JPH0997094A - Onboard voice synthesizer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To convert a teletext sent from a base station to a voice easy to listen of a message style by a start instruction by the voice from a user such as a driver, etc., and to output it. SOLUTION: When the teletext is received by a receiver 101, is stored in a storage 103, and is informed the driver by a display device 112 or a speaker 110. When the driver depresses a button 115, or performs voice instruction by a microphone 114, a central processing unit ill operates a language processor 104 and a voice synthesizer 107 to convert character information in the storage 103 to the voice. The conversion to the voice is performed so that the information is analyzed to a clause so as to become a message style, and its sound recorded voice is outputted when the clause coincides with the contents beforehand registered in a fixed sound recorded voice data base 108, and a synthesis voice or a fixed sound recorded voice using a synonym is outputted when they disagree. Further, a proper noun such as a place name, an institution name, etc., in the clause is voice converted automatically by twice each together with an attached noun.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an in-vehicle voice synthesizing device for converting teletext such as traffic information into voice and providing it to a user such as a driver.

[0002]

2. Description of the Related Art In recent years, services for providing traffic information to drivers by text broadcasting from social infrastructure base stations are increasing, and information is displayed in kanji / kana mixed sentences on a display device such as an in-vehicle television. An example of such a device will be described with reference to the drawings. FIG. 7 shows the configuration of a conventional teletext display device. In FIG. 7, 701 is an antenna, 702 is a receiver, 703 is a storage device, 704 is a central processing unit, 705 is a display device, and 706 is a tuning button. When the driver presses the tuning button 706 to tune the receiver 702 to the teletext and the teletext is received, the received kanji / kana mixed text information is once stored in the storage device 703, and then the central processing unit. It is converted into an electric signal for screen display by 704 and is displayed on the display device 705 as a kanji-kana mixed sentence. Information on teletext is transmitted in frame units, and since an ID part is provided at the beginning of the frame, the central processing unit 704
Compares this ID part with the ID of the information frame currently being displayed, and when different, displays the information of a new frame and updates the screen.

As described above, the conventional teletext display device can also provide the teletext to the driver as visual information.

[0004]

However, gazing at the screen while driving is dangerous and undesirable. If the information can be provided to the driver after being converted into auditory information, that is, voice, instead of visual information, driving safety can be greatly improved. Further, if the operation can be performed by using the voice of the driver, it is not necessary to release the hand from the steering wheel, and the safety can be further improved.

The present invention solves such a conventional problem, and an object of the present invention is to provide an in-vehicle voice synthesizer capable of transmitting traffic information to a driver by voice.

[0006]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention is a language processing means for parsing received character information mixed with kanji and kana and breaking it down into clauses so as to give a message-like voice that is easy to hear. And a voice synthesizing means for creating an output voice using a prerecorded fixed voice or an atypical voice synthesized by rule synthesis based on the result of the language processing. It is designed to be converted into high-quality sound.

Further, the voice recognition means is provided so that the driver can instruct the timing of starting voice conversion by voice without using a button or a touch panel.

[0008]

According to the present invention, the character information received is decomposed into phrase units by the language processing means and prerecorded by the voice synthesizing means when an instruction such as a voice from the driver is given. It is assigned to the synonymous one with the speech, and if there is no synonymous one, it is assigned to the synthetic speech synthesized by rule synthesis or the like, and the conversion into the speech can be realized. Also,
By the voice recognition means, the driver can perform voice conversion by only voice instructions without releasing the hand from the steering wheel, so that driving safety can be improved.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows the configuration of an in-vehicle voice synthesizer according to an embodiment of the present invention. In FIG.
101 is an antenna, 102 is a receiver, 103 is a storage device, 104 is a language processor, 105 is a dictionary database,
106 is a synonym table, 107 is a voice synthesizer, 108 is a fixed recording voice database, 109 is a rule synthesizer, 11
Reference numeral 0 is a speaker, 111 is a central processing unit, 112 is a display device, 113 is a voice recognition device, 114 is a microphone, and 115 is a button.

Next, the operation of the above embodiment will be described. Conventionally, the teletext has been received by the user by matching the channels, but in the present embodiment, the user is always tuned to a traffic information channel set in advance,
Even when the display device 112 is displaying another screen such as navigation, each time the receiver 102 receives a new frame through the antenna 101, the central processing unit 111 detects this (detection is performed in the same manner as in the conventional art. The voice synthesizer 107 sends a message such as “Traffic information received” through the speaker 110 to notify the driver (or user). If the driver does not feel the need, there is no need to operate anything,
The received frame information is stored in the storage device 103.
It only replaces the previously received frame information.

