JP2007333851A - Speech synthesis method, speech synthesizer, speech synthesis program, speech synthesis delivery system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a speech synthesis method for performing suitable hearing restriction or preventing unauthorized use of speech data without replacing contents of input texts, and to provide a speech synthesizer, a speech synthesis program and a speech synthesis delivery system thereof. <P>SOLUTION: The speech synthesis method has an authentication step for checking whether a listener is a registered user based on identification information of the listener hearing output speech; a step for reading a list in a storing means for storing the list holding a table of suppression words; a determination step for determining whether the suppression words are included in the input text; and a synthesizing step for synthesizing speech data based on the input text when the suppression words are not included in the input text or when the listener is the registered user. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a speech synthesis method, a speech synthesizer, a speech synthesis program, and a speech synthesis distribution system.

Conventionally, as a technique of “not outputting a specific word or a combination of words at the time of voice output”, “from a suppressed word list 107 in which words that are prohibited from voice reproduction are registered as registered words and an input document file, Extraction means for extracting registered words from the suppression word list 107, replacement means for replacing the registered words extracted by the extraction means with a predetermined character string in the document file, and a document file replaced by the replacement means And voice output means (104, 105) for outputting voice based on the above. Is proposed (Patent Document 1).
JP 2002-221981 (Abstract)

However, there are cases where it is not always preferable to output the voice by replacing the contents of the input text depending on the use and usage form of the synthesized voice.
In order to cope with such a case, a speech synthesis method, a speech synthesizer, a speech synthesis program, and a speech that can appropriately restrict listening or prevent unauthorized use of speech data without replacing the contents of the input text. A composite distribution system was desired.

A speech synthesis method according to the present invention includes:
When synthesizing and outputting input text as speech, if the text contains a phrase that should suppress speech output (hereinafter referred to as a suppression phrase), a method of outputting with restricted listening is used. There,
An authentication step for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Reading the list from storage means storing a list holding a list of suppression words;
A determination step of determining whether or not a suppression word is included in the input text;
A synthesis step of synthesizing speech data based on the input text only if the input text does not contain a suppression word or if the listener is a regular user;
It is characterized by having.

The speech synthesizer according to the present invention
When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Authentication means for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Storage means for storing a list for holding a list of suppression words;
A determination unit that reads a list holding a list of suppression words from the storage unit, and determines whether or not the suppression word is included in the input text; and
A synthesis means for synthesizing speech data based on the input text only when the input text does not contain a suppression word or when the listener is a regular user;
It is characterized by having.

In addition, the speech synthesis distribution system according to the present invention includes:
A speech synthesis distribution system having a speech synthesis server and a distribution server for distributing speech data output by the speech synthesis server,
The speech synthesis server
Computing means and storage means for storing a speech synthesis program for causing the computation means to execute the speech synthesis method described above,
The storage means
A user table holding the identification information of the listener;
A table representing the relationship between the listener's identification information and the level of suppression allowed for that listener;
The present invention is characterized in that a list holding a set of suppression words and suppression levels of the suppression words is stored.

  According to the present invention, when a voice message to be synthesized and output includes an expression that touches public order and morals, an inappropriate phrase such as a broadcast-prohibited term or a discrimination term, or a person who uses a specific service. A voice message providing service that can only be listened to is provided with a mechanism for restricting listening to the voice, and an appropriate distribution voice can be synthesized according to the attributes of the listener to cope with it appropriately.

Embodiment 1 FIG.
FIG. 1 is a functional block diagram of the speech synthesizer according to Embodiment 1 of the present invention.
A speech synthesizer 100 shown in FIG. 1 includes an input unit 101, an authentication unit 102, a suppression word determination unit 103, a suppression word list storage unit 104, and a speech synthesis unit 105.
The input unit 101 receives the reading target text 110 as an input and outputs it to the suppression word determination unit 103.
The authentication unit 102 receives information for identifying the user of the speech synthesizer 100 as an input, performs authentication processing, and outputs the result to the speech synthesizer 105.
The suppression word determination unit 103 reads the suppression word list from the suppression word list storage unit 104, determines whether or not the suppression word phrase is included in the content of the text to be read 110, and the result and the content of the text 110 to be read out. Is output to the speech synthesizer 105.
The suppression word list storage unit 104 stores a suppression word list shown in FIG.
The voice synthesis unit 105 synthesizes and outputs the voice 111 based on the outputs of the authentication unit 102 and the suppression word determination unit 103. As an output method, a method of writing to a storage means as an audio data file can be used.

FIG. 2 shows a configuration example of the suppression word list stored in the suppression word list storage unit 104.
The suppression word list holds a list of words and phrases that should be suppressed in voice output, and is stored in the suppression word list storage unit 104 in a text file format such as CSV (Comma Separated Values) or a relational database table format. Can be configured to.

FIG. 2 shows a configuration and data example when stored in a table format. Hereinafter, each column will be described.
The “index” column is a column that holds the first character of the suppression word as an index, and is provided for convenience of search and the like. In this Embodiment 1, since the suppression word expressed in Japanese is assumed, the value of this column stores "an" of 50 tones, but holds the English suppression word. In some cases, an alphabetic index may be used, and the data may be useful for organizing and searching.
The “suppression word” column is a column holding words / phrases for which voice output should be suppressed. The values in this column are summarized and organized by the “index” column.

FIG. 3 explains the overall operation flow of the speech synthesizer 100 of FIG. Hereinafter, each step will be described.
(S301)
The input unit 101 receives 0 content of the text 110 to be read out. The received content is output to the suppression word determination unit 103.
(S302)
The authentication unit 102 receives information for identifying the user of the speech synthesizer 100 as an input and performs an authentication process.
The contents of the authentication process can be realized by, for example, a method in which a pair of a regular ID and a password is stored in a storage area in the authentication unit 102 and the consistency with the contents is confirmed.
The result of the authentication process is output to the speech synthesizer 105.
(S303)
The suppression word determination unit 103 reads the suppression word list from the suppression word list storage unit 104.
(S304)
The suppression word determination unit 103 determines whether or not the content of the text to be read 110 received from the input unit 101 includes a phrase held in the suppression word list. The determination result is output to the speech synthesizer 105.
If it is determined that it is included, the process proceeds to step S305. If it is determined that it is not included, the process proceeds to step S306.
The content of the text to be read 110 is output to the speech synthesizer 105.
(S305)
The speech synthesizer 105 determines whether the user of the speech synthesizer 100 is a regular user based on the output of the authentication unit 102.
If it is determined that the user is an authorized user, the process proceeds to step S306. If it is determined that the user is not an authorized user, the process is terminated.
(S306)
The speech synthesis unit 105 performs speech synthesis processing based on the content of the text to be read 110 output from the suppression word determination unit 103.
The content of the synthesis process is obtained by performing a morphological analysis of the text 110 to be read out and combining the speech segments in the speech corpus constructed in advance, or by synthesizing in accordance with a predetermined probability model. Can be realized.
(S307)
The voice synthesizer 105 outputs the voice synthesized in step S306.

