JP4895759B2 - Voice message output device - Google Patents

Description

  The present invention relates to an apparatus capable of outputting a voice message, and is used as, for example, a television broadcast receiving apparatus or a mobile phone.

2. Description of the Related Art Conventionally, a television broadcast receiving and recording device equipped with a voice message output device that performs voice operation guidance has been proposed. In such a television broadcast receiving and recording apparatus, for example, a setting procedure relating to recording reservation is voice-guided (see Patent Document 1 and Patent Document 2).
Japanese Patent Laid-Open No. 11-69283 JP 2005-338665 A

  However, in the conventional voice message output, the voice guidance is output according to the voice (timbre) registered in advance, so that the voice guidance desired by the user cannot be created. As voice guidance, there are those that prepare “standard words” and “local languages” and allow the user to select arbitrary words (see Patent Document 2), but those that have already been registered are simply used. The user can simply select or select (registered) voice data existing on the net, and cannot listen to the user's favorite voice guidance edited by the user.

  In view of the above circumstances, an object of the present invention is to provide a voice message output device that allows a user to listen to a favorite voice guidance edited by the user.

  In order to solve the above problems, the voice message output device according to the present invention selects a voice message corresponding to an event that has occurred and outputs the selected voice message. Means for presenting the user with an event that is currently selected or for which an audio message can be set, means for generating character data or phrase data from speech data by speech recognition, and generated characters Means for associating the data or phrase data with the voice data that caused the generation thereof, means for displaying a character or phrase based on the generated character data or phrase data for user selection, and selected by the user Character string generation means for generating a character string from characters or phrases, and registration for the generated character string A means for determining whether or not an operation has been performed by the user, and when the registration operation has been performed, voice output is performed by joining the voice data based on a character string created by the user with respect to a corresponding event. And a means for executing change registration of voice messages.

  With the above configuration, for example, characters and phrase data are obtained from voice data of a singer that the user likes, a character string (message sentence) is created by combining the characters and phrases, and based on this character string The voice output can be output as a voice message. That is, the user can listen to a favorite voice message edited by the user.

  The voice message output device having the above-described configuration may be configured to acquire the voice data by any one or more of communication, broadcast reception, data transfer, and reading from a mounted memory.

  Also, the voice message output device of the present invention is a voice message output device that selects and outputs a voice message corresponding to an event that has occurred, and the user can determine what voice message is currently selected for each event. Means for presenting an event that can be presented to the user or setting a voice message to the user, sound source means for generating and outputting voice data based on the character data or phrase data, and accepting input of characters or phrases by the user Character string generating means for generating a character string, means for determining whether or not a registration operation for the generated character string has been performed by the user, and when the registration operation has been performed, the corresponding event is generated by the user. So that voice output is performed by joining the voice data based on the character string It means for executing the change registration message, characterized by comprising a.

  With the above configuration, for example, the user can create a character string (message sentence) by freely combining characters and phrases, and output a voice output based on the character string as a voice message. That is, the user can listen to a favorite voice message edited by the user.

  In the voice message output apparatus having the above-described configuration, a Japanese syllabary table storing voice data of at least 47 types of syllables is held in advance as the sound source means, or the Japanese syllabary table is used for data communication, data broadcast reception, data It may be configured so that it can be held later by one or more of transfer and memory mounting, and the voice generation based on the character string may be performed based on the voice data of the Japanese syllabary table. With such a configuration, it is possible to listen to a favorite voice message edited by the user himself / herself using the singer / talent 50-sound table of the user's favorite.

  It is good also as carrying out the audio | voice generation | occurrence | production based on the said character string using the said 50 sound table selected by the user by holding at least 2 or more of the said 50 sound tables.

  Further, a numeric keypad in which one or a plurality of characters are assigned to each numeric key may be provided, and the character input may be performed using the numeric keypad.

  Further, these voice message output devices may be configured as a mobile phone, and a telephone ringtone and / or a mail ringtone may be output by voice generation based on the character string.

  According to the present invention, there is an effect that a user can listen to a favorite voice message edited by himself.

  Hereinafter, an embodiment of the present invention will be described with reference to FIGS.

  FIG. 1 is a block diagram of a cellular phone (with digital terrestrial broadcast receiving capability) 2 equipped with a voice message output device of this embodiment.

