JPS6286396A - Rule synthesization voice communication system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a rule-synthesized speech communication system in which a sentence is transferred from a transmitting device and speech is synthesized by rule-based synthesis on a receiving device.

[Conventional technology]

As a conventional general rule-synthesized voice communication system, the transmitter f
Figure 6 (a) shows the case where the f1t side transmits accented kana sentences and the receiver side performs rule synthesis of the accented kana sentences.
).

In Figure 6 (a), 1 is the accented kana literary ability;
2 is an accented kana sentence transmitter, 3 is a line, 4 is an exchange, 5 is a line, and 6 is a built-in rule synthesizer that can set the voice pitch and speaking speed to change the tone of the rule-synthesized voice. In the accented kana sentence receiving device, 7 is a speaker, and 8 is an input for controlling the pitch and speed of voice for changing the timbre of the rule-synthesized speech.

In the example shown in FIG. 6(a), on the transmitting device side, the accented kana sentence human power 1 created using a keyboard or the like is input to the accented kana sentence transmitting device 2. The accented kana sentence transmission device 2 transmits the accented kana sentence human power 1 to the line 3 using five frame frames as shown in FIG. 6(b).

In FIG. 6(b), A is an accented kana sentence identifier, and B is an accented kana sentence. In Figure 6(b),
Only the parts related to rule synthesis are shown, and headers necessary for synchronization and data transmission are omitted, and the same applies to the FuV-Mufu Saiichi mat shown below. On the other hand, on the receiving device side, the control manual 8
The voice pitch and speaking speed for changing the voice tone of the rule synthesizer in the accented kana sentence receiving device ff1S are set using It is sent to the accented kana sentence receiver f16 and synthesized into speech.

In addition to the above-mentioned rule-based synthesis voice communication system, there is a rule-based synthesis voice communication system that uses a single rule synthesizer that can synthesize multiple voice tones. By changing the pitch of the voice using formant information, the voices of three men (normal voice, low voice, old man's voice), two women (normal voice, low voice), and a boy are synthesized. The tone of voice is specified in the input text to the rule synthesizer. Therefore, by transmitting a sentence with a specified tone of voice on the transmitting device side and outputting a rule-synthesized speech based on the specified tone of voice of the sentence on the receiving device t side, the voice can be heard in the tone specified by the transmitting device. can be synthesized.

[Problem that the invention seeks to solve]

In the conventional general rule-synthesized voice communication system as described above, there is no way to directly specify the tone, tone quality, volume, etc. of the rule-synthesized voice on the receiving device side, so the settings on the receiving device side Rules were synthesized based on voice tone, tone quality, volume, etc., and it was not possible to output to the receiving device with the characteristics specified by the sender. Furthermore, in the rule-synthesized voice communication system that uses a single rule-synthesizer, it is possible to synthesize voice with the voice tone specified by the transmitting device, but it is only possible to synthesize voice tones that have been registered in advance, and the voice tone cannot be specified in the text. Therefore, it is necessary to specify the tone of voice when inputting a sentence, and furthermore, if the user wants to change the tone of voice later, there is a problem in that it is necessary to modify the voice tone specified part in the sentence.

It is an object of the present invention to provide a rule-synthesized voice communication system in which voice tone, tone quality, volume, and voice speed can be specified from a transmitting device side, and the specified information can be transmitted separately from the transmitted sentence.

[Means for solving problems]

The first invention of the rule-synthesized voice communication system according to the present invention is as follows:
At least one of the voice color, tone quality, sound 1, and xA voice speed of the rule-synthesized voice from the transmitting device side to the receiving device side
A specification information sentence that specifies a message is added to the frame to be transmitted, and when the receiving device synthesizes the received sentences according to the rules,
It outputs rule-synthesized speech with the tone of voice, tone quality, volume, and speaking speed specified by the specified information.

