JP2002366175A - Device and method for supporting voice communication - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice communication support device and its method which can reduce apparently improve a voiceless state and reduce the stress to an opposite side by sequentially outputting a synthesized voice according to user's sentence input operation and automatically outputting a redundant word voice when the voiceless state continues. SOLUTION: This device is equipped with a character output part which temporarily stores the character string of a sentence inputted through sentence input operation and outputs it in response to output operation, a voice synthesis part which outputs synthesized voice data from the output character string, a redundant word storage part which previously stores a proper number of pieces of redundant word voice data, and an output control part which takes an output voice out of the synthesized voice data and outputs selected redundant voice data when no synthesized voice data are not inputted for a certain time.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice communication support apparatus for voice-synthesizing and outputting an input sentence in order to convey the intention of a language-disabled person. TECHNICAL FIELD The present invention relates to a voice communication support device and a method thereof capable of reducing the stress of a communication partner that occurs during communication.

[0002]

2. Description of the Related Art There is a method in which a speech-disabled person does not use speech such as facsimile or e-mail when a partner is near, and communicates with another person when the partner is near. However, none of these methods can alert like a voice or communicate to many people at the same time. Further, in order to communicate by facsimile or e-mail, the other party also needs to have a facsimile or mail terminal, and their penetration rate is lower than that of a telephone. Therefore, the scenes and the opponents who can communicate by the above method are limited.

As a means for solving such a problem, a voice communication support device has been developed in which an input device such as a keyboard and a voice synthesizer are combined. As an example of the voice communication support device, there is a conversation support device in which a deaf person performs character input and voice synthesis output by key operation.
No. 08 is also described.

FIG. 18 is a block diagram showing the configuration of a conventional voice communication support apparatus, showing parts related to a user's input operation and voice synthesis. In the figure, 1 is a keyboard, 2 is a character key, 3 is a function key, 4
Is a text input processing circuit, and 5 is a speech synthesis circuit. The keyboard 1 includes character keys 2 for inputting characters and function keys 3 for giving an utterance command or the like, and inputs a conversation character string corresponding to a key operated by the user. Sentence input processing circuit 4
Generates a character code string corresponding to the character key 2 operated by the keyboard 1, and generates a voice synthesizing circuit when a key corresponding to a voice command (hereinafter referred to as a voice key) among the function keys 3 is operated. 5 is output. Voice synthesis circuit 5
Synthesizes and outputs speech based on the input character code string.

[0005]

Since the conventional voice communication supporting device is configured as described above,
The silent state continues until the user completes the input of the conversation sentence and operates the utterance key. For this reason, there is a problem that if it takes a long time to input a conversational sentence, a silence time becomes long, and a great stress is applied to the other party. In particular, when communicating with a distant partner by combining such a conventional device with a telephone, since the situation of the user cannot be seen from the partner, the influence on the partner as described above is serious. There were issues such as becoming.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and outputs synthesized speech sequentially according to a user's sentence input operation and automatically outputs redundant word speech when silence continues. By doing so, it is an object of the present invention to obtain a voice communication support device and a method thereof capable of apparently improving a silent state and reducing stress of the other party.

[0007]

A voice communication support apparatus according to the present invention is a voice communication support apparatus for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, wherein the sentence is input by a user's sentence input operation. An input unit, a character string output unit that temporarily stores a character string of an input sentence, and outputs the character string when an output operation is performed by a user; and a speech synthesis unit that synthesizes a character string output from the character string output unit. A speech synthesizer for outputting synthesized speech data, a redundant word storage for storing an appropriate number of predetermined redundant word speech data in advance, and selecting and outputting one redundant word speech data from the redundant word storage. When the redundant word selection unit reads the synthesized voice data, extracts the output voice, monitors the input state of the synthesized voice data, and detects that the synthesized voice data is not input for a certain period of time. Is obtained by an output control section for the redundant word selection unit outputs a redundant word speech from the selected redundant word speech data.

A voice communication support device according to the present invention analyzes a character string of a sentence input from a sentence input unit instead of a character string output unit, extracts synthesis units, and sequentially extracts the extracted synthesis units. It is provided with a combining unit extracting unit for outputting.

A voice communication support device according to the present invention is a voice communication support device for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, and a sentence input unit for inputting a sentence by a user's sentence input operation; A character string of a sentence input from the input unit is analyzed to extract a synthesis unit, and a language category of the synthesis unit is determined.
A synthesis unit for sequentially outputting a synthesis unit and a language category; a voice synthesis unit for voice-synthesizing a character string of the synthesis unit output from the synthesis unit analysis unit to output synthesized voice data; A redundant word storage unit for storing redundant word voice data in advance, a redundant word selection unit for selecting and outputting one redundant word voice data from the redundant word storage unit, and a redundant word corresponding to the input language category. A timing determination unit that outputs a timing signal for determining the output timing of the word voice data, reads out the synthesized voice data, extracts the output voice, monitors the input state of the synthesized voice data, and receives the timing signal during the sentence input operation. In such a case, there is provided a timing output control section for outputting a redundant word voice based on the redundant word voice data selected by the redundant word selecting section.

In the voice communication support apparatus according to the present invention, instead of the timing determining section, a time length-specific timing signal for determining the output timing of the redundant word voice data generated according to the time length of the immediately preceding synthesized voice data is provided. A timing determining unit for outputting by time length is provided, and a timing output control unit outputs redundant word voice data by a timing signal by time length.

The voice communication support apparatus according to the present invention reads out synthesized voice data instead of the output control unit, extracts the output voice, monitors the input state of the synthesized voice data, and monitors the previous synthesized voice data during the sentence input operation. If the next synthesized voice data is not input within a predetermined time after the input, the preparation for outputting the redundant word voice data is performed, and when the next synthesized voice data is input, the output voice is output. And a redundant word prefix output control unit for reading and connecting the redundant word voice data selected by the redundant word selecting unit before the next synthesized voice data.