  As shown in FIG. 1, the cellular phone 2 uses encoded video / audio data obtained by broadcast reception by the terrestrial digital tuner 230 or encoded video / audio data read from the memory card 3 or the like inserted into the slot 201. As a result, the image is displayed on the liquid crystal display panel 202 and the sound is output from the speaker 203. Also, encoded video / audio data obtained by broadcast reception by the terrestrial digital tuner 230 can be recorded in the memory card 3 or the like. In this embodiment, the encoded video / audio data is H.264. Suppose that it is H.264 data.

  H. read from the memory card 3 The H.264 data is supplied to the video decoder 204 via the PCMCIA interface 220 and the system bus 213. The tuner 230 that receives the terrestrial digital broadcast takes out the bit stream data broadcast by the terrestrial digital broadcast and supplies it to the video decoder 204 and the audio decoder (MPEG4-AAC) 206.

  The video decoder 204 obtains quantization coefficients and motion vectors by decoding the bit stream data, and supplies video data obtained by performing inverse DCT transform, motion compensation control based on motion vectors, and the like to the graphics controller 205. The graphics controller 205 performs processing such as color adjustment on video data (for example, R, G, B data). The graphics controller 205 also performs OSD (on-screen display) processing for displaying characters or the like output from the CPU 209 on the liquid crystal display panel 202. By this OSD processing, a menu screen, a character input screen, and the like can be generated. The audio decoder 206 decodes the audio code data in the bit stream to generate audio data. The SDRAM 210 is used in the above processing of the video decoder 204.

  The LCD controller 207 drives the LCD panel 202 based on the video data supplied from the graphics controller 205. The D / A converter 208 receives the audio data output from the audio decoder 206, performs D / A conversion, generates an analog signal, and supplies the analog signal to the speaker 203.

  Operation information for the main body key 214 is given to the CPU 209 via the interface 215 and the system bus 213. The CPU 209 executes necessary processing based on the operation information. The main body key 214 is provided with a TV viewing start key, a recording (recording) start key, a direction key (arrow key), etc. in addition to a numeric keypad.

  A communication block 216 and an interface 217 are provided to enable a short-range wireless network. Further, a flash ROM 218 and an SDRAM 219 are also provided. The battery 221 is a secondary battery, and obtains power from a charger (not shown) and stores it. The mobile phone unit 231 performs processing related to telephone calls, e-mails, and the like. A numeric keypad on the main body key 214 is used for inputting a telephone number and characters in the cellular phone unit 231. On the other hand, when the television viewing mode is selected, the numeric keypad in the main body key 214 functions as a key for inputting a channel number.

  The voice recognition unit 232 receives the voice data output from the audio decoder 206 and performs voice recognition processing. Details will be described later. The voice recognition result is stored in a rewritable nonvolatile memory such as the FlashROM 218, for example. Further, when voice data serving as a voice message is supplied to the D / A converter 208, voice is output via the speaker 203.

  The CPU 209 also performs processing for a wireless network, control of each functional unit based on received data, read / write control of the Flash ROM 218 and the SDRAM 219, and the like. A boot program and application program for operating the CPU 209 are stored in a rewritable nonvolatile memory, and these boot program and application program are expanded on the volatile memory. A voice data file serving as a voice message (voice guidance) is also stored in the rewritable nonvolatile memory and can be rewritten in units of voice data files (message units). In this embodiment, as will be described later, since the current selection guide sentence is presented (displayed), not only the voice data file but also the character string file (text file) of the voice message is non-volatile. Stored in memory. As will be described later, in the change registration of the voice message, the character string file of the changed voice message and the voice data file obtained by the voice reading process of the character string may be registered. It is also possible to omit the registration of the audio data file. That is, if there is a character string file, the voice of the message can be output by performing the voice reading process at the required time using the character string.

  In addition to basic processing such as broadcast reception, the CPU 209 performs voice guidance output processing, editing processing for voice guidance, utterance data generation processing for editing, voice recognition processing for generating utterance data, and the like.

[Voice guidance output processing]
When an event requiring voice guidance (voice message) output occurs, the CPU 209 obtains an identification number associated with the event, and obtains a file indicated by this identification number. The contents of the file indicated by the identification number may be voice data created (created) based on a character string, or may be character string data before being made voice data. The audio data is supplied to the D / A converter 208 and becomes audio. As an example of voice guidance, for example, when a command to record a received program is given by the user or a music video playback command is given by the user, the CPU 209 determines whether or not the memory card 3 is installed in the slot 201. If the memory card 3 is not attached at this time, voice guidance such as “Please set the memory card” is performed.