The second invention of the rule-synthesized voice communication system according to this invention is as follows:
Prior to the transmission of a sentence from the transmitting device, the audio parameters of each voice from the transmitting device are transmitted and stored in the storage section of the rule synthesis means in the receiving device, and then the receiving device performs rule synthesis of the received sentence. , outputs rule-synthesized speech using the speech parameters of each speech unit stored in the storage section.

The third invention of the rule-synthesized voice communication system according to this invention is as follows:
Prior to the transmission of the text from the transmitting device t side, at least one type of voice parameter for each voice unit from the transmitting device side is transmitted and stored in the storage unit of the rule synthesis means in the receiving device, and then the transmitting device side The voice tone and quality of the rule-synthesized speech on the receiving device side.

Specifying information specifying at least one of the volume and the speaking speed is added to the frame for transmitting the sentence and transmitted, and the receiving device side uses the voice parameter Z of the voice unit stored in the storage unit when performing rule synthesis of the received sentence. The voice tone and tone quality specified by the specified information.

It outputs synthetically synthesized speech according to the volume and speaking speed.

[Effect]

In a first aspect of the present invention, a receiving device outputs a regular synthesized speech based on designation information sent from a transmitting device.

Further, in a second aspect of the present invention, audio parameters of each audio unit sent from the transmitting device are stored on the receiving device side, and based on the audio parameters, the receiving device outputs regular synthesized speech.

Furthermore, the third invention disclosed in this Tsumaki accumulates the audio parameters of each audio unit sent from the transmitting device, and then performs rule synthesis on the receiving device side based on the specified information sent and the audio parameters. Outputs audio.

〔Example〕

FIGS. 1(a) and 1(b) are diagrams showing the configuration of the first invention of the present invention and a fume mat containing text and designation information transmitted from the transmitting device side, and FIG. a), (b)
The same reference numerals indicate the same parts, 11 is a transmitting device, 12 is a switching device, 13 is a receiving device, and 13a is a device for which at least one type of tone of voice can be set for the rule-synthesized voice, and the tone quality, volume, voice speed, etc. can be set arbitrarily. The configurable rule synthesis means 20.21 is a line. Also, C is a voice color specification command, D is a voice color value, E is a tone quality specification command, F is a tone quality value, and G
is a volume designation command, H is a volume value, WORK is a voice speed designation command, and J is a voice speed value.

For the voice color value, tone quality value, volume value, and generation speed value, the voice color, tone quality, volume, and voice speed are determined for each value.

Next, the operation will be explained.

First, on the transmitting device side, the voice tone of the rule-synthesized voice.

Specification information specifying sound quality, volume, and speaking speed and sentences are input to the transmitting device 11, and the transmitting device 11 adds the specified information to the frame for transmitting the sentences, as shown in FIG. 1(b).
The data is transmitted using the configuration shown in FIG. On the receiving device side, within the receiving device 13, the rule synthesizing means 13a performs rule synthesizing of the sentences transmitted simultaneously with the voice tone, tone quality, volume, and speaking speed specified by the control information, and synthesizes the rules of the sentence input from the speaker 7. Synthesized speech is output.

NrI2 Figures (a), (b), and (c) show the structure of the second invention of the present invention, the frame format of the audio parameters of the audio unit transmitted from the transmitter side prior to the transmission of the text, and the text to be transmitted. In this figure, the same reference numerals as in FIGS. 1(a) and 1(b) indicate the same parts, 11a is the audio parameter storage section provided in the transmitting device 11, 14 is the receiving device, and 14a is the rule synthesis section. The apparatus is equipped with a storage section 14b that stores voice parameters of the extracted voice unit using the rule-synthesized voice ability. K is a voice parameter identifier, and L is a voice parameter.

Next, the operation will be explained.

First, on the transmitting device side, prior to transmitting the text, the data input to the transmitting device 11 or the voice parameter storage section 11 is
The voice parameters of the voice unit having the configuration shown in FIG. 2(b) stored in the voice unit a are transmitted and inputted to the reception/transmission f1t14 via the exchange 12, and are stored in the storage section 14b in the rule synthesis means 14a.