  According to the method shown in the flowchart of FIG. 3, the listener is authenticated in advance, and when the suppression target phrase is included in the text to be read 110, the listening restriction is performed. Can be implemented.

FIG. 4 shows a modification of the flowchart shown in FIG.
In FIG. 3, if at least one suppression word / phrase is included in the content of the text to be read out 110, the listening restriction is added to the entire sound.
The flowchart shown in FIG. 4 is such that listening restriction is applied only to the portion including the suppression word / phrase. Hereinafter, each step will be described.
(S401) to (S403)
This is the same as steps S301 to S303 in FIG.
(S404)
The following steps S405 to S409 are repeatedly executed until the end of the text to be read 110 is reached.
(S405)
The suppression word determination unit 103 divides the content of the text to be read 110 into predetermined block units in advance.
Next, the suppression word determination unit 103 reads the first sound reproduction block of the text to be read 110. At the second and subsequent executions of this step, the process moves to the next processing target block of the reading target text 110.
(S406)
The suppression word determination unit 103 determines whether or not the content of the text to be read 110 received from the input unit 101 includes a phrase held in the suppression word list. The determination result is output to the speech synthesizer 105.
If it is determined that it is included, the process proceeds to step S407. If it is determined that it is not included, the process proceeds to step S408.
The content of the text to be read 110 is output to the speech synthesizer 105.
(S407)
The speech synthesizer 105 determines whether the user of the speech synthesizer 100 is a regular user based on the output of the authentication unit 102.
If it is determined that the user is an authorized user, the process proceeds to step S408. If it is determined that the user is not an authorized user, the process returns to step S405.
(S408)
The speech synthesis unit 105 performs speech synthesis processing based on the content of the text to be read 110 output from the suppression word determination unit 103.
The content of the synthesis process is the same as that in step S306 in FIG.
(S409)
The voice synthesizer 105 outputs the voice synthesized in step S408.

According to the method shown in the flowchart of FIG. 4, even if a suppression word / phrase is included in the content of the text 110 to be read out, a portion other than the suppression word / phrase can be heard regardless of the listener's attributes. Increase convenience.
However, in this case, although the content of the text to be read 110 is not changed, the audio output content that the listener hears is different from the content of the text to be read 110. An appropriate method may be used according to individual cases, such as selecting an appropriate method of FIG. 3 or FIG. 4 or inserting a silent part.

  In the first embodiment, the authentication unit 102 determines whether or not the listener is a legitimate person based on the identification information. However, the present invention is not limited to this. For example, an external authentication server or the like is used. May be configured to receive only the authentication result, or a database storing the listener's ID / password pair may be stored in the speech synthesizer 100, and the database may be searched. Authentication may be performed.

As described above, according to the speech synthesizer according to the first embodiment,
When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Authentication means for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Storage means for storing a list for holding a list of suppression words;
A determination unit that reads a list holding a list of suppression words from the storage unit, and determines whether or not the suppression word is included in the input text; and
A synthesis means for synthesizing speech data based on the input text only when the input text does not contain a suppression word or when the listener is a regular user;
Therefore, appropriate restrictions can be implemented according to the attributes of the listener.
Further, if the presence or absence of a suppression word is determined for each predetermined block of the input text, even if the suppression text is included in the content of the text to be read 110, the part other than the suppression word depends on the attribute of the listener. Therefore, convenience for the user is increased.

Embodiment 2. FIG.
In the first embodiment, the configuration of a speech synthesis method and apparatus for preventing a person who is not a regular listener from hearing the synthesized speech has been described. That is, at the stage of speech synthesis, the speech synthesizer itself limits listening.
On the other hand, as a method for restricting listening, a method for delegating whether or not to limit an external application that performs voice reproduction is delegated, and at the stage of speech synthesis, a method for stopping information from being used for the determination is stopped. It is also possible to take.
The speech synthesizer according to Embodiment 2 of the present invention provides information used for such external application determination at the time of speech synthesis.

FIG. 5 is a functional block diagram of the speech synthesizer according to the second embodiment.
A speech synthesis apparatus 500 shown in FIG. 5 includes an input unit 501, a suppression word list storage unit 504, a speech synthesis unit 505, and a calculation unit 506.
The input unit 501 receives the reading target text 110 as an input and outputs it to the calculation unit 506.
The suppression word list storage unit 504 stores a suppression word list shown in FIG.
The calculating unit 506 reads the suppression word list from the suppression word list storage unit 504, determines whether or not the suppression word phrase is included in the content of the text to be read 110, and the result and the content of the text 110 to be read out are voiced. The data is output to the combining unit 505. If a suppression word / phrase is included, the degree of suppression of the speech output (hereinafter referred to as a suppression level) is calculated, and the calculation result is output to the speech synthesis means 505.
The voice synthesizing unit 505 synthesizes and outputs the voice 111 based on the output of the calculating unit 506. As an output method, a method of writing to a storage means as an audio data file can be used.

FIG. 6 shows a configuration example of the suppression word list stored in the suppression word list storage unit 504. As in the first embodiment, the suppression word list can be configured to be stored in the suppression word list storage unit 504 in a text file format such as CSV or a table format of a relational database.
Hereinafter, each column in FIG. 6 will be described.
The contents of the “index” column and the “suppression word” column are the same as those shown in FIG.
The “inhibition degree level” column expresses the inhibition degree level of the inhibition phrase in numerical values. The level classification can be appropriately set in advance according to the content of the suppression word and the assumed usage scene.

