JP2000293187A - Device and method for synthesizing data voice - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a data voice synthesizing device and a data voice synthesizing method in which a user is able to listen to the synthesized voice of a text without delay and a voice file is provided in accordance with the desire of the user. SOLUTION: The device is provided with a text input section 111 which extracts text data that are the voice synthesis object from a text database 12, a text dividing process section 112 which divides the data of the voice synthesis object extracted from the database 12 in accordance with punctuation marks, and a voice synthesis processing section 113 which conducts voice synthesis for every text data of the divided voice synthesis object.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data speech synthesizing apparatus and a data speech synthesizing method, and more particularly to a data speech synthesizing apparatus and a data speech synthesizing method suitable for quickly synthesizing a long text sentence into speech and reading it out. About.

[0002]

2. Description of the Related Art Conventionally, there has been developed a technique for synthesizing text data by voice, converting the text data into a voice file, and outputting the voice as voice. In the conventional method of speech synthesis of text data, text data is passed to a speech synthesis engine as it is, and is converted into a voice file or converted into voice data in a memory.

However, as the amount of information increases and the text data size increases, the time required for speech synthesis processing becomes very long. In addition, the operation cannot be performed while taking the user's intention into account due to a further delay.

As a conventional example relating to the above-described speech synthesis, for example, a technique described in Japanese Patent Application Laid-Open No. 9-307658 has been proposed. The gazette is intended to freely read out an e-mail by a simple operation, and utilizes a message data file for storing a response message described in a predetermined format. An information processing method is disclosed in which an output instruction is input and a response message corresponding to the input output instruction is voice-synthesized and output.

As another conventional example relating to the above-described speech synthesis, a technique described in, for example, Japanese Patent Application Laid-Open No. H10-149273 has been proposed. The gazette aims to realize an utterance that makes it easy to hear the content and meaning of the e-mail, stores the e-mail, analyzes the text of the stored e-mail, and performs the analysis. There is disclosed an information processing method, wherein an utterance attribute is determined according to a result, and the text is speech-synthesized based on the determined utterance attribute.

[0006]

However, the above-described prior art has the following problems.

The first problem is that since the size of text data to be subjected to speech synthesis processing is large, it takes a very long time to perform speech synthesis processing on long text data.

[0008] The second problem is that the processing cannot be interrupted until the reproduction of the text is completed, or even if the processing can be interrupted, the speech synthesis processing of the next text data is not completed, so that it is not ready for reproduction. Another problem is that the user's instruction cannot be reflected during the speech synthesis processing and the reproduction.

SUMMARY OF THE INVENTION An object of the present invention is to provide a data speech synthesizer and a data speech synthesis apparatus which enable a user to listen to the result of text-to-speech synthesis without delay and provide a speech file according to the user's intention. It provides a method.

[0010]

According to the present invention, there is provided a data-speech synthesizing device for synthesizing data, comprising: a storage means for storing a plurality of data; an extraction means for extracting data to be synthesized from the storage means; A dividing unit that divides the speech synthesis target data extracted from the storage unit by the extraction unit based on a predetermined division condition; and a speech synthesis unit that performs speech synthesis for each of the speech synthesis target data divided by the division unit. It is characterized by having.

The present invention also relates to a data speech synthesizing method for synthesizing data, in which specific speech synthesis target data is extracted from a storage unit storing a plurality of speech synthesis target data, and the storage unit is used in the extracting step. The speech synthesis target data extracted from the above is divided based on a predetermined division condition, and speech synthesis is performed for each of the speech synthesis target data divided in the dividing step.

The data-speech synthesizing apparatus according to the present invention, which will be described with reference to FIG. 1, is a data-speech synthesizing apparatus for synthesizing data. Extracting means (111 in FIG. 1) for extracting speech synthesis target data from the storage means; and dividing means for dividing the speech synthesis target data extracted from the storage means by the extraction means based on predetermined division conditions ( A speech synthesis unit (112 in FIG. 1) for performing speech synthesis for each of the speech synthesis target data divided by the division unit (113 in FIG. 1).
Is provided.