[Editing process for voice guidance]
For example, an item such as “edit voice guidance” is prepared on the menu screen. FIG. 2 shows an example of the edit screen. This screen displays the event contents and the currently selected voice guidance. In this state, for example, when the CPU 209 detects the operation of the up / down direction key, the CPU 209 scrolls the screen content. As a result, event contents other than the currently displayed contents and the currently selected voice guidance sentence can be displayed on the screen. When the CPU 209 detects that the enter key is pressed, the event content on which the cursor is placed and the currently selected voice guidance are the event content that the user wishes to edit and the currently selected voice guidance. It is judged that. When the determination key is pressed, the screen is switched to the screen shown in FIG.

  In the display screen shown in FIG. 3, the selected event content and the currently selected voice guidance are displayed on the upper half side of the screen, and a message text input area is formed on the lower half side of the screen. In this state, the CPU 209 accepts pressing of the numeric keypad on the main body key 214 as a character input operation. The CPU 209 performs a process of displaying characters input by the user on the screen (liquid crystal display panel 202) in the same manner as when creating (editing) an e-mail message. In the example shown in FIG. 2, the event is “Warning guide when card is not installed”, the currently selected guide sentence is “Please set memory card”, and the guide sentence input by the user is “ Is it a card? " When the CPU 209 detects pressing of the enter key in this state, the process of changing the guidance sentence to “insert card” for the event “Warning guide when no card is inserted”, that is, storing the guide character string and the event Is associated with the guide character string. This association is also a process of changing the content of the file whose identification number is currently being edited ([0001] in the example of the figure) to the guide character string ("Please enter the card"). As described above, since the guide character string can be converted into voice data, the contents of the file of the identification number currently being edited ([0001] in the example in the figure) can be changed to the voice data file. Good. The same applies to the examples shown below. On the other hand, if the guide character string is stored, information about which of the later-described Japanese syllabary tables used in the reading process is also a file of the identification number currently being edited ([0001] in the example in the figure). To keep it.

  The voice output corresponding to the guide character string can be realized by using so-called reading software. The utterance of each word in the guide sentence may be a voice synthesized sound, or may be performed based on, for example, a Japanese syllabary table input to a television talent. The Japanese syllabary table only needs to store audio data of at least 47 types of Japanese syllables, but preferably includes audio data of muddy sounds, semi-voiced sounds, and stuttering sounds. Further, for example, it is desirable to have voice data of “de” in the up tone and “de” in the down tone even for the same “de”. Furthermore, by having long sound data such as “ah”, “i”,..., The “50” table can utter “ka” on the card. For example, if you use the talent's XX Hanako syllabary table, you will hear as if XX Hanako is saying "Please put a card".

  The 50-sound table may be downloaded from a site on the Internet using the communication block 216, may be downloaded from a site on the Internet using a file transfer function in the mobile phone unit 231, or in digital broadcasting Data for download provided by data broadcasting may be downloaded. Of course, the provision of the Japanese syllabary table may be charged. In addition, a maker and a talent may partner to store a table of some talents in a memory or a memory card in advance. In addition, the manufacturer should be able to provide a table of about tens of talents, and store the table of about 50 talents desired by the user in the memory when delivering the product. It may be. Such processing can be performed at the time of data transfer such as an e-mail address at the time of changing the model of the mobile phone. In order to determine whether a certain file is a file in the Japanese syllabary table, a unique identifier indicating that it is a Japanese syllabary table may be provided as an identifier in the file name.

  FIG. 4 shows another example of the edit screen. The screen of FIG. 4 displays, for example, the item “Edit from the Japanese syllabary table” and the item “Edit from the user recording” before the transition to the editing screen shown in FIG. This is displayed when "Edit from user recording" is selected. When “Edit from the Japanese syllabary table” is selected, a. XX Hanako, b. XXX Taro, c. As in the case of Goro, the main display of the Japanese syllabary table may be performed, and after the user is selected by the user, the screen shown in FIG.