Next, the transmitting device 11 transmits the input text having the structure shown in FIG.
input. On the receiving device side, the rule synthesizing means 14a in the receiving device 14 performs rule synthesis using the voice unit parameters of the voice units stored in the storage unit 14b, and the rule synthesized speech of the sentence input from the speaker 7 is obtained. .

FIGS. 3(a) to 3(a) show the configuration of the third aspect of the present invention and the frame former of the voice harameter transmitted from the transmitter side prior to the transmission of the text, and the frame former containing the text and control information. Figure 1 (a)
), (b) and Figure 2 (a).

The same reference numerals as in (b) indicate the same parts 7, 15 is a receiving male device, 15a is a rule synthesis means, and the voice tone of the rule synthesis voice can be set to at least one level or higher, and the sound quality.

A storage unit 15b whose volume and speaking speed can be arbitrarily set and which stores audio parameters of extracted audio units.
'Y are prepared. 15c is the audio parameter of the audio unit,
This is the specification information and accented kana sentence identification part.

Next, the operation will be explained. In addition, Fig. 3(a) to (C
The embodiment shown in ) includes the embodiments shown in FIGS. 1(a) and 1(b), so it will be explained incorrectly.

An accented kana sentence is used as the sentence to be transmitted, and the transmitting device transmits the voice parameters of the voice unit extracted for the voices of the two people prior to transmitting the sentence.

When transmitting the text, the tone of the rule-synthesized voice on the receiving device side (assuming that the receiver device 11 has stored in advance the voice parameters of the voice units extracted for the voices of three people,
Prior to the transmission of the text, the transmission E- is combined with the audio parameters of each audio unit extracted for the voices of two people, and the audio parameters of each audio unit extracted for a total of five voices! Six poles of tones! (to be synthesized), specification information t for specifying sound quality, volume, speaking speed, etc., and added to the 7V-me for transmitting text.

On the receiving device side, after accumulating the voice parameters of the voice unit extracted from the two received voices, when performing rule synthesis of the received sentence, the newly accumulated voice parameters of the voice unit are combined with the voice parameters of the voice unit previously accumulated. Audio parameters extracted for each voice of the three voices! and the tone of voice specified by the specified information.

A case will be explained in which a rule-synthesized speech is output based on sound quality, volume, speaking speed, etc.

First, the operation on the transmitter side will be described.

First, the audio parameters of each audio unit extracted for the voices of three people are transmitted from the audio parameter storage 511a in the configuration shown in FIG.
Store it in 5b. Then, the 2 parameters stored in the audio parameter storage section 11a, which are necessary prior to the transmission of the text, are
The audio parameters in units of audio for one person or the audio parameters in units of audio for two people input into the transmitter 11 from the outside are stored in the storage section 15b. Next, specification information for specifying the tone, tone quality, volume, voice speed, etc. of the rule-synthesized speech on the receiving device side and accented kana sentences created using a keyboard or the like are input to the transmitting device 11.

The transmitting device 11 sends the accented kana sentence and the specification information in the same frame as shown in FIG. 3(e) or in the same frame as shown in FIG. 4(a).
) and (b) are transferred in separate frames as shown in (b). If you want to change the tone of voice, tone quality, volume, speaking speed, etc. midway through, please select the third V-frame following the previously transferred frame.
The frame shown in FIG. 4(c) or the frames shown in FIGS. 4(a) and 4(b) are transferred.

Next, the operation of the receiving device 911 will be described.

Figures 3(b) and (C) received via exchange 12Y.
) or frames as shown in FIGS. 4(a) and 4(b) are input to the identification unit 15c in the receiving device t15. aQ
The separate section 15e analyzes 7 the received data, and performs 7 identification of voice parameters, specification information, and accented kana sentences for each voice. First, the speech parameters of the speech units sent prior to the transmission of a sentence are stored in the storage section 15b within the rule synthesis means 15a.

Through this accumulation, the audio parameters of each audio unit extracted for a total of five voices, including the audio variations of each audio unit extracted for three voices that have been stored in advance, are stored in the storage unit 15.
It is stored in b.

