KR100275529B1 - Text/voice converter and method using trick mode - Google Patents

Abstract

PURPOSE: A text-to-voice converter using a trick mode and a method thereof are provided to generates a composite sound by enabling a user to designate an arbitrary sentence using a trick mode and a slide bar and to determine interface between information and a text-to-voice converter. CONSTITUTION: A graphic user interface module(230) sets a file using a file button, and transmits information on the file to an input text analysis module(210). The input text analysis module(210) receives and analyzes the information on the file, and transmits the result. A user input analysis module(220) receives the information on sentence analysis from the input text analysis module(210). The user input analysis module(220) analyzes a user input command inputted from the graphic user interface module(230). In addition, the user input analysis module(220) transmits the user input information analyzed. A synthesizer module(240) converts a text into the voice using the user input information analyzed. The synthesizer module(240) reports information on the sentence outputted as a composite sound to a user through the graphic user interface module(230).

Description

Text-to-speech converter using trick mode and method

The present invention provides a text-to-speech converter (TTS: text-to-voice converter) that converts text to speech by setting the position of a sentence, a syllable or a word in a sentence using a trick mode and a slider bar of a graphical user interface. -speech conversion system) and a method thereof.

Conventional text / voice converters have a simple function of outputting data input in sentence units as synthesized sounds. Therefore, in order to continuously output the sentences stored in the file as a synthesized sound, a main control program which reads the sentences in the file and passes them to the input of the text / voice converter is required. In such a main control program, there is a method of simply outputting the synthesized sound once from the beginning to the end of the file. However, since there is no method of outputting the synthesized sound from the user's desired position, user convenience is hardly considered. Alternatively, it is possible to select a region to be synthesized using the text editor function, but it is less convenient to generate a synthesis sound in conjunction with a text editor. Recently, the user's convenience has been improved by implementing a sequential search and playback function as in a video player (VTR) using a graphical interface to the main control program. However, no function has yet been implemented for the user to arbitrarily select any sentence in the file or any position in the sentence and output the synthesized sound.

In addition, since the existing text / sound converter generates a compound sound based on the input sentence, if a compound sound is to be generated from an arbitrary position in the sentence, the sentence except for the text before the position must be transmitted as an input to the text / sound converter. . However, since the text / voice converter analyzes the input sentence and controls the rhyme, when the input sentence changes, the rhyme pattern also changes. Therefore, even if the original sentence is the same, if the synthesized sound is generated from an arbitrary position in the sentence, the rhyme pattern in the original sentence cannot be maintained.

Therefore, in order for a user to conveniently use a text / to-speech converter that supports the trick mode, the user can easily select an arbitrary sentence in the file, or an arbitrary position in the sentence, and output a composite sound, and optionally It is necessary to implement a function that maintains the rhyme pattern in the original sentence even when the synthesized sound is generated from the position of.

Thus, the present invention allows the user to specify arbitrary sentences in the file using trick modes and file slidebars, arbitrary positions within sentences using sentence slider bars, and define the interface between these information and the text / voice converter. To provide a text-to-speech converter and a method for generating the synthesized sound

1 is a configuration diagram of a system to which the present invention is applied;

2 is a block diagram of a text-to-speech converter using a trick mode according to the present invention;

3 is an exemplary diagram of a graphical user interface using a trick mode according to the present invention;

4A and 4B are flowcharts illustrating a text / voice conversion method using a trick mode according to the present invention.

* Explanation of symbols for main parts of the drawings

101: user input device 102: central processing unit

103: text / voice converter 104: digital / analog converter

105: output device 210: input text analysis module

220: user input analysis module 230: graphical user interface module

240: synthesizer module 250: speaker

310: trick mode control window 320: additional option window

330: Lips animation window 340: Text output?

