JP4319334B2 - Audio / image processing equipment - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention provides a code conversion unit that converts sound data into a sound code image that is encoded so as to be optically readable, and an image image based on the sound code image and the image data to create an image sheet with sound. The present invention relates to an audio / image processing apparatus provided with a printing unit for printing.
[0002]
[Prior art]
In recent years, with the advent of the multimedia era, it has been actively attempted to use the auditory sense instead of relying solely on the visual transmission of information. For example, in Japanese Patent Application Laid-Open No. 6-231466 and Japanese Patent Application Laid-Open No. 7-181606, a dot code obtained by converting a voice in addition to a figure, a photograph, and a character to be optically readable An image sheet with sound is disclosed, in which (sound code image) is printed on the same sheet and sound is heard by a dedicated scanner that reads the dot code. Such an image sheet with sound is used for special occasions such as language teaching materials for repetitive study of pronunciation, photo pictorial books that record animal calls, picture books that make sounds, or commemorative events such as weddings, adult ceremonies, and Shichigosan. It is applied as a photograph to record with accompanying sound.
[0003]
[Problems to be solved by the invention]
Recently, a new photo card with a voice-coded message attached to an appropriate photo has been attracting attention. If you request a DP shop to create such a message card, print it out. In addition to submitting photographic film and digital camera recording media that contain the images you want, you also need to send voice messages through the microphone at the store. This is a very embarrassing act for the general public regardless of the content of the message. For this reason, many people are hesitant to create a message card. To avoid blowing voice messages at the store, you may bring a cassette tape or MD that has been voiced at home in advance, but there are many cases to play for confirmation, and for simple messages each time. Recording at home is annoying.
In view of the above situation, an object of the present invention is to improve an input method of audio data when requesting creation of an image sheet with audio.
[0004]
[Means for Solving the Problems]
In order to solve the above-described problems, a code conversion unit that converts audio data into an audio code image that is encoded so as to be optically readable, and the audio code image and the image data to generate an image sheet with audio. In a voice / image processing apparatus provided with a printing unit for printing an image based on the text input processing unit for processing input text data according to the present invention A speech synthesizer that generates synthesized speech data based on the text data processed by the text input processor; Is provided, The code converter uses the synthesized voice data generated by the voice synthesizer as source voice data for the voice code image, and Voice code image for playing text data voice The Generation You It is characterized by that.
[0005]
In this configuration, text data is input as the source data of the audio code image necessary for creating the image sheet with audio. Synthesized speech data using speech synthesis technology The customer does not need to blow the message content with the voice. As the text data input, for example, the message content may be directly typed using a keyboard connected to the sound / image processing apparatus, or a message previously created using a word processor or the like is recorded on a recording medium such as a floppy. And you can bring it into the store. Furthermore, it is also possible to send the message content to the store via e-mail. At that time, if the image data for the sound-added image sheet to be created is sent as an attached file, the customer can order the image sheet with sound. No longer need to go to the store.
[0008]
As an example of a speech synthesizer, in a preferred embodiment of the present invention, a phoneme sequence obtained by analyzing input text data using a text analysis dictionary and identifying its reading and further setting accents and prosody A text-to-speech synthesizer for generating synthesized speech data using a synthesized speech element dictionary. In this configuration, a text analysis dictionary is created for text data input by reading text data mixed with kana and kanji input from the keyboard, text documents saved on recording media, and text documents sent via e-mail. Use it to give accents to readings and phrases, and to adjust the sound power and fundamental frequency through the prosody parameter editing process while accessing the synthesized speech element dictionary, it is possible to set a certain tone. Accordingly, final synthesized voice data can be created by selecting a female tone, a male tone, or an angry tone or a joyful tone according to a customer's request. As a more preferable form in this regard, if the synthesized speech element stored in the synthesized speech element dictionary is created based on the personal voice data registered individually, the synthesized speech is familiar to the customer's real voice. It is also possible to use audio.
[0009]
Since the text-to-speech synthesizer as described above requires a high level of technology and is a heavy burden both in terms of apparatus and operation, as a simple speech synthesizer to replace this, another text-to-speech synthesizer is provided. In some embodiments, the input text data is provided with a voice editing / synthesizing unit that sequentially converts the input text data into synthesized voice data using a registered voice element dictionary that stores pre-registered vocabulary and phrase real voice data. is there. This is a method called edit synthesis that generates sentence speech from limited vocabulary / phrase fragments. Synthetic speech data is generated by registering text data fragments using a registered speech element dictionary. Therefore, high-speed processing can be performed and the apparatus cost can be reduced.