A voice communication support device according to the present invention is a voice communication support device for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, and a sentence input unit for inputting a sentence by a user's sentence input operation; A character string of a sentence input from the input unit is analyzed to extract a synthesis unit, and a language category of the synthesis unit is determined.
A synthesis unit that sequentially outputs a synthesis unit and a language category, a speech synthesis unit that synthesizes a character string of the synthesis unit output from the synthesis unit analysis and outputs synthesized speech data, and a language of the synthesis unit A category redundant word storage unit for storing redundant word voice data in advance for each category, and a category for selecting and outputting one corresponding redundant word voice data from the category redundant word storage unit according to the language category from the synthesis unit analysis unit Another redundant word selection unit, reads out the synthesized voice data, extracts the output voice, monitors the input state of the synthesized voice data, and monitors the input state of the next synthesized voice data during the sentence input operation within a predetermined time after the input of the previous synthesized voice data. When the synthesized speech data is not input, the preparation for outputting the redundant word speech is performed, and the next synthesized speech is output as the next synthesized speech data is input. It is obtained by a redundant word before 置出 force control unit that connects reads the redundant word speech data corresponding to the next synthesized speech category redundant word selection unit selects the previous data.

The speech communication support apparatus according to the present invention includes a redundant word storage unit for each tail vowel which stores redundant word voice data which can be output for each tail vowel of the synthesis unit in advance, instead of the redundant word storage unit. Instead of the word selection unit, the type of the tail vowel of the input synthesis unit is identified, and one redundant word voice data that matches the type of the tail vowel is selected from the tail vowel-based redundant word storage unit. It has a word selection unit.

The speech communication support apparatus according to the present invention includes a redundant word storage unit for each time length for storing redundant word voice data that can be output for each time length of the synthesis unit, instead of the redundant word storage unit. Instead of the selecting unit, a redundant word selecting unit for each time length for selecting one redundant word voice data matching the time length of the input synthesized voice data from the redundant word storing unit for each time length is provided.

In the voice communication support apparatus according to the present invention, instead of the output control unit, the synthesized voice data is read in, the output voice is extracted, the output synthesized voice data is temporarily stored, and the input state of the synthesized voice data is monitored. If it is detected that the synthesized speech data has not been input for a certain period of time, the redundant word speech is output from the redundant word speech data selected by the redundant word selection unit, or the stored synthesized speech data is output again. It has a repeat output control unit.

According to the voice communication support method of the present invention, in a voice communication support apparatus for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, a sentence is input by a user's sentence input operation, and the input sentence is converted into a voice. Synthesize to generate synthesized voice data, store an appropriate number of predetermined redundant word voice data in advance, take out the output voice by the generated synthesized voice data, monitor the input state of the synthesized voice data, and perform input operation. When it is detected that the synthesized speech data has not been input for a certain period of time, the stored redundant word speech data is selected and the redundant word speech is output.

[0017]

Embodiments of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is a block diagram showing the configuration of a voice communication support apparatus according to Embodiment 1 of the present invention. In the figure, 11 is a sentence input unit, 12 is a character string output unit, 13 is a speech synthesis unit, and 14 is a redundant word storage. , 15 is a redundant word selection unit, and 16 is an output control unit. Next, the operation will be described. The sentence input unit 11 inputs a sentence by a user's sentence input operation. The sentence input operation in this case is, for example, a kana character input operation from a keyboard and a kana-kanji conversion operation. Therefore, for example, if the user performs kana-kanji conversion for each phrase, the output of the sentence input unit 11 is a phrase character string mixed with kana-kanji characters. It is assumed that the dictionary required for the kana-kanji conversion is included in the sentence input unit 11. The character string output unit 12 temporarily stores a character string input from the sentence input unit 11 and outputs it when a user performs an output operation for voice conversion. The speech synthesis unit 13 synthesizes a voice of the character string input from the character string output unit 12 by, for example, text-to-speech conversion, and outputs the synthesized voice data. The redundant word storage unit 14 stores in advance a plurality of redundant word voice data of redundant words such as “er” and “a”. The redundant word selecting unit 15 selects one piece of voice data from the redundant word storage unit 14, for example, randomly. The output control unit 16 reads the synthesized voice data, extracts the output voice, monitors the input state of the synthesized voice data, and monitors the input state of the synthesized voice data. A redundant word voice is output from the word voice data.

FIG. 10 is an explanatory diagram showing a specific example of synthesized speech and redundant word speech output according to the first embodiment, showing the user's operation, the output of speech, and the timing of the operation of the output control unit 16. The horizontal axis represents time. Suppose the sentence is "Please give me one curry, one mixed fry set meal and one fried rice." The user inputs kana characters from the keyboard and performs kana-kanji conversion. When the input and conversion of the character string to be output at once are completed, an output operation such as pressing a return key is performed.
In the sentence input operation in the figure, a white portion indicates a time zone of Kana character input, and a shaded portion indicates a time zone of Kana-Kanji conversion. The arrow of the output operation indicates the timing of pressing the return key.

The output control unit 16 starts monitoring the synthesized speech data input from the speech synthesis unit 13 at the same time when the user starts a sentence input operation. The output control unit 16 constantly measures the time when there is no input of synthesized speech data, and when this time reaches a certain time (hereinafter referred to as a silence allowable time), reads out the redundant word speech data selected by the redundant word selection unit 15, and The redundant word voice is output. In this example, the silence allowable time is 4
Seconds. Next, when an output operation is performed about 3.5 seconds after the start of the sentence input operation, the character string “one curry” is input to the voice synthesis unit 13 and the synthesized voice data is output. The time from the start of the sentence input operation to the output operation is the silence allowable time (4
Second), the redundant word voice data is not read out, and the output control unit 16 outputs the synthesized voice “one curry”.

During the above-mentioned audio output, the user continues to input the next character string “one mixed fry set meal”. Here, it takes time to enter this next string,
It is assumed that the output operation has not been performed even if 4 seconds have elapsed since the end of the immediately preceding audio output. At this time, the output control unit 16 causes the redundant word selecting unit 15 to randomly select one redundant word voice data from the redundant word storage unit 14 and outputs the voice data of the selected redundant word “Eto” from the redundant word selecting unit 15. Read and output audio. Thereafter, when an output operation is performed, a synthesized voice “one mix fly set meal” is output as a voice.

Further, during the above-described voice output, the user continues to input the next character string "Please give me one fried rice." If it is less than 4 seconds of the silence allowable time from the previous synthesized voice output to the output operation of "Please give me one fried rice", the redundant word will not be read out, and the output control unit will make the synthesized voice "Please give me one fried rice." Is output as audio. By the above operation, the entire audio output is as shown in FIG.
"One curry, um, one mixed fry set meal,
Please give me one fried rice "and the silence will not continue for more than 4 seconds. If redundant word output is not made, "one curry" and "mix fry set meal 1"
A silent state of about 6 seconds occurs between the two.

