JPH02165324A - Document reader - Google Patents

Abstract

PURPOSE:To smoothly proofread a document without interrupting the reading of a document by storing a voicing position as a marking position by operating a switch when a character misinput position or the contents to be corrected appears in a reading document. CONSTITUTION:When a document misinput position or the contents to be corrected in the document appears at the time of checking the contents of the document by the output of a synthesized sound, the address of a document voiced at the time of depressing a marking switch 10 and the position in the document are stored in a marking memory 11 as marking positions in each depression of the switch 10. Thus, the proofreading or confirmation of an input sentence or the like can be smoothly executed while listening the voice without interrupting the generating voice.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Industrial field of application J#ζ The invention relates to a speech synthesis device. In particular, the present invention relates to a text reading device that is equipped with a rule synthesis device that can synthesize speech from an arbitrary sentence a, and that allows proofreading and checking of human-written texts while listening to the speech.

(b) 6J of conventional technology, rule synthesis for synthesizing speech from the text of α
There is a lot of research into this, and there are even prototype proofreading devices for newspapers, readers for the blind, and one machine that has been put into practical use.

These devices determine the pronunciation, accent, etc. for an input sentence, and connect pre-prepared speech segments to create an RF voice corresponding to the input sentence [for example, Japanese Patent Publication No. 62-8813 No. Publication].

With such conventional text reading devices, when proofreading text input using Wharf 0, etc., whenever a part that requires correction due to an input error is discovered, the audio being generated is interrupted each time. Then, the correction process was carried out.

(c) Problems to be Solved by the Invention As mentioned above, in order to perform sentence σ proofreading with a conventional text reading device, the operator must alternately perform sentence checking work by operating the speech synthesis device and sentence correction work. This has the disadvantage that these switching operations are complicated. That is, for example, even if there is a part of the text that you would like to spend some time checking, it is necessary to interrupt the audio and check at that point, which causes problems such as making it difficult to grasp the flow of the text. .

(d) Means for Solving the Problems The sentence reading device of the present invention is capable of reading aloud aloud when a portion of the sentence that needs to be rechecked, such as a revised portion, is discovered during reading aloud. A marking memory is provided to perform marking. The device also includes a display device that displays the marked portions of the text after the text has been read aloud.

(E) Function According to the text reading device of the present invention, when checking the content of a text using the synthesized voice 11 and force, if there is a human error in the text or a part of the text that requires correction, marking is performed. By operating the switch, the position during utterance can be memorized as a marking point, so the correction process can be postponed and the II'(k) recognition of this sentence can be performed while continuing to read out the sentence. Since the parts to be corrected and parts of the text that require reconsideration can be displayed separately from other parts of the text on the display device, it is easy to search for the parts to be corrected during the correction process after finishing reading the text.

(F) Embodiment FIG. 1 is a block diagram showing the configuration of an example of a text reading device of the present invention, which inputs an arbitrary character code, stores it in a text buffer memory (1), 1)
The text is read out by synthesized speech output according to the character code string.

In the figure, (2) is the generation switch, (3) is the intermediate switch, (4) is the address pointer for the text buffer memory (1), (5) is the display buffer, (6 is the rule Synthesis circuit, (7) is CRT display, (8) is busy signal, (
9) is the display pointer, (10) is the marking switch/h,
(11) indicates a marking memory, (12) an address pointer for the marking memory (11), and (13) a shift switch.

Figure 2 shows an example of a sentence stored in the text buffer memory (1), in which the 1.2 characters "overseas" in the 4th sentence amount are underlined in the figure, and the 51st sentence is underlined. The 8.9 character ``toko'' for eyes is described as a place that requires correction.

Figure 3 shows the marking information (in this case, [text address S] and [occurring position C]) in the marking memory <U).
) is exemplified.

The apparatus of the present invention and its operation will be explained below based on these figures.

First, as shown in FIG. 2, the text created using a text processing device such as a word processor is
) is stored as a character code string.

If you now suppress the generation switch (2), the intermediate switch will
When the chain (3) is closed, the character code string of the sentence in the text buffer memory (1) is sent to the display buffer (5) at around 1 text lit, and the address pointer (4) is set. The character code string in the memory (1) is sequentially sent to the rule synthesis circuit (6) at real-time speaking speed.

As a result, sentences are displayed on the CRT display 15 (7) according to the character strings in the display buffer (5), and at the same time, in the rule synthesis circuit (6), a 6°C synthetic U voice is added to the sequentially input character strings. A signal is output. During the operation of this standard 1111 synthesis circuit (6), that is, during vocalization, the ToG signal (8
) is output, the intermediate switch (3) is opened in response to this busy signal (8), and the busy signal (8) is output.
), the address pointer (.I) is incremented and the next sentence is specified. And the above movement f
Repeat ya again.