[0010]
In the present invention, the registered voice element dictionary stores the voice data registered for each individual so that the synthesized voice data of lower quality can be heard more intimately than the real voice obtained by such editing synthesis. Proposed to do. That is, the voice elements necessary for editing and synthesis are registered in advance with the customer's own voice for the customer of the image sheet with voice. When ordering an image sheet with audio, if you submit text data and appropriate image data as the source of the audio message, synthesized audio data is created by editing synthesis using the registered audio element of the person, so you can use fluent audio Even if the person's real voice fragment is used, it will be played as familiar voice.
[0011]
Furthermore, as a preferred embodiment of the present invention, if the speech synthesizer includes a voice quality transformation unit that transforms the voice quality of the synthesized speech data, it is possible to create voice data that does not exist in the world, It is suitable for the purpose such as exchanging image sheets with sound by feeling. Such sound quality deformation can be easily performed by, for example, linearly converting the frequency of the audio data. At this time, if the sound quality modification parameters are managed for each customer, the customer can secure the sound data modification parameters having unique sound characteristics as their own.
[0012]
As a measure for a customer who does not like to directly type in message contents using a keyboard, in one of the preferred embodiments of the present invention, a character recognition device is additionally provided, which is output by the character recognition device. The text data is used for voice code image conversion. Here, the character recognition device is a general term for devices that read characters handwritten on paper with OCR and convert them into text data, or read characters written with a predetermined pen on the touch panel into text data. The audio data input method for requesting creation of an image sheet with audio is further simplified, and the input forms are diversified.
[0013]
As is clear from the above description, an important feature of the present invention is that the text data is voiced. However, the text data handled in the present invention includes characters, numbers and symbols as printed text. It is interpreted in a broad sense, as represented by a collection of data read by a scanner for printed materials, a collection of characters, numbers, and symbols as digitized text, and a character code group that is sequentially input through an input device. It should include all data that is recognized in some way as character information in the media conversion technology of computers.
Other features and advantages of the present invention will become apparent from the following description of embodiments with reference to the drawings.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
One embodiment of a sound / image processing apparatus according to the present invention for creating a sound-added image sheet is shown by the external view of FIG. 1 and the functional block diagram of FIG. The core of this sound / image processing apparatus is a general-purpose computer 1, and various functions required for creating the sound-added image sheet shown in FIG. 2 are constructed by hardware and software. This audio / image processing apparatus is of a type that is arranged at a store such as a DP shop and provides a service for creating a photo with audio at the request of a customer.
[0015]
Various input devices and output devices are connected to the computer 1 via the I / O interface unit 10. As an output device, a silver halide photographic printer 3 (which is also used for printing a silver halide photographic film, etc.) is used as a print unit for finally outputting a photograph 2 with sound as an image sheet with sound. Examples include a monitor 4 for confirming an image inside and a speaker 5 for checking input voice data. Input devices include a microphone 6a and a cassette player 6b for directly inputting sound into the computer 1, a card reader 7a for capturing a captured image by a digital camera, and a captured image from a silver halide film. Film scanner 7b. Furthermore, examples of devices for inputting text data to the computer include a keyboard 8a, a flatbed scanner 8c for reading handwritten or printed characters, and a communication device 8d for receiving text data sent through the Internet.
[0016]
The computer 1 also includes a floppy drive 8e and an MO drive 8f that are often used for inputting and outputting audio data and image data. The keyboard 8a is used together with the mouse 8b to give commands to the functions shown in FIG. 3, and the communication device 8d can naturally receive not only text data but also image data.
[0017]
The mechanism for creating the photograph 2 with sound using the input image data and sound data will be described in detail later. In the photograph 2 with sound output from the silver halide photographic printer 3, the photograph image is shown in FIG. When the audio code image area 2b arranged around the area 2a is scanned by the dedicated reading scanner 90, an audio signal corresponding to the audio code image is output by the operation of the audio reproduction circuit built in the reading scanner 90. For example, the sound can be heard with the earphone 91 or the like.