  In FIG. 6, the suppression words and their suppression levels are described as being held in the same table. However, the present invention is not limited to this. The suppression levels are stored in different tables, and the relationship between the two is external. You may comprise so that it may represent by key restrictions etc.

FIG. 7 shows an internal configuration image of synthesized audio data in the second embodiment.
In the second embodiment, the synthesized voice data includes information indicating the suppression level obtained by the calculation means. The information indicating the suppression level may be configured to have one common value for the entire audio data, for example, for each audio frame, information indicating the suppression level of the suppression word / phrase in the frame. You may comprise as follows.

(1) in FIG. 7 shows an image of audio data when the audio data is configured to have one common value.
In this case, it is preferable to provide a header portion at the beginning of the audio data and to have information indicating the suppression level of the entire audio data in a part of the header portion.

(2) of FIG. 7 shows an image of audio data in a case where each audio frame is configured to have information indicating the suppression level of the suppression word / phrase in the frame.
In general, since the degree to which suppression words are included in audio data varies from part to part, if information indicating the suppression level is provided for each frame, frames with a suppression level of a certain value or less are attributed to the listener. However, it is possible to make it possible to listen, so that convenience for the user is increased.
In this case, each frame of the audio data is divided into a frame header part and a frame data part, and the frame header part has information indicating the suppression level together with the audio attributes such as the byte length and bit rate of the frame. It is good to make it.

FIG. 8 illustrates an overall operation flow of the speech synthesizer 500 of FIG. Hereinafter, each step will be described.
In addition, in FIG. 8, as shown to (2) of FIG. 7, the flowchart in the case of calculating a suppression degree level for every predetermined | prescribed block is shown.
(S801)
The input unit 501 receives 0 content of the text 110 to be read out. The received content is output to the calculation means 506.
(S802)
The calculation unit 506 reads the suppression word list from the suppression word list storage unit 504. At this time, together with the suppression word / phrase, the suppression level of the suppression word / phrase is read as a set.
(S803)
The following steps S804 to S807 are repeatedly executed until the end of the text to be read 110 is reached.
(S804)
The calculating means 506 divides the contents of the text to be read 110 into predetermined block units in advance.
Next, the calculation means 506 reads the first sound reproduction block of the text 110 to be read out. At the second and subsequent executions of this step, the process moves to the next processing target block of the reading target text 110.
(S805)
The calculating means 506 calculates the inhibition level of the current block.
In the calculation, any calculation standard can be used, such as summing up the suppression levels of the suppression words included in the current block, or obtaining an average value.
The calculation result of the suppression level and the contents of the text 110 to be read out are output to the speech synthesizer 505.
(S806)
The voice synthesizing unit 505 performs a process of synthesizing voice data including information representing the calculation result based on the content of the text to be read 110 output from the calculation unit 506 and the calculation result of the suppression level. As the contents of the audio data, any of the formats (1) and (2) in FIG. 7 can be used.
(S807)
The voice synthesizer 505 outputs the voice synthesized in step S806.

  As shown in (1) of FIG. 7, when calculating the suppression level of the entire audio data, steps S806 and S807 of FIG. 8 are taken out of the loop, and the suppression level of the entire audio data is calculated. What is necessary is just to comprise so that speech synthesis may be performed later.

According to the method shown in the flowchart of FIG. 8, since the suppression level can be calculated at the same time when the suppression word list is read and matching processing with the input text is performed, as shown in FIG. Can create audio data with various levels of suppression.
In other words, listening suppression processing is delegated to an application that reproduces speech, and only information to be used for suppression determination is given at the time of speech synthesis, so speech synthesis is performed with a simple method with less calculation load and is appropriate. Can be suppressed.

As described above, according to the second embodiment,
When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Storage means for storing a list that holds a combination of a suppression word and a suppression level of the suppression word;
From the storage means, a calculation means for reading a list holding a set of suppression words and suppression levels of the suppression words, and calculating a suppression level in the input text;
A synthesizing unit that synthesizes voice data including information representing the calculation result based on the input text and the processing result of the calculating unit;
So that
Speech synthesis can be performed by a simple method with a small calculation load, and appropriate listening suppression can be performed.

The calculating means calculates a suppression level for each predetermined block of the input text,
Since the synthesizing unit synthesizes audio data including information indicating the degree of suppression for each predetermined block,
A frame with a suppression level equal to or less than a certain value can be made audible regardless of the attributes of the listener, which increases the convenience for the user.

Embodiment 3 FIG.
In the first and second embodiments, the configuration has been described in which it is confirmed whether or not the listener is a legitimate user, or information indicating the suppression level is embedded in the audio data. This uses the identification information of the listener and the suppression level as separate information without associating them at all.
In general, however, the degree of suppression allowed for a user may vary from one to another depending on an attribute such as the user's age.
In the third embodiment of the present invention, the suppression level allowed for each user is determined, and the synthesized voice is output only when it is within the range of the appropriate suppression level.

FIG. 9 is a functional block diagram of the speech synthesizer according to the third embodiment.
The speech synthesizer 900 shown in FIG. 9 includes an input unit 901, an authentication unit 902, a suppression word list storage unit 904, a speech synthesis unit 905, a calculation unit 906, and a suppression level determination unit 907.
The input unit 901 receives the reading target text 110 as an input and outputs it to the calculation unit 906.
The authentication unit 902 receives information for identifying the user of the speech synthesizer 900 as an input, performs authentication processing, and outputs the result to the suppression level determination unit 907.
The suppression word list storage unit 904 stores a suppression word list shown in FIG. 6 and a tolerance table shown in FIG. 10 described later.
The calculation unit 906 reads the suppression word list from the suppression word list storage unit 904, determines whether or not the suppression word phrase is included in the content of the text to be read 110, and suppresses the result and the content of the text 110 to be read out. It outputs to the degree level judgment means 907. If a suppression word is included, the suppression level is calculated and the calculation result is output to the suppression level determination means 907.
The suppression level determination unit 907 receives the calculation result of the calculation unit 906 and the authentication result of the authentication unit 902, and determines whether or not the suppression level is within a range allowed by the user of the speech synthesizer 900. . The determination result and the content of the text 110 to be read out are output to the speech synthesizer 905.
The voice synthesizer 905 synthesizes and outputs the voice 111 based on the output of the suppression level determination unit 907. As an output method, a method of writing to a storage means as an audio data file can be used.

