JPH10222192A - Substitute reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a substitute reader which can generate a synthesized voice corresponding to text data by using the voice of an arbitrary voice input person. SOLUTION: A speech recognition device 22 extracts phoneme data by analyzing the voice that an arbitrary voice input person has inputted by using a microphone 21. The phoneme data are stored in a phoneme data storage part 13. A voice synthesis part 10 composes a text by using text data read out of a CD-ROM by using a CD-ROM drive 11 and the phoneme data stored in the phoneme data storage part 13 and outputs a synthesized voice from a speaker 14. Therefore, the text can be read aloud with the voice of the arbitrary voice input person.

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 reading apparatus for reading out given text data with the voice of an arbitrary voice input person.

[0002]

2. Description of the Related Art Conventionally, in the field of speech synthesis technology, there is a technique called text synthesis in which synthesized speech corresponding to text data is output by performing speech synthesis according to given text data. Are known. In text synthesis, speech units (for example,
(B, c,...) Are applied to each phoneme data from text data by using phoneme data (a combination of a plurality of parameters and a PCM sound source are known). It produces synthesized sounds. The phoneme data determines the sound quality of the synthesized sound. Further, when generating a synthetic sound, there is a method that uses prosody data for giving inflection according to a word unit or a connection between words.

[0003]

In the above-described apparatus for synthesizing text, generally, the maker creates phonemic data, and the apparatus uses an appropriate storage medium such as a ROM or a magnetic disk. Supplied.
Therefore, the user can generate the synthesized sound only in the sound quality in the range prepared by the manufacturer.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to provide a substitute reading apparatus capable of generating a synthesized sound equivalent to text data using a voice of an arbitrary voice input person. It is in.

[0005]

According to the first aspect of the present invention, there is provided text input means for inputting text data, and phonological data of a voice input person by sampling voice input by an arbitrary voice input person through a voice input device. , A phoneme data storage unit that stores the phoneme data extracted by the speech recognition unit for each voice unit, and text data input by the text input unit and phoneme data stored in the phoneme data storage unit. And a voice synthesizing means for synthesizing voice corresponding to the text data based on the voice data and outputting a synthesized voice from the voice output device.
According to this configuration, since the phonetic data is extracted from the voice input by the voice input user through the voice input device and stored in the phonemic data storage unit, it is possible to generate the voice of any voice input user as a synthesized voice. As a result, it becomes possible to read out text corresponding to text data with various human voices.

According to a second aspect of the present invention, in the first aspect of the present invention, an instruction means for prompting a voice input person to speak is added for each voice unit stored in the phoneme data storage unit. According to this configuration, the instruction means instructs how to input the phoneme data, so that appropriate phoneme data can be given. According to a third aspect of the present invention, in the first or second aspect of the present invention, there is provided a reading device for reading text data from a recording medium recording text data corresponding to the content of a separately prepared print medium, Are used as text input means. According to this configuration, the print medium on which the text is printed is paired with the recording medium on which the text data is recorded, and the text data on the recording medium can be read aloud by an arbitrary voice input person while watching the print medium. Things. For example, if a picture book for a child is prepared as a print medium and the phoneme data of the parent's voice is registered in the phoneme data storage unit, the content of the picture book is read out by the voice of the parent when the child views the picture book. Such a use becomes possible, and the picture book can be viewed while listening to the voice familiar to the child.

According to a fourth aspect of the present invention, in the first or the second aspect of the present invention, there is provided a display means capable of displaying characters and graphics, and image data and still image data of a still image or a moving image to be displayed on the display means. And a reading device for reading image data and text data from a recording medium on which the text data is recorded. The reading device is also used as text input means. According to this configuration, it is possible to hear the voice of the voice input person corresponding to the text data while watching the image displayed on the display means. Therefore, if the image to be displayed on the display means is an image for a child such as a picture book, and the phonemic data of the parent's voice is registered in the phonemic data storage unit, the image can be viewed while listening to the parent's voice, The child can view the image while listening to a voice that is comfortable. By the way, in recent years, an electronic picture book or the like that displays an image like a picture book on a screen of a display device and interactively switches images and generates sound related to the image has been marketed. The invention of Item 4 is different from this type in that a voice of an arbitrary voice input person is generated.