[0018]
As can be understood from FIG. 2, the sound / image processing apparatus is an image input processing unit 21 that receives image data as a source of a photographic image in the sound-added photo 2, and the sound in the sound-added photo 2 as main functional units. A voice input processing unit 22 that directly receives voice data as a source of a code image from outside, a text input processing unit 23 that receives text data to be a source of voice data to be converted into a voice code image, and based on the input text data A speech synthesis unit 30 for generating synthesized speech data, a code conversion unit 40 for converting speech data into a speech code image coded so as to be optically readable, an image data storage unit 51, and a speech code image storage unit 52. And audio from properly processed image data and audio code images And an image voice synthesizing section 60 for generating print data for photographic 2 come.
[0019]
The image input processing unit 21 includes an image editing unit 21a and an image selection unit 21b, and image data input from the card reader 7a, the film scanner 7b, the communication device 8d, the floppy drive 8e, the MO drive 8f, or the like is necessary. Accordingly, the image selection unit 21b selects the selected image data, and the image editing unit 21a performs editing processing such as color correction and resolution conversion.
[0020]
The audio input processing unit 22 is used when audio data as a source of an audio code image formed in the photo 2 with audio is directly given from a customer. The audio input processing unit 22 uses a microphone 6a, a cassette player 6b, and a card reader 7a. Audio data input from (when using a digital voice recorder memory card) is selected by the audio selection unit 22b as necessary, and editing processing is performed by the audio editing unit 22a.
[0021]
The text input processing unit 23 is used when a customer gives text data as a source of an audio code image formed in the sound-added photo 2, and is a text file or e-mail stored on a floppy disk brought by the customer. After the text data sent in the form of is received through the floppy drive 8e or the communication device 8d, the text editing unit 23a performs necessary text editing. Further, text data directly input by the customer or operator through the keyboard 8a is also processed by the text editing unit 23a. Furthermore, as an option, a character recognition unit 24 can be provided to provide an OCR function. When the OCR function is provided, the message document presented by the customer is read by the flatbed scanner 8c, and then converted into text data by the character recognition unit 24. That is, the flat bed scanner 8c and the character recognition unit 24 constitute a character recognition device.
[0022]
A speech synthesizer 30 that converts text data that has undergone the editing processing required by the text input processor 22 into synthesized speech data includes a text speech synthesizer 31, a text analysis dictionary 32, and a synthesized speech element dictionary 33. The text-to-speech synthesizer 31 analyzes the input text data using the text analysis dictionary 32 to identify the reading, and further uses the synthesized speech element dictionary 33 from the phoneme sequence obtained by setting the accent and prosody. Generate synthesized speech data. The voice as the source of the synthesized voice element dictionary 33 may be either a female voice or a male voice, but it is also possible to select both. Furthermore, it is possible to prepare a large number of synthesized speech element dictionaries 33 that use a specific person's voice as a source and use them by arbitrarily switching them.
[0023]
Furthermore, the voice synthesis unit 30 is also accompanied by a voice quality transformation unit 34 that transforms the voice quality of the synthesized voice data created as described above. By changing the frequency linearly by down-sampling and adjusting the time axis, the tape recorder is transformed and output in a manner similar to fast or slow rotation. The voice transformation unit 34 can perform voice transformation processing on the voice data sent from the voice input processing unit 22.
[0024]
The voice code conversion unit 40 that converts the voice data of the real voice sent from the voice input processing unit 22 and the synthesized voice data sent from the voice synthesis unit 30 into a voice code image includes waveform coding, analysis synthesis coding, and the like. An audio data compression encoding unit 41 constructed by appropriately selecting from known encoding methods, and an audio code image generation unit 42 for expanding the audio code data encoded thereby into a two-dimensional code image; A pre-calculation of the size (outside dimension) of the audio code image formed on the audio-attached photo 2 is convenient for layout editing of the audio-attached photo of the image image and audio code image based on the image data to be performed later. And a code image generation unit 43.
[0025]
The image data edited by the image input processing unit 21 is temporarily stored as an image image in the image data storage unit 51, and the voice code image converted by the code conversion unit 21 is temporarily stored in the voice code image storage unit 52. The processing unit 60 converts the data into print data so as to be printed out by the printer 3 with a desired layout. For this reason, the image / speech synthesis processing unit 60 performs an image / sound code image layout editing unit 61 that performs layout processing of the image image stored in the image data storage unit 51 and the sound code image stored in the sound code image storage unit 52. And an image / audio code image synthesis processing unit 62 for synthesizing both images with the determined layout to generate print data. At the time of editing the layout, a dummy box area based on the size of the audio code image calculated by the precode image generation unit 43 is displayed on the monitor 4, and an accurate comparison is made while comparing the positional relationship with the displayed image image. Layout work is possible.
