JPH06149284A - Text speech synthesizing device - Google Patents

Abstract

PURPOSE:To vary the synthesizing speed of speech synthesis without using any special symbol and vary the synthesizing speed in real time halfway during the speech synthesis of a document after a text analyzing process and a synthetic parameter generating process. CONSTITUTION:The text speech synthesizing device has a speed setting switch 10 for setting the speed of speech synthesis, a mode changeover switch 20 for specifying which of an internal set value and the set value set with the mode setting switch 10 is employed as the speed set value of speech synthesis, and a speed parameter buffer part 30 in which the values set with the speed setting switch 10 and mode changeover switch 20 are stored; and the setting of the speed of speech synthesis is controlled with the set values in the speed parameter buffer part 30.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a text-to-speech synthesizer for synthesizing text into speech.

[0002]

2. Description of the Related Art Heretofore, there has been known a voice synthesizing device for synthesizing text-form words into voice. A conventional text-to-speech synthesizer will be described below by taking Japanese speech synthesis as an example. FIG. 9 is a block diagram of conventional text-to-speech processing.

In FIG. 9, A0 is a Japanese text, A
Reference numeral 1 is word concatenation information, A2 is text analysis, A3 is a Japanese word dictionary, B1 phonological discipline addition, C1 is waveform synthesis, C2 is a waveform dictionary, and C3 is synthesized speech. The speech synthesis is composed of three stages of text analysis processing A, synthesis parameter generation processing B and speech synthesis processing C. Japanese text A0 is synthesized through text analysis processing A, synthesis parameter generation processing B and speech synthesis processing C. The sound C3 is output to the outside.

The Japanese text A0 is a character string and is input to the text-to-speech synthesizer via an input device (not shown). The text analysis processing A performs analysis processing with reference to the word concatenation information A1 of the character string of the input Japanese text A0 and the Japanese word dictionary A3. In the text analysis process A, for example, a word or a phrase is identified, the reading of the word or the phrase is added, the accent of the phrase is determined, and rendaku or phonophore processing is performed.

Next, in the synthesis parameter generation process B, various parameters such as pitch and power when synthesizing a voice are determined. This synthesis parameter generation process B is performed by prosody rule assignment B1. Prosody rule addition B1
Performs the unit determination, the insertion of a pause, the phoneme duration control, the amplitude control, the pitch control, etc. when performing voice synthesis according to the processing result of the text analysis A2.

Finally, in the voice synthesizing process C, a synthetic sound is generated. This speech synthesis C is the waveform synthesis C1 according to the processing of the synthesis parameter generation processing B in the waveform dictionary C.
With reference to 2, the waveform of the synthesized voice C3 is generated. Next, the configuration of the conventional text-to-speech synthesizer will be described with reference to the block diagrams of the conventional text-to-speech synthesizers shown in FIGS.

In FIG. 10, 1 is a character character output unit, 2 is a character character input unit, 3 is an internal bus, 4 is a control unit, 5 is a dictionary storage unit, 6 is an internal buffer, 7 is a voice synthesis unit, and 8 is The voice output unit 40 is a higher-level device. The upper device 40 is connected to the text-to-speech synthesizer and has the character / character output unit 1 as one of its functions. The character / character output unit 1 is a port for sending out character characters. Therefore, the text to be voice-synthesized is input from the character / character output unit 1 of the host device 40 to the character / character input unit 2 of the text-to-speech synthesizer.

When the character / character input unit 2 in the text-to-speech synthesizer receives the character / character sent from the host device 40, it sends the character / character to the control unit 4 through the internal bus 3. The control unit 4 executes the text analysis process A and the synthesis parameter generation process B while accessing the dictionary storage unit 5 and the internal buffer 6. The generated synthesis parameter is temporarily stored in the internal buffer 6 and then sent to the voice synthesis unit 7.

The voice synthesizing unit 7 performs a voice synthesizing process C and outputs a voice from the voice output unit 8. Functions of a general text-to-speech synthesizer include switching of voice quality of male and female voices, change of voice synthesis speed, change of pitch, change of accent size, and change of volume. Character characters input to drive these functions are defined as special character strings. Then, this special character string is generally used by embedding it in a sentence in a host device.

