JPS6288028A - Text vocalizing device - Google Patents

Abstract

PURPOSE:To improve the understandability of a test by providing plural vocalizing devices that prepare different vocalizing forms with each other at a vocalizing means and switching them from a specified character combination. CONSTITUTION:A text vocalizing device is constituted with an overall controller 1, a KANJI(Chinese character/KANA(Japanese syllabary) character string input device 2, a character string attribute detecting device 3, a KANJI/KANA conversion device 4 that forms read information, three kinds of vocalizing devices 11-13, and a vocalizing device controller 5 that selects one out of the above vocalizing devices. In the device, the presence of the character combination that regulates an attribute in a KANJI/KANA character string inputted from the input device 2 is investigated at the attribute detecting device 3 and also, the read information of the character string is formed at the conversion device 4. The read information and an attribute information are sent to the vocalizing device controller 5, and one of the vocalizing devices 11-13 is selected by the attribute information, the reading information being vocalized.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a text utterance device that utters input Kanji-Kana character strings, and more specifically, the present invention relates to a text utterance device that utters input Kanji-Kana character strings. The present invention relates to a text utterance device that makes the content of the utterance more clear by changing the utterance of the text and annotation portions to other parts.

[Conventional technology and its problems]

A text utterance device that inputs kanji and kana character strings and outputs audio can be used to great effect when used to read input text aloud, query databases, etc. Conventional sentence utterance devices convert the input kanji-kana character string to kanji-kana to obtain the reading character string, and then utter the previous syllable in the early - position. 1 - Adding intonation information. This makes it possible to make the vocalizations more natural. Documents related to such prior art include:
For example, Proceedings of the 1981 National Conference of the Institute of Electronics and Communication Engineers 56
-3 [Speech synthesis from Japanese text] etc. However, with the above technology, there is no information such as whether the part to be uttered is in the middle of a conversation or an annotation, so for example, he says, "Thank you."

Although the sentence ``Thank you'' could be paused, it was not possible to make it clear that it was a conversation by changing the sound quality. (The underline indicates a pose.) Similarly, in 1985, the sentence ``Showarokujuunen-senkyuhiyakuhachijiyugo-ni'' is pronounced as ``1985''. °゛ga゛Showa C
) It is not possible to clearly know that it is an annotation for year 0,
As a result, the intelligibility of the text was significantly reduced.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a text utterance device that avoids the drawbacks of the conventional techniques described above and allows a user to clearly hear a conversation part, annotation part, etc. during the utterance of a text.

[Structure of the invention]

According to the present invention, there is provided a human-powered device for inputting a Kanji-Kana character string, a Kanji-Kana conversion device that converts the Kanji-Kana character string into Kanji-Kana character strings to form reading information of the Kanji-Kana character string, and the reading information. A text utterance device having a utterance means for uttering a utterance based on the utterance, the utterance means comprising a plurality of utterance devices each having a different utterance form, and wherein the utterance device includes a plurality of utterance devices each having a different utterance form, and the utterance device is configured to emit a utterance based on the existence of a specific character combination in the kanji-kana character string. an attribute detection device that detects an attribute of a portion surrounded by the specific character combination by knowing the character combination; and a voice device control device that switches the voice device based on the attribute obtained by the attribute detection device. A sentence utterance device is obtained.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.6 Figure 1 shows an embodiment of the present invention. 1 is an overall control device for controlling the overall flow, 2 is a kanji-kana character string input device for inputting kanji-kana character strings, and 3 is a device for detecting the existence of a specific character combination within a kanji-kana character string. 4 is an attribute detection device that detects the attribute of the part surrounded by the specific character combination by knowing the character combination; 4 analyzes the Kanji-Kana character string obtained by the Kanji-Kana character string input device 2;
It is a kanji-kana conversion device that forms reading information, and 5 is a vocalization device that selects one of the three types of vocalization devices described below based on the attributes of the kanji-kana character string obtained by the attribute detection device 3. ], 1.12.13 is a voice device for vocalizing the reading information obtained by the Kanji-Kana conversion device 4 under the control of the voice device control device 5. FIG. 3 shows the flow of the processing procedure in this embodiment. In this embodiment, the specific character combinations detected by the attribute detection device 3 are "°" and "1°'" indicating a conversational part, and "°("' and ゛)'' indicating an annotation part. Assume that the conversation part voicing device 11 is used, the voicing device 12 is used for voicing the annotation part, and the voicing device 13 is used for other parts (hereinafter referred to as normal parts).

Here we actually ``Thank you,'' he said.

We will explain the processing flow when the sentence is written manually.

