JP4684609B2 - Speech synthesizer, control method, control program, and recording medium - Google Patents

Description

  The present invention relates to a speech synthesizer, a control method, a control program, and a recording medium, and more particularly to a speech synthesizer that performs speech synthesis by inputting text information, a control method thereof, a control program, and a recording medium.

Conventionally, in a navigation device that guides the guidance route from the current location to the destination,
It is known that a TTS (Text To Speech) controller is mounted, and speech synthesis is performed based on text information corresponding to guidance route guidance and arbitrarily specified text information (VICS information, e-mail, etc.). (For example, refer to Patent Document 1).
In this case, the synthesized speech quality (voice quality) is determined by using phonemes such as speech speed, voice pitch, voice thickness, etc., set in advance in the navigation device as fixed values. ) Has been performed, and there has been a problem that it becomes a certain thing.
Japanese Patent Application Laid-Open No. 5-113395

By the way, when speaking in general, depending on the content, the speaker may partially change the way of speaking (speaking speed, accent, loudness, etc.) according to the importance of the content. In this way, the listener may be able to hear the content more easily or can quickly grasp the content.
However, since the quality of the synthesized voice is always constant in the conventional navigation system described above, for example, a user's desire to read a long email quickly and a unique name (location name, building name, facility name, etc.) However, there was a problem that it was not possible to respond to requests such as slowing down the utterance speed or requesting a loud voice.
SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a speech synthesizer capable of appropriately changing the mode of speech synthesis based on the type and content of text information to be synthesized in the speech synthesizer, a control method therefor, and a control program. And providing a recording medium.

In order to solve the above problems, in the speech synthesizing apparatus for performing speech voice line speech synthesis based on the input text information, mail reading mode, reading lengthy mail faster in response to the input source type of the text information A mode setting unit for setting the voice reading mode to any of the mode, traffic information reading mode or route guidance mode, and a combination of the voice synthesis control parameters corresponding to the set voice reading mode, and the mail reading mode or In the long mail high-speed reading mode, proper nouns are pronounced greatly clearly. In the traffic information reading mode, proper names such as place names and interchange names are pronounced clearly. In the route guidance mode, distances , Direction, placemark Is a parameter specifying unit to allow increased clearly pronounce the voice synthesizing unit for outputting as the speech by performing speech synthesis based on speech synthesis parameters identified, comprising the.
According to the above configuration, the mode setting unit sets the voice reading mode to one of the mail reading mode, the long text high speed reading mode, the traffic information reading mode or the route guidance mode according to the type of the text information input source.
Thus, the parameter specifying unit specifies the combination of the speech synthesis control parameters corresponding to the determined speech reading mode, and in the mail reading mode or the long text high speed reading mode, the proper noun is pronounced largely and clearly. In the traffic information reading mode, unique names such as place names and interchange names are pronounced clearly, and in the route guidance mode, the distance, direction, and landmarks are pronounced clearly. Then, speech synthesis is performed based on the identified speech synthesis parameters and output as the speech.

In this case, the input source of the text information is a mail application, a traffic information processing application or a navigation application, and the mode setting unit, when the input source of the text information is a mail application, Is set to the mail reading mode or the long text high speed reading mode, and when the input source of the text information is a traffic information processing application , the voice reading mode is set to the traffic information reading mode, and the text information When the input source is a navigation application, the voice reading mode may be set to the route guidance mode .

A method of controlling a speech synthesis apparatus for performing speech voice I line speech synthesis based on the input text information, in response to the input source type mail reading mode of the text information, long sentence mail fast reading mode, reading traffic information A combination of a mode setting process for setting the voice reading mode to either the mode or the route guidance mode and a voice synthesis control parameter corresponding to the set voice reading mode, and the mail reading mode or the long sentence high speed reading mode In the traffic information reading mode, proper names such as place names and interchange names are pronounced clearly, and in the route guidance mode, distances, directions, and landmarks are pronounced. Big A parameter specifying the process to be sound, and characterized in that and a speech synthesis step of outputting as the speech by performing speech synthesis based on speech synthesis parameters specified.
In this case, the input source of the text information is a mail application, a traffic information processing application, or a navigation application. If the input source of the text information is a mail application in the mode setting process, the voice reading mode is set. Is set to the mail reading mode or the long text high speed reading mode, and when the input source of the text information is a traffic information processing application , the voice reading mode is set to the traffic information reading mode, and the text information When the input source is a navigation application, the voice reading mode may be set to the route guidance mode.