[Operation] The data-speech synthesizing apparatus of the present invention comprises:
It controls to divide text and synthesize speech. For this reason, the speech synthesis processing of the divided text can be performed in a short time, and the user can hear the result of the speech synthesis of the text without delay. Also,
The voice-synthesized data is controlled so as to follow a voice synthesis schedule in which a user's instruction is read ahead. For this reason, it is possible to provide the user with an audio file such as “next e-mail” according to the user's intention. Further, the speech synthesis target data can handle not only text data but also a wide range of data. Therefore, the user can obtain an advantage that data is not limited.

[0014] In addition, the data-speech synthesizing apparatus of the present invention converts the speech-synthesis target data extracted from the storage means to a period /
Divide into several data based on division conditions such as reading points,
The feature is that the divided data is synthesized and output as a voice, which is different from the above-described conventional voice synthesizing method.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

(1) Description of Configuration FIG. 1 is a block diagram showing a configuration example of a text-to-speech synthesis processing system according to a first embodiment of the present invention. In FIG. 1, the text-to-speech synthesis processing system according to the first embodiment of the present invention includes a speech synthesis processing device 11, a text database 12, an instruction device 13, and a voice output device 14. Further, the speech synthesis processing device 11 includes the text input unit 1
11, text division processing unit 112, speech synthesis processing unit 11
3 is provided.

More specifically, the speech synthesis processor 11 divides the text extracted from the text database 12, performs speech synthesis for each of the divided texts, and stores the synthesized speech as a file. (Not shown), and play it back at the occasion of reading out.
The speech synthesis processing device 11 is specifically a device applied to a speech synthesis engine that performs speech synthesis processing in software in a computer. The text input unit 111 of the speech synthesis processing device 11 extracts a text to be speech-synthesized from the text database 12 based on a speech synthesis instruction from the instruction device 13. The text segmentation processing unit 112 of the speech synthesis processing device 11 divides the text to be speech-synthesized into several texts within a predetermined size with a break at a reading point / punctuation mark or the like. In the actual use, the predetermined size is, for example, 500 kbytes held as an initial value, and the division size is up to the reading point and the period beyond 500 kbytes. The speech synthesis processing division processing unit 113 of the speech synthesis processing device 11 performs a speech synthesis process for each divided text.

The text database 12 is specifically provided in an e-mail server and stores text (e-mail). The instruction device 13 is specifically a numeric button of a telephone or a keyboard of a computer, and is used to input a user ID, a password, and a voice synthesis instruction. The voice output device 14 is, specifically, a telephone receiver or a speaker connected to a computer, and outputs voice of a voice file synthesized by the voice synthesis processing device 11.

(2) Description of Operation Next, the operation of the first embodiment of the present invention will be described with reference to FIGS.
This will be described in detail with reference to FIG. FIG. 2 is a flowchart showing the flow of a speech synthesis process according to the first embodiment of the present invention, FIG. 3 is an explanatory diagram showing a speech synthesis schedule according to the first embodiment of the present invention, and FIG. FIG. 4 is an explanatory diagram illustrating a specific example of a speech synthesis process.

First, the overall operation of the first embodiment of the present invention will be described in detail with reference to the flowchart of FIG.

First, the text input unit 111 of the speech synthesis processing device 11 extracts a text to be speech-synthesized based on the user ID, password, and speech synthesis instruction input by the user from the instruction device 13. The text is referred to from the text database 12 (Step S21).
Next, the text input unit 111 determines whether there is text (step S22). In this case, since the text has been extracted from the text database 12, next,
The text division processing unit 112 performs text division processing, and delivers the divided text to the speech synthesis processing unit 113 (speech synthesis engine) (step S23). Thereby, the voice synthesis processing unit 113 performs voice synthesis for each of the divided texts to create a voice file (step S24).