The change of the characteristic of the conventional speech synthesis by the special character string will be described below by taking the speech synthesis speed as an example. FIG. 11 is an example of a sentence that changes the voice synthesis speed using a conventional special character string. In FIG. 11, the character string enclosed by the special symbols {} defines a special character string. For example, the special character string represented by {Sφ} 51 indicates the designation of the voice synthesis rate by S, and the subsequent φ indicates the value of the voice synthesis rate. And this special character string {Sφ}
"AIUEOKAKUKUKEKO" 52 following 51 indicates a sentence to be speech-synthesized.

Further, the special character string {S1} 53 is used for designating the composition speed similarly to the special character string {Sφ} 51. The composition speed of {S1} 53 is greater than the composition speed of {Sφ} 51. Can also indicate that the speed is fast. Note that the sentence of "Aiueo Kaki Kuke." 54 following this special character string {S1} 53 represents a sentence to be synthesized at the voice synthesis speed determined by this special character string {S1} 53.

FIG. 12 is a diagram showing the passage of time when the sentence of FIG. 11 is speech-synthesized. "Aiue Okaki Kukeko" 52 following {Sφ} 51 is synthesized more slowly than "Aiue Okaki Kukeko." 54 following {S1} 53. Here, the description has been given by taking the speed of voice synthesis as an example, but not only the speed of voice synthesis but also other volume and voice quality can be designated by special character strings in the same manner.

Conventionally, these processes are realized by performing control in accordance with the contents of the special character example in the synthesis parameter generation process B.

[0014]

However, the above-mentioned conventional text-to-speech synthesizer has the following problems. The problems will be described below with reference to FIGS. 9 to 11. (1) In the conventional text-to-speech synthesizer, since the Japanese text A0 sentence must be surrounded by a special symbol as shown in {Sφ} 51, the Japanese text A0 already sent to the text-speech synthesizer. For, you cannot change the synthesis speed.

(2) Further, since the text analysis processing A and the synthesis parameter generation processing B in the text-to-speech synthesizer are high-speed processing, while the voice synthesis processing C is low in speed, the Japanese text is output when the voice output starts. A0 ends the text analysis process A and the synthesis parameter generation process B, and may be stored in the internal buffer 6. Therefore, when the synthesized voice is heard, a special character string for speed change is sent to the text-to-speech synthesizer. But it doesn't change. (3) When the input is too long to fit in the internal buffer 6, the combination speed does not change until all the internal buffers 6 are combined. There is a big time difference.

The present invention solves the above-mentioned conventional problems and can change the synthesis speed of speech synthesis without using a special symbol, and a sentence for which text analysis processing and synthesis parameter generation processing have been completed. Another object of the present invention is to provide a text-to-speech synthesizer having an excellent man-machine interface that can change the synthesis speed in real time from the middle of speech synthesis.

[0017]

In order to achieve the above object, the present invention provides a text-to-speech synthesizer with a speed setting switch for setting the speed of speech synthesis, and a speed setting value of speech synthesis as an internal setting value. Mode switch that specifies the speed mode to be set with the speed setting switch or the speed setting switch, and the speed parameter buffer that stores the value set by the speed setting switch and mode switching switch. The setting value controls the setting of the speed of voice synthesis.

[0018]

The present invention uses the speed parameter buffer different from the internal buffer that stores the duration control parameter, which is one of the results of the synthesis parameter generation process sent to the waveform synthesis processing unit. By providing a changeover switch for switching, it is possible to change the speed of speech synthesis in real time even in the middle of speech synthesis even for a sentence for which text analysis processing and synthesis parameter generation processing have been completed.

As a result, when changing the speed of speech synthesis of Japanese text which is about to be speech-synthesized, it is necessary to use a special symbol such as {Sφ} as in the conventional case, or a special symbol is used. It is possible to eliminate the problem that the speech synthesis speed cannot be changed in real time during speech synthesis for unwritten sentences.