  In the editing screen shown in FIG. 4, the event content and the currently selected voice guidance are displayed on the upper half side of the screen. A screen for selecting an audio source is displayed on the lower half side of the screen. In this screen display example, as a voice source, a voice recognition file with the title “XX Goro / User Recording / Year Year / Month / Day / Music Song Battle” and “XX Hanako / User Recording / Year There is a voice recognition file with the title of “Month Day Day / Drama XX Yamakawa”. The CPU 209 performs processing for opening the file on which the cursor is placed when the enter key is pressed.

  In the selected file, as shown in FIG. 5, for example, “Yes”, “No”, “Operation”, “Menu”, “Digital”, “Setting”, “Card”, “Please”, There are phrases (character strings) such as “letter”, “coming”, “ga”, and voice data corresponding to them. On the editing screen, the user sequentially selects words and creates a guidance sentence. For example, if you select the words "I can't do", "Card", and "Please" in order, the guidance sentence "I can't do it" will be completed, and the voice output of "I can't do it" will be output in the voice of XX Goro It will be possible. Since it is a guidance sentence in a device that the user personally uses, if the user can understand it, the guidance sentence is sufficiently established. The word can be selected by operating the direction key for moving the cursor and the operation of the enter key.

  FIG. 6 shows an example of an edit screen in the mobile phone guidance. The selected event content and the currently selected voice guidance are displayed on the upper half side of the screen, and a message input area is formed on the lower half side of the screen. In this state, the CPU 209 accepts pressing of the numeric keypad on the main body key 214 as a character input. The CPU 209 performs a process of displaying characters input by the user on the screen in the same manner as when creating (editing) an e-mail message. In the example shown in FIG. 6, the event is “email ringtone output”, the message text does not exist, and XX is set as the ringing melody. The user inputs, for example, a word such as “I received an e-mail. When the CPU 209 detects that the determination key is pressed in the above state, the CPU 209 performs a process of changing the event to “email ringtone output” and the voice message to “I received a mail now.

  As described above, the utterance of each word in the message sentence may be performed based on the Japanese syllabary table input by the television talent. For example, if you use the entertainer table of the talent XX Hanako, you will hear as if XX Hanako is saying "I've got an email. In addition, as shown in FIG. 4 and FIG. 5, the title of “XX Goro / User Recording / Year Year / Month Day / Music Song Battle” is added as an audio source in the editing in “Guide for Mobile Phone”. A message sentence can be created using the created file. For example, if a message sentence such as “letter has arrived” is created, the voice of “Tegami-gai” will be output with the voice of XX Goro.

  In addition, a mail ring tone for a mail sender registered under a certain condition (group 1 mail ring tone), a mail ring tone for a mail sender registered under other conditions (a group 2 mail ring tone) ), Or different incoming voice messages may be created for each individual sender. In the case of an individual unit, as shown in FIG. 7, for example, it is conceivable to create a guidance sentence such as “E-mail from Jiro”. These can be similarly applied to incoming calls. When such a process is executed, an identification number (for example, Jiro XX) with a branch number is added to the audio data file (for example, [0110]) indicated by the basic (common) identification number. [0110-1]) will be co-existing. When the mail arrives (when an event occurs), the CPU 209 determines who the mail is based on the sender mail address and the mail list, and identifies the voice guidance for the identified partner. It is determined whether or not the number exists (for example, if the mail is from Jiro XX, the above identification number [0110-1] is obtained). If the identification number exists, the voice data file of the identification number is obtained. Will play. In order to facilitate editing in units of individuals, it may be possible to shift to a screen for the above-described mail ringing tone editing process after selecting a mail address (via a mail list screen).

  In the case of a configuration with a digital camera, it is possible to output a voice message such as “I will shoot” when the shutter is pressed. Then, instead of “shooting”, it is also possible for the user to create and edit a message sentence such as “shoot it” and to read it out using the above-mentioned 50-sound table.

[Speech data generation processing for editing]
The process of creating the voice recognition file of “XX Goro / User Recording / Year Year / Month / Day / Music Song Battle” will be described. The mobile phone 2 receives and records a program called “music song battle” by digital broadcast reception based on a user operation. Thereby, a recording file is created. That is, the user records the sound of a favorite talent from the TV program. Of course, an audio file provided on the Internet may be downloaded. Alternatively, the “music song battle” may be received by another broadcast receiving device, recorded on a memory card, and the memory card may be inserted into the slot 201.

