JPH06131148A - Speech synthesizing device - Google Patents

Abstract

PURPOSE:To provide a speech synthesizing device in which time required for outputting a speech can be confirmed easily in respect of the speech synthesizing device to output an inputted sentence by making it into a voice. CONSTITUTION:A means 10 to generate a phonetic character string from the inputted sentence, the means 12 to generate a vocal sound string obtained by converting each phonetic character of the generated phonetic character string into a vocal sound, the means 14 to calculate total time required for the sound output of the generated vocal sound string, the means 16 to output the calculated total time, and the means 18 to output the generated vocal sound string by converting it into a voice signal are provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice synthesizing device for converting input text into voice and outputting it.

This type of device is used in fields such as reading of prepared texts during proofreading, presentation of products at exhibitions and stores, various voice guidance, and education.

[0003]

2. Description of the Related Art A speech synthesizer is connected to a computer serving as its host, and text data is given to the speech synthesizer from the host side.

When the voice signal output processing is started in the voice synthesizer, a busy signal is output from the voice synthesizer to the computer, and the termination of the voice synthesize operation is confirmed in the host computer from the output stop of the busy signal.

[0005]

In the case of a presentation, since voice output should be performed in accordance with the display of sentences, graphs, and animations, actual document data is given to the voice synthesizer to output a busy signal. The time to stop the sound is measured, and the work of editing the text of the voice output and changing the speed of the voice output is repeated in consideration of the measured time. Therefore, this work requires labor and time. Was there.

The present invention has been made in view of the above conventional problems, and an object thereof is to provide a voice synthesizing device capable of easily confirming the time required for voice output.

[0007]

In order to achieve the above object, the apparatus according to the first invention is configured as shown in FIG. 1, and the apparatus shown in the figure generates a phonetic character string from an input sentence. The pronunciation character string generating means 10, the pronunciation character string analyzing means 12 for analyzing the generated pronunciation character string, and the sentence reading time calculating means for calculating the total time required for voice output of the pronunciation character string based on the analysis result. 14, a sentence reading time output means 16 for outputting the calculated total time, and a voice signal output means 18 for converting the generated phonetic character string into a voice signal and outputting the voice signal.

The apparatus according to the second invention is configured as shown in FIG. 2, and has a pronunciation character string generating means 10 for generating a pronunciation character string from an input sentence and an analysis of the generated pronunciation character string. The pronunciation character string analysis means 12, the syllable reading time calculation means 20 for converting the pronunciation character string into a syllable string and calculating the time required for the voice output of each syllable based on the analysis result, and each calculated time correspond Syllable reading time output means 22 for outputting corresponding syllables, and voice signal output means 1 for converting the generated syllable string into a voice signal and outputting the voice signal.
8 and.

Further, the apparatus according to the third invention is configured as shown in FIG. 3, and the apparatus shown in the figure has a sounding character string generating means 10 for generating a sounding character string from an inputted sentence, and the generated sound character string generating means 10. Pronunciation character string analysis means 12 for analyzing a pronunciation character string
And a phoneme reading time calculating means 24 for converting a phonetic character string into a phoneme string and calculating a time required for voice output of each pronunciation based on the analysis result, and each calculated time corresponding to a corresponding phoneme. Phonological reading time output means 26 for outputting
And a voice signal output means 18 for converting the generated phoneme sequence into a voice signal and outputting the voice signal.

The apparatus according to the fourth aspect of the invention is configured as shown in FIG. 4, and the apparatus shown in the figure has a sounding character string analyzing means 12 for analyzing the inputted sounding character string and a sound of the sounding character string. Sentence reading time calculation means 14 for calculating the total time required for output based on the analysis result, sentence reading time output means 16 for outputting the calculated total time, and converting the input pronunciation character string into a voice signal. And an audio signal output means 18 for outputting the output.

The apparatus according to the fifth aspect of the present invention is configured as shown in FIG. 5, and the apparatus shown in the figure is based on a pronunciation character string analysis means 12 for analyzing an inputted pronunciation character string and the analysis result. , A syllable reading time calculating means 20 for converting a pronunciation character string into a syllable string and calculating a time required for voice output of each syllable, and a syllable reading time outputting means for outputting each calculated time in association with the corresponding syllable. 22 and a voice signal output means 1 for converting the generated syllable string into a voice signal and outputting the voice signal.
8 and.

The device according to the sixth aspect of the invention is configured as shown in FIG. 6, and the device shown in FIG. 6 has a phonetic character string analysis means 12 for analyzing the input phonetic character string and a phoneme for the phonetic character string. Phoneme reading time calculating means 24 for converting the time into a string and calculating the time required for outputting each phoneme based on the analysis result, and phoneme reading time outputting means for outputting each calculated time in association with the corresponding phoneme. 26, and a voice signal output means 18 for converting the generated phoneme sequence into a voice signal and outputting the voice signal.