As described above, according to the first embodiment, a synthesized speech can be output in accordance with the sentence input operation of the user, and when the silent state continues for a certain period of time during the sentence input operation, the redundant word sound is automatically output. By outputting and interpolating and apparently reducing the silence time generated by the user's sentence input operation, the effect of reducing the stress felt by the other party communicating with the user can be obtained.

Embodiment 2 FIG. FIG. 2 is a block diagram showing a configuration of a voice communication support device according to a second embodiment of the present invention. In the drawing, elements denoted by the same reference numerals as those in FIG. 1 are the same as those in the first embodiment. Omitted. In the different part, reference numeral 21 denotes a combination unit extraction unit, which is replaced by the character string output unit 12 in FIG. Next, the operation will be described. Synthesis unit extraction unit 21
Analyzes a character string of a sentence input from the sentence input unit 11, extracts a character string that is collectively output during speech synthesis, that is, a synthesis unit, and sequentially outputs the extracted synthesis units to the speech synthesis unit 13. To operate. Here, the composition unit is extracted by regarding punctuation marks, question marks, exclamation marks, and the like as boundaries of the composition unit.

FIG. 11 is an explanatory diagram showing a specific example of a synthesized speech and a redundant word speech output according to the second embodiment, showing timings of user operations, speech output, and operations of the output control unit 16. The sentence example to be input is the same as in the first embodiment. The user inputs kana characters from the keyboard and performs kana-kanji conversion. The determined kana-kanji character strings are sequentially input to the synthesis unit extraction unit 21. Composition unit extraction unit 2
1 analyzes an input character string and, for example, when detecting a punctuation mark, a question mark, or an exclamation mark in the character string, regards the character string as a boundary of a synthesis unit and sequentially outputs the input character string to the speech synthesis unit. . Here, three character strings of "one curry,""one mixed fry set meal," and "one fried rice please" are sequentially input by the kana-kanji conversion operation. The combining unit extraction unit 21 regards the ending punctuation mark of "one curry,""one mixed fry set meal," and the ending point of "one fried rice please." The same three as the column, "one curry,""one mixed fry set meal," and "Please give me one fried rice." Are sequentially output to the speech synthesizer 13 as synthesis units. Therefore, since the synthesis unit is automatically output when extracted, the user can omit the output operation performed in the first embodiment. The output control unit 16
Although the input state of the synthesis unit is monitored, the subsequent operation, that is, the operation of inserting a redundant word voice into a silent state is the same as in the first embodiment, and a description thereof will be omitted.

As described above, according to the second embodiment, the user's operation can be reduced by outputting the synthesized speech sequentially for each synthesis unit in accordance with the user's sentence input operation, and the silent state can be reduced. By continuously outputting and interpolating redundant word voices for a certain period of time, the silence time associated with the user's sentence input operation is apparently shortened, thereby reducing the stress felt by the other party communicating with the user. .

Embodiment 3 FIG. 3 is a block diagram showing a configuration of a voice communication support device according to a third embodiment of the present invention. In the drawing, elements denoted by the same reference numerals as those in FIG. 1 are the same as those in the first embodiment. Omitted. 1, reference numeral 31 denotes a synthesis unit analysis unit, 32 denotes a timing determination unit, and 33 denotes a timing output control unit.

Next, the operation will be described. The synthesis unit analysis unit 31 analyzes a character string input from the sentence input unit 11, sequentially extracts and outputs synthesis units, and determines and outputs a language category of the synthesis unit. This composite unit is
For example, it is determined by regarding punctuation marks, question marks, exclamation marks, and the like as boundaries of the synthesis unit, as in the second embodiment.
Further, the language category is, for example, five types such as continuous modification, continuous modification, parallel, end of sentence, and other five types, and is determined by applying a rule based on the use of particles or words at the end of the synthesis unit. For example, if the end of the composition unit is the adjunct form of the adjunct particle "no" or a verb, the language category is adnominal modification.

The timing determining section 32 outputs a timing signal generated according to the input language category. This timing signal determines the output timing of the redundant word voice data. The output timing is the time length from the end of the immediately preceding synthesized voice output, that is, the silence allowable time, and its value is determined according to the input language category.
For example, the determination is made with reference to a table as shown in FIG. The timing determination unit 32 outputs a timing signal to the timing output control unit 33 every time the time specified by the output timing elapses after the immediately preceding output of the synthesized speech. Note that the output timing when no language category is input is 7 seconds corresponding to the language category "end of sentence" in the table of FIG. If the language category of the immediately preceding synthesis unit is adjoint modification or parallel, it can be inferred that there is a subsequent synthesis unit, and thus it is considered that if the silent state continues for a long time, the partner will be greatly stressed.
For this reason, the silence permissible time corresponding to continuous modification or parallel is set short.

The timing output control unit 33 reads the synthesized voice data to extract the output voice, and when there is no input of the synthesized voice data until the timing signal is input from the timing determination unit 32 during the sentence input operation. The redundant word voice is output from the redundant word voice data selected by the redundant word selection unit 15. For example, in the above example, if the language category of the immediately preceding synthesis unit is “end of sentence”, the redundant word data is read out and output as speech if no speech input is performed even after 7 seconds have elapsed since the completion of the output of the synthesized speech.

A specific example of determining the redundant word output timing will be described below. Assume that the input sentence is "Please give me one set of low skatatsu and one zaru soba. How much is it?" The composition unit is three, "One set of roast cutlet", "Please give me one zaru soba", and "How much is it?". Each language category is "Parallel", "End of sentence",
"End of sentence". The timing determination unit 32 determines the permissible silence time for each category from the end of the speech synthesis to the output of the redundant word voice according to the table of FIG. 12, as shown in the explanatory diagram of FIG. Therefore, the timing output control unit 33 monitors the input state of the synthesized voice data according to the silence allowable time, and the redundant word selecting unit 15 selects the non-input state of the synthesized voice data exceeding each silence allowable time. The redundant word voice data can be read and output as redundant word voice.

Here, the output timing of the redundant word voice is
We mentioned that we decided only on language categories,
Alternatively, it may be determined by two of the language category and the synthetic speech time length. For example, the allowable silence time length for each language category shown in FIG. 12 is determined by further changing the time length of the synthesized speech by using an equation.

As described above, according to the third embodiment, by sequentially outputting synthesized speech for each synthesis unit in accordance with the user's sentence input operation, the operation of the user can be reduced, and the silent state can be reduced. When continued, redundant word speech is automatically output and interpolated at the output timing according to the language category of the synthesis unit that was input immediately before, and the silence time accompanying the user's sentence input operation is apparently shortened. The effect of reducing the stress felt by the partner communicating with the user can be obtained.

