JP2013068665A - Speech recognition device - Google Patents

Abstract

To provide a voice recognition device capable of properly using a voice recognition unit to be used even in the same scenario.
A speech recognition apparatus according to the present invention is a speech recognition apparatus that performs speech recognition by being connected to one or a plurality of connected devices, and is connected to a microphone that captures sound and the one or more connected devices. A connection unit, a storage unit for storing character data for speech recognition obtained from the one or more connected devices, and conversion from the character data for speech recognition in the storage unit to phonemes for speech recognition processing A conversion unit; a speech recognition dictionary that includes phonemes converted by the conversion unit and is used for speech recognition; a speech recognition unit that performs speech recognition using speech captured from the speech recognition dictionary and the microphone; It is determined whether or not the voice recognition target character data is acquired from one or a plurality of connected devices, and whether or not the voice captured from the microphone is recognized by the voice recognition unit based on the determination result And a control unit for controlling.
[Selection] Figure 1

Description

  The present invention relates to a voice recognition device using a voice recognition technology for controlling a device by voice.

  As a method of operating a device, a device that realizes an operation by recognizing a voice spoken by a human and converting it into a command or the like is generally popular. Particularly in an in-vehicle environment, it is effective as a method for operating an in-vehicle terminal while a driver is driving.

  On the other hand, in an embedded software environment such as an in-vehicle terminal, resources such as memory and CPU power tend to be restricted. Therefore, the “command” to be recognized in speech and the recognition target candidates are limited to a specific pattern or set, not an arbitrary character string. By doing so, it is possible to realize resources at low cost by suppressing resources such as memory and CPU power.

  In a speech recognition interactive system, it is also important to present the recognition result early, that is, to speed up the response. In order to improve the response with limited resources, it is very effective to limit the candidates for recognition.

  As a conventional voice recognition device, according to the selector described in the voice dialogue scenario, the “voice recognition unit of the terminal (self)” or the “voice recognition unit of the center (remote device connected via a communication line)” There has been disclosed a voice dialogue system that includes a voice dialogue control unit that selects one of them, and that can properly use the voice recognition unit according to the difficulty level of voice recognition processing (see, for example, Patent Document 1).

JP 2005-37662 A

  However, in the conventional voice dialogue system, for example, in the case of the same scenario “selection of music to be played back”, there are cases where the voice recognition unit to be used cannot be properly used. In the method of describing which voice recognition unit is used for a scenario, since the voice recognition unit is described for each scenario, a plurality of voice recognition units can be flexibly selected for one scenario. Absent. For example, in the scenario where the recognition unit of the own device is used for the song name and the recognition unit of the center is used for the place name, the recognition unit of the own device is always used for the song name.

  An object of the present invention is to provide a voice recognition device that can use different voice recognition units according to the arrangement of candidate character strings for voice recognition even in the same usage (eg, selection of music to be played). is there.

  As one aspect of the present invention, a speech recognition apparatus is a speech recognition apparatus that performs speech recognition by being connected to one or a plurality of connection devices, and is connected to a microphone that captures sound and the one or more connection devices. A connection unit, a storage unit for storing character data for speech recognition obtained from the one or more connected devices, and conversion from the character data for speech recognition in the storage unit to phonemes for speech recognition processing A conversion unit; a speech recognition dictionary that includes phonemes converted by the conversion unit and is used for speech recognition; a speech recognition unit that performs speech recognition using speech captured from the speech recognition dictionary and the microphone; Whether or not the voice recognition target character data is acquired from one or more connected devices, and whether or not the voice recognition unit recognizes the voice captured from the microphone based on the determination result And a control unit for controlling.

  According to the present invention, even in the same usage (for example, selection of music to be played back), it is possible to use a different voice recognition unit to be used according to the arrangement of candidate character strings to be recognized.

Block diagram of voice recognition device 10 and connected device 20 in an embodiment of the present invention The figure which shows an example of the data accumulate | stored in the memory | storage part 204 of the connection apparatus 20. An example of data accumulated in the storage unit 104 of the speech recognition apparatus 10 (1) An example of the speech recognition dictionary 106 of the speech recognition apparatus 10 (1) Example of data accumulated in storage unit 104 of speech recognition apparatus 10 (2) An example (2) of the speech recognition dictionary 106 of the speech recognition apparatus 10 An example (3) of the speech recognition dictionary 106 of the speech recognition apparatus 10 Block diagram showing a modification of the present embodiment The figure for demonstrating the structure of the memory | storage part of each connection apparatus 20A, 20B, 20C.

