0, 1, or multiple flags are added to each of the multiple pieces of language information corresponding to one text registered in the language dictionary 16A.

[0069] In addition, the display control unit 20B may control the display screen 30G to rearrange the language information among the multiple language information corresponding to one text in the language dictionary 16A in order from the language information to which the greater number of flags are assigned, and to arrange and display the language information on the display screen 30G so that the language information to which the greater number of flags are assigned has a higher priority when the display screen 30G is next displayed.

[0070] Furthermore, when the correction completion button 40D on the display screens 30E to 30G is operated, the display control unit 20B may display the display screen 30C on the display unit 14A (see FIG. 2C) in the same manner as described above, and execute the same process as described above.

[0071] Returning to Figure 1, we will continue with the explanation.

[0072] When the voice synthesis unit 20E receives an instruction to reproduce the language information, it generates synthetic voice data using the language information and the synthetic voice dictionary 16B.

[0073] When only one language information display field 40B is displayed on the display screen 30 where the play button 40C is displayed in response to a user's instruction to operate the input unit 14B, the voice synthesis unit 20E generates synthetic voice data using the language information displayed in the language information display field 40B.

[0074] Specifically, for example, when the play button 40C is operated while the display screen 30F shown in FIG. 3B is displayed on the display unit 14A, the voice synthesis unit 20E generates synthetic voice data using the language information displayed in the language information display field 40B of the display screen 30F.

[0075] In addition, when multiple language information display fields 40B are displayed on the display screen 30 where the play button 40C operated by the user's operation instruction of the input unit 14B is displayed, the voice synthesis unit 20E generates synthetic voice data using the language information displayed in the language information display field 40B in which the operated play button 40C is located.

[0076] Specifically, for example, assume that the playback button 40C2 is operated while the display screen 30G shown in Fig. 3C is displayed on the display unit 14A. In this case, the voice synthesis unit 20E generates synthetic voice data using the language information displayed in the language information display field 40B2 of the display screen 30G.

[0077] The voice synthesis unit 20E generates synthetic voice data using the language information to be reproduced and the synthetic voice dictionary 16B.

[0078] The synthetic speech dictionary 16B is a dictionary used for speech synthesis, and includes an acoustic model that models voice quality, a prosodic model that models prosody such as intonation and rhythm, and various other information necessary for speech synthesis. The prosodic model is a learning model that takes linguistic information as input and outputs prosodic patterns such as intonation and rhythm. The prosodic pattern is represented, for example, by a time change pattern (also called pitch envelope) of the voice pitch (fundamental frequency) and the length of each phoneme. The acoustic model is a learning model that takes a prosodic pattern as input and outputs synthetic speech data represented by a speech waveform.

[0079] The speech synthesis unit 20E obtains a prosodic pattern by inputting the linguistic information to be reproduced into a prosodic model included in the synthetic speech dictionary 16B. The speech synthesis unit 20E then obtains synthetic speech data by inputting the obtained prosodic pattern into an acoustic model. Through these processes, the speech synthesis unit 20E generates synthetic speech data.

[0080] The playback unit 20F plays back the synthetic voice data generated by the voice synthesis unit 20E. Playing back the synthetic voice data means outputting the voice represented by the voice data from the voice output unit 14C.

[0081] In this way, the voice synthesis unit 20E generates synthetic voice data using the language information specified by the user, and the playback unit 20F plays back the synthetic voice data. This allows the user to input modifications to the language information while previewing the synthetic voice data generated from the selected or modified language information.

[0082] Next, an example of the flow of information processing executed by the speech synthesis support device 10 of this embodiment will be described.

[0083] FIG. 4 is a flowchart showing an example of the flow of information processing executed by the speech synthesis support device 10 of this embodiment.

[0084] The conversion unit 20A analyzes the text using the language dictionary 16A, and converts the text into language information (step S100).

[0085] The display control unit 20B displays the display screen 30 including the text analyzed in step S100 and the converted language information on the display unit 14A (step S102). By the process of step S102, for example, the display screen 30A shown in FIG. 2A is displayed on the display unit 14A.