Embodiment 4 FIG. FIG. 4 is a block diagram showing a configuration of a voice communication support device according to a fourth embodiment of the present invention. Elements denoted by the same reference numerals as in FIGS. 1 to 3 are the same as those in the respective embodiments. Description is omitted. In the different part, reference numeral 41 denotes a time length-based timing determination unit.

Next, the operation will be described. The time length-specific timing determination unit 41 outputs a time length-specific timing signal generated according to the time length of the synthesized voice data. The timing signal for each time length is generated, for example, when the silence time following the synthesized speech data in the synthesis unit which is the output of the speech synthesis unit 13 reaches the silence allowable time calculated using the time length of the immediately preceding synthesized speech data as a parameter. You. That is, the timing signal for each time length determines the output timing of the redundant word voice data. Therefore, when the timing signal for each time length is input during the sentence input operation, that is, when the time exceeds the silence allowable time during which the input of the next synthesized voice data is not performed, the redundant word selecting unit 33 15 reads out the selected redundant word voice data and interpolates a silent portion exceeding the allowable silent time as a redundant word voice. Note that the time length-specific timing determination unit 41 determines the output timing of the redundant word voice based on the calculated silence permissible time. Instead, the time length range of the synthesized voice data and the time length from the end of the immediately preceding voice output are determined. May be determined with reference to a table stored in advance.

As described above, according to the fourth embodiment, by sequentially outputting synthesized speech for each synthesis unit in accordance with the user's sentence input operation, the operation of the user can be reduced, and the silent state can be reduced. If it continues, the redundant word voice is automatically output and interpolated at the output timing determined according to the time length of the synthesis unit input immediately before, and the silence time associated with the user's sentence input operation is apparently shortened By doing so, the effect of reducing the stress felt by the other party communicating with the user can be obtained.

Embodiment 5 FIG. 5 is a block diagram showing a configuration of a voice communication support device according to a fifth embodiment of the present invention. In the figure, elements denoted by the same reference numerals as those in FIGS. 1 and 2 are the same as those in the respective embodiments. , And the description is omitted. In a different part, 51 is a redundant word prefix output control unit. Next, the operation will be described. The redundant word prefix output control unit 51 reads the synthesized voice data, extracts the output voice, monitors the input state of the synthesized voice data, and monitors the input state of the synthesized voice data within a predetermined time after the input of the previous synthesized voice data during the sentence input operation. When the next synthesized voice data is not input, preparations are made to output redundant word voice data, and when the next synthesized voice data is input, output is performed before the next synthesized voice data as the output voice. The redundant word voice data selected by the redundant word selection unit 15 is read and connected.

FIG. 14 is an explanatory diagram showing a specific example of synthesized speech and redundant word speech output according to the fifth embodiment, showing the timing of user operation, speech output, and operation of the redundant word prefix output control unit 51. Things. The sentence to be input is the same as in the first embodiment, "Please give me one curry, one mixed fry set meal, and one fried rice." When a predetermined time, for example, 4 seconds, has elapsed after the speech synthesis output of “one curry” has been completed, it is determined that a redundant word is to be output before the next synthesized speech output,
For example, a redundant word output flag is set. At this point, redundant word speech is not yet output. When the next synthesized voice data “one mixed fly set meal” is input, the redundant word output flag is set, so the redundant word prefix output control unit 5
1 reads the redundant word voice data from the redundant word selection unit 15. As a result, for example, the redundant word “erto” is read. The redundant word prefix output control unit 51 connects the redundant word voice before the next synthesized voice, and outputs “Um, mixed fly set meal 1”.
Is output as a voice. As a result of this operation, the time between “one mixed fry set meal” and “one fried rice please” is reduced.

As described above, according to the fifth embodiment, the user operation can be reduced by outputting synthesized speech sequentially for each synthesis unit in accordance with the user's sentence input operation, and the silence state can be reduced. If the speech continues for a certain period of time, the redundant speech is automatically output immediately before the next synthesized speech and interpolated, realizing the natural speech feature that the redundant speech is placed immediately before a meaningful utterance unit. By reducing the silence time associated with the input operation on average, the effect of reducing the stress felt by the partner communicating with the user can be obtained.

Embodiment 6 FIG. FIG. 6 is a block diagram showing a configuration of a voice communication support device according to a sixth embodiment of the present invention. Elements denoted by the same reference numerals as in FIGS. 1, 3 and 5 are equivalent to those of the respective embodiments. Yes, and the description is omitted. In different parts, 61
Is a category-based redundant word storage unit, and 62 is a category-based redundant word selection unit.

Next, the operation will be described. In the sixth embodiment, it is assumed that the language category of the synthesis unit determined by the synthesis unit analysis unit 31 is composed of three types: a question sentence end, a sentence other than a question sentence, and a sentence end. The category-specific redundant word storage unit 61 stores in advance redundant word speech data that can be connected for each language category of the synthesis unit, for example, as shown in the table of FIG. The category-based redundant word selection unit 62 identifies the language category of the synthesis unit from the synthesis unit analysis unit 31 and selects one corresponding redundant word voice data from the category-based redundant word storage unit 61. In this case, the redundant word prefix output control unit 51 outputs the redundant word if the next synthesized voice data is not input within a predetermined time after the input of the previous synthesized voice data during the sentence input operation. Prepare to output voice, and as the output voice when the next synthesized voice data is input, correspond to the next synthesized voice selected by the category redundant word selection unit before the next synthesized voice data The redundant word voice data is read and connected.

Here, a specific example of determining a redundant word will be described below. The input sentence is the same as in Embodiment 3, "Please give me one set of roasted cutlet and one zaru soba. How much is it?"
And The redundant words by category that can be selected when the language category is input from the synthesis unit analysis unit 31 to the redundant word by word selection unit 62 are as shown in FIG. 16 with reference to the table of FIG. Therefore, for example, the redundant word "um" is not output immediately before "how much". The category-based redundant word selection unit 62 selects one of the category-based redundant words that can be selected as shown in FIG.
For example, it will be selected at random.

As described above, according to the sixth embodiment, by sequentially outputting synthesized speech for each synthesis unit in accordance with the user's sentence input operation, the operation of the user can be reduced, and the silent state can be reduced. After a certain period of time, the redundant word speech judged to be appropriate from the language category of the succeeding synthetic unit is connected before the next synthetic speech and automatically output and interpolated, and the redundant word By realizing the characteristic of the natural voice that is placed immediately before, and by shortening the silence time associated with the user's sentence input operation on the average, the effect of reducing the stress felt by the partner communicating with the user can be obtained.