However, in order to prevent the user from not listening, at the same time as the voice message, as shown in FIG.
The reception mark 202 is displayed in one corner of the display screen 201 of FIG. A display time of about 5 seconds is suitable, and it goes without saying that blinking the reception mark 202 will further improve the effect.

The driver who knows the reception pushes the button 115 to give an instruction, or gives a voice instruction to the microphone 114 to start voice conversion. The utterance content at this time is something like "traffic information display", which is set in advance at the time of factory shipment, but can be re-registered by the user if desired. When registering, all the Japanese syllables are displayed on the screen of the display device 112, and the user selects and registers one sound and one sound, or after the user switches the in-vehicle voice synthesizer to the registration mode, Register by actually speaking the utterance content. In the latter case, it is difficult to react to a voice other than the user who registered the utterance content, which is also effective in crime prevention.

Based on this instruction, the central processing unit 111
Displays the received contents on the display device 112 as usual, and operates the language processor 104 and the voice synthesizer 107 to cause the storage device 103 to convert the character information containing the kanji and kana of the newly received frame into voice. When converting character information mixed with kanji and kana into speech, the length of the phrase is determined and a pause is automatically inserted between the phrases so that the message tone is easy to hear.

The length of the phrase is determined according to the flowchart shown in FIG. In FIG. 3, first, the language processor 1
04 divides the kanji / kana mixed character information received from the storage device 103 into words using the word information stored in the dictionary 105 (steps 301 and 302). Specifically, well-known parsing algorithms (eg Okada, p.
p.10-19, "Introduction to Natural Language Processing", Kyoritsu Shuppan Co., Ltd., 1991), and the part of speech of a word and dependency information (
Both are stored in the dictionary database 105 for each word), the connection relation between the words is determined, and the clause is provisionally determined (step 303). In this way, the number of words and the number of syllables in the determined syllable are checked. If the number of words exceeds 5 words (step 304), or if the entire syllable consists of 13 syllables or more (step 305), this syllable is selected. Divide by breaking words (step 30)
7). The division is basically immediately after the word in the middle of the phrase,
If this word is an independent word accompanied by an adjunct word, a split point follows the adjunct word. For clauses divided in this way,
When further division is to be performed with the same criterion, division is performed (step 308), and such processing is performed for all clauses (steps 306 and 309).

FIG. 4 shows an example of division of the broadcast content "Meiji Dori is about 5 km of traffic congestion due to an accident in the up lane.""MeijiDori""Uphill diagonal line""Due to accident""About 5 km It is divided into "it is a traffic jam".

An exception is made when one word has 13 syllables or more, and a syllable of only that word and an annex is adopted. Further, it may be better to use different values for the maximum number of words 5 and the maximum number of syllables 13 depending on the speech speed of the synthesized speech. The number of words and the number of syllables should be increased in the case of fast talk, and the number of words and the number of syllables should be decreased in the case of late talk.

The phrase is rewritten into an intermediate language by using the reading from the word dictionary and the accent information. An example is shown in FIG.
, The accent strength of the accent (for example, 0
And 1)).

After that, the phrase is checked by searching the synonym table 106 to find out whether or not there is a phrase that matches the content of the standard recorded voice registered in the standard recorded voice database 108 in advance. Accent information is ignored when searching. If there is a match, the fixed recording voice number is added after the intermediate language of the syllable together with the control symbol (&), and nothing is added to the non-match. In this case, the rule synthesizing unit 109 creates a synthesized voice using the reading of Kana and the accent information. In this way, message-like voice conversion is realized.

It should be noted that instead of the rule synthesizer 109, all Japanese syllables are recorded in advance, and the synthesized speech can be realized by sending out the recorded one note and one note as the reading kana. You can All Japanese syllables are loud,
Accent can also be realized by recording in three levels, medium and low.

Further, as shown in FIG. 5, even if the phrase does not necessarily match the content registered in the standard recording voice database 108, the synonym table 106 is referred to,
If there is a match, the standard recording sound is output.

Further, as shown in FIG. 6, proper nouns such as place names and facility names in a sentence containing kanji and kana are automatically assigned an intermediate language twice together with a noun attached word, and a voice is output twice. In this way, we try to prevent listening.