Next, the specified information output is input to the rule synthesis means 15a,
The tone of voice, tone quality, volume, speaking speed, etc. specified by the transmitting side are set, and then the identifying section 15C sends the identified accented kana sentence to the rule synthesizing means 15a to synthesize it into speech.

In this embodiment, a rule synthesis method has been described; however, it can be similarly applied to a monosyllable editing method in which monosyllables are used as speech units and are connected discretely. Although the case where the 0th sentence is an accented kana sentence is shown, the same applies to other sentences as well. ■Sound units include phonemes, monosyllables, and VCV.

Possible options include CVC.

In addition, the audio generated by rule synthesis is combined with other media,
Often used as voice guidance. Even when a sentence to be synthesized by rules is compounded with other speech, the output state of the synthesized speech by rules can be controlled by configuring the parts related to rule synthesis as in this example. can.

Further, in the above embodiment, an end-to-end communication mode is shown, but an end-to-center-end communication mode is also conceivable, in which one terminal is used as a center and a general telephone is connected to the center.

FIG. 5 shows an example of an end-center-end configuration.

In FIG. 5, 20 to 23 are lines, 24 to 26 are telephones, 27 is a center, and 28 is a terminal on the transmitter side. The terminal 28 is a terminal of an information provider, etc., and transmits phoneme file information, specification information, and sentences for rule synthesis in FIGS. 3(b) and 3(c).
), the frame format shown in FIGS. 4(a) and 4(b) is transferred to 11M2O. The center 27 synthesizes speech based on the phoneme file information, specification information, and sentences for rule synthesis received through the line 20. The center 2T transfers the synthesized voice to telephones 24-26 through seven lines 21-23.

〔Effect of the invention〕

A first aspect of the present invention relates to the tone and quality of regularly synthesized speech from the transmitting device side to the receiving device side.

Specified information such as volume and speaking speed is added to the frame for transmitting the text and transmitted, and when the receiving device synthesizes the received sentence according to the rules, the receiving device uses the tone and quality of voice specified by the specified information transmitted together with the received sentence. , it is possible to output the rule-synthesized voice according to the volume, voice speed, etc., so it is possible to have the receiver @ output the rule-synthesized voice according to the voice tone, tone quality, volume, voice speed, etc. intended by the sender. It also has the advantage of being able to separate sentences from specific information such as tone of voice, tone quality, volume, and speaking speed.

Further, in the second invention according to the present invention, the transmitter side transmits the audio parameters of each voice extracted for one person's voice prior to transmitting the text, and the receiver side transmits the voice parameters of the received new voice. Accumulate the audio parameters of the unit, and the audio parameters of this newly accumulated audio unit! It has the advantage of using the rules of the received text.

Furthermore, the third invention according to the present invention is such that, prior to sending a text, audio parameters for each voice unit extracted from one or more people's voices are transmitted, and when sending the text, the receiving device side The tone and quality of the rule-synthesized speech,
Specification information specifying the volume, speaking speed, etc. is added to the frame for transmitting the text and transmitted, and the receiving device ft1
In step 111, the voice parameters of the newly received voice unit are accumulated, and when the received sentence is synthesized according to the rules, the voice parameters of the newly accumulated voice unit and the voice parameters of the voice unit that have been registered in advance are used, and the voice parameters of the voice unit that have been registered in advance are used. Since the rule-synthesized voice is output with the specified voice tone, tone quality, volume, and speaking speed, it is possible to output the voice tone that the sender intended,
It is not a scale that allows the receiving device to output a synthesized voice according to the sound quality, volume, and speaking speed, but also the sentence and tone of voice,
It has the advantage of being able to separate specified information such as sound quality, volume, and speaking speed.