The voice synthesis processing unit 113 outputs the generated voice file via the voice output device 14 as voice. This allows the user to hear the completed audio file as audio (step S25). Thereafter, the text input unit 111 determines whether or not the user has given a speech synthesis instruction from the instruction device 13 (step S26). If there is a next user instruction, the process returns to step S21 to repeat the above series of processing. On the other hand, if there is no next user's instruction, this processing ends.

Next, a specific example of the first embodiment of the present invention will be described with reference to FIG. As shown in FIG. 1, the user makes a telephone call from the pointing device 12 (telephone in this example) to the voice synthesizer 11 and sends a text (e-mail).
Give a listening instruction. When the user inputs a user ID and a password from the instruction device 12, the speech synthesis processing device 1
1 text input unit 111 is a text (email)
From the text database 12 of the e-mail server. The text division processing unit 112 divides the text (e-mail) obtained by the text input unit 111 according to a period or a reading point. Do. The voice synthesized data follows the voice synthesis schedule.

The speech synthesis schedule arranges the divided texts as shown in FIG. 3 in the order required by the user and reproduces the text as instructed by the user. Eleven voice synthesis processing units 113 are created and stored in advance based on the input from the instruction device 13. The rules for the speech synthesis schedule are shown below.

First, speech synthesis is performed on text 1-division 1. Next, when the user instructs the "next text", the text 2 which is immediately required is the text 2-division 1, so this is subjected to speech synthesis next. Next is text 1-split 1
If the two voice synthesis processes of text 3 and division 1 and text 1 and division 2 are completed within the playback time of text 3
-Division 1. However, when two speech synthesis processes cannot be performed, speech synthesis of text 1-split 2 is performed first.
At this time, it is necessary to take into consideration that the time required for the speech synthesis is reduced when the user performs “fast-forward” or “skip” to reduce the reproduction time. As described above, by performing the voice synthesis processing by pre-reading the behavior of the user based on the voice synthesis schedule, it is possible to perform processing without delay.

The created audio files are reproduced for the user in order. During that time, the user will receive the "Next Email"
It is possible to send an instruction such as “fast forward” or “rewind” to the voice synthesizing processing device 11, and it is possible to immediately reproduce the audio file according to the instruction.

As described above, when an instruction to perform speech synthesis is given from the instruction device 13 for transmitting the user's instruction to the text input unit 111 of the speech synthesis processing device 11, the text input unit 111 extracts the speech from the text database 12. The long text is divided by the text division processing unit 112 into several divided texts with a break at a reading point / punctuation mark or the like within a predetermined size. The divided texts are successively delivered to the speech synthesis processing unit 113. The speech synthesis processing of the text divided into small pieces is processed in a short time, converted into a speech file, and output to the speech output device 14 as speech.

Further, by adopting the speech synthesis schedule as shown in FIG. 3, it is possible to read out without delay. Even if the user instructs "skip", "fast forward", or "read the next text", it is possible to immediately read out.

As described above, according to the first embodiment of the present invention, since the text is divided and the voice is synthesized, the voice synthesis processing of the divided text can be performed in a short time. Can hear the result of text-to-speech synthesis without delay. Further, since the voice-synthesized data follows the voice synthesis schedule in which the user's instruction is read ahead, a voice file according to the user's intention, such as “next e-mail”, can be provided to the user.

[Second Embodiment] Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

(1) Description of Configuration FIG. 5 is a block diagram showing a configuration example of a text-to-speech synthesis processing system according to a second embodiment of the present invention. In FIG. 5, a text-to-speech processing system according to a second embodiment of the present invention includes a speech
2, an instruction device 53 and an audio output device 54 are provided.
Further, the speech synthesis processing device 51 includes a text input unit 51.
1. Text division processing unit 512, speech synthesis processing unit 513
Is provided. Further, the text input unit 511 includes a data search unit 511A and a text generation unit 511B.

The difference between the second embodiment and the first embodiment in the above configuration will be described.
Stores ordinary data that is not in text format.
The data search unit 511A of the speech synthesis processing device 51 searches the general-purpose database 52 for data to be subjected to speech synthesis based on the speech synthesis instruction from the instruction device 53. The text generation unit 511B of the speech synthesis processing device 51 converts the data searched by the data search unit 511A into text by performing processing such as merging. The other configuration is the same as that of the first embodiment, and the description is omitted.