According to a fifth aspect of the present invention, in the first to fourth aspects of the present invention, there is provided a pitch control section for controlling a pitch of the synthesized sound in accordance with an external signal. According to this configuration, it is possible to control the pitch. In addition to simply changing the pitch of the sound, if the pitch data for controlling the pitch is associated with the text data, a pitch is assigned to the text data. This makes it possible to output songs.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION

(Embodiment 1) This embodiment shows an example in which text data is stored in a CD-ROM as a storage medium. However, as the storage medium, an IC card, a flexible disk, or the like can be used.

As shown in FIG. 1, as a reading device for reading text data recorded on a CD-ROM, a CD is used.
A ROM drive 11 is provided, and text data read by the CD-ROM drive 11 is input to the speech synthesis unit 10 through the text selection unit 12. The speech synthesis unit 10 performs text synthesis corresponding to the text data using the phoneme data registered in the phoneme data storage unit 13, and outputs a synthesized sound from a speaker 14 as a speech output device. That is, the speech synthesis unit 10 generates a synthesized speech using phoneme data for each speech unit in the text data. The speech synthesis unit 10 is provided with a dictionary for generating prosody data by performing semantic analysis of text data, and intonation is determined by the prosody data. The technique of this type of text synthesis (rule synthesis based on text data) is well known in the art and will not be described in detail.

The text selection unit 12 inputs to the speech synthesis unit 10 either text data read from the CD-ROM in the CD-ROM drive 11 or text data of a message prepared in advance in the message storage unit 15. When the message storage unit 15 is selected, various messages can be output from the speaker 14 as audio.

By the way, the present invention is characterized in that the phoneme data recorded in the phoneme data storage section 13 is prepared by an arbitrary voice input person, and therefore, the voice input device is used as a voice input device for inputting voice. The microphone includes a microphone 21 and a voice recognition unit 22 that analyzes voice input through the microphone 21 and generates phoneme data. The voice recognition unit 22 samples voice input through the microphone 21 and extracts parameters necessary for rule synthesis from the sampled voice. Such parameters are registered in the phoneme data storage unit 13 as phoneme data.

It is necessary to register phoneme data for each voice unit. The voice unit includes 50 sounds such as a, a, c,..., And a phoneme chain (vowel-consonant-vowel connection).
Additional information such as is used as needed. By using such additional information as phoneme data, the clarity of the synthesized sound can be increased. In addition, since the phoneme data must be registered in association with the voice input from the microphone 21 and the unit of the voice, the voice input user needs to input the voice according to a predetermined procedure. For this,
For example, it is conceivable to read out a predetermined text at a specified timing, but it is necessary to repeat the work several times until the correspondence between the voice unit and the phoneme data is correctly performed. Is troublesome.

Therefore, in the present embodiment, a message instructing the procedure of phoneme data registration work is registered in the message storage unit 15 so that the phoneme data can be registered relatively easily by working in accordance with the voice message. It is. For example, after a voice to be pronounced by the voice input person is transmitted from the speaker 14, the voice is repeated and the correspondence between the voice unit and the phoneme data is facilitated. In addition, the same sound is repeatedly generated a plurality of times, an abnormal value is discarded, and an average value is extracted, so that appropriate phoneme data can be set. That is, the message storage unit 15 is the main configuration of the instruction unit. Note that required phoneme data is registered in the phoneme data storage unit 15 as a default in advance, and when generating a synthesized sound corresponding to the message stored in the message storage unit 15,
Use phoneme data registered as default.

The control of each unit described above is performed by a control unit 20 composed of a microcomputer, and the control unit 20 instructs the timing of generating phoneme data and synthesizing speech according to text data. That is, at the time of creating the phonemic data, the speech is synthesized using the text data corresponding to the message stored in the message storage unit 15.
An instruction is given to the message storage unit 15, the text selection unit 12, and the speech synthesis unit 10, and the parameters obtained by the speech recognition unit 22 are associated with the units of speech and the phoneme data storage unit 1 is used.
3 to the phoneme data storage unit 13.