Embodiment 7 FIG. FIG. 7 is a block diagram illustrating a configuration of a voice communication support device according to a seventh embodiment of the present invention. In the drawing, elements denoted by the same reference numerals as those in FIG. 2 are the same as those in the second embodiment. The description is omitted. In different parts, reference numeral 71 denotes a tail vowel-based redundant word storage unit, and reference numeral 72 denotes a tail vowel-based redundant word selection unit.

Next, the operation will be described. The tail-vowel-based redundant word storage unit 71 stores in advance a tail vowel-based redundant word that can be output for each tail vowel of the synthesis unit. An example of the correspondence between the tail vowel of this synthesis unit and the redundant word is shown in the table of FIG.
For example, if the last vowel is / u /, the redundant word "um"
"Hmm" can be output. Here, a redundant word that can be output is set so that the vowel at the beginning of the redundant word and the vowel at the end of the synthesis unit are as similar as possible. When the synthesis unit is input from the synthesis unit extraction unit 21, the tail vowel-specific redundant word selection unit 72 checks the type of the vowel at the end of the unit, and finds a redundant word that matches the vowel type and stores a redundant word that matches the vowel type. From 1
Choose one. For example, if the synthesis unit is “one curry”, the tail vowel is / u /, so the tail word-by-vowel redundant word selection unit 72 selects one from the redundant words “um”, “um”, etc.
One is selected at random, for example. Therefore, the output control unit 16 monitors the input state of the synthesized voice data according to the silence allowable time, and when the non-input state of the synthetic voice data continues beyond the silence allowable time, the tail word-based vowel-based redundant word selection unit 72. Automatically outputs the selected redundant word voice data to form a voice output of a sentence with a good flow.

As described above, according to the seventh embodiment, the user operation can be reduced by sequentially outputting the synthesized speech for each synthesis unit in accordance with the user's sentence input operation, and the silence state can be reduced. When a certain period of time continues, redundant words voice data that matches the last vowel of the synthesis unit is automatically output and interpolated, and the silence time associated with the user's sentence input operation is apparently shortened, so that the other party communicating with the user can An effect that can reduce feeling stress can be obtained. Note that the tail word-by-vowel redundant word storage unit 71 and the tail-vowel-by-tail redundant word selection unit 72 can be applied to the arrangements of FIGS.

Embodiment 8 FIG. FIG. 8 is a block diagram showing a configuration of a voice communication support device according to an eighth embodiment of the present invention. In the figure, elements denoted by the same reference numerals as those in FIG. 2 are the same as those in the second embodiment. The description is omitted. In different parts, reference numeral 81 denotes a time-based redundant word storage unit, and reference numeral 82 denotes a time-based redundant word selection unit.

Next, the operation will be described. The redundant word storage unit 81 for each time length stores in advance redundant word voice data that can be output for each time length of the synthesis unit. For example, redundant word voice data in which the time length of the long sound of the redundant word “er” is changed to several types is prepared, and the time length of a long synthesis unit is associated with the redundant word voice data having a long time length of the long sound. When the synthesized speech data is input, the redundant word selecting section 82 checks the time length and selects one redundant word from the redundant word storing section 81 according to the time length of the synthesis unit. If the time length of the immediately preceding synthesis unit is long, redundant word voice data having a long sound length is selected. Therefore, the output control unit 16 monitors the input state of the synthesized speech data according to the permissible silence time, and when the non-input state of the synthesized speech data continues beyond the permissible silence time, the redundant word selection unit 82 for each time length. Automatically outputs the selected redundant word voice data to form an output voice of a sentence with a good flow.

As described above, according to the eighth embodiment, the user operation can be reduced by outputting synthesized speech sequentially for each synthesis unit in accordance with the user's sentence input operation, and the silence state can be reduced. When a certain period of time continues, redundant words voices that match the time length of the synthesis unit are automatically output and interpolated, and the silence time associated with the user's sentence input operation is apparently shortened, so that the person who communicates with the user feels The effect of reducing stress can be obtained. The redundant word storage unit 81 according to time length and the redundant word selection unit 82 according to time length can be replaced and applied to each of the arrangements of FIGS.

Embodiment 9 FIG. FIG. 9 is a block diagram showing a configuration of a voice communication support device according to a ninth embodiment of the present invention. In the figure, elements denoted by the same reference numerals as those in FIGS. 1 and 2 are the same as those in the respective embodiments. Yes, and the description is omitted. In different parts,
Reference numeral 91 denotes a repetition output control unit.

Next, the operation will be described. The repetition output control unit 91 reads the synthesized voice data, extracts the output voice, temporarily stores the output synthesized voice data, monitors the input state of the synthesized voice data, and checks that the synthesized voice data is not input for a certain period of time. Is detected,
It reads out redundant word voice data from the redundant word selection unit 15 and outputs it as voice, or outputs stored synthetic voice data again. Which of the redundant word voice data and the stored synthesized voice data is output as speech is determined, for example, by the number of times the redundant word voice data or the stored synthesized voice data has been output since the immediately preceding voice synthesized output. For example, immediately after the output of the synthesized voice, the stored synthesized voice is repeated, and if a silent state continues thereafter, a redundant word voice is output. Therefore, as a whole, an audio output of a sentence with a good flow is formed.

Here, when determining which of the redundant word voice data and the stored synthesized voice is to be output,
It has been described that it is determined by the number of times that the redundant word voice data or the stored synthesized voice has been output since the immediately preceding voice synthesis output, but this is instead determined by the time length of the synthesized voice. Is also good. For example, repetition is performed only when the time length of the synthesized speech is equal to or less than a certain value, and otherwise, a redundant word is output. This can also be determined by the language category of the composition unit. For example, if the synthesis unit contains a number or proper noun, the repetition is performed, and otherwise, redundant word voice data is output.

As described above, according to the ninth embodiment, the user operation can be reduced by sequentially outputting synthesized speech for each synthesis unit according to the user's sentence input operation, and the silent state can be reduced. After a certain period of time, the synthesized speech output immediately before is automatically re-output and interpolated, and the silence time associated with the user's sentence input operation is apparently shortened, thereby reducing the stress felt by the person communicating with the user. The effect that can be obtained is obtained. The repetition output control section 91 can be adapted by replacing the arrangement in each of FIGS.