In addition, in order to determine the relationship between the displayed sentence and the utterance position of this sentence, the rule synthesis circuit (6) sends the position information of the character code being uttered to the display pointer (9), and displays the character at this position. Make the display different from other positions. For example, reverse display is performed.

While continuing the reading process as described above, for example, if you realize that you mistakenly entered "domestic" while pronouncing the fourth sentence and entered "overseas", press the marking switch (10) before operation. Manipulate.

Incidentally, at the time of this switch operation, the generated characters have reached the third character "de" after "domestic".

As described above, when the marking switch (10) is operated, the value of the address pointer (4) (ie, address information 5 of the sentence being uttered = 4) is sent to the marking memory (11). Furthermore, at the same time, the display pointer (9
) value (i.e. character position information in the sentence being uttered C=3)
is sent to the marking memory (11). Then, these image information S and C are stored in a marking memory (11) as shown in FIG. 3 as marking information formed in a tree. Furthermore, at the same time, the address pointer (12) for the marking memory (11) is incremented.

Through the above operations, each time the marking switch (lO) is operated, the address of the sentence being uttered at the time the marking switch (lO) is pressed and the position in the sentence are stored in the marking memory (11).

After finishing reading the text, the correction [
When the switch (13) is pressed, the address pointer (12) of the marking memory (11) first points to the address of the sentence in the marking memory (11) [5=
=1] is sent to the address pointer (4), and
Similarly, the position information Mj [C=
3] is sent to the display pointer (9), the character code string of the fourth sentence including the marked character position is sent to the display buffer (5), and the CRT display (7) displays the message "Even overseas... "Now on the market." is displayed, and the marked character "de" is displayed in reverse or as an underline. By displaying the masking characters that have been decorated in this way and distinguishing them from the text of other characters, the operator can easily identify the marking characters, and as a result, the characters that need to be corrected several characters before the marking characters column, in this case "Useful for detecting overseas J.

By detecting this character string that requires modification, the operator can check its character notation and expression content, and can make corrections by inputting new input from the keyboard (14). That is,
In this case, "overseas" should be corrected to "domestic." In addition, the address pointer (12) can be changed by operating the correction switch (13).
) is incremented and the correction swifter (13) is operated (
Each time you press the button, the marked location will be displayed in sequence. The correction process is then repeated. With this, 51
The 8.9 character [toko J in the second sentence can also be correctly corrected as ``koto''.

In the following explanation, the position of the reading character when the marking switch (10) is operated is marked, but if the character R is marked several characters before the marking switch (10) is operated, The correction position and mark position can be almost matched.

In addition, in the above-mentioned embodiment, since the mark character is displayed in units of sentences and further modified, it is easy to detect characters that require correction based on the expressions before and after the mark character. For example, display page +4 in places,
It is possible to display the sentences that require correction and the mark characters within them.

In the device according to the embodiment described in 1-1, even if you notice a part that requires correction or reconsideration while checking the input sentence using the "Y" voice, there is no need to interrupt the utterance and you can mark that part. By storing them in advance, you can make corrections all at once later, making smooth writing +E possible.

(G) 5F, Effect 4 of the Invention: When the text reading device of the invention contains a character input error or a part of the text that requires correction, by operating a switch, the text reading device of the invention can Since the position can be memorized as a marking point, the text can be checked smoothly without interrupting the text reading.

Also, during the correction process after the text is read out,
? i: Since the parts of the content that need to be revised and the parts that need to be reconsidered are displayed in an easy-to-understand manner, it is easy to search for Mineshosha.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of the configuration of an apparatus for implementing the present invention, FIG. 2 is a content diagram of a text buffer memory, and FIG. 3 is a content diagram of a marking memory. (1)...Text buffer memory, (5)...Display buffer memory, (6)...Rule synthesis circuit, (7)...
・CRT, (10)...Marking switch, (
11)...Marking memory. 1st @

Claims (2)

[Claims] (1) A text-to-speech device reads out a text using a speech synthesis circuit that reads out a character code string from the text buffer memory that stores the character code string and converts this character code string into speech, by operating a switch. A text reading device comprising: a marking memory that writes marking information corresponding to the character code position of the text being read out at the time of this operation. (2) A text-to-speech device reads out a text by reading a character code string from the text buffer memory that stores the character code string of the text, and converts this character code string into speech by means of a speech synthesis circuit that reads the text by operating a switch. At the time of this operation, there is a marking memory that writes marking information corresponding to the character code position of the sentence being read aloud, and a sentence part near the character code position corresponding to the marking information in the marking memory can be distinguished from other sentence parts. What is claimed is: 1. A text reading device characterized by comprising a display device that displays a message.