In the control program for controlling a computer speech synthesis apparatus for performing speech voice line speech synthesis based on the input text information, in response to the input source type mail reading mode of the text information, long sentence email fast reading mode, the traffic or information reading mode or route guidance mode is set the speech output mode, and specifying a combination of speech synthesis control parameters corresponding to the speech output mode set, the mail reading mode or the lengthy In the mail high-speed reading mode, proper nouns are pronounced greatly clearly. In the traffic information reading mode, proper names such as place names and interchange names are pronounced clearly. In the route guidance mode, distance The direction and the mark are pronounced largely and clearly , and speech synthesis is performed based on the specified speech synthesis parameters to output the speech.
In this case, the input source of the text information is a mail application, a traffic information processing application, or a navigation application. When the input source of the text information is a mail application, the voice reading mode is set to the mail reading mode or When the long text mail high speed reading mode is set and the text information input source is the traffic information processing application , the voice reading mode is set to the traffic information reading mode, and the text information input source is the navigation application. In some cases, the voice reading mode may be set to the route guidance mode.
The control programs may be recorded on a computer-readable recording medium.

  According to the present invention, it is possible to appropriately change the mode of speech synthesis based on the type and content of text information to be synthesized.

Embodiments of the present invention will be described below with reference to the drawings. In the following description, an in-vehicle navigation device (so-called car navigation device) will be exemplified as the navigation device.
FIG. 1 is a block diagram showing a functional configuration of navigation device 100 according to the present embodiment. As shown in this figure, the navigation device 100 includes an absolute position / orientation detection unit 1, a relative orientation detection unit 2, a vehicle speed detection unit 3, a main control unit 4, a ROM 5, a DRAM 6, an SRAM 7, and a VRAM 8. A user interface unit 9, a display unit 10, an input unit 11, a disk control unit 12, an FM multiplexed signal processing unit 13, an external recording device control unit 14, and a voice data generation unit (speech synthesis output unit) 15. And.

  The absolute position / orientation detection unit 1 includes a receiver (including an antenna) that receives GPS radio waves transmitted from a GPS (Global Positioning System) satellite, and is the current location of the automobile on which the navigation device 100 is mounted, that is, the own vehicle position. Is calculated based on GPS radio waves and output to the main control unit 4. The relative azimuth detecting unit 2 has a gyro sensor, detects a relative azimuth of the own vehicle position, and outputs it to the main control unit 4. The vehicle speed detection unit 3 processes vehicle speed pulses obtained from the automobile and outputs the speed of the host vehicle to the main control unit 4.

  The main control unit 4 executes various processes such as control of each part of the navigation device 100 and processes for the navigation function, and includes a CPU as arithmetic means and other peripheral circuits. The ROM 5 stores various programs such as a control program, a basic input output system (BIOS), a boot program for starting the apparatus, and a program for realizing a navigation function, and is accessible by the main control unit 4. ing. The DRAM 6 is a volatile memory and is used as a work area for the main control unit 4. The SRAM 7 is a non-volatile memory, and is supplied with power from a main power source (not shown) such as an accessory power source of an automobile, and a standby power source (not shown) such as a battery while the main power source is off. Is configured so that the stored contents can be held at all times by supplying power from the storage device, and functions as a backup memory. The VRAM 8 is a buffer memory in which screen data displayed on the display unit 10 is written.

  The display unit 10 displays various information such as a map for navigation, a vehicle position, and an operation menu under the control of the user interface unit 9. For example, the display unit 10 is an LCD (Liquid Crystal Display) or an EL (Electro Luminescent). ) A display device such as a display is provided. The input unit 11 receives a user's instruction operation and outputs it to the user interface unit 9. The input unit 11 is used for a plurality of operators arranged on the front portion of the navigation device 100 or a display device provided in the display unit 10. A provided touch panel (not shown) is provided. Note that the input unit 11 may be configured such that the navigation device 100 can be remotely operated by a remote controller or the like.