[0020]

Embodiments of the present invention will now be described in detail with reference to the drawings. The configuration of the text-to-speech synthesizer of the present invention will be described with reference to the block diagram of the text-to-speech synthesizer of the present invention shown in FIG. In FIG. 1, 1 is a character character output unit, 2 is a character character input unit, 3 is an internal bus, 4 is a control unit, 5 is a dictionary storage unit, 6 is an internal buffer,
Reference numeral 7 is a voice synthesis unit, 8 is a voice output unit, 10 is a speed setting switch, 20 is a mode changeover switch, 30 is a speed parameter buffer unit, and 40 is a host device.

The host device 40 is connected to the text-to-speech synthesizer and has the character / character output unit 1 as one of its functions. The character / character output unit 1 is a port for sending out character characters.
The text to be voice-synthesized is input from the character / character output unit 1 of the host device 40 to the character / character input unit 2 of the text-to-speech synthesizer.

When the character / character input unit 2 in the text-to-speech synthesizer receives the character / character sent from the higher-level device 40, it sends the character / character to the control unit 4 through the internal bus 3. Then, the control unit 4 executes the text analysis process and the synthesis parameter generation process while accessing the dictionary storage unit 5 and the internal buffer 6.

The generated synthesis parameter is temporarily stored in the internal buffer 6 and then sent to the voice synthesis unit 7. The voice synthesizing unit 7 performs a voice synthesizing process and outputs a voice from the voice output unit 8. Further, the text-to-speech synthesizer of the present invention has a speed setting switch 10, a mode changeover switch 20 and a speed parameter buffer section 30 in addition to the conventional text-to-speech synthesizer.

The text-to-speech synthesizer of the present invention is constructed so that two kinds of speed modes for speech synthesis can be selected, a soft mode speed and a hard mode speed. The speed of this soft mode is set in the synthesis parameter generation process, and the set value of the speed is stored in the internal buffer 6. The speed in the hard mode is set by the speed setting switch 10, and the set value of the speed is stored in the speed parameter buffer unit 30. Further, the mode changeover switch 20 designates whether to use the speed parameter generated by the synthesis parameter generation processing in the internal buffer 6 or the speed parameter set in the speed parameter buffer unit 30 as the stage of the speed of voice synthesis. It is a switch to do.

The mode using the speed parameter of the internal buffer 6 is called a soft mode, and the mode using the speed parameter of the speed parameter buffer section 30 is called a hard mode. The speed parameter buffer unit 30 has a speed parameter buffer 31 shown in FIG. 2, and for example, 4 of bit3 to bit0.
Composed of bits. In FIG. 2, each bit of the speed parameter buffer unit 30 indicates the speed mode 3
3 and the speed setting value 32, for example, the speed mode 33 is assigned to the most significant bit bit3, and the speed setting value 32 is assigned to the lower bits bit2 to bit0 and stored.

Here, for example, when the bit value of the speed mode 33 is "1", it corresponds to the hard mode, and when the bit value of the speed mode 33 is "0", it corresponds to the soft mode. Also, 3 of speed setting value 32
It is possible to set and store the voice synthesis speed value of 8 stages of 0 to 7 by the bit. Next, referring to FIGS. 1 and 3 showing the flow of changing the speed of speech synthesis of the text-to-speech synthesizer of the present invention, the timing of changing the speed of speech synthesis and the synthesis speed and speed parameter buffer unit 30 at that time. The state will be explained.

The speed of the voice synthesis is changed when the mode change switch 20 is turned ON / OFF, and when the mode change switch 20 is ON and the speed setting switch 1 is set.
This is done by inserting an interrupt into the control unit 4 when 0 changes. Then, the control unit 4 accepts the interrupt, and if the interrupt is recognized as a speed setting related interrupt, the speed setting process is started.

In FIG. 3, when the voice synthesizing process is performed with the speed value S in the soft mode as the synthesizing speed, the uppermost bit of the speed parameter buffer 31.
The value of 3 is "0". At this time, when the mode changeover switch 20 is switched, the speed setting process is performed by the interrupt, and the speed is set to the speed set in the hardware mode.