Although the above-mentioned embodiment shows an example of receiving the teletext of the traffic information, the same applies to other broadcasts such as train / ship / aircraft departure / arrival guidance broadcasts and restaurant service announcements. Is easy to apply to.

In addition, in cases other than guide broadcasting, message-style voice conversion may be inadequate, so that the user removes the phrase length restriction by using a selection button or the like, cancels the allocation of fixed-form recording voice, and uses proper nouns. Also, it is easy to stop the voice conversion twice and perform normal reading voice conversion.

Even if the character broadcast is not a mixed kanji / kana sentence but only kana, the same process can be performed by changing the word dictionary to a kana word dictionary.

[0025]

As described above, according to the present invention, if the driver indicates the content of the teletext with a voice when necessary, an excellent in-vehicle voice synthesis can be heard as a voice of a message tone that is easy to hear. The device can be realized.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an in-vehicle voice synthesizer according to an embodiment of the present invention.

FIG. 2 is a reception notification screen diagram according to an embodiment of the present invention.

FIG. 3 is a flowchart showing a method of decomposing a received sentence into clauses for creating a message tone voice according to an embodiment of the present invention.

FIG. 4 is an exploded view of a received phrase for creating a message tone voice according to an embodiment of the present invention.

FIG. 5 is a conversion diagram of a phrase and a recorded fixed voice according to an embodiment of the present invention.

FIG. 6 is a data diagram of automatically converting a proper noun and its adjunct nouns twice in one embodiment of the present invention.

FIG. 7 is a block diagram showing a configuration of a conventional teletext display device.

[Explanation of symbols]

 101 antenna 102 receiver 103 storage device 104 language processor 105 dictionary database 106 synonym table 107 voice synthesizer 108 fixed recording voice database 109 rule synthesizer 110 speaker 111 central processing unit 112 display device 113 voice recognition device 114 microphone 115 button (voice (Start conversion button)

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Takeshi Takeshi 4-3-1, Tsunashima-higashi, Kohoku-ku, Yokohama-shi, Kanagawa Matsushita Communication Industrial Co., Ltd.

Claims (8)

[Claims] 1. An in-vehicle voice synthesizing device comprising means for receiving teletext of traffic information and converting the content into voice and outputting the voice. 2. The in-vehicle voice synthesizing apparatus according to claim 1, further comprising means for receiving a text broadcast of traffic information, converting the content into a voice and outputting the voice, and displaying the voice on a display device. 3. A driver is notified that a teletext is received, the received information is temporarily stored in a storage device, and the reception is made by a synthesized voice, a mark display on a display device, or both. The vehicle-mounted speech synthesizer according to claim 2, further comprising: 4. When converting the received character information mixed with kanji and kana into speech, a syntactic analysis is performed and a maximum of 5 words in one syllable or a maximum of 13 syllables in one syllable.
We set the condition of up to individual pieces, determine the range of bunsetsu according to the lesser one, and automatically insert a certain silent interval between bunsetsu and decompose it into bunsetsu to give a message-like voice. 4. The in-vehicle voice synthesizer according to claim 1, further comprising a language processing unit for performing the processing. 5. When converting the received kanji / kana mixed character information into voice, the prerecorded standard voice is compared with the contents of a database in which it is stored, and if the contents match, the recorded voice is recorded. 5. The in-vehicle voice synthesizing apparatus according to claim 1, further comprising a voice synthesizing unit that uses a voice and, when there is no matching content, converts the voice into a voice by a synthesized voice such as rule synthesis. 6. When converting the received kanji-kana mixed character information into voice, the prerecorded standard voice is compared with the contents of a database in which it is stored, and even if they do not match,
The in-vehicle voice synthesizing apparatus according to claim 5, wherein the recorded voice is used in the case of synonym, and when there is no synonym, the voice is converted into voice by voice synthesis such as rule synthesis. 7. When converting the received kanji-kana mixed character information into voice, in the case of proper nouns such as place names and facility names, the proper nouns and the annexes of the nouns attached to the proper nouns are included. 7. The in-vehicle voice synthesizing apparatus according to claim 5 or 6, wherein voice conversion is performed twice for each part with a constant silence interval. 8. The in-vehicle voice synthesizing apparatus according to claim 1, wherein the user instructs the start of voice conversion by a button, a touch panel, or voice recognition.