  The user interface unit 9 includes an input / output (I / O) control circuit and a driver circuit, and functions as an interface that connects the display unit 10 and the input unit 11 to the main control unit 4. Specifically, the user interface unit 9 executes display control of the display unit 10 under the control of the main control unit 4 and outputs an operation of the input unit 11 to the main control unit 4.

  The disk control unit 12 controls storage devices such as a CD-ROM drive, a DVD-ROM drive, and a hard disk drive that store map data used for navigation and various data. The FM multiplex signal processing unit 13 receives FM multiplex broadcast waves, extracts VICS (road traffic information communication system) information indicating information such as traffic jams, accidents, and traffic regulations, and outputs the information to the main control unit 4. is there. The external recording device control unit 14 performs recording and reading of data with respect to an external recording medium such as a memory card, a Memory Stick (registered trademark), and a compact flash (registered trademark) card.

FIG. 2 is a schematic configuration block diagram of the audio data generation unit.
The voice data generation unit 15 includes a synthesis unit series conversion unit 15A to which a phonetic symbol string is input.
When a speech notation symbol string is input, the synthesis unit sequence converter 15A analyzes the speech notation symbol string, converts it into a synthesis unit sequence symbol string that is a symbol string of a speech synthesis unit, and a speech synthesis control parameter generation unit Output to 15B.
The voice synthesis control parameter generation unit 15B refers to a preset TTS parameter and a voice synthesis rule database (DB) 15C, and sets voice synthesis control parameters for controlling the sound source unit 15D and the voice synthesis filter 15E in time series. To generate.

In this case, the speech synthesis control parameter generation unit 15B sets a text-to-speech mode according to the content of text information or the type of text information input source (mail application, navigation application, traffic information processing application, etc.). Function as. Accordingly, the speech synthesis control parameter generation unit 15B reads out the speech according to the content of the text information or the type of the input source of the text information among a plurality of TTS parameters set in advance corresponding to the plurality of speech reading modes. The mode is automatically generated and the corresponding TTS parameter is referred to.
The generated speech synthesis control parameters include a parameter that determines the transfer characteristics of the vocal tract and parameters (pitch (fundamental frequency), sound source amplitude, voiced / unvoiced, etc.) related to the sound source characteristics. These parameters are set in correlation with each other within a certain time frame.

As a result, the sound source unit 15D and the speech synthesis filter 15E synthesize continuous speech based on the set parameters, and the synthesized speech is output from the speaker 15F.
In the above configuration, prior to performing speech synthesis and reading out various kinds of information, the user performs a TTS parameter setting process for presetting a plurality of sets of TTS parameters corresponding to a plurality of speech reading modes. It is necessary to go.

FIG. 3 is a flowchart of TTS parameter setting processing.
First, the user selects a speech reading mode N in which a TTS parameter is to be set (step S1).
In this case, examples of the voice reading mode N include a mail reading mode, a long mail high speed reading mode, a traffic information (VICS) reading mode, a route guidance mode, and the like. In the mail reading mode, normal speech synthesis is performed, or proper nouns and the like are pronounced largely clearly. In the long mail high-speed reading mode, in order to grasp the contents of the mail in a short time, the proper nouns are pronounced largely and clearly, and the speech speed is increased. In the traffic information (VICS) read-out mode, processing is performed such that a unique name such as a place name or interchange name is pronounced largely and clearly. In the route guidance mode, a process such as making the distance, direction (traveling direction, etc.), and a mark (signal, building, etc.) loud and clear is performed.

Subsequently, the user sets TTS parameters P1 to Pn for the selected speech reading mode (step S2).
Here, the types of TTS parameters P1 to Pn include pitch, speech speed, voice thickness, whether special symbols can be read out, male / female, voice pitch, voice volume, accent size, vowel The possibility of devoicing is mentioned.
Subsequently, the main control unit 4 stores the set TTS parameters P1 to Pn in the DRAM 6 and the SRAM 7 in association with the reading mode N, and ends the setting process (step S3).