[0026]
A typical procedure for creating the photograph 2 with sound by the sound / image processing apparatus described above will be described with reference to the flowchart of FIG. Here, it is assumed that the order of the photo with sound 2 is made by e-mail.
When the e-mail arrives (# 1), image data as an attached file of the e-mail is input to the image input processing unit 21 (# 11), and the image data is operated by the image editing unit 21a, and the monitor 4 Then, editing processing such as color tone / gradation conversion and enlargement / reduction is performed through the operation of the operator while confirming the image (# 12). When there are a plurality of input images, the editing process is performed after the images are selected by the image selection unit 21b. The edited image data is temporarily stored in the image data storage unit 51 (# 13).
[0027]
On the other hand, the e-mail file including the text data as the source of the audio code image is sent to the text editing unit 23a of the text input processing unit 23 (# 14). The text data including only the message to be fetched is cut out (# 15).
[0028]
The text data sent to the speech synthesizer 30 as text mixed with kanji characters is analyzed by the text speech synthesizer 31 while accessing the text analysis dictionary 32 (# 21), read while identifying words, and accented. (# 22). Next, the breath connection position is set and the intonation of the whole sentence is determined, and a phoneme sequence including phoneme symbols and prosodic parameters is created (# 23). The synthesized speech elements are sequentially connected to the generated phoneme sequence while accessing the synthesized speech element dictionary 33 to generate synthesized speech data (# 24).
[0029]
When voice quality modification processing is required for this synthesized voice data (# 25 YES branch), frequency linear transformation or the like is performed by the voice quality transformation unit 34 (# 26), and when not required (# 25 NO branch), synthesized voice The data is sent to the code conversion unit 40 as it is.
[0030]
First, the synthesized speech data is sent to the speech data compression encoding unit 41, subjected to compression processing, and then converted into an optically readable speech code image by the speech code image generation unit 42 (#). 31). Further, the size (outer dimension) of the voice code image is calculated by the precode image generation unit 43 (# 32), and the size data together with the voice code image data is temporarily stored in the voice code image storage unit 52 (# 33). ).
[0031]
The image data stored in the image data storage unit 51 and the audio code image stored in the audio code image storage unit 52 are respectively taken into the image / audio code image layout editing unit 61 of the image audio synthesis processing unit 60 to generate an image. Image and voice code image layout editing processing is performed (# 40). In the actual layout editing process, a layout editing screen is displayed on the screen of the monitor 4, and a dummy box area that shows a pseudo image image and audio code image is edited in accordance with an instruction from the cursor. This layout editing can adopt a method of automatically flowing an image image and a voice code image using a preselected template. At this time, for example, when the length of the voice code image exceeds the printable length, re-editing is performed such that the voice code image is separated into two parts to form a two-stage configuration.
[0032]
Upon receiving the layout information from the image / sound code image layout editing unit 61, the image / sound code image synthesis processing unit 62 transmits a request signal to the image data storage unit 51 and the sound code image storage unit 52, respectively. Receive data and voice code image data. The received image image data and audio code image data are integrated based on the layout information and generated as print data (# 41). When this print data is transmitted to the printer 3, the image image and the audio code image are exposed on the photographic paper, and the exposed photographic paper is developed to produce a photo 2 with audio as shown in FIG. (# 50).
[0033]
[Another embodiment]
In the functional block diagram of another embodiment of the present invention shown in FIG. 5, the speech synthesizer 30 replaces the text-to-speech synthesizer 31 with a speech editing synthesizer, as compared with the previous embodiment shown in FIG. It is different in that it is constituted by the portion 35.
[0034]
In order to generate synthesized speech data using this method, which is called edit synthesis that generates sentence speech from limited vocabulary / phrase fragments, registered speech that stores pre-registered vocabulary and phrase speech data The element dictionary 36 is necessary, and the speech editing / synthesizing unit 35 fragments the text data sent from the text input processing unit 23 and replaces the fragment with a fragment of speech data using the registered speech element dictionary.
[0035]
In this embodiment, the registered voice element dictionary 36 can store real voice data registered for each individual. In other words, the voice elements required for editing and synthesis are registered and stored in advance with the customer's own voice for the customer of the image sheet with sound, and the person is registered when ordering the image sheet with sound. Synthetic speech data is created by editing synthesis using speech elements. For customers who are not registered, the standard voice elements are used.
