JPH04175046A - Audio response equipment - Google Patents

Abstract

PURPOSE:To make the speech synthesis service efficient while skipping hearing of undesired part by revising man/woman voice synthesis and talking speed in matching with the preference of the user. CONSTITUTION:An NCU section 2 detects a request by the user about the revision of man/woman voice synthesis and talking speed even on the way of the ruled synthesis of an electronic mail and the revision of the man/woman voice synthesis and talking speed is implemented on the request of the preference of the user. Since the service is received by the ruled synthesis sound in matching with the user's preference, the utility of the device is improved, the advertising capability of the device is improved and the efficient speech synthesis service is received.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a voice response device that applies voice rule synthesis for generating voice from arbitrary characters according to predetermined rules to telephone services.

(Prior Art) There is a device that uses speech rule synthesis for telephone services. In speech rule synthesis, an arbitrary input character string is analyzed, phonetic information and prosody information are obtained, and synthesized speech corresponding to the input character string is output based on predetermined rules. As an example of application to telephone services, it is possible to convert the contents of a user's mail (text file) sent to a center via PC communication into speech using voice rule synthesis, and then provide telephone services to the user. This service selects a voice output email by inputting a button according to a guidance voice, and synthesizes the selected email content according to voice rules.

In order to make this service easier to listen to, for example, the guide voice for services such as e-mail selection is set to a synthesized female voice, and the audio output for e-mails addressed to the user is output to a synthesized male voice. The audio to be output is determined on the side. Furthermore, the speech rate is set at a normal speech rate (about 4 moles per second).

However, some users who listen to sent e-mails using regularly synthesized sounds have a desire to have their e-mails synthesized into a voice that is easy to hear and corresponds to their own ears. To do this, a user who is listening to an e-mail over the phone must change the male/female sound or change the speech rate midway through the e-mail. However, with the current telephone service system, the user is listening to an e-mail when a man/
There was a problem in that it was not possible to change the female sound or the rate of speech, and therefore it was not possible to respond to this request.

(Issues to be solved by the invention) The present invention has been made in consideration of these circumstances, and
The purpose of this is to have a means to detect when a user listens to an e-mail over the phone and changes his/her male/female tone or rate of speech midway through, and to make it possible to change the male/female tone or speech rate depending on the user's preference. /An object of the present invention is to provide a voice response device for telephone service that can change the female synthesized voice and the speaking rate.

[Structure of the Invention] (Means for Solving the RS Problem) In order to achieve this object, the present invention provides a device for converting the contents of an e-mail into speech by a speech rule synthesis device and providing a telephone service. It detects whether a user who is listening to an e-mail requests a male/female sound or a change in speech rate, and adjusts the male/female tone according to the user's preference.
It allows you to synthesize female sounds and change the speech rate.

(Operation) According to the present invention, in an apparatus for converting the content of an e-mail into a voice using a voice rule synthesis device and providing a telephone service,
Even if e-mail rule composition is in progress, the user's male/
It is possible to detect a request to change the female sound or speech rate, and change the male/female sound or speech rate according to the user's preference. Therefore, compared to the past, users can receive services with regular synthesized sounds that suit their tastes, the convenience of the device is improved, and the communication and advertising capabilities of the device can be greatly increased.

(Example) Hereinafter, details of an example of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of an e-mail listening device according to the present invention. In the figure, 1 is a speech rule synthesis section, 2 is an NCU section, 3 is an email, and 4 is a main control section. Here, details of the speech rule synthesis section 1 will be explained using the block diagram of FIG. 2.

Reference numeral 11 denotes a character string analysis unit that analyzes a character string, verifies the accent position with reference to the pronunciation 1f12, and obtains a phonetic symbol string and prosody information. The phoneme symbol string is input to the phoneme parameter generation device ff114, which generates a phoneme parameter string by referring to the speech segment file 15. On the other hand, the attt information is
The signal is supplied to a prosodic parameter generation device 13, and a prosodic parameter sequence is generated. The synthesizer 16 generates and outputs a synthesized sound according to a predetermined synthesis rule according to the phonetic parameter string and the prosody parameter string thus obtained.

Next, the NCU section will be explained using the block diagram of FIG. The NCU unit is connected to a telephone line through a modular jack 27, receives the synthesized sound output from the voice rule synthesis unit 1 from the synthesized sound input unit 21, and sends it out to the telephone line via a transformer 28.

In addition, 22 is a 400Hz detection unit that detects busy sound;
3 is a PB detecting section for detecting a bush tone, 24 is a dial section, and 25 is a calling signal and polarity reversal detecting section.