(2) Description of Operation Next, the operation of the second embodiment of the present invention will be described with reference to FIGS.
This will be described in detail with reference to FIG. FIG. 6 is a flowchart showing the flow of the speech synthesis processing according to the second embodiment of the present invention, and FIG. 7 is an explanatory diagram showing a specific example of the speech synthesis processing according to the second embodiment of the present invention.

First, the overall operation of the second embodiment of the present invention will be described in detail with reference to the flowchart of FIG.

First, the data retrieval unit 511A of the text input unit 511 of the speech synthesis processing unit 51 sends a speech synthesis target from the general-purpose database 52 based on the user ID, password, and speech synthesis instruction input by the user from the instruction device 53. Data search is performed (step S61). Next, the text generation unit 511B of the text input unit 511 converts the data searched by the data search unit 511A into text (Step S62). Next, the text division processing unit 511B of the text input unit 511 determines whether there is a text (Step S63). In this case, since the data in the general-purpose database 52 is converted into text, the text division processing unit 512 next performs text division processing, and delivers the divided text to the speech synthesis processing unit 513 (speech synthesis engine). (Step S64). Thus, the speech synthesis processing unit 513 performs speech synthesis for each of the divided texts to create a speech file (step S65).

The voice synthesis processing unit 513 outputs the generated voice file via the voice output device 54 as voice. This allows the user to hear the completed audio file as audio (step S66). Thereafter, the text input unit 511 determines whether or not the user has given a speech synthesis instruction from the instruction device 53 (step S67). If there is a next user instruction, the process returns to step S61 to repeat the series of processes. On the other hand, if there is no next user's instruction, this processing ends.

Next, differences between the second embodiment and the first embodiment will be described. The text-to-speech synthesis system of the first embodiment shown in FIG. 1 has an aspect that the speech synthesis process cannot be performed unless the original text is prepared in advance.

On the other hand, in the text-to-speech processing system of the second embodiment, data is retrieved from the general-purpose database 52 by the data retrieval processing by the data retrieval unit 511A of the text input unit 511 of the speech synthesis processing unit 51 (see above). Step S61 in FIG. 6), text generator 511
Convert the data to text by B (step S62 in FIG. 6)
This makes it possible to perform a speech synthesis process and reproduce the speech.

Next, a specific example of the second embodiment of the present invention will be described with reference to FIG. General purpose database 5
Let us consider the following case, assuming that the data structure in 2 is as follows.
[Key: Numeric string], [Setting value 1: Numeric string], [Setting value 2: Numeric string], [Reference DB key: Key of another DB] In FIG. When a data search is performed, the data after the search is “3, 4, 3”.
321 ”. In this state, the voice synthesis processing unit 51
Even if speech synthesis processing is performed in step 3, the speech information is not understandable to the user. Therefore, the text input unit 511
The general-purpose database 52 is searched by the data search unit 511A, the [reference DB key] is searched, and by applying the actual data, it can be provided as effective information to the user.

In another DB, if the data retrieved with key: “3321” is text data “title: effective use of database ...”, the data necessary for the user is “set value 1 is 3, set value 2 is 4, title: About effective use of database ... ", it is possible to provide useful information to the user.

As described above, even if the data in the general-purpose database 52 is not in the text format, the data search unit 511A of the text input unit 511 must be passed before the data is passed to the speech synthesis processing unit 513 of the speech synthesis processing unit 51. If the text generation unit 511A performs processing such as retrieval or merging to convert the text into text, the voice synthesis processing unit 513 can perform voice synthesis. Further, the instruction from the instruction device 53 can also include a database operation. This enables speech synthesis of dynamically generated text as well as fixed text.

As described above, according to the second embodiment of the present invention, not only the text data of the text database 12 but also the wide data Therefore, the user can obtain an advantage that data is not limited.