First, the above-mentioned Kanji/Kana character string inputted from the Kanji/Kana character string human power device 2 is sent to the attribute detection device 3 under the control of the overall control device 1 . The attribute detection device 3 checks the existence of a character combination that defines an attribute in the input Kanji/Kana character string. In this example, it can be seen that "Thank you" surrounded by "r" and J"' is a conversational part. On the other hand, the input Kanji/Kana character string is sent to the Kanji/Kana conversion device 4 under the control of the overall control device 1, where it is converted into Kanji/Kana and the reading information necessary for vocalization is formed. In this example, "thank you" translates into "thank you for reading."

The reading information obtained by the kanji-kana conversion 'A device 4 and the attribute information obtained by the attribute detection device 3 are used as a whole system;
1 to the voice generating device control device 5, and the voice generating device control device 5 controls the voice generating device 11 according to the attribute information.
]3. When the vocalization devices 11 to 13 receive the reading information, they vocalize the given reading information under the control of the overall control device 1. In this example, Kanji-kana conversion′! The reading "Thank you" sent from A4 is sent to the name 5 voice device 11, which is the conversation part, and is uttered.

In the same way, "and he" are uttered from the voice device 13 as normal parts, and the "fat part" is surrounded by "(°° and °゛)", so it is annotated. ``Taro'', which is the reading, is uttered from the vocal device I2, and ``'', which is the reading of ``, which ゛ said'' was said.''
is uttered from the vocalization device 13 as a normal part.

Through this series of first order, the user can say ``Arigatou ° is the conversation part, ``Tarou'' or ``Arigatou ° is the conversational part,'' while pausing the second utterance.
You can see that the t-interpreted part is the normal part, and the other part is the normal part.

Note that each device shown in FIG. 1 can be realized using conventional techniques. Also.

The different vocal forms of the voice generating devices 11 to 13 can be easily realized by, for example, changing the volume or pitch.

FIG. 2 shows another embodiment of the present invention. The first [in the seventh embodiment, the attribute detection device 3] specifically>It is: *
Although the configuration is such that the gender information and the reading information obtained from the Kanji/Kana conversion device 4 are sent separately to the voice control device 5, in this embodiment, the attribute information? Ga is added to the kanji-kana character string and sent to the voice-producing device control device 5 via the kanji-kana conversion device 1.

That is, in this case, the input kanji/kana character string inputted from the kanji/kana character string input device 2 is sent to the attribute detection device 3 under the control of the overall control device 1, and is searched for a specific character combination. The attributes of the part surrounded by are detected.Next, the input kanji-kana character string is sent to the kanji-kana conversion device 4 under the control of the overall control device 1, where it is converted into kanji-kana and the reading information necessary for utterance is generated. Next, the reading information obtained by the kanji-kana conversion device 4 and the attribute information obtained by the attribute detection device 3 are sent to the utterance device control device 5 under the control of the overall control device 1. , is outputted by the voice device control device 5 to any of the voice devices 11 to 13 according to the attribute information.

FIG. 3 is a flowchart showing the flow of processing in the embodiment shown in FIG. When a kanji/kana character string is input, a conversation part, an annotation part 9, and a normal part are detected by attribute detection. On the other hand, the reading is obtained by kanji-kana conversion, and utterances are performed according to the attributes. In addition, the flow shown in Figure 3 is to input Kanji/Kana character strings all at once, detect attributes, convert to English/Kana, and then utter the words. It can be modified in various ways, such as by repeating ``+7''.

[Effect of the invention J As described above, when the present invention is used, the user can easily distinguish between the conversation part, the annotation part 1 and the normal part just by listening to the utterance Jh, so the intelligibility of the user can be improved. can be greatly improved.

[Brief explanation of drawings]

1 and 2 are for explaining the present invention J) Block diagram 4 FIG. 3 shows the processing in FIG. 1? This is a flowchart 1 showing FL deviation. 1... Overall control device, 2... Kanji-kana character string input device, 3... Attribute detection device, 4... Kanji-kana conversion device,
5...Voice device control device, 11.12.13...Voice device. ,,,:',, L Acting Patent Attorney Uchihara Former゛6゜+ ,'! ! ,,,. Kaya 3 figure

Claims (1)

[Scope of Claims] 1. An input device for inputting a Kanji-Kana character string, a Kanji-Kana conversion device that converts the Kanji-Kana character string to Kanji-Kana to form reading information of the Kanji-Kana character string, and In a text utterance device, the utterance device includes a plurality of utterance devices each having a different utterance form, and the utterance means includes a plurality of utterance devices that make utterances based on reading information, and the utterance means includes a plurality of utterance devices that take different utterance forms, and the utterance device is configured to utter a specific character combination in the Kanji-Kana character string. An attribute detection device that detects the attribute of the part surrounded by the specific character combination by knowing the existence thereof, and a voice device control device that switches the voice device based on the attribute obtained by the attribute detection device. Characteristic text production device. 2. The text utterance device according to claim 1, wherein the specific character combination is a character combination that defines at least one of a conversation part and an annotation part.