FIG. 4(a) is an example of an e-mail message synthesized by voice rules in a telephone service. Figure 4(b) is the same as Figure 4(a).
This is an intermediate analysis result that is output when an e-mail is input to the character string analysis section 11. "-" is an accent mark,
"," is a pause symbol, "/" is an accent phrase boundary symbol,
"//" represents a phrase end symbol, and katakana represents a phoneme to be synthesized.

Here, an aspect of the e-mail telephone service using the voice rule synthesis device equipped with the NCU function will be explained using the flowchart shown in FIG.

When a call is received, it is detected by the NCUC2O4 output signal detection section 25, the telephone line is closed, and telephone service is started (ST51).

When starting the service, the telephone service guidance voice will be output, and if necessary, enter your member number and PIN number here.
(push button) to input (ST52).

The number of emails addressed to the user making the call and the title attached to the email are read out loud, and the email number that you want to hear is entered in PB (ST53).Next, the email selected by PB is synthesized using voice rules, and the message is sent to the phone. Flow (ST54).

At this point, if the other party has not hung up, (ST53
) and continue with telephone service.

During this series of operations, the user who made the call (ST54)
While listening to the synthesized e-mail sound, you may want to change the male/female sound or the speech rate.

In the present invention, in the process (ST54), the user's request for switching between male/female sounds and changing the speech rate is determined by a specific PB multiplication signal as shown in FIG. 7, and the voice rule synthesis unit 1 It has a clause end symbol detection unit for the intermediate analysis result output when an email to be converted into speech is input to the character string analysis unit 11,
When a bunsetsu end symbol is detected during speech synthesis, PB double signal detection is performed for about 2 seconds to check whether the PB shown in FIG. 7 has been input. Here, if the PB shown in FIG. 7 is detected, switching between male/female sounds and changing the speech rate are performed.

Next, the e-mail voice synthesis process will be explained in more detail using the flowchart of FIG.

First, an e-mail to be synthesized according to voice rules is selected and read and opened (ST61).

Next, the character string of the voice-converted email is stored in a buffer until the clause end symbol "//" of the voice-converted email is detected (ST62).

Each time this storage is performed, it is determined whether or not a clause end symbol is detected (5T63), and if it is not detected, storage of step 5762 is further performed. On the other hand, when a bunsetsu end symbol is detected, the character string stored in the buffer is sent to the speech rule synthesis section and synthesized into speech.
It is sent to the telephone line (ST64).

Next, detect if it is the end of the email (5T65),
If it is not the end, the NCUC2O4B detection unit 23 detects the seventh
Detects the PB double signal to see if the PB double signal shown in the figure is input (
ST66). The PB detection time here is, for example, 2
Possibly about seconds. If "*J" is detected in this PB detection (ST67), the male/female voice of the regular synthesized voice is switched (ST70) as shown in FIG. If a male voice was being output before J detection, the telephone service is performed by immediately switching to a female voice.On the other hand, if a female voice was being output before ``1'' was detected, the voice output is switched to a male voice. Provide telephone services.

Also, if "1" is detected in PB detection (ST68), the speech rate is reduced by one step less than L (ST71), and if r2J is detected in PB detection (ST69), the speech speed is slowed down by one step. (ST72). The speech rate is about 4 moras per second at normal speed, and can be set in about 5 to 9 steps in the speech rule synthesis section.

Speech synthesis processing continues, and if it is detected in (ST65) that it is the end of the e-mail, (ST73)
) to close the read open file and finish reading the email.

Note that in this embodiment, there are two voice types: male/female.
The speech rate can be changed in 5 levels.

The default mode for reading e-mails is
Male voice, normal speech rate.

Next, explanation will be given using a specific e-mail.

In the case of the e-mail shown in Figure 4(b), "Higashi...
The part //J (4a) is voice synthesized using a male voice at a normal speech rate, and then the PB double signal is detected.

Here, if the PB double signal in FIG. 7 is not detected,
Baira...J (4b) is voice synthesized using a male voice at a normal speech rate, and the PB double signal is detected. In PB detection before speech synthesis in (4b), PB rlJ
If this is detected, "Baira..." (4b) is synthesized into a female voice with a normal speech rate. Also, (4b
), if PB rlJ is detected in the PB detection before speech synthesis of ``Baira...J (4b), the speech speed is increased by one step, the voice is synthesized with a male voice, and PB rlJ is detected.
If 2J is detected, the speech rate is slowed down by one step and a male voice is synthesized. In this way, the PB is detected every time the phrase ends, and speech synthesis is performed according to the detection result.