While the embodiments of the present invention have been described above, these embodiments can cope with some variations as described below. In the first to ninth embodiments, the character input operation using a keyboard and the Kana-Kanji conversion operation have been described for the sentence input operation. Kana-kanji conversion operation, selection of a sentence or phrase stored in advance, selection of a sentence template which is a sentence example in which a word can be filled in, and word input operation can also be performed.

In the second to ninth embodiments,
In particular, when the sentence input operation is the selection of a sentence template and the word input operation, the determination of the synthesis unit can be performed by previously describing the synthesis unit boundary in the sentence template.

In the first to ninth embodiments,
Regarding selection of redundant word data, language category, time length,
It has been described that one word is randomly selected from redundant words that meet conditions such as a tail vowel, but this is done by registering in advance the redundant words used by the user and selecting from them unless they violate the conditions. You can choose to do so.

In the third and sixth embodiments, the linguistic category of the composition unit is, for example, the part of speech at the end or the beginning of the composition unit, the type of particle, the inflected form of the word, the
Categories can be subdivided according to the type of sentence such as question or negation. In addition, the language category may be described in advance in the synthesis unit in the sentence template, particularly when the sentence input operation is the selection of a sentence template and the word input operation, and the synthesis unit boundary is described in the sentence template. it can. In this case, since it is not necessary to extract and analyze the synthesis unit, a configuration including the character string output unit 12 described in the first embodiment instead of the synthesis unit analysis unit 31 can be adopted.

In the third and fourth embodiments, the timing determining section 32 or the time length-based timing determining section 41 determines the time length from the immediately preceding voice output, which is the output timing of the redundant word voice output, to the language category. Or it was a fixed value determined by the last synthesized speech time length,
Instead, it may be output as a random time length having a certain time width based on the certain value.

In the seventh to ninth embodiments,
By providing a unit corresponding to the synthesis unit analysis unit 31 and the timing determination unit 32 described in the third embodiment in place of the synthesis unit extraction unit 21, the redundant word voice data or the synthesized voice output immediately before can be output at regular intervals. Instead, it is also possible to output at different timings according to the language category of the composition unit output immediately before. In addition, by providing the equivalent of the time length-based timing determination unit 41 described in the fourth embodiment, the redundant word voice data or the synthesized voice output immediately before can be combined with the synthesized unit output immediately before instead of at regular intervals. May be output at different timings according to the time length of the data.

[0059]

As described above, according to the present invention, in a voice communication support apparatus for converting a sentence inputted by a user into a synthesized speech and outputting the synthesized speech, a sentence inputting section for inputting a sentence by a user's sentence inputting operation is provided. A character string output unit that temporarily stores a character string of an input sentence and outputs the character string when the output operation is performed by a user, and synthesizes speech data by speech-synthesizing the character string output from the character string output unit. , A redundant word storage unit for storing an appropriate number of predetermined redundant word voice data in advance, and a redundant word selection unit for selecting and outputting one redundant word voice data from the redundant word storage unit. Department and
The synthesized speech data is read in, the output speech is extracted, and the input state of the synthesized speech data is monitored. If it is detected that the synthesized speech data is not inputted for a certain period of time, the redundant word selecting unit selects the redundant word from the selected redundant word speech data. Since it is configured to include an output control unit that outputs a voice, it is possible to output a synthesized voice according to a user's sentence input operation,
If silence continues for a certain period of time during sentence input operation, redundant words are automatically output and interpolated, and the other party communicating with the user can be apparently reduced in silence time generated by the user's sentence input operation. It has the effect of reducing the stress you feel.

According to the present invention, in place of the character string output unit, a character string of a sentence input from the sentence input unit is analyzed to extract a composite unit, and the extracted composite unit is sequentially output. Since the system is equipped with an extraction unit, synthesized speech can be sequentially output for each synthesis unit according to the user's sentence input operation to reduce the user's operation, and redundant speech is automatically generated when silence continues for a certain period of time. And the interpolation is performed, and the silence time associated with the user's sentence input operation is apparently shortened, thereby reducing the stress felt by the other party communicating with the user.

According to the present invention, in a voice communication support device for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, a sentence input unit for inputting a sentence by a user's sentence input operation, and an input from the sentence input unit A synthesis unit that analyzes a character string of a sentence, extracts a synthesis unit, determines a language category of the synthesis unit, and sequentially outputs the synthesis unit and the language category; A speech synthesizing unit for speech-synthesizing the character string of the synthesized unit and outputting synthesized speech data; a redundant word storage unit for storing an appropriate number of predetermined redundant word speech data in advance; A redundant word selecting unit for selecting and outputting one word voice data, and a timing outputting a timing signal for determining an output timing of the redundant word voice data according to the input language category The decision unit reads the synthesized speech data, extracts the output speech, monitors the input state of the synthesized speech data, and uses the redundant word speech data selected by the redundant word selection unit when a timing signal is input during the sentence input operation. Since it is configured to include a timing output control unit that outputs a redundant word voice, the user operation can be reduced by sequentially outputting synthesized voice for each synthesis unit according to the user's sentence input operation, When silence continues, redundant words are automatically output and interpolated at the output timing according to the language category of the synthesis unit that was input immediately before, and the silence time associated with the user's sentence input operation is apparently reduced. This has the effect of reducing the stress felt by the partner communicating with the user.

According to the present invention, instead of the timing determining unit, the time length for outputting the timing signal for each time length for determining the output timing of the redundant word voice data generated according to the time length of the immediately preceding synthesized voice data A separate timing decision unit is provided, and the timing output control unit is configured to output redundant word voice data by a time length-specific timing signal, so that a synthesized voice is sequentially output for each synthesis unit according to a user's sentence input operation. When the silent state continues, the redundant word voice is automatically output at the output timing determined according to the time length of the synthesis unit input immediately before, and interpolation is performed. In addition, the silence time associated with the user's sentence input operation is apparently shortened, so that the person who communicates with the user can feel the stress. There is an effect that can be reduced.

According to the present invention, instead of the output control unit,
The synthesized voice data is read, the output voice is extracted, and the input state of the synthesized voice data is monitored. During the sentence input operation, the input of the next synthesized voice data is performed within a predetermined time after the input of the previous synthesized voice data. If not, make preparations for outputting redundant word voice data. When the next synthesized voice data is input, redundant word voice data selected by the redundant word selection unit before the next synthesized voice data as output voice Is configured to include a redundant word prefix output control unit that reads and connects a redundant word voice according to a user's sentence input operation. Automatically outputs and interpolates immediately before the next synthesized speech, realizing the characteristic of natural speech that a redundant word is placed immediately before a meaningful utterance unit, and silence associated with the user's sentence input operation There is an effect that the user can reduce the stress party feels that performs communication by shortening between average.