[0086] The reception unit 20C judges whether or not a correction of at least a part of the language information displayed in the language information display field 40B of the display screen 30 has been received (step S104). The user changes at least a part of the language information displayed in the language information display field 40B to correct information by operating the input unit 14B. The reception unit 20C makes the judgment of step S104 by judging whether or not the corrected language information input by the user's operation instruction has been received. If the judgment is negative in step S104 (step S104: No), the process proceeds to step S108 described later. If the judgment is positive in step S104 (step S104: Yes), the process proceeds to step S106. If the judgment is positive in step S104, for example, as shown in the display screen 30B of FIG. 2B, at least a part of the language information displayed in the language information display field 40B is corrected. In FIG. 2B, the underlined parts of the language information displayed in the language information display field 40B indicate the parts to be corrected.

[0087] Next, the reception unit 20C judges whether or not a correction completion instruction has been received (step S106). The reception unit 20C executes the judgment of step S106 by determining whether or not an input of a correction completion instruction has been received from the input unit 14B by operating the correction completion button 40D according to an operation instruction of the input unit 14B by the user. If the judgment in step S106 is negative (step S106: No), the process proceeds to step S108.

[0088] In step S108, the reception unit 20C determines whether or not a playback instruction has been received (step S108). The reception unit 20C performs the determination in step S108 by determining whether or not an input of a playback instruction signal has been received by an operation instruction of the playback button 40C by the user.

[0089] If the determination in step S108 is negative (step S108: No), the process returns to step S104. If the determination in step S108 is positive (step S108: Yes), the process proceeds to step S110.

[0090] In step S110, the voice synthesis unit 20E generates synthetic voice data using the language information and synthetic voice dictionary 16B displayed at a position corresponding to the playback button 40C operated by the user's operation instruction on the input unit 14B (step S110). The playback unit 20F plays the synthetic voice data generated in step S110 (step S112). Then, the process returns to step S104.

[0091] On the other hand, if it is determined in step S106 that a correction completion instruction has been received (step S106: Yes), the process proceeds to step S114.

[0092] In step S141, the display control unit 20B displays on the display unit 14A a second display screen showing the correspondence between the text to be converted in step S100 and the corrected language information for which correction was accepted in step S104 (step S114). By the process of step S114, for example, the display screen 30C shown in FIG. 2C is displayed on the display unit 14A.

[0093] When the user operates the register button 40H on the display screen 30C by operating the input unit 14B, the accepting unit 20C accepts a correction completion instruction (step S116), the registering unit 20D associates the text with the corrected language information and registers them in the language dictionary 16A (step S118). Then, this routine ends.

[0094] As described above, the speech synthesis support device 10 of this embodiment includes a conversion unit 20A, a reception unit 20C, and a registration unit 20D. The conversion unit 20A analyzes text using the language dictionary 16A and converts the text into language information to be input to the synthetic speech dictionary 16B used to generate synthetic speech data. The reception unit 20C receives corrections to at least a portion of the language information. The registration unit 20D associates the text with the corrected language information and registers it in the language dictionary 16A.

[0095] In conventional technology, if there was an error in the language information resulting from text analysis, the language information had to be corrected each time analysis was performed through user input, making it difficult to easily convert text into accurate language information.

[0096] On the other hand, in this embodiment, the registration unit 20D associates the text with the corrected language information and registers them in the language dictionary 16A.

[0097] For this reason, even if incorrect language information is registered in language dictionary 16A, language information corrected by the user can be easily registered in language dictionary 16A. Therefore, when generating synthetic speech data from text, it is possible to create synthetic speech data using language information converted by language dictionary 16A in which the corrected language information is registered.

[0098] Therefore, the speech synthesis support device 10 of the present embodiment can easily convert text into accurate linguistic information.

[0099] Moreover, the user can easily correct at least a part of the language information by operating the input unit 14B while visually checking the language information displayed in the language information display field 40B of the display screen 30. Moreover, the speech synthesis support device 10 of the present embodiment A can provide a language dictionary 16A capable of converting text into accurate language information.

[0100] In addition, in the speech synthesis support device 10 of this embodiment, multiple pieces of language information are displayed as candidates for one text, and one piece of language information selected by the user from the multiple pieces of language information is registered in the language dictionary 16A as the correct language information for the text. Therefore, the user can easily select the correct language information from the multiple pieces of language information and give an operation instruction to register it in the language dictionary 16A.

[0101] Furthermore, the voice synthesis unit 20E generates synthetic voice data from the language information selected by the user, and the playback unit 20F plays back the synthetic voice data. This allows the user to easily modify the language information while listening to the synthetic voice data generated from the language information.