Speed parameter buffer section 3 at this time
The lower bits bit2 to bit0 of 0 are, for example, "a, b,
If "c", the voice synthesis is performed by the speed value H1 corresponding to the set value. Next, when the value of the speed setting switch 10 changes, the speed setting processing is performed by the interrupt processing, and the setting is made in the hardware mode. When the lower bits bit2 to bit0 of the speed parameter buffer unit 30 at this time are set to, for example, "d, e, f", voice synthesis is performed by the speed value H2 corresponding to the set value.

FIG. 1 shows the case where the interrupt is performed.
4 and the flowchart of the interrupt of FIG. Step S1: The control unit 4 detects whether or not the mode changeover switch 20 is switched in the state where the voice synthesis is performed at the synthesis speed set by the soft mode or the hard mode, and the mode changeover switch 20 is detected. When the switching of is detected, it is determined that there is a change in the composition speed, and the process proceeds to interrupt processing S4.

Step S2: Next, in order to detect the case where the speed of voice synthesis is changed by changing the speed setting in the hardware mode, the mode changeover switch 2
It is detected whether or not there is a change in the speed setting switch 10 when the switch of 0 has not been performed but in the state of the hard mode. At this time, the switch state of the mode selector switch 20 is set to the hard mode when the switch is ON, and the switch is OFF.
When, it is assumed that the soft mode is set.

Therefore, when the mode changeover switch 20 is ON, the process proceeds to step S3, and when it is OFF, the process returns to step S1 to wait for the speed change of the voice synthesis speed. Step S3: If it is detected that the speed setting switch 10 has changed, it is determined that the composite speed has been changed, and the process proceeds to the interrupt processing S4 to set the speed setting switch 1
If there is no change in 0, it is considered that the synthesis speed has been changed, and the process returns to step S1 to wait for the speed change of the voice synthesis speed.

Step S4: The control unit 4 performs an interrupt process to change the speed of voice synthesis. When the interrupt processing is accepted, the speed setting processing is performed next. Next, the speed setting process will be described with reference to the speed setting flowcharts shown in FIGS.

Step S11: The control section 4 reads the speed parameter stored in the speed parameter buffer 31. Step S12: The bit value of the speed mode 33 of the read speed parameter is checked to determine whether the bit value is "1". As described above, when the bit value of the speed mode 33 is "1", it represents the hard mode, and "0" represents the soft mode. Therefore, when the bit value of the speed mode 33 is "1", it is the soft mode. Is recognized as an interrupt for switching from to hard mode, and the process proceeds to step S13.

Step S13: Speed setting switch 1
Recognizing that 0 has changed, set the speed in hard mode. This speed setting corresponds to the speed parameter buffer 3 for the waveform synthesis C1 of the voice synthesis processing C of FIG.
This is done by setting the speed parameter stored in 1.

The waveform synthesis C1 determines the length of each phoneme for speech synthesis according to the speed parameter. The control unit 4 normally reads the speed parameter in the internal buffer 6 and sets it in the waveform synthesis C1, but in the hard mode, since the speed parameter in the internal buffer 6 is ignored, the speed setting switch 10 is specified. It enables speedy and immediate voice synthesis.

Step S14: In step S12, if the bit value of the speed mode 33 is "0", it means that the mode is the soft mode, and therefore the speed setting process is performed by the speed parameter in the internal buffer 6. Next, the process of changing the speed of voice synthesis will be described with reference to the time charts of the voice synthesizer of the present invention shown in FIGS. 1 and 6 to 8.

First, the case of changing from the soft mode to the hard mode will be described with reference to FIG. When in the soft mode, the mode selector switch 20 turns off.
When the FF is changed to ON and the speed mode in the speed parameter is changed from "0" to "1", the change causes an interrupt to enter the control unit 4 and the speed setting process is performed.