  The speech recognition apparatus according to the embodiment of the present invention determines to which apparatus a character string of a candidate for speech recognition, that is, “database (music name, person name, etc.)” is arranged, and according to the determination result. , It is possible to select which device voice recognition means is used.

  Hereinafter, a speech recognition apparatus according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a speech recognition apparatus 10 according to an embodiment of the present invention. The speech recognition apparatus 10 illustrated in FIG. 1 includes a speech recognition unit 101, a connection unit 102, a control unit 103, a storage unit 104, a G2P conversion unit 105, a speech recognition dictionary 106, and a microphone 107. The connected device 20 illustrated in FIG. 1 includes a voice recognition unit 201, a connection unit 202, and a storage unit 204.

  Specific examples of the speech recognition device 10 include an in-vehicle terminal, a navigation system, and an in-vehicle audio. Specific examples of the connection device 20 include a smartphone, a portable audio device, a telephone, etc. brought into the vehicle. The voice recognition device 10 is connected to the connection device 20 through a wired connection or a wireless communication interface. An example of a wired connection interface is USB (Universal Serial Bus). Wireless communication interfaces include, for example, WiFi (Wireless Fidelity) (registered trademark) and Bluetooth (registered trademark).

  Hereinafter, in the present embodiment, a case where the voice recognition device 10 is an in-vehicle audio and the connection device 20 is a portable audio device will be described as a specific example.

  The connected device 20 includes a voice recognition unit 201, a connection unit 202, and a storage unit 204. As shown in FIG. 2, the storage unit 204 stores “music data (music file)” and “music information data” attached to the music data.

  The music information data shown in FIG. 2 is metadata accompanying the music data. The music information data may include, for example, the name of the album in which the music data is included, the name of the artist who owns the music data, and the genre name of the music data in addition to the music data.

  The connection unit 202 is connected to the connection unit 102 of the voice recognition device 10, transmits data held in the storage unit 204 to the voice recognition device 10, and receives an instruction from the control unit 103 of the voice recognition device 10. The connection unit 202 outputs an instruction from the control unit 103 of the voice recognition device 10 to the voice recognition unit 201. The speech recognition unit 201 can perform speech recognition processing based on an instruction from the control unit 103 of the speech recognition apparatus 10 using a built-in speech recognition dictionary.

  There are devices that can provide “music information data” to the speech recognition apparatus 10 and devices that cannot be provided via the connection unit 202. The reasons include (1) restrictions due to the specifications of the connection device 20 and (2) restrictions due to the interface specifications of the connection between the speech recognition apparatus 10 and the connection device 20.

  Hereinafter, in the present embodiment, when the voice recognition device 10 acquires “music information data” of the connected device 20 <case 1>, the voice recognition device 10 acquires “music information data” of the connected device 20. If not, the operation of the speech recognition apparatus 10 will be described for each case of <Case 2>.

<Case 1>
As Case 1, the operation of each unit of the speech recognition device 10 when the speech recognition device 10 acquires the “music information data” of the connected device 20 will be described.

  The connection unit 102 is connected to the connection device 20 and transmits / receives information (such as “music information data”).

  The storage unit 104 stores “character data for speech recognition” acquired from the connected device 20. Here, as a specific example of “character data subject to speech recognition”, “music information data” stored in the connected device 20 can be cited. The “music information data” includes a music name, an album name, an artist name, a genre name, and the like. When the speech recognition apparatus 10 acquires “music information data” from the connection device 20, the control unit 103 associates “music information data” with the connection device 20 from which the “music information data” is acquired. And stored in the storage unit 104.

  FIG. 3 shows an example of “music information data” stored in the storage unit 104 of the speech recognition apparatus 10. As shown in FIG. 3, the storage unit 104 includes a song name list (“song name 1”, “song name 2”,... “Song name N”) as song information data in addition to “music data (music file)”. As shown in FIG. 3, in the present embodiment, “music information data” stored in the storage unit 104 of the speech recognition apparatus 10 is associated with the connected device 20 from which the “music information data” is acquired. Yes.