FIG. 10 shows the configuration and data example of the tolerance table stored in the suppression word list storage unit 904. Hereinafter, each column will be described.
The “user ID” column corresponds to information that the authentication unit 902 receives and identifies the user of the speech synthesizer 900.
The “allowable restraint level” column represents the maximum restraint level allowed for the user identified by the “user ID”.

FIG. 11 illustrates an overall operation flow of the speech synthesizer 900 of FIG. Hereinafter, each step will be described.
(S1101)
The input unit 901 receives 0 content of the text 110 to be read out. The received content is output to the calculation means 906.
(S1102)
The authentication unit 902 receives information for identifying the user of the speech synthesizer 900 as an input and performs authentication processing.
The contents of the authentication process can be realized by a method of, for example, storing a pair of a regular ID and a password in a storage area in the authentication unit 902 and confirming the consistency with the contents.
The result of the authentication process is output to the calculation unit 906 and the suppression level determination unit 907.
(S1103)
The calculation unit 906 reads the suppression word list from the suppression word list storage unit 904. At this time, together with the suppression word / phrase, the suppression level of the suppression word / phrase is read as a set.
(S1104)
The following steps S1105 to S1110 are repeated until the end of the text to be read 110 is reached.
(S1105)
The calculating means 906 previously divides the content of the text to be read 110 into predetermined block units.
Next, the calculation means 906 reads the first voice reproduction block of the text to be read 110. At the second and subsequent executions of this step, the process moves to the next processing target block of the reading target text 110.
(S1106)
The calculation unit 906 determines whether the user of the speech synthesizer 900 is a regular user based on the output of the authentication unit 902.
If it is determined that the user is an authorized user, the process proceeds to step S1107. If it is determined that the user is not an authorized user, the process returns to step S1105.
(S1107)
The calculating means 906 calculates the suppression level of the current block.
In the calculation, any calculation standard can be used, such as summing up the suppression levels of the suppression words included in the current block, or obtaining an average value.
The calculation result of the suppression level and the content of the text 110 to be read out are output to the suppression level determination unit 907.
(S1108)
The suppression level determination unit 907 receives the calculation result of the calculation unit 906 and the authentication result of the authentication unit 902, and determines whether or not the suppression level is within a range allowed by the user of the speech synthesizer 900. . In the determination, the contents of the tolerance table stored in the suppression word list storage unit 904 are also referred to.
If it is determined that it is within the allowable range, the process proceeds to step S1109. If it is determined that it is not within the range, the process returns to step S1105.
The determination result and the content of the text 110 to be read out are output to the speech synthesizer 905.
(S1109)
The voice synthesizing unit 905 performs a process of synthesizing voice data including information representing the calculation result based on the content of the text to be read 110 output from the suppression level determination unit 907 and the calculation result of the suppression level. As the contents of the audio data, any of the formats (1) and (2) in FIG. 7 can be used.
(S1110)
The voice synthesizer 905 outputs the voice synthesized in step S1109.

Since the suppression level is determined for each user and for each block, sufficient listening suppression processing can be performed without embedding information indicating the suppression level in the audio data, but the sound is reproduced. Depending on the environment, it may be necessary to perform more detailed listening suppression processing.
Therefore, also in the third embodiment, as in the second embodiment, in step S1109, information representing the suppression level is embedded in the audio data.

As described above, according to the third embodiment,
When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Authentication means for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Storage means for storing a list that holds a combination of a suppression word and a suppression level of the suppression word;
From the storage means, a calculation means for reading a list holding a set of suppression words and suppression levels of the suppression words, and calculating a suppression level in the input text;
A degree-of-suppression level judgment means for judging whether or not the degree-of-suppression level calculated by the calculating means is within an allowable range for the listener;
Synthesis that synthesizes speech data including information representing the calculation result based on the input text only when the listener is a regular user and the suppression level is within the allowable range for the listener. Means,
So that
Based on the authentication result of the authentication means 902, it is possible to listen to the synthesized voice only for the authorized user.

In addition, the storage means
Contains a table representing the relationship between the listener's identification information and the level of suppression allowed for that listener,
The suppression level determination means includes
Based on the contents of the table, it is determined whether the suppression level is within the allowable range for the listener.
Since the upper limit value of the suppression level allowed for each user is determined, it is possible to perform listening restriction in accordance with user-specific attributes in units of audio data or predetermined blocks.

Embodiment 4 FIG.
In the configuration described in the first to third embodiments, only the person who can permit listening can be listened to using the listener's identification information and the suppression word list.
However, depending on the usage form of speech data and the rights of speech segment data used for speech synthesis, after an authorized listener obtains speech data, the speech data that can be heard is obtained by an unauthorized listener. In some cases, it is necessary to prevent unauthorized use.
In the fourth embodiment of the present invention, a configuration will be described in which audio data is generated by applying a listening restriction, and then the audio data is encrypted and output to prevent unauthorized use by a person who is not a normal listener.

FIG. 12 is a functional block diagram of the speech synthesizer according to the fourth embodiment.
The speech synthesizer 1200 shown in FIG. 12 includes an input unit 1201, an authentication unit 1202, a suppression word list storage unit 1204, a speech synthesis unit 1205, a calculation unit 1206, a suppression level determination unit 1207, and an encryption unit 1208.
The input unit 1201 receives the reading target text 110 as an input and outputs it to the calculation unit 906.
The authentication unit 1202 receives information for identifying the user of the speech synthesizer 1200 as an input, performs authentication processing, and outputs the result to the suppression level determination unit 1207 and the encryption unit 1208.
The suppression word list storage unit 1204 stores the suppression word list shown in FIG. 6 and the tolerance table shown in FIG.
The calculation unit 1206 reads the suppression word list from the suppression word list storage unit 1204, determines whether or not the suppression word / phrase is included in the content of the text to be read 110, and suppresses the result and the content of the text 110 to be read out. It outputs to the degree level judgment means 1207. If a suppression word is included, the suppression level is calculated, and the calculation result is output to the suppression level determination means 1207.
The suppression level determination unit 1207 receives the calculation result of the calculation unit 1206 and the authentication result of the authentication unit 1202, and determines whether or not the suppression level is within the range allowed by the user of the speech synthesizer 1200. . The determination result and the content of the text 110 to be read out are output to the speech synthesizer 1205.
The voice synthesizing unit 1205 synthesizes voice data based on the output of the suppression level determination unit 1207 and outputs the synthesized voice data. As an output method, a method of writing to a storage means as an audio data file can be used.
The encryption unit 1208 receives the authentication result from the authentication unit 1202 and receives the synthesized voice data from the voice synthesis unit 1205. Next, the audio data is encrypted using the information and output as audio 111. As an output method, a method of writing to a storage means as an audio data file can be used.