Note that although there are two voice types, male and female, it is possible to have multiple voice types. You can also switch between male/female voices by pressing
``One B'', but it is also possible to use ``#'' or a combination of numbers. Further, although the phrase end symbol is "//", it is also possible to use other symbols. Also,
Although the processing is performed in units up to the bunsetsu end symbol "//", it is also conceivable to similarly perform the processing in units up to the accent phrase boundary symbol r/J.

Furthermore, although the PB detection time is set to 2 seconds, it may be possible to make it longer than this to ensure that the user's change request is known. In addition, here, as basic revision items, switching between male/female voices and speech rate were used, but it is also possible to change the pitch of the voice by manipulating the basic pitch of the rule synthesis section in the same way.

According to the embodiment described above, in a device that converts the contents of an e-mail into speech using a voice rule synthesis device and provides telephone service, even if the rule synthesis of the e-mail is in progress,
It is possible to detect a user's request to change the male/female voice or speech rate, and change the male/female voice or speech rate according to the user's preference. Therefore, compared to the past, users can receive services with regular synthesized sounds that suit their tastes. In addition, the convenience of the device is improved, and the transmission and advertising capabilities of the device can be greatly increased, making the voice rule synthesis device easier to use as a telephone service device,
This has the advantage that it can be used as a machine interface.

In addition, in the process of (ST54) in FIG. 5 in the above-mentioned embodiment,
Requests for switching the user's male/female voice or changing the speaking speed,
As shown in FIG. 7, a specific PB multiplier signal is determined, and when PB is detected, the male/female tone is switched and the speech rate is changed as shown in FIG. Similarly, for example, when PBrtiJ is detected, it may be possible to skip it.

That is, to explain using FIG. 4(b), first, the part "Higashi...//J (4a) is synthesized into speech, then PB is detected, and here, when PBr#J is detected. Then, the next part "Baila... //J is skipped, r nonsense...//" is synthesized according to the phonetic rules.

Note that here, speech synthesis is performed up to the bunsetsu end symbol r//'', PB detection is performed, and when PBr#J is detected, the next bunsetsu end symbol r//'' is skipped and the next bunsetsu end symbol is selected. The text up to the phrase ending symbol "//" is processed as the qi position, such as synthesizing speech up to "r//".
Similarly, it is also conceivable to perform processing in units up to the accent phrase boundary symbol "/".

According to this embodiment, a user who is listening to an e-mail by telephone can skip reading it by inputting PB rlJ, so for confirmation, the user can read the e-mail with a blank space.
When listening again, you can extract and listen to only the important parts. Therefore, the user can efficiently receive the e-mail speech synthesis service without hearing unnecessary parts.

[Effects of the Invention] As described in detail above, according to the present invention, it is possible to provide a voice response device for telephone service in which male/female voices and speaking speed can be changed according to the user's preference.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the voice response device of the present invention, FIG. 2 is a block diagram showing details of the voice rule synthesis section shown in FIG. 1, and FIG. 3 is a block diagram showing the details of the voice rule synthesis section shown in FIG. A block diagram showing details of the NCU unit, Figure 4(a) is an example of an email, Figure 4(b) is an analysis result of the email in Figure 4(a), and processing of the fifth fixed story service. FIG. 6 is a flowchart for explaining the present invention, and FIG. 7 is an example of a PB double signal that is regarded as a request for male/female sounds or a change in speech rate. 1...Speech rule synthesis unit 2...NCU unit 3...E-mail 4...Main control department agent Patent attorney Nori Chika Ken Yamashita Butsuma - Fig. 1 Fig. 2 1pPL--J LM Fig. 4 (oL) Fig. 4 (ri Fig. 5 Fig. 6 Fig. 7

Claims (4)

[Claims] (1) An electronic file comprising a rule synthesis means for adding an accent to an input character string and synthesizing speech, a telephone line control means, an electronic file storage means, a clause end symbol detection means, and a push button detection means. This is a telephone service using voice synthesis that detects whether a user who is listening to an e-mail on the phone requests to change the male/female sound, and converts the electronic file into voice with a synthesized voice corresponding to this result. A voice response device characterized by rule synthesis. (2) In the voice response device according to claim 1, it is detected whether or not a user listening to an e-mail by telephone has requested a change in speech rate, and synthesis is performed in response to this result. A voice response device characterized by synthesizing electronic files using sound rules. (3) In the voice response device according to claim 1, it is detected whether or not a user listening to an e-mail by telephone has requested a change in the basic pitch, and synthesis is performed in accordance with this result. A voice response device characterized by synthesizing electronic files using sound rules. (4) In the voice response device according to claim 1, it is detected whether or not a user who is listening to an e-mail by telephone has requested to skip the e-mail, and in response to this result, an electronic file is generated. A voice response device characterized by skipping over and synthesizing voice rules.