  The G2P conversion unit 105 performs conversion from “character data to be recognized by speech recognition” stored in the storage unit 104 to phonemes for speech recognition processing. This is a process of converting from “Grapheme: character or grapheme” to “Phonee: phoneme”, and is called G2P (Grapheme To Phoneme) conversion.

  The speech recognition dictionary 106 describes a dictionary and grammar in speech recognition processing, and basically determines a vocabulary that can recognize the contents described in this dictionary. The voice recognition dictionary 106 can also be changed dynamically while the voice recognition device 10 is operating. If a character string group that is a source of G2P conversion is obtained, this can be realized by G2P converting the character string group and dynamically changing the speech recognition dictionary 106 to change the vocabulary to be recognized.

  For example, when “music information data” is acquired and converted to G2P, the speech recognition dictionary 106 uses the vocabulary “Play Music”, “Stop”, and “Pause” as examples of fixed commands as shown in FIG. Hold. In addition to these fixed commands, the speech recognition dictionary 106 also includes variable commands, that is, “Song Title 1”, “Song Title 2”,... “Song Title N” in the title list which is character data based on information acquired from the connected device 20. May be included. Here, N is a number, which may be thousands to tens of thousands in the case of recent portable audio devices.

  The speech recognition unit 101 performs speech recognition processing based on the dictionary and grammar described in the speech recognition dictionary 106. That is, when the speech recognition unit 101 determines that a vocabulary that matches well with the speech captured from the microphone 107 among the vocabulary described in the speech recognition dictionary 106 is speech uttered by the user, the speech recognition unit 101 considers that this vocabulary has been uttered. Then, voice recognition processing is performed. For example, when the content corresponding to the words “Play Music”, “Stop”, and “Pause” is registered in the speech recognition dictionary 106, the speech recognition unit 101 and the registered content of the speech uttered by the user. Vocabulary that matches well is selected by matching process. Then, the speech recognition unit 101 performs speech recognition processing assuming that the selected vocabulary is spoken.

  The control unit 103 determines whether or not “music information data” stored in the connected device 20 is transferred to and stored in the storage unit 104 via the connection unit 102. When it is determined that the voice recognition device 10 has acquired “music information data” from the connected device 20, the control unit 103 routes or transfers the voice captured from the microphone 107 to the voice recognition unit 101. The control unit 103 instructs the voice recognition unit 101 to perform voice recognition processing based on the transferred or routed voice. When the speech recognition apparatus 10 acquires “music information data” from the connection device 20, the control unit 103 associates “music information data” with the connection device 20 from which the “music information data” is acquired. And stored in the storage unit 104.

  In Case 1, instead of determining whether or not the “music information data” stored in the connected device 20 is transferred to and stored in the storage unit 104, the control unit 103 stores the connected device 20 in the voice recognition dictionary 106. It may be determined whether or not a vocabulary equivalent to “music information data” is added.

<Case 2>
As Case 2, the operation of each part of the speech recognition device 10 when the speech recognition device 10 has not acquired the “music information data” of the connected device 20 will be described. In this case, it is assumed that the speech recognition apparatus 10 cannot acquire the music information data “from the connection device 20 although the connection unit 102 is connected to the connection device 20. That is, the connection device 20 is connected via the connection unit 202. Therefore, it is assumed that the “music information data” cannot be provided to the voice recognition device 10.

  FIG. 5 is a conceptual diagram of the storage unit 104 in Case 2. As described above, the speech recognition apparatus 10 cannot acquire the music information data ”from the connected device 20. For this reason, as shown in FIG. 5, the storage unit 104 has no music title list.

  The G2P conversion unit 105 performs conversion from “character data to be recognized by speech recognition” stored in the storage unit 104 to phonemes for speech recognition processing. In other words, the G2P conversion unit 105 converts the character data stored in the storage unit 104, but the original character data ("music information data") is not stored in the storage unit 104 (empty set), and thus is generated as a result. There is no phoneme (empty set).