FIG. 13 illustrates an overall operation flow of the speech synthesizer 1200 of FIG. Hereinafter, each step will be described.
(S1301)
The input unit 1201 receives 0 content of the text 110 to be read out. The received contents are output to the calculation means 1206.
(S1302)
The authentication unit 1202 receives information for identifying the user of the speech synthesizer 1200 as an input and performs authentication processing.
The contents of the authentication process can be realized by a method of, for example, storing a pair of a regular ID and a password in a storage area in the authentication unit 1202 and confirming consistency with the contents.
The result of the authentication process is output to the calculation unit 1206, the suppression level determination unit 1207, and the encryption unit 1208.
(S1303)
The calculation unit 1206 reads the suppression word list from the suppression word list storage unit 1204. At this time, together with the suppression word / phrase, the suppression level of the suppression word / phrase is read as a set.
(S1304)
The following steps S1305 to S1311 are repeatedly executed until the end of the text to be read 110 is reached.
(S1305)
The calculation means 1206 previously divides the content of the text to be read 110 into predetermined block units.
Next, the calculation means 1206 reads the first voice reproduction block of the text to be read 110. At the second and subsequent executions of this step, the process moves to the next processing target block of the reading target text 110.
(S1306)
The calculation unit 1206 determines whether or not the user of the speech synthesizer 900 is a regular user based on the output of the authentication unit 1202.
If it is determined that the user is an authorized user, the process proceeds to step S1307. If it is determined that the user is not an authorized user, the process returns to step S1305.
(S1307)
The calculation means 1206 calculates the suppression level of the current block.
In the calculation, an arbitrary calculation standard can be used, such as summing up the suppression levels of the suppression words included in the current block, or obtaining an average value.
The calculation result of the suppression level and the content of the text 110 to be read out are output to the suppression level determination unit 1207.
(S1308)
The suppression level determination unit 1207 receives the calculation result of the calculation unit 1206 and the authentication result of the authentication unit 1202, and determines whether or not the suppression level is within a range allowed by the user of the speech synthesizer 1200. . In the determination, the contents of the tolerance table stored in the suppression word list storage unit 1204 are also referred to.
If it is determined that it is within the allowable range, the process proceeds to step S1309, and if it is determined that it is not within the range, the process returns to step S1305.
The determination result and the content of the text 110 to be read out are output to the speech synthesizer 1205.
(S1309)
The voice synthesizing unit 1205 performs a process of synthesizing voice data including information representing the calculation result based on the content of the text to be read 110 output from the suppression level determination unit 1207 and the calculation result of the suppression level. As the contents of the audio data, any of the formats (1) and (2) in FIG. 7 can be used.
The synthesized voice data is output to the encryption unit 1208.
(S1310)
The encryption unit 1208 receives the authentication result from the authentication unit 1202, encrypts the audio data, and outputs it.
As the encryption key used in the encryption process, for example, the identification information received by the authentication unit 1202 may be used as it is, or based on a predetermined generation rule, a new encryption key is generated from the identification information and used. Also good.
(S1311)
The encryption unit 1208 outputs the sound encrypted in step S1310.

  According to the method shown in the flowchart of FIG. 13, the encryption key is generated based on the identification information known only by the authorized listener, and the voice data is encrypted using the encryption key. Unauthorized use by those who are not listeners can be prevented.

As described above, by performing encryption processing on the audio data, an effect of preventing unauthorized use can be expected. However, in order to decrypt the encrypted audio data and reproduce the audio, the algorithm used for the encryption is used. It is necessary to use an audio player or the like that can use a corresponding decoding algorithm. For this reason, the user is forced to use the audio player, which may be an excessive burden on the user.
Therefore, if the configuration is such that the audio watermark information can be restored by embedding predetermined audio watermark information in the audio data instead of encrypting the audio data and using a corresponding restoration method, it is excessive for the user. If it is not burdensome and unauthorized use is made, it will be revealed by the watermark information, so a certain deterrent effect can be expected.

FIG. 14 illustrates the internal structure image of the audio data shown in FIG. 7 further embedded with audio watermark information.
In both (1) and (2) of FIG. 14, the watermark information is embedded in the audio data part itself, not in the header part.
The watermark information may be embedded when the voice synthesizing unit 1205 generates the voice data.

  (1) in FIG. 14 shows an image of audio data when the audio data is configured to have one common audio watermark information. In this case, audio watermark information is embedded in an arbitrary part of the audio data portion.

(2) in FIG. 14 shows an image of audio data in the case where the configuration is such that watermark information is provided for each audio frame.
In this case, the frame data portion is configured to have watermark information. The audio watermark information of each frame may be common to all the frames, or information such as a hash value that can identify the frame may be separately added so that all the frames have different watermark information.
When configured to have different watermark information for each frame, even if someone who is familiar with the frame structure of audio data has illegally copied or altered the data by dividing the frame, etc. It is possible to detect tampering.

As described above, according to the fourth embodiment,
Since the encryption key is generated based on the identification information and the voice data synthesized by the synthesis means is encrypted,
An encryption key is generated based on identification information known only to a legitimate listener, and the voice data is encrypted using the encryption key to prevent unauthorized use by a person who is not a legitimate listener. .

The synthesizing unit synthesizes the audio data including watermark data for identifying the synthesis source of the audio data for each predetermined block of the input text.
If unauthorized use is made, it will be revealed by the watermark information, so a certain deterrent effect can be expected.

Embodiment 5 FIG.
In the first to fourth embodiments, the speech synthesis method provided with predetermined listening restrictions and unauthorized use prevention means and the configuration of the speech synthesis apparatus that implements the speech synthesis method have been described.
In the fifth embodiment of the present invention, a configuration of a system that distributes audio data generated by using these speech synthesis methods or speech synthesis apparatuses to a client terminal as audio content will be described.