The apparatus of the present invention for achieving the above object, to set the file using the file button to start the synthesis and to transmit the information of the sentence, syllable and word count of the set file, using the file slider bar of the trick mode User graphic interface means for setting an arbitrary position of a sentence and transmitting the information by setting an arbitrary position in a sentence using a sentence slider bar; when the user opens a file in the user graphical interface, Input text analysis means for receiving information, analyzing sentences, syllables, and words, and transmitting the result; Receives and uses sentence analysis information from the input text analysis means, analyzes a user input command input from the user graphic interface means, informs the user of the result through the user graphic interface means, and transmits the analyzed user input information. User input analyzing means; And converts the text into speech using analyzed user input information input from the user input analyzing means, and informs the user of the sentence and the word or syllable information in the sentence that are currently output as synthesized sounds through the user graphic interface means. Synthesis means for outputting to an output device.

On the other hand, the method of the present invention, a text / speech conversion method applied to the text / speech converter, comprising: a first step of opening a file selected for operating the synthesizer; A second step of selecting a synthesis start point using a trick mode of a sentence of the selected file; And a third step of synthesizing the corresponding sentences and converting them into speech.

In addition, the method of the present invention comprises: a fourth step of reading a corresponding sentence when the file slider is changed, grasping a synthesis start point after the sentence slider is set, and outputting the selected sentence; And a fifth step of identifying the synthesis starting point and outputting the selected sentence when there is a change in the sentence slider.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a system to which the present invention is applied, in which 101 is a user input device, 102 is a central processing unit, 103 is a text / voice converter, 104 is a digital / analog converter, 105 Indicates an output device, respectively.

The user input device 101 receives a user's input through a keyboard and a mouse and transmits the input to the central processing unit 102.

The central processing unit 102 is a device that performs appropriate processing according to a user's request, and is a part where the algorithm of the present invention is mounted and performed.

Text-to-speech converter 103 is a device that receives a text as an input and converts it to voice.

The digital / analog (D / A) converter 104 converts the synthesized digital data into an analog signal and outputs it to the outside.

The output device 105 displays the contents and the results currently performed by the central processing unit, and provides the information necessary for the user.

2 is a block diagram of a text / voice converter using a trick mode according to the present invention, 210 is an input text analysis module, 220 is a user input analysis module, 230 is a graphic user interface module, 240 is a synthesizer module, and 250 is a speaker. Represent each.

The present invention basically includes a language processing unit 241, a rhyme processing unit 242, a signal processing unit 243, and a synthesis unit database 244 of an existing text / voice converter, an input text analysis module 210, and user input analysis. The module 220, a graphic user interface module 230, and an image output device 105 are further provided.

The graphic user interface module 230 sets a file by using a file button to start synthesis, and transmits the sentence, syllable, and word count information of the set file to the input text analysis module 210, and the file in the trick mode. Use the slider bar to set the position of an arbitrary sentence, and use the sentence slider bar to set the position within a sentence to determine the sentence you want to start synthesizing in the given text and the position of the starting word and syllable within the selected sentence. It makes a selection, and transmits the information to the user input analysis module 220.

The input text analysis module 210 analyzes the input text to find out the composition of the text. The input text analysis module 210 determines how many sentences the input text data is composed of, and each sentence is represented by how many syllables and words. It determines whether it is configured and delivers the result to the user input analysis module 220.

The user input analysis module 220 uses sentence analysis information of the input text analysis module 210 and analyzes a user input command in the user graphic interface module 230 to convert text into speech and output a sentence currently being synthesized. The result of the word or syllable location information in the sentence and the sentence is notified to the user again through the user graphic interface module 230, and when controlling the synthesizer module 240, the text data and various information necessary for driving the synthesizer module 240 In the case of driving the synthesizer, the synthesizer module uses the synthesized start sentence selected by the user through the graphical user interface and the start point information of the sentence, and synthesizes the necessary information such as the selected sentence, the start position in the sentence, and the gender and the voice speed. And passes the analyzed user input information to the synthesizer. Send to module 240.