[0102] Next, the hardware configuration of the speech synthesis support device 10 of this embodiment will be described.

[0103] FIG. 5 is a diagram showing an example of a hardware configuration of the speech synthesis support device 10 of the present embodiment.

[0104] The speech synthesis support device 10 of this embodiment includes a control device such as a CPU 10A, storage devices such as a ROM (Read Only Memory) 10B and a RAM (Random Access Memory) 10C, a HDD (Hard Disk Drive) 10D, an I / F 10E that connects to a network and communicates, and a bus 10F that connects each part.

[0105] The programs executed by the speech synthesis support device 10 of this embodiment are provided in a state that they are pre-installed in the ROM 10B or the like.

[0106] The program executed by the speech synthesis support device 10 of this embodiment may be configured to be provided as a computer program product by being recorded in an installable or executable format on a computer-readable recording medium such as a CD-ROM (Compact Disk Read Only Memory), a flexible disk (FD), a CD-R (Compact Disk Recordable), or a DVD (Digital Versatile Disc).

[0107] Furthermore, the program executed by the speech synthesis support device 10 of the present embodiment may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Also, the program executed by the speech synthesis support device 10 of the present embodiment may be provided or distributed via a network such as the Internet.

[0108] The program executed by the speech synthesis support device 10 of this embodiment can cause a computer to function as each unit of the above-mentioned speech synthesis support device 10. In this computer, the CPU 10A can read the program from a computer-readable storage medium onto a main storage device and execute the program.

[0109] In the above embodiment, the speech synthesis support device 10 has been described as being configured as a single device. However, the speech synthesis support device 10 may be configured as a plurality of devices that are physically separated and connected to each other so as to be able to communicate with each other via a network or the like.

[0110] Furthermore, the speech synthesis support device 10 of the above embodiment may be realized as a virtual machine that operates on a cloud system.

[0111] Although the embodiment of the present invention has been described above, the embodiment is presented as an example and is not intended to limit the scope of the invention. This new embodiment can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the gist of the invention. This embodiment and its modifications are included in the scope and gist of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0112] 10. Voice synthesis support device 20A conversion unit 20B Display control unit 20C Reception 20D Registration Department 20E Voice synthesis unit 20F Playback Department

Claims

1. a conversion unit that analyzes text using a language dictionary and converts the text into language information to be input into a synthetic speech dictionary used to generate synthetic speech data; a reception unit for receiving a correction of at least a part of the language information; a registration unit that associates the text with the corrected language information and registers the text in the language dictionary; A speech synthesis support device comprising:

2. a display control unit that displays a first display screen including the text and the language information converted by the conversion unit; The display control unit is when accepting correction of at least a part of the linguistic information, displaying a second display screen showing a correspondence between the text and the linguistic information after the correction is accepted; The registration unit is when a registration instruction is received via the second display screen, the text included on the second display screen is associated with the corrected language information and registered in the language dictionary. The speech synthesis support device according to claim 1 .

3. The registration unit is registering the text included in the second display screen, the corrected language information, and a category to which the text belongs in the language dictionary in association with each other; The speech synthesis support device according to claim 1 .

4. The display control unit is Selectably displaying one of a plurality of pieces of language information corresponding to the text, the plurality of pieces of language information being different from one another; The registration unit is registering the text in the language dictionary in association with one of the displayed pieces of language information selected from the plurality of pieces of language information; The speech synthesis support device according to claim 1 .

5. a voice synthesis unit that generates synthetic voice data using the language information and the synthetic voice dictionary when an instruction to reproduce the language information is received; a playback unit for playing back the synthetic voice data; The speech synthesis support device according to claim 1 .

6. A speech synthesis support method executed by a speech synthesis support device, comprising: a conversion step of analyzing the text using a language dictionary and converting the text into language information for input to a synthetic speech dictionary used to generate synthetic speech data; a receiving step of receiving a correction to at least a part of the language information; a registration step of registering the text and the corrected language information in the language dictionary in association with each other; A speech synthesis support method comprising:

7. a conversion step of analyzing the text using a language dictionary and converting the text into language information for input to a synthetic speech dictionary used to generate synthetic speech data; a receiving step of receiving a correction to at least a part of the language information; a registration step of registering the text and the corrected language information in the language dictionary in association with each other; A speech synthesis support program for running the above on a computer.

Images