By this speed setting process, the mode is changed from the soft mode to the hard mode, and the voice synthesis speed is changed from the soft mode speed S to the hard mode speed H1. As mentioned above,
The speed S in the soft mode is set by the speed parameter in the internal buffer 6, and the speed H1 in the hard mode is set by the speed setting value 32 in the speed parameter.

Next, the case of changing from the hard mode to the soft mode will be described with reference to FIG. When in the hard mode, the mode switch 20 is turned on.
To OFF, and when the speed mode in the speed parameter changes from "1" to "0", this change causes an interrupt to enter the control unit 4 to perform speed setting processing. In this hardware mode, the speed setting value in the speed parameter is set to, for example, the voice synthesis speed H1.

By this speed setting processing, the mode is changed from the hard mode to the soft mode, and the voice synthesis speed is changed from the hard mode speed H1 to the soft mode speed S. As mentioned above,
The speed S in this soft mode is set by the speed parameter in the internal buffer 6. Next, a case where the speed setting value in the speed parameter is changed in the hard mode state will be described with reference to FIG.

In the hard mode, the mode changeover switch 20 is kept ON, and when the speed mode in the speed parameter remains "1", the speed setting in the speed parameter is set. When the value changes, the change causes an interrupt to enter the control unit 4, and the speed setting process is performed. When the speed setting value in this speed parameter changes the voice synthesis speed from H1 to a value corresponding to H2, the voice synthesis speed becomes H.
Changed from 1 to H2.

Further, in the above description, the case where the speed of voice synthesis of the synthesis parameters is changed is described, but other powers, pitches, voice qualities, etc. determined in the synthesis parameter generation processing are also applicable. . It should be noted that the present invention is not limited to the above-described embodiments, and various modifications can be made based on the spirit of the present invention, and these are not excluded from the scope of the present invention.

[0044]

As described above in detail, according to the present invention, (1) a mode changeover switch and a speed parameter buffer are added to the speed setting switch, and the speed set value is used as an internal buffer according to the speed mode in the control unit. Since it can be specified from either of the speed parameter buffers, it is possible to change the synthesis speed in real time. (2) Further, even when the text analysis process and the synthesis parameter generation process have already been completed and the synthesis parameter is stored in the internal buffer, the voice synthesis speed of the synthesis parameter can be changed. (3) Therefore, improvement of the man-machine interface can be expected.

[Brief description of drawings]

FIG. 1 is a block diagram of a text-to-speech synthesizer of the present invention.

FIG. 2 is a configuration diagram of a speed parameter buffer.

FIG. 3 is a diagram showing a flow of changing the speed of speech synthesis of the text-to-speech synthesizer of the present invention.

FIG. 4 is a flowchart of interruption.

FIG. 5 is a flowchart of speed setting.

FIG. 6 is a time chart of the speech synthesizer of the present invention.

FIG. 7 is a time chart of the speech synthesizer of the present invention.

FIG. 8 is a time chart of the speech synthesizer of the present invention.

FIG. 9 is a block diagram of conventional text-to-speech synthesis processing.

FIG. 10 is a block diagram of a conventional text-to-speech synthesizer.

FIG. 11 is an example of a sentence that changes the voice synthesis speed using a conventional special character string.

FIG. 12 is a diagram showing passage of time when a sentence is speech-synthesized by a conventional device.

[Explanation of symbols]

 1 character character output unit 2 character character input unit 3 internal bus 4 control unit 5 dictionary storage unit 6 internal buffer 7 voice synthesis unit 8 voice output unit 10 speed setting switch 20 mode selection switch 30 speed parameter buffer unit 32 speed setting value 33 speed Mode 40 Upper device

Claims (1)

[Claims] 1. An apparatus for synthesizing voice from text, comprising: (a) a speed setting switch for setting the speed of the voice synthesis; and (b) setting the speed setting value of the voice synthesis as an internal setting value or the speed setting. A mode changeover switch for designating a speed mode as a set value by a switch; (c) a speed parameter buffer for storing a value set by the speed setting switch and the mode changeover switch; A text-to-speech synthesizer characterized in that the setting of the speed of the speech synthesis is controlled by a set value in a speed parameter buffer.