Even if the phoneme data is registered in the phoneme data storage unit 13 as described above, a case may be considered in which the phoneme data is such that a clear synthesized speech cannot be obtained. Therefore, after registering as much phoneme data as necessary in the phoneme data storage unit 13, a synthesized sound corresponding to a required message stored in the message storage unit 15 is generated using the phoneme data. The voice singer checks whether all messages are clear or not, and if there is an unclear message, the operation unit 2
Operate 3 to indicate which message was ambiguous. As a result, the phonemic data is re-registered for the unit of voice included in the message.

As described above, if the phoneme data is stored in the phoneme data storage unit 15, a synthesized sound corresponding to text data can be generated using phoneme data based on the voice of the voice input person. Thus, the text data can be read aloud by the voice of the voice input person. A CD-ROM as a storage medium used in the above device is a CD-R
If the content is sold in a pair with a print medium having the content corresponding to the text data registered in the OM, the text can be read aloud by the voice of the voice input person while looking at the print medium. For example, if a print medium is a picture book and text data corresponding to the contents of the picture book is stored in a CD-ROM, text corresponding to the contents of the picture book can be read out by the voice of the voice entrant. A parent can register a voice for a picture book for a child and read it to a child in the voice of the parent. In addition, there is no need for the parent to record the entire text as in a tape recorder, etc., the text data is prepared by the manufacturer, and the parent only registers the voice required to generate phoneme data. Since the same phonological data can be used even if the data changes, the work performed by the voice input person is significantly reduced as compared with the case where the voice input person records all of the plurality of text data.

In the above-described example, all the messages stored in the message storage unit 15 are given by synthetic sounds. However, display means capable of displaying at least characters is provided.
Some messages may be displayed on the display means. That is, the display means may be used as the instruction means to give a message in characters. Such a message can be used when creating phonemic data.

In the above embodiment, the parameters for rule synthesis are used as phoneme data.
It is also possible to A / D convert the voice input from 1 and sampled by the voice recognition unit 22 and use it for phoneme data. That is, a so-called PCM sound source is used as the phoneme data. Further, it is possible to apply dedicated hardware to a part of the above-described configuration, or to combine software with a general-purpose computer device without using dedicated hardware.

(Embodiment 2) In Embodiment 1, text data is provided on a storage medium, and a separately prepared print medium is associated with the text data of the storage medium. Is used, a sufficient amount of image data other than text data can be stored at the same time. Therefore, in this embodiment, C is used as the storage medium.
Using a D-ROM, image data of a still image or a moving image;
The text data corresponding to the image data and the CD-
5 shows an example of storing in a ROM. As this kind of image data, image data such as a picture book for children, which is known as an electronic picture book or the like, can be used.

In this embodiment, as shown in FIG. 2, a display device 16 is provided as a display means for displaying image data among data read by the CD-ROM drive 11 as a reading device. Display device 1
6 is a CRT or a liquid crystal display capable of displaying characters and figures. It is desirable that the image displayed on the display device 16 be one that interactively changes the image. That is, by providing a pointing device such as a mouse on the operation unit 23, or by providing a touch panel on the screen of the display device 16, a desired position on the screen can be designated, and according to the designated position. Various processes such as changing a part of an image, outputting a sound corresponding to the position, or switching a scene are performed. The output voice is a synthesized voice generated by the voice synthesis unit 10 using the phoneme data stored in the phoneme data storage unit 13.
The text data used by the voice synthesizer 10 is read from the CD-ROM according to the designated location on the screen. In the present embodiment, text may be displayed on the screen of the display device 16 as necessary.

The present embodiment is different from the first embodiment in that image data is stored together with text data on a CD-ROM and is displayed on a display device 16. However, other components having the same reference numerals are the same as those in the first embodiment. It works. (Embodiment 3) In the present embodiment, as shown in FIG. 3, a pitch control unit 17 capable of adjusting the pitch of the phoneme data read from the phoneme data storage unit 13 (that is, the pitch of the synthesized sound) is provided. The pitch of the phoneme data is specified by the control unit 20. The pitch instruction can be set in a stepwise manner by the operation unit 23, or can be stored in a CD-ROM together with text data, and the pitch can be changed by pitch information associated with the text data. I have.