  For example, an item such as “Create voice recognition file” is prepared on the menu screen. When the CPU 209 detects that the “voice recognition file creation” item has been selected, it displays a recording file list screen. The CPU 209 opens a recording file selected by the user, generates voice data, applies the voice data to voice recognition processing, and records the voice recognition file obtained thereby on a rewritable nonvolatile memory. That is, the user can record the voice of the favorite talent from the TV program, and can acquire the speech data (character data and voice data) of the favorite talent from the recorded data. Voice recognition is a process of converting voice data into character data, which is also called “transcription”. In addition to such processing, the CPU 209 adds character data obtained by voice recognition and the character data. A process of associating the voice data part corresponding to the data (corresponding to the voice data part causing the voice recognition result and the voice recognition result (character data)) is also performed. However, in the case of so-called speech recognition, identification of the meaning of words (identification of words (units of plural characters)) is basically required, but the identification of the meaning of words is not particularly necessary in the speech recognition of the present application. It is sufficient to associate the voice data portion with the character in units of one character. For example, if the voice recognition result “I will return from now on” is obtained, the voice data part of “ko”, the voice data part of “re”, the voice data part of “ka”, etc. It is only necessary to be able to take correspondence with audio data.

  For example, the character code of “ko” and the voice data of “ko” are stored as a pair of files in the recognition file of “XX Goro / User Recording / Year / Month / Day / Music Song Battle” Each file is assigned an identifier for identifying the data type. The character code of “ko” is used for character presentation (or phrase presentation, see FIG. 5) when the user creates a character string, and the voice data of “ko” is added to the created character string. ”Is used in voice output of voice guidance.

  When the user designates a series part of “from now on”, the CPU 209 combines a plurality of voice data of “ko”, “re”, “ka”, and “ra” as a group, so that “ The phrase “from now on” (character string) and voice data corresponding to them can also be present. Of course, by providing dictionary data, speech recognition can be performed in units of words.

  In the so-called telephone voice recognition system, not all the voice of a telephone is converted into character data, but only a keyword prepared in advance is searched (recognized) in view of the difficulty of voice recognition. . In the present invention, phrases (character strings) such as “Shi”, “I can't”, “Sosa”, “Menyu”, “Dijitaru”, “Seite”, “Kado”, “Please” are used. It may be registered as a keyword, and a phrase (character string) and voice data corresponding to these may be stored based on only the recognition result that matches them.

  Further, by recognizing a break in the audio part (can be determined if the sound is analog broadcast sound), the sound (sentence) can be divided. For example, it is possible to divide and recognize as “From now on” and “Return from home” instead of “Return from now on”.

  Note that the present invention is not limited to applying voice data (uncompressed) obtained by reproducing a recording file to voice recognition processing. It is also possible to apply voice data obtained by broadcast reception to voice recognition processing in real time. In addition, when subtitle sentence data is obtained, the accuracy of voice recognition can be improved by referring to the subtitle sentence data. For example, if the speech recognition result is “Wait at the station today”, the subtitle sentence data output at the time when the speech recognition result was obtained or within a predetermined time range before and after the speech recognition result was obtained. Is present. The CPU 209 executes the character match determination, thereby obtaining a match result in “Wait at the station” and determining that “Today” and “Tokyo” do not match. The CPU 209 corrects the character recognition result on the assumption that the phrase indicated by the caption text data is correct for the mismatched portion.

  In the above example, an example is shown in which a TV talent provides a Japanese syllabary table. However, the present invention is not limited to this, and a TV talent can perform “Yes”, “No”, “Operation”, “Menu”, “ It is also possible to provide audio data of phrases such as “digital”, “setting”, “card”, “please”.

  In the above example, the voice message currently selected for each event is presented to the user. However, presentation of the currently selected voice message may be omitted. That is, each event (event content) for which a voice message can be set may be presented to the user.

  Further, it may be possible to adjust the reading speed when reading a character string to be the voice message.

  In the above example, a mobile phone is illustrated, but the present invention is not limited to this, and it is configured as a vehicle-mounted television receiver, a car navigation system with a communication function, a recorder or a player, and a stationary television receiver. Can do. Moreover, it is good also as an apparatus which performs not only digital broadcast reception but analog broadcast reception. For data registration, analog voice is converted into a digital signal.