[0036]
The audio / image processing apparatus of this embodiment has a box-shaped appearance such as an ID photo apparatus or a photo booth apparatus as shown in FIG. After charging the fee, in accordance with the instruction message displayed on the monitor 4, take a picture with the provided digital camera and input using the touch panel keyboard 8 a provided with text data to be converted into a voice message Or, the voice is input through the microphone 6a. Further, a sublimation type thermal transfer printer is employed as the printing unit 3.
[0037]
A typical procedure for creating the photograph with sound 2 by the sound / image processing apparatus according to another embodiment will be described with reference to the flowchart of FIG. Here, the image source for the photograph 2 with sound is captured image data of the digital camera, and the sound source is text data directly input from the provided keyboard 8a.
[0038]
A customer who wishes to create a photograph 2 with sound inserts a designated coin into the coin slot (# 101), and according to the menu displayed on the monitor 4, first the ID photograph apparatus or the The subject is photographed in the same procedure as the photo booth apparatus (# 110). Since this digital camera is directly connected to the I / O interface 10, the image data acquired by the digital camera is immediately transferred to the image input processing unit 21 (# 111). The image data transferred to the image input processing unit 21 can be subjected to editing processing such as trimming and enlargement / reduction while confirming the image image on the monitor 4 by the function of the image editing unit 21a (# 112). The edited image data is temporarily stored in the image data storage unit 51 (# 113).
[0039]
Subsequently, since text data by keyboard input is selected as source data for converting into a voice code image this time, a voice message to be incorporated into a photograph with sound is input from the keyboard 8a as a sentence (# 114). Since the text editing unit 23a has a function as a text editor, a text is created from data input through the keyboard 8a, and finally the text data is converted into a format suitable for editing and synthesis and sent to the speech synthesis unit 30. (# 115).
[0040]
In the editing / synthesizing process, first, it is checked whether or not the customer has registered in advance (# 121). If the voice is registered, the registered voice element file of the customer is loaded (# 122). There are various modes for loading the registered voice element file. Here, two typical modes are introduced.
[0041]
The first one was obtained by the customer registering the necessary vocabulary / phrase with the voice element registration device in advance with the voice, and file the registered vocabulary / phrase in a format suitable for editing and synthesis. In this mode, an audio element file is recorded on a memory card. By inserting the corresponding memory card into the card reader 7a at the stage of checking whether or not the voice is registered, the registered voice element file is loaded into the registered voice element dictionary 36 of the voice synthesizer 30. The second is a form in which a voice element file created in advance by a voice element registration device is stored in the registered voice element dictionary 36 using a customer ID as a key. Is set so that the registered voice element file of the customer is used as the registered voice element dictionary 36 in the subsequent editing and synthesis work. The registered voice element dictionary 36 is preferably not provided in the voice / image processing apparatus but provided in a server connected by a communication line. That is, when the customer ID is input, the registered voice element file of the corresponding customer is loaded into the registered voice element dictionary 36 of the voice synthesizer 30 through the communication line.
[0042]
When the voice is not registered, the standard voice element file stored in the registered voice element dictionary 36 is set to be used as the registered voice element dictionary 36 in the subsequent editing / synthesis work. (# 123).
[0043]
In any case, in the editing / synthesizing process, the message sentence represented by the text data to be processed is first decomposed into vocabulary / phrases (# 124), and each is extracted from the speech element file as the registered speech element dictionary 36. The fragmented speech elements are assigned to generate synthesized speech data (# 125).
[0044]
When voice quality modification processing is required for this synthesized voice data (# 25 YES branch), frequency linear transformation or the like is performed by the voice quality transformation unit 34 (# 26), and when not required (# 25 NO branch), synthesized voice The data is sent to the code conversion unit 40 as it is, and the voice code image of the synthesized voice data is converted as described above in steps # 31 to # 33, and the generated voice code image is temporarily stored in the voice code image storage unit 52. Is done.
[0045]
The image data stored in the image data storage unit 51 and the audio code image stored in the audio code image storage unit 52 are subjected to layout editing processing by the image / audio code image layout editing unit 61 using a template selected in advance. (# 40).