In the former case, it is possible to adjust the clarity of the synthesized sound by changing the pitch. In the latter case, if a pitch is given as pitch information, the text is given a pitch. You can sing a song. The other components denoted by the same reference numerals function in the same manner as in the first embodiment.

[0024]

According to the first aspect of the present invention, there is provided a text input means for inputting text data, and a voice for extracting phoneme data of the voice input person by sampling a voice input by an arbitrary voice input person through a voice input device. A recognition unit, a phoneme data storage unit that stores phoneme data extracted by the speech recognition unit for each unit of speech, and a text data input by the text input unit and a phoneme data stored in the phoneme data storage unit. Speech synthesis means for synthesizing speech equivalent to text data and outputting a synthesized sound from a speech output device, the phoneme input unit extracts phoneme data from speech input through the speech input device, and stores the phoneme data in the phoneme data storage unit. Because it is stored, it is possible to generate the voice of any voice input person as a synthesized voice,
As a result, there is an advantage that a text corresponding to the text data can be read out by the voice of the voice input person.

According to the second aspect of the present invention, in the case where an instruction means for prompting a speech input person to speak is added for each unit of speech stored in the phoneme data storage unit, the instruction means instructs how to input the phoneme data. Therefore, there is an advantage that appropriate phoneme data can be given. According to the third aspect of the present invention, there is provided a reading device for reading text data from a recording medium recording text data corresponding to the content of a separately prepared printing medium, and using the reading device as text input means. A print medium on which text is printed and a recording medium on which text data is recorded are paired, and text data on the recording medium can be read aloud by an arbitrary voice input person while looking at the print medium.

According to a fourth aspect of the present invention, there is provided a display means capable of displaying characters and figures, and a recording medium which records still image data or moving image image data to be displayed on the display means and text data related to the image data. A reading device for reading image data and text data, wherein the reading device is also used as the text input means, so that the voice of the voice input person corresponding to the text data can be heard while watching the image displayed on the display means. .

According to the fifth aspect of the present invention, in a device having a pitch control unit for controlling the pitch of a synthesized sound in accordance with an external signal, the pitch can be controlled, and the pitch of the sound is simply changed. In addition, if pitch data for controlling the pitch is associated with text data, there is an advantage that a song can be output by attaching a pitch to the text data.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a block diagram showing Embodiment 2 of the present invention.

FIG. 3 is a block diagram showing a third embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Speech synthesis part 11 CD-ROM drive 12 Text selection part 13 Phoneme data storage part 14 Speaker 15 Message storage part 16 Display device 17 Pitch control part 20 Control part 21 Microphone 22 Voice recognition part 23 Operation part

Claims (5)

[Claims] 1. Text input means for inputting text data, voice recognition means for extracting phoneme data of a voice input person by sampling voice input by an arbitrary voice input person through a voice input device, and voice recognition means A phoneme data storage unit that stores the phoneme data extracted by the above for each voice unit, and a speech corresponding to text data based on the text data input by the text input unit and the phoneme data stored in the phoneme data storage unit. A voice synthesizing means for synthesizing and outputting a synthesized sound from a voice output device. 2. The substitute reading device according to claim 1, further comprising an instruction means for instructing a procedure for storing phoneme data based on the voice of the voice input person in the phoneme data storage unit. 3. A reading device for reading text data from a recording medium recording text data corresponding to the content of a separately prepared printing medium, wherein the reading device is used as text input means. The substitute reading device according to claim 1 or 2. 4. Display means capable of displaying characters and graphics,
A reading device for reading image data and text data from a recording medium on which image data of still or moving images to be displayed on the display means and text data related to the image data are recorded, wherein the reading device is also used as the text input means. The substitute reading device according to claim 1 or 2, wherein 5. The substitute reading apparatus according to claim 1, further comprising a pitch control unit that controls a pitch of the synthesized sound in accordance with an external signal.