The synthesizer module 240 processes the text to convert the text into speech using the information in the user input analysis module 220 and informs the user of the location of the text through the user graphic interface module 230. . In addition, the lip shape corresponding to the phoneme being synthesized is transmitted to the graphic user interface module 230 to enable lip animation.

To implement this feature, two slide bars (317, 318) for each sentence and position within the sentence and two radio buttons for selecting whether the position within the selected sentence should be represented by syllables or words. 326 is used. In addition, the user interface module allows a user to select various functions related to the synthesizer, perform a desired function, and view the result.

The speaker 250 outputs the synthesized voice from the synthesizer module.

3 is an exemplary diagram of a graphical user interface using a trick mode to which the present invention is applied, where 310 is a trick mode control window, 320 is an additional option window, 330 is a lip animation window, and 340 is a text output window.

The configuration of the graphical user interface is a graphical user interface module that represents the output of the central processing unit, and accepts all inputs of the user and shows the result of the processing.

First, the trick mode control window 310 is a control button such as a video player play 311, pause 312, resume 313, stop 314, forward ( forward 315, and backward 316 buttons.

Play 311 drives the synthesizer, pause 312 pauses the synthesizer, resume 313 drives the synthesizer in its paused state from that position, and stops. Stop 314 terminates the synthesizer, backward 315 moves the sentence to be synthesized to the previous sentence, and forward 316 moves the sentence to be synthesized to the next sentence.

In addition, the trick mode control window 310 informs the user of the start position of the synthesizer by indicating how many sentences the text to be used for the current composition and how many syllables or words are included in each sentence, A file slider bar 317 indicating the number of sentences and a sentence slider bar 318 indicating the structure of the sentence selected from the file slider bar are provided to inform the currently synthesized sentence and the position within the sentence. You can use these two slider bars to select the sentence you want to combine and to precisely adjust the starting position within the selected sentence. Then, when the slider bar is moved to select a sentence for composition, a sentence corresponding to the position is displayed on the text output window 340.

The additional option window 320 may include a file open button 321 and a file close button 322 for opening and closing a sentence to be synthesized, an on / off button 323 for lip animation, There is a male / female sound selection button 324, a synthesized speech rate selection slider bar 325, and a start unit selection button 326 that allow selection of items necessary for operating the synthesizer.

When the selection of the start unit selection button 326 in the sentence is changed, the display of the sentence slider bar 318 of the trick mode control window 310 is changed accordingly.

The lip animation window 330 shows a lip-shaped picture made by standard patterning of human lip movements during speech in consideration of the phoneme environment of what is being synthesized in the synthesizer, thereby giving an effect as if a person is talking. This lip movement is linked to the lip animation on / off button 323 of the additional option window 320.

The text output window 340 shows text data that is currently prepared for compositing. The text output window 340 includes a full sentence window showing the entire text and a current sentence window showing only the sentences synthesized by the current synthesizer. Is linked with the file slider bar 317 of the trick mode control window 310, and the backward 315 and forward 316 buttons.

4A and 4B are flowcharts of a text / voice conversion method using a trick mode according to the present invention, and show a flow of performance of a corresponding function when a user controls a graphical user interface using a keyboard / mouse as a user input device. .

First, it checks whether there is a user input (401) and performs a corresponding function. Here, two variables, a play identification flag (Flag_Play) and a file identification flag (Flag_File), are used. The play identification flag indicates whether the synthesizer is currently operating and the file identification flag indicates whether a target file to be synthesized exists.

In order for the synthesizer to work, you must first open the target file for synthesis. When the user selects Open File (402), the user opens the desired file (413), changes the file identification flag indicating whether the target file to be synthesized to 1 (414), and then inputs the contents of the file into the input text analysis module (415). Send to. The input text analysis module 415 analyzes how many sentences the input text is composed of and how many words and how many words each sentence is, and then initializes the file slider bar and sentence slider bar in the trick mode control window. do. When the file is opened, the text output window appears and the first sentence is displayed in the current sentence window.