[0013]

In the first and fourth inventions, the total time required for voice output is output, in the second and fifth inventions, the time required for voice output of each syllable is output, and in the third and sixth inventions. The time required for voice output of each phoneme is output.

However, in the first invention, the second invention, and the third invention, the pronunciation character string is generated by inputting a sentence such as an input kanji / kana mixed sentence, and the fourth invention, the fifth invention, and the sixth invention. In, the phonetic character string is input.

The pronunciation character string can be composed of information about phonemes (phonemes, syllables, etc.) and information about accents and intonations (prosodic information), and the phoneme contains only phonemes or phonemes and prosody.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 7 shows the configuration of a system used in a presentation, and FIG. 8 shows an example of the use of the system with a flowchart.

Further, FIG. 9 describes the processing contents of the voice synthesis with the use of a flowchart, and FIG. 10 describes the processing contents with respect to the sentence reading time output using the flow chart.

Further, FIG. 11 illustrates the required time output operation in the second mode and the third mode. In the first mode in which the user is notified of the total time required for voice output, the time required for voice output of each syllable. Second mode for notifying the user, and a third mode for notifying the time required for the voice output of each phoneme (phoneme)
You can specify the mode of.

In FIG. 7, the host computer has a main body 40, a keyboard 42 (used for voice output instruction, voice output speed designation, voice output required time notification instruction, required time notification mode designation), display 44 ( It is used to display the time required for audio output).

Then, the content of the sentence to be voice output, the voice output instruction, the voice output required time notification instruction, and the data regarding the mode designation of the required time notification are sent from the computer main body 40 in accordance with the keyboard operation, and the voice synthesizer operates. It is provided to the interface unit 50.

When these are given to the interface section 50, the chapter data of the sentence to be output as voice is inputted to the sentence analysis section 52 and converted into pronunciation character string data using the contents of the word dictionary memory 54.

Further, the speech production speed setting unit 58 of the voice processing unit 56.
Then, the process of setting the utterance speed to the one specified by the user given to the interface unit 50 is performed, and the default speed is selected when the user does not specify.

Further, the phoneme length calculation unit 6 of the voice processing unit 56.
At 0, a process of generating a phoneme string in which each phonetic character of a phonetic character string is converted into a phoneme by using the stored contents of the voice unit memory 62 and the output (phonetic speed) of the phonetic speed selection unit 58, and the voice of each phoneme Calculation processing of the time required for output is performed.

At this time, when the voice output instruction is given from the host side, the spectrum parameter generating unit 64, the pitch parameter generating unit 66 of the voice processing unit 56,
The waveform synthesizing unit 68 is activated to convert the phoneme sequence of the phoneme length calculating unit 60 into a voice signal, and the voice signal is output to the speaker 72 via the amplifier 70.

When the host gives an instruction to notify the required time for voice output and an instruction to specify the required time notification mode, the output of the phoneme length calculator 60 is output to the reading time calculator 74. It

The reading time calculation unit 74 notifies the host of the total time required for voice output when the first mode is designated. When the second mode is designated, each syllable and its voice output are processed. A process of notifying the host side of the contents corresponding to the required time, and a process of notifying the host side of the contents corresponding to each phoneme and the required time of its voice output when the third mode is designated.

When the above system is used for presentations at academic conferences and the like, the user operates the host computer in advance to edit the text data of the manuscript, and outputs it to the speech synthesizer (step 500 in FIG. 8).

If the default speaking rate is not selected, the speaking rate is designated (steps 502 and 50 in FIG. 8).
4) Instructing the notification of the voice output required time (step 506 in FIG. 8), and instructing any one of the above-described first mode, second mode, and third mode (step 508 in FIG. 8).

Then, the content notified by the voice synthesizer and displayed on the display 44 is confirmed (step 510 in FIG. 8), and if the content is not satisfactory, the original is reedited or the speaking speed is changed (step 512 in FIG. 8). , 514).

Here, in the voice processing unit 56, the processing shown in FIG. 9 (steps 600, 602, 604, 606, 60) is performed.
8,610,612,614,616,618,62
0,622), but the reading time calculation unit 74 performs the processing of FIG. 10 (steps 700, 702, 704, 70).
6, 708) is performed and the first mode is designated,
The total time required for voice output of a sentence is displayed (when a voice output instruction is given, voice output of the sentence is also read out). Therefore, it is possible to easily adjust the voice output time so that the reading of the sentence ends at an appropriate timing in accordance with the switching of the sentence display and the like.