According to the present invention, in a voice communication support apparatus for converting a sentence inputted by a user into a synthesized speech and outputting the synthesized speech, a sentence input section for inputting a sentence by a user's sentence input operation, and an input from the sentence input section A synthesis unit that analyzes a character string of a sentence, extracts a synthesis unit, determines a language category of the synthesis unit, and sequentially outputs the synthesis unit and the language category; A speech synthesizing unit for synthesizing the character string of the synthesized unit to output synthesized speech data, a redundant word storage unit for each category that stores redundant word speech data in advance for each language category of the synthesis unit, and a synthesis unit analyzing unit. A redundant word selection unit by category for selecting and outputting one corresponding redundant word voice data from a redundant word storage unit by category according to the language category of It monitors the input state of synthesized voice data while extracting the output voice only, and if the next synthesized voice data is not input within a predetermined time after the input of the previous synthesized voice data during the sentence input operation Prepares for the output of redundant word speech, and as the output speech when the next synthesized speech data is input, the next synthesized speech selected by the category-based redundant word selection unit before the next synthesized speech data A redundant word prefix output control unit for reading out and connecting the corresponding redundant word voice data, so that the user operation can be performed by sequentially outputting synthesized voice for each synthesis unit in accordance with the user's sentence input operation. When silence continues for a certain period of time, redundant word speech judged appropriate from the language category of the subsequent synthesis unit is connected before the next synthesized speech and automatically Output to the interpolation,
Realizes the feature of natural speech that a redundant word is placed immediately before a meaningful utterance unit, and apparently shortens the silence time associated with the user's sentence input operation on average, so that the user with whom the user communicates feels It has the effect of reducing stress.

According to the present invention, instead of the redundant word storage unit, a redundant word storage unit for each tail vowel is provided which stores in advance redundant word voice data that can be output for each tail vowel of the synthesis unit. Alternatively, the type of the tail vowel of the input synthesis unit is identified, and a tail word-based redundant word selection unit that selects one redundant word voice data that matches the type of the tail vowel from the tail vowel-based redundant word storage unit is provided. With this configuration, the operation of the user can be reduced by sequentially outputting synthesized speech for each synthesis unit in accordance with the user's sentence input operation, and when silence continues for a certain period of time, the end of the synthesis unit Redundant word voice data that matches vowels is automatically output and interpolated, and the silence time associated with the user's sentence input operation is apparently shortened so that the user with whom the user communicates feels There is an effect that can reduce stress.

According to the present invention, instead of the redundant word storage unit, there is provided a redundant word storage unit by time length for storing redundant word voice data that can be output for each time length of the synthesis unit, and the redundant word selection unit is replaced. And a redundant word selecting unit for each time length for selecting one redundant word voice data corresponding to the time length of the input synthesized voice data from the redundant word storage unit for each time length. The user operation can be reduced by sequentially outputting synthesized speech for each synthesis unit according to the operation, and when silence continues for a certain period of time, a redundant word voice that matches the time length of the synthesis unit is automatically generated. By outputting and interpolating and apparently shortening the silence time associated with the user's sentence input operation, there is the effect that the stress felt by the user with whom the user communicates can be reduced.

According to the present invention, instead of the output control unit,
When the synthesized voice data is read, the output voice is extracted, and the output synthesized voice data is temporarily stored, and the input state of the synthesized voice data is monitored. Since the redundant section voice data is output from the redundant word voice data selected by the selecting section, or a repetition output control section for re-outputting the stored synthesized voice data is provided, the synthesized voice is output in accordance with the user's sentence input operation. When silence continues for a certain period of time, the synthesized speech output immediately before is automatically re-output and interpolated, and the silence time accompanying the user's sentence input operation is apparently shortened. This has the effect of reducing the stress felt by the communication partner of the user.

According to the present invention, in a voice communication support apparatus for converting a sentence inputted by a user into a synthesized speech and outputting the synthesized speech, a sentence is inputted by a sentence input operation of the user, and the inputted sentence is synthesized by speech synthesis. Voice data is generated, an appropriate number of predetermined redundant word voice data is stored in advance, an output voice based on the generated synthesized voice data is taken out, an input state of the synthesized voice data is monitored, and a synthesized voice is input during the input operation. When it is detected that data is not inputted for a certain period of time, the stored redundant word voice data is selected and the redundant word voice is output, so that the synthesized voice can be output according to the user's sentence input operation. When silence continues for a certain period of time during sentence input operation, redundant word sounds are automatically output and interpolated, and the silence time generated by the user's sentence input operation is checked. There is an effect that can reduce the stress the other party feel to perform user and the communication by the above reduction.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a voice communication support device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a voice communication support device according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a voice communication support device according to a third embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a voice communication support device according to a fourth embodiment of the present invention.

FIG. 5 is a block diagram showing a configuration of a voice communication support device according to a fifth embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a voice communication support device according to a sixth embodiment of the present invention.

FIG. 7 is a block diagram showing a configuration of a voice communication support device according to a seventh embodiment of the present invention.

FIG. 8 is a block diagram showing a configuration of a voice communication support device according to an eighth embodiment of the present invention.

FIG. 9 is a block diagram showing a configuration of a voice communication support device according to a ninth embodiment of the present invention.

FIG. 10 is an explanatory diagram showing a specific example of synthesized speech and redundant word speech output according to the first embodiment of the present invention;

FIG. 11 is an explanatory diagram showing a specific example of synthesized speech and redundant word speech output according to Embodiment 2 of the present invention;

FIG. 12 is an explanatory diagram showing a relationship between a language category and a silence allowable time according to Embodiment 3 of the present invention.

FIG. 13 is an explanatory diagram illustrating an operation of a timing determining unit according to the third embodiment of the present invention.

FIG. 14 is an explanatory diagram showing a specific example of synthesized speech and redundant word speech output according to Embodiment 5 of the present invention.

FIG. 15 is an explanatory diagram showing an example of correspondence between language categories and redundant words according to Embodiment 6 of the present invention;

FIG. 16 is an explanatory diagram illustrating an operation of selecting a redundant word according to a sixth embodiment of the present invention.

FIG. 17 is an explanatory diagram showing an example of correspondence between a synthesized unit tail vowel and a redundant word according to the seventh embodiment of the present invention.