If the file close 403 is selected, the file currently used is closed, the file identification flag indicating whether the file to be synthesized is changed to 0 (417), and the value already given by the input text analysis module is initialized to 0. (415).

When play 404 is selected, the file slider bar and the sentence slider bar are examined to determine the start point of the synthesis (418), then the synthesis is started (419), and the play identification flag variable indicating whether the synthesizer is operating is changed to 1 (420). At this time, the sex of the synthesized sound is examined through the additional option window (432), and after checking the speech speed (433), the synthesizer is operated through the synthesizer control routine (434) (435).

When the synthesizer starts running, the synthesizer starts synthesizing from the given starting point, and displays the position of the sentence currently synthesized by the synthesizer as a change of the sentence slider bar in the trick mode control window (436), and also the lip sync option. In the on state (437), a standard patterned lip shape considering the phoneme environment is displayed on the lip animation window (438).

After the synthesis of the given sentence is finished, if there is no user request (401), the file slider bar is automatically changed to synthesize the next sentence (440). If the just synthesized sentence was the last sentence in the file, the file slider bar changes the play identification flag to 0 to indicate that there will be no further change 441 and that the synthesis is all finished (442).

When the position of the file slider bar is changed, the sentence at the position is read (427), the sentence slider bar is changed according to the selected sentence (428), and the currently selected sentence is displayed in the text output window (430). The file slider bar and the sentence slider bar are automatically changed during synthesizer operation, and the user can change it to select the starting point of the synthesis before the synthesizer operation. Adjust the processing flow for (431).

If the synthesizer was in operation, the synthesizer continued to synthesize the newly changed sentence through the male / female test 432 and the speech rate test 433.

In order to maintain the same rhyme pattern as when a user synthesizes a whole sentence when the starting point of the synthesis is not the beginning of the sentence, the synthesizer first calculates a rhyme parameter for the entire sentence and then generates a synthesis sound from the starting point in the given sentence. .

When pause 406 is selected, the synthesizer is temporarily stopped, and when resume 405 is selected, synthesis resumes from the current standing position. When the stop 407 is selected, the operation of the synthesizer is terminated. When the forward 408 is selected, the sentence to be synthesized is moved to the next one, and when the backward 409 is selected, the synthesizer is synthesized. Move a sentence to the previous one. The selection of forward 408 and backward 409 acts as a direct move of the file slider bar, but the forwarder 408 or backward 409 is disabled while the synthesizer is running. Once selected, stop the synthesis once as stop 407 is selected and then move the file slider bar.

When the file slider bar for selecting a sentence or the sentence slider bar for selecting a position in a sentence is moved without manipulation of various buttons, the composition starts to be detected by detecting a position change 410 of the file slider bar or a position change 411 of the sentence slider bar. The starting point is calculated and the selected sentence is displayed in the text output window.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the spirit of the present invention for those skilled in the art to which the present invention pertains, and the above-described embodiments and accompanying It is not limited to drawing

According to the present invention, the user can set the position of a sentence, a syllable or a word in a file using a graphic user interface including a file slider bar, a sentence slider bar, and a trick mode of a screen output device. By starting the voice conversion, the following effects can be obtained.

First, through the graphical interface, the user can easily understand the composition of the text of how many sentences and text each syllable is composed of the text to be synthesized.

Second, through the graphical interface, it is possible to easily set the position to start synthesis through the sentence selection and the words and syllables in the sentence.When the synthesizer is in operation, the sentence currently synthesized and the position in the sentence Tells the user exactly.

Third, even if the starting position of the synthesis set by the user is any point in the sentence other than the beginning of the sentence, it creates a natural synthesis sound while maintaining the rhyme pattern for the whole sentence.