FIG. 15 shows an example of a Web page including synthesized audio content as a screen image.
A client terminal 1501 is a computer used by a user for browsing a Web page and listening to audio content. The user activates the web browser software of the client terminal 1501 and browses a web page as shown in a screen configuration example 1502 of the web page.
(1) The user clicks a link button.
The user clicks a link to audio content displayed in the screen configuration example 1502 of the Web page.
(2) The player starts up.
When the user clicks a link to audio content, audio player software corresponding to the audio content format is activated. The user listens to the linked audio content via the audio player software.
In addition, when the audio | voice data contain the information which shows a suppression word and its suppression degree level, it can also be comprised so that the information may be displayed on a player screen.

FIG. 16 illustrates the configuration of a speech synthesis and delivery system for delivering a Web page including link information to audio content as shown in FIG. 15 and linked audio content to client terminals.
In FIG. 16, the client terminal 1601 and the server center 1602 are connected via a network 1603. The user accesses the server center 1602 using the client terminal 1601.
Hereinafter, a communication procedure when a user views a Web page or audio content will be briefly described.
(1) The user transmits a login request to the server center 1602 using the client terminal 1601.
(2) The server center 1602 returns an authentication result to the client terminal 1601.
(3) The user uses the client terminal 1601 to request the server center 1602 to distribute a web page or audio content.
(4) The server center 1602 delivers the requested Web page and audio content.

  A detailed processing sequence will be described later with reference to FIG.

Next, the internal configuration of the server center 1602 will be described with reference to FIG.
In the server center 1602, a distribution server 1604 and a speech synthesis server 1605 are installed.
The distribution server 1604 provides a function of receiving a synthesized voice distribution request from the client terminal 1601 and transferring the request to the voice synthesis server 1605. In addition, a function of delivering voice data generated by the web page and the voice synthesis server 1605 to the client terminal 1601 is also provided.
The voice synthesis server 1605 provides a function of dynamically synthesizing voice based on the request transferred from the distribution server 1604 and returning it to the distribution server 1604. In the fifth embodiment, “dynamically” voice synthesis means that voice data is not prepared in advance, but voice synthesis is started when a request is received, and the generated voice data is output. It means that.

Next, the configuration of the speech synthesis server 1605 will be described.
The speech synthesis server 1605 includes a calculation unit (not shown) and a storage unit 1606.
The storage unit 1606 stores a speech synthesis program 1610, a user table 1620, a suppression word list 1621, and a tolerance table 1622.
The speech synthesis program 1610 can be configured as a program that causes the calculation means to execute any of the flowcharts (FIGS. 3, 4, 8, 11, and 13) described in the first to fourth embodiments. In the fifth embodiment, it is assumed that the program is a program for causing the arithmetic means to execute the processing of the flowchart of FIG.
The user table 1620 holds user identification information. Details will be described later with reference to FIG.
The suppression word list 1621 has the same configuration as that shown in FIG.
The tolerance table 1622 has the same configuration as that shown in FIG.
Details will be described later with reference to FIG.

With the above configuration, the speech synthesis server plays the role of an audio data supplier, that is, an encoder. In addition, the distribution server serves as a streaming server that distributes the voice data generated by the voice synthesis server serving as an encoder to the client terminal.
However, the method for distributing the audio data is not limited to the streaming method, and it may be configured to distribute by the download method.

FIG. 17 explains the configuration and data example of the user table 1620. Hereinafter, each column will be described.
The “user ID” column is an ID for identifying a person who views a Web page and audio content using the client terminal 1601.
The “password” column is a login password assigned to the person identified by the user ID.
The user transmits a login request to the distribution server 1604 using a set of values in the “user ID” column and “password” column (corresponding to step (1) in FIG. 16).
The “session ID” column is a field for storing a unique random alphanumeric string assigned to the user after the login process is completed. Details will be described later with reference to FIG.

FIG. 18 illustrates a detailed processing sequence when the user views a Web page or audio content in FIG. Hereinafter, each step will be described.
(S1801)
The user uses the client terminal 1601 to transmit a login request to the server center 1602 (distribution server 1604). At this time, the user ID and password assigned to the user in advance are transmitted together.
(S1802)
The distribution server 1604 receives the ID / password pair transmitted from the client terminal 1601, and issues a query to the speech synthesis server 1605 using the pair as a key.
(S1803)
The computing means of the speech synthesis server 1605 performs user authentication processing based on instructions from the speech synthesis program 1610.
Specifically, the user table 1620 is searched with the ID / password combination transmitted from the client terminal 1601, and it is determined whether or not the user is a legitimate person based on whether or not there is corresponding data. be able to.
The computing means of the speech synthesis server 1605 returns the result of the authentication process to the distribution server 1604 based on the instruction of the speech synthesis program 1610.
Hereinafter, the description will be continued assuming that the user is authenticated as a regular user.
(S1804)
The distribution server 1604 generates a session ID consisting of a random alphanumeric string. The value of the generated session ID is stored in Cookie and transmitted to the client terminal 1601 and stored in the “session ID” column of the corresponding row of the user table.
(S1805)
The user uses the client terminal 1601 to request the distribution server 1604 to distribute the audio content, for example, by clicking a link to the audio content on the screen as shown in FIG.
When the client terminal 1601 issues a request to the distribution server 1604, the session ID received in step S1804 is transmitted to the distribution server 1604 together with the request.
(S1806)
The distribution server 1604 issues an inquiry query to the speech synthesis server 1605 using the session ID transmitted by the client terminal 1601 as a key.
(S1807)
Based on an instruction from the speech synthesis program 1610, the computing means of the speech synthesis server 1605 searches the user table 1620 using the session ID transmitted by the client terminal 1601 as a key. If there is data having a value in the corresponding “session ID” column, it can be determined that the user who transmitted the session ID has been authenticated.
Next, the computing means of the speech synthesis server 1605 reads the value of “user ID” from the corresponding data in the user table 1620 based on the instruction of the speech synthesis program 1610 and searches the tolerance table 1622 using that value as a key. To do. The value of the suppression level permitted by the user is read from the corresponding data.
(S1808)
The computing means of the speech synthesis server 1605 performs speech synthesis processing based on instructions from the speech synthesis program 1610.
The composition process executes the contents shown in steps S1304 to S1309 in FIG. 13 in the fourth embodiment. The input text can be stored in advance in the storage unit 1606 and used.
(S1809)
Based on an instruction from the speech synthesis program 1610, the arithmetic unit of the speech synthesis server 1605 encrypts the speech data generated in step S1808 using the session ID transmitted by the distribution server 1604 in step S1806 as a key.
(S1810)
The calculation means of the voice synthesis server 1605 returns the generated voice data to the distribution server 1604 based on the instruction of the voice synthesis program 1610.
(S1811)
The distribution server 1604 distributes the voice data received from the voice synthesis server 1605 in step S1810 to the client terminal 1601.
At the time of distribution, processing such as dynamically increasing or decreasing the distribution bit rate according to the line speed of the network 1603 may be added.