  In the above example, characters are input using the numeric keypad. However, the present invention is not limited to this. For example, when a relatively large display is provided, a keyboard is displayed on the screen, and a direction key and a determination key are displayed. The character may be specified and input can be performed by the above operation. Of course, a configuration including a keyboard can also be employed.

  Incidentally, a device that records voice guide data as fixed data on a hard disk or the like is equipped with a file system that manages the voice guide data (fixed data) as a file. The application software accesses the hard disk using an API (Application Programming Interface) provided by the file system. The file system manages files by dividing them into a general partition that allows application software to read and write, and a special partition that can be written only by special operations and that application software cannot normally read. However, even special partitions can be written by changing to the write-permitted mode. Therefore, when voice guide data is written to the special partition, the special partition is changed to the write permission mode, the data is written, and after the data is written, the special partition is again prohibited from being written.

It is the block diagram which showed the structure of the mobile telephone (voice message output device) of embodiment of this invention. It is explanatory drawing which showed the example of the edit screen of embodiment of this invention. It is explanatory drawing which showed the example of the edit screen of embodiment of this invention. It is explanatory drawing which showed the example of the edit screen of embodiment of this invention. It is explanatory drawing which showed the example of the edit screen of embodiment of this invention. It is explanatory drawing which showed the example of the edit screen of embodiment of this invention. It is explanatory drawing which showed the example of the edit screen of embodiment of this invention.

Explanation of symbols

2 Mobile phone 202 LCD panel 205 Graphics controller 207 LCD controller 209 CPU
230 Terrestrial digital tuner 231 Mobile phone unit 232 Voice recognition unit

Claims (7)

In a voice message output device that selects and outputs a voice message corresponding to an event that has occurred, it is possible to present to the user what voice message is currently selected for each event or to set the voice message Means for presenting the event to the user, means for generating character data or phrase data from the voice data by voice recognition, means for associating the generated character data or phrase data with the voice data that caused the generation A means for displaying a character or phrase based on the generated character data or phrase data for user selection, a character string generation means for generating a character string based on the character or phrase selected by the user, and Means for determining whether or not a registration operation has been performed on the character string by the user; Means for registering the change of the voice message so that the voice output is performed by joining the voice data based on the character string created by the user in response to the corresponding event. A voice message output device characterized by the above. 2. The voice message output device according to claim 1, wherein the voice data is acquired by one or more of communication, broadcast reception, data transfer, and reading from a mounted memory. Voice message output device. In a voice message output device that selects and outputs a voice message corresponding to an event that has occurred, it is possible to present to the user what voice message is currently selected for each event or to set the voice message Means for presenting an event to the user, sound source means for generating and outputting voice data based on character data or phrase data, means for generating character data or phrase data from voice data by voice recognition, and generated characters Means for associating the data or phrase data with the voice data that caused the generation thereof, means for displaying a character or phrase based on the generated character data or phrase data for user selection, and selected by the user the characters or phrases for inputting text or phrase by either or user generates a character string A character string generating means for generating a character string by dragging, a means for determining whether or not a registration operation for the generated character string has been performed by the user, and for the corresponding event when the registration operation is performed, the user A voice message output device comprising: means for performing change registration of a voice message so that voice output is performed by joining the voice data based on the character string created by the above. 4. The voice message output device according to claim 3, wherein a syllabary table storing at least 47 types of syllable speech data is stored in advance as the sound source means, or the syllabary table is used for data communication and data broadcasting. It is configured so that it can be held later by one or more of reception, data transfer, and memory mounting, and voice generation based on the character string is performed based on voice data of the Japanese syllabary table Voice message output device. 5. The voice message output device according to claim 4, wherein at least two or more of the Japanese syllabary tables are held, and voice generation based on the character string is performed using a Japanese syllabary table selected by a user. Voice message output device. The voice message output device according to claim 4 or 5, further comprising a numeric keypad in which one or more characters are assigned to each numeric key, and is configured to allow character input using the numeric keypad. A voice message output device. The voice message output device according to any one of claims 1 to 6, wherein the voice message output device is configured as a mobile phone, and outputs a telephone ringtone and / or a mail ringtone by voice generation based on the character string. Voice message output device.

Images