[0046]
The image / audio code image synthesis processing unit 62 integrates the image image data and the audio code image data based on the layout information, and generates print data (# 41). When this print data is transmitted to the printer 3, the image image and the voice code image are printed on a dedicated sheet, and discharged as a photograph 2 with sound as shown in FIG. (# 50).
[0047]
In the above-described embodiment, the image data and the voice code image are synthesized by the image / speech synthesis processing unit 60. However, the image / speech synthesis processing unit 60 is omitted and printed out by the printer 3. The image data and the voice code image may be printed out by different printers. At this time, it is preferable that the sticker on which the voice code image is printed is pasted on the sheet on which the voice code image is formed, for example, a photographic print, with a sticker printer.
[0048]
Furthermore, in all the embodiments described above, the input text data is once converted into synthesized speech data by the speech synthesizer 30 and this synthesized speech data is converted into a speech code image. It is also possible to convert the processed text data directly into a voice code image. As shown in FIG. 8, such a voice / image processing apparatus converts the text data obtained by fragmenting text data into predetermined elements in the code conversion unit 40 instead of omitting the voice synthesis unit 30. A text / voice code image replacement unit 44 for sequentially assigning corresponding voice code images and a voice code image dictionary 45 in which voice code images corresponding to text elements are registered are provided. In other words, a final speech code image is created by sequentially joining together while applying speech code images corresponding to vocabulary and phrases constituting the text.
[Brief description of the drawings]
FIG. 1 is an external view showing one embodiment of an audio / image processing apparatus according to the present invention.
2 is a functional block diagram of the audio / image processing apparatus according to FIG. 1;
FIG. 3 is an explanatory diagram showing a state in which sound is reproduced from a sound-added photo created by the sound / image processing apparatus.
FIG. 4 is a flowchart showing a procedure for creating a photograph with sound using the sound / image processing apparatus shown in FIG. 2;
FIG. 5 is a functional block diagram showing another embodiment of the sound / image processing apparatus according to the present invention.
6 is an external view of the sound / image processing apparatus according to FIG. 5;
FIG. 7 is a flowchart showing a procedure for creating a photograph with sound using the sound / image processing apparatus shown in FIG. 5;
FIG. 8 is a functional block diagram showing still another embodiment of the sound / image processing apparatus according to the present invention.
[Explanation of symbols]
2 Image sheet with sound (photo with sound)
3 Print section (silver salt photo printer, sublimation thermal transfer printer)
21 Image input section
22 Voice input part
23 Text input processing section
24 character recognition part
30 Speech synthesis unit
31 Text-to-speech synthesizer
32 Text analysis dictionary
33 Synthetic Speech Element Dictionary
34 Voice quality transformation
35 Voice Editing / Synthesizer
36 Registered Voice Element Dictionary
60 Image / sound synthesis processor

Claims (7)

Code conversion unit for converting audio data into an audio code image encoded so as to be optically readable, and a print for printing an image image based on the audio code image and the image data in order to create an image sheet with audio In the sound / image processing apparatus provided with
A text input processing unit for processing input text data ;
A speech synthesizer that generates synthesized speech data based on the text data processed by the text input processor ,
Using the synthetic speech data to which the code converter is generated by the speech synthesizer as the source audio data for the audio code image, characterized that you generate the audio code image for reproducing audio of the text data An audio / image processing apparatus. The speech synthesizer analyzes input text data using a text analysis dictionary to identify the reading, and further uses a synthesized speech element dictionary from a phoneme sequence obtained by setting accents and prosody. The speech / image processing apparatus according to claim 1 , further comprising a text-to-speech synthesis unit that generates data. The voice / image processing apparatus according to claim 2, wherein the synthesized voice element stored in the synthesized voice element dictionary is created based on real voice data registered for each individual. The speech synthesizer includes a speech editing / synthesizing unit that sequentially converts input text data into synthesized speech data using a registered speech element dictionary storing pre-registered vocabulary and phrase real voice data. The audio / image processing apparatus according to claim 1 . 5. The sound / image processing apparatus according to claim 4 , wherein the registered sound element dictionary stores real voice data registered for each individual. The voice / image processing apparatus according to claim 1 , wherein the voice synthesizer includes a voice quality transformation unit that transforms voice quality of the synthesized voice data. Character recognition apparatus is provided with additionally, speech and image processing according to any one of claims 1 to 6, the text data outputted by the character recognition apparatus, characterized by being used in the voice code image converting apparatus.

Images