According to the processing sequence shown in FIG. 18, when the delivery server 1604 receives a request from the client terminal 1601, the so-called “Pull-type delivery” in which voice data is extracted from the voice synthesis server 1605 and delivered to the client terminal 1601. Can be realized.
According to this distribution method, since it is easy to generate voice data dynamically according to the content of the request, for example, the input text used for voice synthesis is input by the user, and the content is voice-synthesized and distributed. It is also possible.
In order to realize this, the content of the input text may be transmitted by the client terminal 1601 in step S1805 in FIG. 18, and the content may be transmitted to the speech synthesis server 1605 in step S1806. In this case, it is not necessary to store the contents of the input text in the storage unit 1606 in advance.

FIG. 19 shows another configuration example of the Web page shown in FIG.
Since the voice data can be embedded with information indicating the suppression level, if the voice player has a function of displaying the suppression level, the user can set the suppression level of the voice data after the voice player is activated. Can know.
However, if the suppression level can be known before starting the audio player and playing the audio data, it turns out that the audio data cannot be viewed until the user acquires the audio data. It is possible to avoid the trouble of performing twice.
Therefore, if the suppression level is displayed at the link destination to the audio content as in the screen configuration example shown in FIG.

  The distribution server 1604 distributes the Web page displaying the suppression level as shown in FIG. 19 (1) to the client terminal 1601. Web page data to be distributed may be created in advance and stored in the distribution server 1604, or may be configured to be dynamically generated by reading the header portion of the audio data (see FIG. 7). Good.

As described above, according to the fifth embodiment,
Since the program for causing the calculation means to execute any of the flowcharts (FIGS. 3, 4, 8, 11, and 13) described in the first to fourth embodiments is configured,
The speech synthesis method represented by these flowcharts can be realized by software, and the software can be applied to various applications.

Further, according to the fifth embodiment,
A speech synthesis distribution system having a speech synthesis server and a distribution server for distributing speech data output by the speech synthesis server,
The speech synthesis server
It has a calculation means and a storage means for storing the speech synthesis program described above,
The storage means
A user table holding the identification information of the listener;
A table representing the relationship between the listener's identification information and the level of suppression allowed for that listener;
Since we stored a list that holds pairs of suppression words and suppression levels of the suppression words,
It is possible to provide a dynamic speech synthesis distribution service that starts speech synthesis at the stage of receiving a request and outputs the generated speech data.

In addition, the distribution server
Accepts a synthesized speech delivery request from a client and forwards the request to the speech synthesis server;
The computing means is
Upon receiving the request transferred from the distribution server, based on the instructions of the voice synthesis program, read the table and list stored in the storage means, perform the voice synthesis and return to the distribution server,
The distribution server
Since the returned audio data is delivered to the client,
When a request is received from the client terminal 1601, so-called “Pull type distribution” in which voice data is extracted from the voice synthesis server 1605 and distributed to the client terminal 1601 can be realized.
According to this distribution method, since it is easy to generate voice data dynamically according to the content of the request, for example, the input text used for voice synthesis is input by the user, and the content is voice-synthesized and distributed. It becomes possible.

In addition, the distribution server
A web page including link information to voice data output by the voice synthesis server is delivered to the client;
At the time of the delivery of the web page, information representing the suppression level of the linked voice data is delivered together.
Only after the user has acquired the voice data can it be avoided that it is found that the voice data cannot be viewed.

1 is a functional block diagram of a speech synthesizer according to Embodiment 1. FIG. The example of a structure of the suppression word list which the suppression word list memory | storage means 104 stores is shown. An overall operation flow of the speech synthesizer 100 of FIG. 1 will be described. 6 shows a modification of the flowchart shown in FIG. FIG. 3 is a functional block diagram of a speech synthesizer according to the second embodiment. The example of a structure of the suppression word list which the suppression word list memory | storage means 504 stores is shown. 6 shows an internal configuration image of synthesized audio data in the second embodiment. The overall operation flow of the speech synthesizer 500 in FIG. 5 will be described. FIG. 9 is a functional block diagram of a speech synthesizer according to a third embodiment. The structure and data example of the tolerance table stored in the suppression word list storage unit 904 are shown. An overall operation flow of the speech synthesizer 900 in FIG. 9 will be described. FIG. 10 is a functional block diagram of a speech synthesizer according to a fourth embodiment. The overall operation flow of the speech synthesizer 1200 of FIG. 12 will be described. In the image of the internal structure of the audio data shown in FIG. 7, the audio data embedded with audio watermark information will be described. An example of a Web page including synthesized audio content is shown as a screen image. A configuration of a voice synthesis distribution system for distribution to a client terminal will be described. The configuration and data example of the user table 1620 will be described. A detailed processing sequence when the user views a Web page or audio content in FIG. 16 will be described. Fig. 16 shows another configuration example of the Web page shown in Fig. 15.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Speech synthesizer, 101 Input means, 102 Authentication means, 103 Suppression word judgment means, 104 Suppression word list storage means, 105 Speech synthesizer, 110 Text to be read out, 111 Speech, 500 Speech synthesizer, 501 Input means, 504 Suppression Word list storage means, 505 speech synthesis means, 506 calculation means, 900 speech synthesizer, 901 input means, 902 authentication means, 904 suppression word list storage means, 905 speech synthesis means, 906 calculation means, 907 suppression degree level determination means, 1200 speech synthesizer, 1201 input unit, 1202 authentication unit, 1204 suppression word list storage unit, 1205 speech synthesis unit, 1206 calculation unit, 1207 suppression level determination unit, 1208 encryption unit, 1501 client terminal, 1502 Web page screen Configuration example 1 03 Voice player screen configuration example, 1601 client terminal, 1602 server center, 1603 network, 1604 distribution server, 1605 speech synthesis server, 1606 storage means, 1610 speech synthesis program, 1620 user table, 1621 suppression word list, 1622 tolerance table 1901 Client terminal, 1902 Web page screen configuration example 2, 1903 Audio player screen configuration example.