FIG. 4 is a flowchart of the audio output request process.
Thereafter, when a voice output request is made from the mail application, traffic information processing application or navigation application being executed by itself, the main control unit 4 acquires or sets the text data T to be output and the voice reading mode ( In step S11), the TTS output request information S is set and output to the audio data generation unit 15 (step S12).

FIG. 5 is a process flowchart of the TTS audio output process.
When the voice data generation unit 15 acquires the TTS output request information S from the main control unit 4 (step S21), the voice data generation unit 15 sets TTS parameters P1 to Pn (step S22).
Further, the voice data generation unit 15 sets text data T (step S23).
Thereby, the voice data generation unit 15 converts the text data T into voice data (step S24).

FIG. 6 is a process flowchart of the audio data generation unit.
First, the speech data generation unit 15 analyzes the input text data T, converts it into a sequence of phonetic symbols, and performs morphological analysis and syntactic analysis to automatically generate prosodic features (step S31). ).
Specifically, the text input with reference to the word search table 15G and the dictionary 15H defining the types of words that can be grammatically linked to a certain word is divided into words and morpheme sequences.

Subsequently, based on the divided words and the morpheme sequence, the reading information, the grammatical information, the accent information, and the word / sentence accent information are extracted as speech synthesis information and output to the synthesis unit sequence conversion unit ( Step S32).
As a result, the synthesis unit sequence conversion unit 15A analyzes the information for speech synthesis, converts it into a synthesis unit sequence symbol string that is a symbol string of a speech synthesis unit, and outputs it to the speech synthesis control parameter generation unit 15B.

The voice synthesis control parameter generation unit 15B refers to the set TTS parameters P1 to Pn and the voice synthesis rule database (DB) 15C, and sets voice synthesis control parameters for controlling the sound source and the voice synthesis filter in time series. Generate (step S34).
Here, as speech synthesis control parameters, there are a parameter that determines the transfer characteristics of the vocal tract and parameters (pitch (fundamental frequency), sound source amplitude, voiced / unvoiced, etc.) related to the sound source characteristics. These parameters are set in correlation with each other within a certain time frame.

As a result, the sound source unit 15D and the speech synthesis filter 15E synthesize continuous speech based on the set parameters, and the synthesized speech is output from the speaker 15F (step S25).
At this time, the synthesized speech output from the speaker 15F follows the set TTS parameters P1 to Pn.
For example, if the mail volume is a long text mail that is larger than a predetermined reference capacity, the reading mode is set to the long text high speed reading mode, the reading speed is increased, and the content of the mail is grasped in a shorter time. be able to.

In addition, when reading out route guidance information and traffic information, it becomes possible to make it easier to hear, for example, slowly pronounce a distance or place name (unique name).
As described above, according to the present embodiment, the main control unit 4 automatically sets the speech reading mode according to the content of the text information or the input source type of the text information, and the set speech reading mode is set. Since the synthesized speech is output by performing speech synthesis based on the set of TTS parameters corresponding to, synthesized speech that is easier to hear can be obtained according to the reading target, and usability is improved.

In the above description, the main control unit 4 automatically sets the voice reading mode. However, the user can arbitrarily set the voice reading mode via the input unit 11. .
In the above description, the case where the text information input source is a mail application, a traffic information processing application, or a navigation application executed on the navigation device has been described, but various other applications may be used.
Furthermore, text information may be input not only from an application but also from an external device such as an external traffic information processing device or an Internet terminal.
In the above description, the navigation device has been described. However, any device can be applied as long as it can provide a speech synthesizer.

It is a block diagram which shows the functional structure of the navigation apparatus which concerns on embodiment of this invention. It is a general | schematic block diagram of an audio | voice data generation part. It is a flowchart of a setting process of a TTS parameter. It is a flowchart of an audio | voice output request | requirement process. It is a process flowchart of a TTS audio | voice output process. It is a flowchart which shows operation | movement of a navigation apparatus. It is a process flowchart of an audio | voice data generation part.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Navigation apparatus 1 Absolute position / direction detection part 2 Relative direction detection part 3 Vehicle speed detection part 4 Main control part (mode setting part, parameter specific part)
5 ROM
6 DRAM (parameter storage unit)
7 SRAM (parameter storage unit)
9 User interface part 10 Display part 11 Input part (mode designation part)
12 Disc control unit 13 FM multiplexed signal processing unit 14 External recording device control unit 15 Voice data generation unit (speech synthesis output unit, voice synthesis unit)
15A Synthesis unit sequence conversion unit 15B Speech synthesis control parameter generation unit 15C Speech synthesis rule database (DB)
15D sound source unit 15E voice synthesis filter 15F speaker