FIG. 18 is a block diagram showing a configuration of a conventional voice communication support device.

[Explanation of symbols]

11 sentence input unit, 12 character string output unit, 13 speech synthesis unit, 14 redundant word storage unit, 15 redundant word selection unit, 16
Output control unit, 21 synthesis unit extraction unit, 31 synthesis unit analysis unit, 32 timing determination unit, 33 timing output control unit, 41 time length-dependent timing determination unit, 51 redundant word prefix output control unit, 61 redundant word storage by category Part, 62
Redundant word selecting unit by category, 71 Redundant word storing unit by last vowel, 72 Redundant word selecting unit by last vowel, 81 redundant word storing unit by time length, 82 Redundant word selecting unit by time length, 91 Repetition output control unit.

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Tadashi Suzuki 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Mitsubishi Electric Corporation F-term (reference) 5D045 AA11 DA11

Claims (10)

[Claims] 1. A voice communication support apparatus for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, wherein a sentence input unit for inputting a sentence by a user's sentence input operation, and a character string of the input sentence are temporarily stored. A character string output unit that outputs when an output operation is performed by the user; a voice synthesis unit that voice-synthesizes the character string output from the character string output unit and outputs synthesized voice data; A redundant word storage section for storing predetermined redundant word voice data in advance; a redundant word selecting section for selecting and outputting one redundant word voice data from the redundant word storage section; And monitoring the input state of the synthesized speech data, and when it is detected that the synthesized speech data has not been input for a predetermined time, the redundant word speech data selected by the redundant word selection unit. And an output control unit for outputting a redundant word voice from the voice communication support device. 2. A system according to claim 1, further comprising a synthesis unit extracting a synthesis unit by analyzing a character string of a sentence input from the sentence input unit, and sequentially outputting the extracted synthesis unit, instead of the character string output unit. The voice communication support device according to claim 1, wherein: 3. A voice communication support device for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, wherein: a sentence input unit for inputting a sentence by a user's sentence input operation; Analyzing a character string to extract a synthesis unit, determining a language category of the synthesis unit, and sequentially outputting the synthesis unit and the language category; and a synthesis unit analysis unit that sequentially outputs the synthesis unit and the language category. A speech synthesizer for speech-synthesizing a character string of a synthesis unit and outputting synthesized speech data; a redundant word storage unit for storing an appropriate number of predetermined redundant word speech data in advance; A redundant word selecting unit that selects and outputs one word voice data, and outputs a timing signal that determines an output timing of the redundant word voice data according to the input language category. A timing determining unit that reads the synthesized voice data, extracts an output voice, monitors the input state of the synthesized voice data, and selects the redundant word selection unit when the timing signal is input during a sentence input operation. And a timing output control unit for outputting a redundant word voice based on the redundant word voice data. 4. A time length-dependent timing determining unit that outputs a time-length-specific timing signal that determines the output timing of redundant word voice data generated according to the time length of the immediately preceding synthesized voice, instead of the timing determining unit. 4. The voice communication support device according to claim 3, wherein a timing output control unit outputs redundant word voice data according to the time length-specific timing signal. 5. In place of the output control unit, read synthesized voice data to extract output voice, monitor the input state of the synthesized voice data, and determine a predetermined state after inputting the previous synthesized voice data during a sentence input operation. When the input of the next synthesized voice data is not performed within a predetermined time, preparations are made for outputting the redundant word voice data, and when the next synthesized voice data is input, the next synthesized voice data is used as the output voice. 3. The voice communication support device according to claim 1, further comprising a redundant word prefix output control unit that reads and connects the redundant word voice data selected by the redundant word selection unit before the voice data. 6. A voice communication support device for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, wherein: a sentence input unit for inputting a sentence by a user's sentence input operation; Analyzing a character string to extract a synthesis unit, determining a language category of the synthesis unit, and sequentially outputting the synthesis unit and the language category; and a synthesis unit analysis unit that sequentially outputs the synthesis unit and the language category. A speech synthesis unit that synthesizes a character string of a synthesis unit and outputs synthesized voice data, a category-based redundant word storage unit that stores redundant word voice data in advance for each language category of the synthesis unit, and a synthesis unit analysis unit. A category-based redundant word selection unit that selects and outputs one corresponding redundant word voice data from the category-based redundant word storage unit according to the language category of The synthesized voice data is read, the output voice is extracted, and the input state of the synthesized voice data is monitored. During the sentence input operation, the input of the next synthesized voice data is performed within a predetermined time after the input of the previous synthesized voice data. If not performed, the preparation for outputting the redundant word voice is performed, and as the output voice when the next synthesized voice data is input, the category-based redundant word selecting unit is arranged before the next synthesized voice data. And a redundant word prefix output control unit for reading and connecting redundant word voice data corresponding to the selected next synthesized voice. 7. A redundant word storage unit for each tail vowel that stores in advance redundant word voice data that can be output for each tail vowel of the synthesis unit, instead of the redundant word storage unit, and And a tail word-by-speech redundant word selection unit for identifying a tail vowel type of the synthesis unit to be selected and selecting one of the redundant word voice data matching the tail vowel type from a tail vowel-by-tail redundant word storage unit. The voice communication support device according to any one of claims 2 to 5, wherein: 8. A redundant word storage unit for storing a redundant word voice data which can be output for each time length of a synthesis unit in place of the redundant word storage unit. 6. A redundant word selecting unit for each time length which selects one redundant word voice data matching the time length of the synthesized voice data from the redundant word storing unit for each time length. Any one of them
The voice communication support device according to the item. 9. In place of the output control unit, the synthesized speech data is read in, the output speech is extracted, and the output synthesized speech data is temporarily stored. The input state of the synthesized speech data is monitored, and the synthesized speech data is monitored. If it is sensed that no input is made for a certain period of time, the redundant word selection unit outputs a redundant word voice from the selected redundant word voice data, or a repetition output control unit that re-outputs the stored synthesized voice data. The voice communication support device according to any one of claims 1 to 8, further comprising: 10. A voice communication support apparatus for converting a sentence input by a user into a synthesized voice and outputting the synthesized voice, stores an appropriate number of predetermined redundant word voice data in advance, and inputs a sentence by a user's sentence input operation. Generating synthesized speech data based on the input sentence, extracting an output speech based on the generated synthesized speech data, and monitoring the input state of the synthesized speech data. A speech communication support method, wherein when it is sensed that no input is made, the stored redundant word voice data is selected and the redundant word voice is output.