Claims (18)

When synthesizing and outputting input text as speech, if the text contains a phrase that should suppress speech output (hereinafter referred to as a suppression phrase), a method of outputting with restricted listening is used. There,
An authentication step for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Reading the list from storage means storing a list holding a list of suppression words;
A determination step of determining whether or not a suppression word is included in the input text;
A synthesis step of synthesizing speech data based on the input text only if the input text does not contain a suppression word or if the listener is a regular user;
A speech synthesis method characterized by comprising: When the input text is synthesized and output as speech, if the text contains suppression words, it is a method of outputting with the restriction of listening,
A step of reading the list from the storage means storing the list holding the set of the suppression word and the suppression level of the suppression word;
A calculation step for calculating a suppression level in the input text based on the list;
A synthesis step of synthesizing speech data including information representing a calculation result based on the input text and the result of the calculation step;
A speech synthesis method characterized by comprising: When the input text is synthesized and output as speech, if the text contains suppression words, it is a method of outputting with the restriction of listening,
An authentication step for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
A step of reading the list from the storage means storing the list holding the set of the suppression word and the suppression level of the suppression word;
A calculation step for calculating a suppression level in the input text;
Determining whether or not the suppression level calculated in the calculating step is within an allowable range for a listener;
Synthesis that synthesizes speech data including information representing the calculation result based on the input text only when the listener is a regular user and the suppression level is within the allowable range for the listener. Steps,
A speech synthesis method characterized by comprising: The storage means
Contains a table representing the relationship between the listener's identification information and the level of suppression allowed for that listener,
4. The speech synthesis method according to claim 3, wherein it is determined whether or not the suppression level is within a range allowed for the listener based on the contents of the table. In the calculating step, a suppression level is calculated for each predetermined block of input text,
5. The speech synthesis method according to claim 2, wherein in the synthesis step, speech data is synthesized for each predetermined block including information indicating a suppression level.   6. The speech synthesis method according to claim 1, further comprising an encryption step of generating an encryption key based on the identification information and encrypting the speech data synthesized in the synthesis step. .   7. The voice data is synthesized by including the watermark data for identifying the voice data synthesis source for each predetermined block of the input text in the synthesis step. The speech synthesis method described in 1. When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Authentication means for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Storage means for storing a list for holding a list of suppression words;
A determination unit that reads a list holding a list of suppression words from the storage unit, and determines whether or not the suppression word is included in the input text; and
A synthesis means for synthesizing speech data based on the input text only when the input text does not contain a suppression word or when the listener is a regular user;
A speech synthesizer characterized by comprising: When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Storage means for storing a list that holds a combination of a suppression word and a suppression level of the suppression word;
From the storage means, a calculation means for reading a list holding a set of suppression words and suppression levels of the suppression words, and calculating a suppression level in the input text;
A synthesizing unit that synthesizes voice data including information representing the calculation result based on the input text and the processing result of the calculating unit;
A speech synthesizer characterized by comprising: When the input text is synthesized and output as speech, if the text contains a suppression phrase, it is a device that outputs with restriction of listening,
Authentication means for confirming whether or not the listener is an authorized user based on the identification information of the person who listens to the output sound;
Storage means for storing a list that holds a combination of a suppression word and a suppression level of the suppression word;
From the storage means, a calculation means for reading a list holding a set of suppression words and suppression levels of the suppression words, and calculating a suppression level in the input text;
A degree-of-suppression level judgment means for judging whether or not the degree-of-suppression level calculated by the calculating means is within an allowable range for the listener;
Synthesis that synthesizes speech data including information representing the calculation result based on the input text only when the listener is a regular user and the suppression level is within the allowable range for the listener. Means,
A speech synthesizer characterized by comprising: The storage means
Contains a table representing the relationship between the listener's identification information and the level of suppression allowed for that listener,
The suppression level determination means includes
11. The speech synthesizer according to claim 10, wherein it is determined whether or not the suppression level is within an allowable range for the listener based on the contents of the table. The calculation means calculates a suppression level for each predetermined block of input text,
12. The speech synthesizer according to claim 9, wherein the synthesis unit synthesizes speech data including information indicating a suppression level for each predetermined block.   13. The speech synthesizer according to claim 8, further comprising an encryption unit that generates an encryption key based on the identification information and encrypts the speech data synthesized by the synthesis unit. .   14. The voice data according to any one of claims 8 to 13, wherein the synthesizing unit synthesizes voice data including watermark data for identifying a voice data synthesis source for each predetermined block of input text. The speech synthesizer described.   A speech synthesis program for causing a computing means to execute the speech synthesis method according to claim 1. A speech synthesis distribution system having a speech synthesis server and a distribution server for distributing speech data output by the speech synthesis server,
The speech synthesis server
Calculation means, and storage means for storing the speech synthesis program according to claim 15;
The storage means
A user table holding the identification information of the listener;
A table representing the relationship between the listener's identification information and the level of suppression allowed for that listener;
A speech synthesis and distribution system characterized by storing a list that holds a combination of a suppression word and a suppression level of the suppression word. The distribution server
Accepts a synthesized speech delivery request from a client and forwards the request to the speech synthesis server;
The computing means is
Upon receiving the request transferred from the distribution server, based on the instructions of the voice synthesis program, read the table and list stored in the storage means, perform the voice synthesis and return to the distribution server,
The distribution server
The speech synthesis delivery system according to claim 16, wherein the returned speech data is delivered to the client. The distribution server
A web page including link information to voice data output by the voice synthesis server is delivered to the client;
18. The voice synthesizing and distributing system according to claim 17, wherein when the Web page is distributed, information indicating a suppression level of the linked voice data is also distributed.

Images