When the second or third mode is designated,
As shown in FIG. 11, a string in which a syllable and its output time are words or a string in which a phoneme and its output time are units is displayed. Therefore, voice output should be performed in accordance with the motion of the animation (for example, the movement of the mouth).
It is possible to finely adjust the audio output time.

[0032]

As described above, according to the present invention, the total time required for voice output, the time required for output of each syllable, and the time required for output of each phoneme can be confirmed. Therefore, switching between sentence display and graph display is possible. ， You can easily adjust the audio output time and timing according to the movement of animation.

Therefore, the preparation work for the presentation can be performed extremely efficiently.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the first invention.

FIG. 2 is a diagram illustrating the principle of the second invention.

FIG. 3 is a diagram illustrating the principle of the third invention.

FIG. 4 is an explanatory view of the principle of the fourth invention.

FIG. 5 is an explanatory view of the principle of the fifth invention.

FIG. 6 is a diagram illustrating the principle of the sixth invention.

FIG. 7 is an explanatory diagram of the configuration of the embodiment.

FIG. 8 is a diagram illustrating a configuration of a usage example of a system by a user.

FIG. 9 is a flowchart illustrating processing contents of a voice processing unit.

FIG. 10 is a flowchart illustrating processing contents of a reading time calculation unit.

FIG. 11 is an explanatory diagram of a second mode and a third mode.

[Explanation of symbols]

 40 computer body 42 keyboard 44 display 50 interface unit 52 sentence analysis unit 54 word dictionary memory 56 speech processing unit 60 phoneme length calculation unit 62 speech unit memory 64 spectrum parameter generation unit 66 pitch parameter generation unit 68 waveform synthesis unit 70 amplifier 72 speaker 74 Reading time calculator

Claims (6)

[Claims] 1. A pronunciation character string generation means (10) for generating a pronunciation character string from an input sentence, a pronunciation character string analysis means (12) for analyzing the generated pronunciation character string, and a sound of the pronunciation character string. Sentence reading time calculating means (14) for calculating the total time required for output based on the analysis result
A text reading time output means (16) for outputting the calculated total time, and a voice signal output means (18) for converting the generated pronunciation character string into a voice signal and outputting the voice signal. Voice synthesizer. 2. A pronunciation character string generation means (10) for generating a pronunciation character string from an input sentence, a pronunciation character string analysis means (12) for analyzing the generated pronunciation character string, and based on the analysis result. , A syllable reading time calculating means (20) for converting a pronunciation character string into a syllable string and calculating a time required for voice output of each syllable, and a syllable reading time for outputting each calculated time in association with the corresponding syllable. A voice synthesizer comprising: an output unit (22); and a voice signal output unit (18) for converting the generated syllable string into a voice signal and outputting the voice signal. 3. A pronunciation character string generation means (10) for generating a pronunciation character string from an input sentence, a pronunciation character string analysis means (12) for analyzing the generated pronunciation character string, and a phoneme for a pronunciation character string. Phoneme reading time calculating means (24) for converting the time into a sequence and calculating the time required for the voice output of each phoneme based on the analysis result, and the phoneme reading time for outputting each calculated time in association with the corresponding phoneme. A voice synthesizer comprising: an output unit (26); and a voice signal output unit (18) for converting the generated phoneme sequence into a voice signal and outputting the voice signal. 4. A pronunciation character string analysis means (12) for analyzing an input pronunciation character string, and a sentence reading time calculation means (14) for calculating a total time required for voice output of the pronunciation character string based on the analysis result. )
A text reading time output means (16) for outputting the calculated total time, and a voice signal output means (18) for converting the input pronunciation character string into a voice signal and outputting the voice signal. Voice synthesizer. 5. A phonetic character string analysis means (12) for analyzing an input phonetic character string, and a time required for converting the phonetic character string into a syllable string and outputting voice of each syllable based on the analysis result. Syllable reading time calculation means (20), syllable reading time output means (22) for outputting each calculated time in association with the corresponding syllable, and converting the generated syllable string into a voice signal for output. A voice synthesizer, comprising: a voice signal output means (18). 6. A phonetic character string analysis means (12) for analyzing an input phonetic character string, and a time required for converting the phonetic character string into a phoneme string and outputting voice of each phoneme based on the analysis result. A phoneme reading time calculating means (24), a phoneme reading time outputting means (26) for outputting each calculated time in association with the corresponding phoneme, and a generated phoneme sequence converted to a voice signal for output. A voice synthesizer, comprising: a voice signal output means (18).