Claims (10)

To start compositing, set the file using the file button, transmit the sentence, syllable and word count information of the set file, set the arbitrary sentence position using the file slider bar in trick mode, and use the sentence slider bar. User graphic interface means for setting an arbitrary position in the sentence and transmitting the information; Input text analysis means for receiving all information of sentences contained in the file when the user opens the file in the user graphic interface, analyzing the sentences, syllables, and words, and transmitting the result; Receives and uses sentence analysis information from the input text analysis means, analyzes a user input command input from the user graphic interface means, informs the user of the result through the user graphic interface means, and transmits the analyzed user input information. User input analyzing means; And The text is converted into speech using analyzed user input information input from the user input analyzing means, and the user is informed of the sentence and the word or syllable position information in the sentence and the speech output as the current synthesized sound through the user graphic interface means. Text-to-speech converter including synthesizing means for output to output The method of claim 1, The synthesis means, After calculating the prosody parameter for the entire sentence, the user generates the synthesized sound from the starting point in the sentence designated by the user through the graphic user interface means, and thus synthesizes the synthesized sound from the beginning of the sentence even if it generates the synthesized sound from any position in the sentence. Text to speech converter characterized by having a rhyme. The method according to claim 1 or 2, The graphical user interface means, A trick mode for a user to designate an arbitrary sentence position using a selection button; A file slider bar that precomputes the total number of sentences in the file using sentence punctuation and informs the user when a target file of text / voice conversion is opened in order to allow the user to easily specify a starting point of the synthesis; And A text / voice converter comprising a sentence slider bar for calculating a syllable number, which is a unit of position movement within a sentence, to inform the user in advance of the sentence to be synthesized by the user. The method of claim 3, wherein The graphical user interface means, And text input means for allowing a user to specify an arbitrary position of a sentence to be synthesized and an arbitrary position in a sentence based on the total sentence number information in the file and the total syllable information in the sentence displayed on the screen. / Voice converter The method according to claim 1 or 2, The graphical user interface means, A trick mode for a user to designate an arbitrary sentence position using a selection button; A file slider bar that precomputes the total number of sentences in the file using sentence punctuation and informs the user when the target file of text / voice conversion is opened so that the user can easily specify the starting point of the synthesis; And A text / voice converter, comprising: a sentence slider bar for pre-calculating the number of words that are position movement units within a sentence specified by the user in a sentence to be synthesized by the user and informing the user of the sentence number; The method of claim 5, The graphical user interface means, The apparatus further comprises input means for allowing a user to specify an arbitrary position of a sentence to be synthesized and an arbitrary position in a sentence based on the total number of sentence information in the file and the total number of words in the sentence displayed on the screen. / Voice converter In the text / voice conversion method applied to the text / voice converter, A first step of opening the selected file to operate the synthesizer; A second step of selecting a synthesis start point using a trick mode of a sentence of the selected file; And A text / voice conversion method comprising a third step of synthesizing a corresponding sentence into a voice The method of claim 7, wherein A fourth step of reading a corresponding sentence when the file slider is changed, identifying a synthesis start point, and outputting the selected sentence after the sentence slider is set; And And a fifth step of identifying the synthesis starting point and outputting the selected sentence when there is a change in the sentence slider. The method according to claim 7 or 8, The second step, Selecting play to determine synthesis starting point and then starting synthesis; Selecting synthesis to restart synthesis; Stopping the synthesis for a moment when selecting a pause; Selecting synthesis stops the synthesis; Selecting forward to move the object to be synthesized to the previous sentence; And Selecting backward, the object to be synthesized comprises the step of moving to the back sentence text / voice conversion method comprising the The method of claim 9, The third step, In order to control the lip animation window of the graphic user interface, gender is examined, voice speed is examined, and synthesized through a synthesizer control routine, when a human lip movement is standardized and converted into an image form. Text to speech conversion method characterized in that output to the user as a lip animation using the conversion rule between

Images