[Brief explanation of drawings]

FIGS. 1(a) and 1(b) are diagrams showing the structure of a first embodiment of the invention and a frame format including designation information, and FIGS. 2(a) and 2(b) are diagrams showing a frame format including designation information. 4(c) is a diagram showing a frame format according to the present invention (including the configuration and configuration of the second invention and designation information), and FIGS. 4(a) and 4(b) are diagrams showing other frame formats. )
The figure shown in Y, Figure 5, shows this invention! Figure 6 (a) is a diagram showing the configuration when used in an end-center-end type of communication form.
) and (b) are diagrams showing a configuration diagram and a frame format of a conventional rule-based synthetic voice communication system. In the figure, 11 is a transmitter, 11a is a voice parameter i′ product unit, 12 is an exchange, 13 is a receiver, 130 is a rule synthesis means, 14 is a reception device, 14 & is a rule synthesis means, 14b
1 is a storage unit, 15 is a receiving device, 15a is a rule synthesis means, 1
5b is a storage section, 15c is an identification section, and 20.21 is a line. Figure 5 Figure 6 Procedural amendment (method) % formula % 1. Indication of the case Japanese Patent Application No. 60-225979 2. Title of the invention Rule-synthesized voice communication system 3. Person making the amendment Relationship with the case Patent applicant Address 1-1-6 Uchisaiwai-cho, Chiyoda-ku, Tokyo Name (4
22) Nippon Telegraph and Telephone Corporation Kobayashi Patent Office Telephone 0
3 (496) No. 1256 5, Date of amendment order January 28, 1985 6, Drawing subject to amendment 7, Contents of amendment Figures 1 to 3 will be amended as shown in the attached sheet. Written amendment to the above procedures (voluntary) April 30, 1985

Claims (3)

[Claims] (1) Speech is synthesized from accented kana sentences input from the transmitter side, sentences with kanji and kana mixed in the orthography normally used, or foreign language sentences, and at least one type of tone of voice is set for the synthesized voice according to rules. Possible, sound quality,
In a receiving device having a rule-synthesizing means in which a volume and a voice speed can be arbitrarily set, a specification for specifying at least one of the tone, tone quality, volume, and voice speed of a rule-synthesized voice from the transmitting device side to the receiving device side. Information is added to a frame for transmitting a sentence and transmitted, and the receiving device side outputs a rule-synthesized voice with a tone of voice, tone quality, volume, and speaking speed specified by the specified information when performing rule synthesis of the received sentence. A rule-synthesized voice communication system featuring the following. (2) Synthesize speech from accented kana sentences input from the transmitter side, sentences with kanji and kana mixed in the commonly used orthography, or foreign language sentences, and check the tone, quality, volume, and utterance of the synthesized voice according to rules. At least one of the speeds can be arbitrarily set, and the transmitting device uses a receiving device that has a rule synthesis means that is equipped with a storage unit that stores audio parameters for each audio unit, and the transmitting device After transmitting the voice parameters of the voice units from the side and storing them in the storage unit of the rule synthesis means in the receiving device, the receiving device side uses the voice parameters accumulated in the storage unit when performing rule synthesis of the received sentence. A rule-synthesized speech communication method characterized by outputting rule-synthesized speech using speech parameters. (3) Synthesize speech from accented kana sentences inputted from the transmitting device side, sentences with kanji and kana mixed in the commonly used orthography, or foreign language sentences, and set at least one type of tone of voice for the synthesized voice according to rules. Possible, sound quality,
Using a receiving device that has a rule synthesis means in which the volume and speaking speed can be arbitrarily set and is equipped with a storage unit that stores audio parameters for each audio unit, at least one type of text is synthesized prior to transmission of the text from the transmitting device side. After transmitting the voice parameters for each voice from the transmitting device side and storing them in the storage unit of the rule synthesis means in the receiving device, the voice tone and quality of the rule-synthesized voice from the transmitting device side to the receiving device side, At least one of volume and speaking speed
The receiving device side uses the audio parameters of each audio unit stored in the storage unit to synthesize the received sentences by adding designation information to the frame for transmitting the text. A rule-synthesized voice communication method characterized by outputting rule-synthesized voice with specified voice tone, tone quality, volume, and speaking speed.