[0043]

As described above, according to the present invention, the text is divided and the voice is synthesized, so that the voice synthesis processing of the divided text can be performed in a short time, and the user is not delayed. You can hear the result of text-to-speech synthesis. Further, since the voice-synthesized data follows the voice synthesis schedule in which the user's instruction is read ahead, a voice file according to the user's intention, such as “next e-mail”, can be provided to the user. Further, since the voice synthesis target data can handle not only text data but also a wide range of data, the user can obtain an advantage that data is not limited.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of a text-to-speech synthesis processing system according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing a flow of a speech synthesis process according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing an example of a speech synthesis schedule according to the first embodiment of the present invention.

FIG. 4 is an explanatory diagram showing a specific example of a speech synthesis process according to the first embodiment of the present invention.

FIG. 5 is a block diagram illustrating a configuration example of a text-to-speech synthesis processing system according to a second embodiment of the present invention.

FIG. 6 is a flowchart illustrating a flow of a speech synthesis process according to a second embodiment of the present invention.

FIG. 7 is an explanatory diagram showing a specific example of a speech synthesis process according to the second embodiment of the present invention.

[Explanation of symbols]

 11, 51 voice synthesis processing device 12 text database 13, 53 pointing device 14, 54 voice output device 52 general-purpose database 111, 511 text input unit 112, 512 text division processing unit 113, 513 voice synthesis processing unit 511A data search unit 511B text Generator

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI theme coat ゛ (reference) H04L 12/58 H04L 11/20 101B H04M 3/50 F term (reference) 5B089 GA21 GB04 JA31 JB05 KB04 KH14 LB13 5D045 AA20 AB02 5K015 AA00 GA00 GA12 5K030 GA18 HA06 KA04 KA20 LB16 LD17 LE14 9A001 BZ03 FF03 HH18 JJ14 JZ19

Claims (7)

[Claims] 1. A data-speech-synthesizing apparatus for performing data-speech-synthesis, comprising: a storage unit that stores a plurality of data to be voice-synthesized; an extraction unit that extracts data to be voice-synthesized from the storage unit;
A dividing unit that divides the speech synthesis target data extracted from the storage unit by the extraction unit based on a predetermined division condition; and a speech synthesis unit that performs speech synthesis for each of the speech synthesis target data divided by the division unit. A data speech synthesizer comprising: 2. An apparatus according to claim 1, further comprising: instruction means for instructing speech synthesis of the speech synthesis target data; and speech output means for outputting a speech synthesis result as speech. The voice synthesis target data is extracted from the storage unit, and the dividing unit divides the voice synthesis target data extracted from the storage unit by the extraction unit into breaks, punctuation marks, and the like, and the voice synthesis unit 2. The data speech synthesizer according to claim 1, wherein speech synthesis is performed for each of the speech synthesis target data divided by the division means, and the speech synthesis result is outputted as speech from the speech output means. 3. The data speech synthesizer according to claim 2, wherein the speech synthesis means causes the speech output means to output the speech synthesis result according to a speech synthesis schedule. 4. The voice synthesizing schedule according to claim 3, wherein the voice synthesizing target data divided by the dividing means is arranged in an order that a user needs next. Data speech synthesizer. 5. The speech synthesis target data is text data such as an electronic mail, and the storage unit stores a plurality of the text data.
Or the data speech synthesizer according to 2. 6. The speech synthesis target data is data that is not in a text format, the storage unit stores a plurality of the data that is not in a text format, and the extraction unit is further configured to output the data that is not in the text format from the storage unit. 3. The data speech synthesizing apparatus according to claim 1, further comprising: a data search unit that searches data; and a text generation unit that converts the non-text data searched by the data search unit into text. 7. A data speech synthesis method for performing speech synthesis of data, comprising: an extraction step of extracting specific speech synthesis target data from a storage unit that stores a plurality of speech synthesis target data; A data speech, comprising: a dividing step of dividing the extracted speech synthesis target data based on predetermined division conditions; and a speech synthesis step of performing speech synthesis for each of the speech synthesis target data divided in the division step. Synthesis method.