Claims (7)

In the speech synthesizing apparatus for performing speech voice line speech synthesis based on the input text information,
Depending on the input source type and e-mail read aloud mode, long sentence e-mail high-speed read-out mode, the mode setting unit that sets a text-to-speech mode to one of the traffic information reading mode or route guidance mode of the text information,
The combination of the speech synthesis control parameters corresponding to the set speech reading mode is specified, and in the mail reading mode or the long sentence high speed reading mode, proper nouns are pronounced largely and clearly, in the traffic information reading mode A parameter specifying unit that clearly pronounces unique names such as place names and interchange names, and pronounces clearly about distance, direction, and landmarks in the route guidance mode ;
A speech synthesizer that performs speech synthesis based on the identified speech synthesis parameters and outputs the synthesized speech;
A speech synthesizer characterized by comprising: The speech synthesizer according to claim 1.
The input source of the text information is a mail application, a traffic information processing application or a navigation application,
When the text information input source is a mail application, the mode setting unit sets the voice reading mode to the mail reading mode or the long text high-speed reading mode, and the text information input source is the traffic information processing If it is an application , the speech reading mode is set to the traffic information reading mode, and if the input source of the text information is a navigation application, the voice reading mode is set to the route guidance mode.
A speech synthesizer characterized by the above. A method of controlling a speech synthesis apparatus for performing speech voice I line speech synthesis based on the input text information,
Depending on the input source type and e-mail read aloud mode, long sentence e-mail high-speed read-out mode, the mode setting process of setting the text-to-speech mode to one of the traffic information reading mode or route guidance mode of the text information,
The combination of the speech synthesis control parameters corresponding to the set speech reading mode is specified, and in the mail reading mode or the long sentence high speed reading mode, proper nouns are pronounced largely and clearly, in the traffic information reading mode In the route guidance mode, a parameter specifying process for clearly pronounced distances, directions, and landmarks is pronounced .
A speech synthesis process in which speech synthesis is performed based on the identified speech synthesis parameters and output as the speech;
A method for controlling a speech synthesizer, comprising: The method of controlling a speech synthesizer according to claim 3,
The input source of the text information is a mail application, a traffic information processing application or a navigation application,
In the mode setting process, when the text information input source is a mail application, the voice reading mode is set to the mail reading mode or the long text high-speed reading mode, and the text information input source is the traffic information processing If it is an application , the speech reading mode is set to the traffic information reading mode, and if the input source of the text information is a navigation application, the voice reading mode is set to the route guidance mode.
A method of controlling a speech synthesizer characterized by the above. In the control program for controlling a computer speech synthesis apparatus for performing speech voice I line speech synthesis based on the input text information,
Depending on the input source type mail reading mode of the text information, long sentence e-mail high-speed read-out mode, to set the text-to-speech mode to one of the traffic information reading mode or route guidance mode,
The combination of the speech synthesis control parameters corresponding to the set speech reading mode is specified, and in the mail reading mode or the long sentence high speed reading mode, proper nouns are pronounced largely and clearly, in the traffic information reading mode Is to pronounce unique names such as place names, interchange names, etc., and in the route guidance mode, it should be pronounced clearly about distance, direction and landmarks ,
Performing speech synthesis based on the identified speech synthesis parameters and outputting as speech.
A control program characterized by that. The control program according to claim 5, wherein
The input source of the text information is a mail application, a traffic information processing application or a navigation application,
If the input source of the text information is a mail application, the text-to-speech mode is set in the e-mail reading mode or a long sentence e-mail high-speed read-out mode, when the input source of the text information is a traffic information processing applications The speech reading mode is set to the traffic information reading mode, and when the input source of the text information is a navigation application, the voice reading mode is set to the route guidance mode.
A control program characterized by that. Computer-readable recording medium characterized by recording a control program according to claim 5 or claim 6.

Images