  As a result, as shown in FIG. 6, for example, the speech recognition dictionary 106 holds the vocabulary “Play Music”, “Stop”, and “Pause” as an example of a fixed command. As an example, a vocabulary (for example, a song name list) based on “music information data” is not held.

  The control unit 103 determines whether or not “music information data” stored in the connected device 20 is transferred to and stored in the storage unit 104 via the connection unit 102. When the voice recognition device 10 determines that the “music information data” has not been acquired from the connected device 20, the control unit 103 uses the voice recognition unit 101 of its own device to pick up the voice captured from the microphone 107. Rather, it is routed or transferred to the voice recognition unit 201 of the connected device 20, and the voice recognition unit 201 of the connected device 20 is controlled to perform voice recognition processing.

  Here, the operation of the voice recognition unit 201 of the connected device 20 will be described. The voice recognition unit 201 performs voice recognition processing on the voice transferred or routed from the microphone 107 of the voice recognition device 10 based on the control of the control unit 103. That is, when the speech recognition unit 201 determines that a vocabulary that matches well with the speech captured from the microphone 107 of the speech recognition device 10 among the vocabulary described in the built-in speech recognition dictionary is speech uttered by the user, The voice recognition process is performed assuming that is spoken. For example, when the content corresponding to the vocabulary “Play Music”, “Stop”, and “Pause” is registered in the built-in speech recognition dictionary, the speech recognition unit 201 registers the speech uttered by the user and these Select matching vocabulary by matching content. Then, the speech recognition unit 201 performs speech recognition processing assuming that the selected vocabulary is spoken.

  In case 2 as well, instead of determining whether or not the “music information data” stored in the connected device 20 is transferred to and stored in the storage unit 104, the control unit 103 connects the connected device to the voice recognition dictionary 106. It can be determined by whether or not a vocabulary equivalent to 20 “music information data” has been added.

  As described above, according to the speech recognition apparatus 10 according to the present embodiment, it is determined whether or not character data is acquired from the connected device 20, and the speech captured from the microphone 107 is determined based on the determination result. A control unit 103 that controls whether the recognition unit 101 recognizes or not is provided. With this configuration, the speech recognition apparatus 10 according to the present exemplary embodiment recognizes speech recognition to be used according to the arrangement of candidate character strings to be recognized even if they have the same usage (for example, selection of music to be played). You can use different parts. Therefore, the speech recognition device 10 according to the present embodiment can select which device the speech recognition unit is to use.

  In the present embodiment, “song name” is described as a specific example of “music information data” as an example of “text data to be recognized by speech”, but the present invention is not limited to the name of a song. For example, an album name, an artist name, a genre name, etc. may be used.

  In the present embodiment, “music information data” is taken as an example of “character data for speech recognition”, but it is obvious that the present invention is not limited to character data music information data. For example, when “character data for voice recognition” is set to “phonebook data”, when the connected device 20 is a telephone or a smartphone, voice recognition for “person name” or “phone number” is performed. Easy to apply. Furthermore, for example, by setting “character data for speech recognition” to “place name data”, the present invention can be applied even when the connected device 20 is a device including place name data.

  The phone book data may include “name (name)” data and at least one or more telephone number information data attached to the “name (name) data”. The telephone number information data may include information such as attribute information (mobile / work / home / other) of at least one or more telephone numbers associated with this name (name). The “name” data included in the phone book data is character string data such as a person name, a company name, and a nickname. In addition, the “name” data is usually used as the “character data for speech recognition”.

  In the present embodiment, the voice recognition dictionary 106 has been described as a part that is based on a fixed command and a part that is based on information obtained from a connected device 20 that is a variable command. However, the present invention is not limited to this. FIG. 7 shows another example of the configuration of the voice recognition dictionary 106. As shown in FIG. 7, for example, when the speech recognition apparatus 10 itself has music data, the speech recognition dictionary 106 uses a fixed command other than the vocabulary “Play Music”, “Stop”, and “Pause”. As a variable command, a song name list (song name 1, song name 2,..., Song name N) based on song information data of the connected device 20 and a song name list (song name A, song name) based on song information data held by the speech recognition apparatus 10 itself. B,..., Song title Z) can also be used.

  In the present embodiment, the interface for connection between the speech recognition apparatus 10 and the connection device 20 is not particularly limited, and a plurality of types of interfaces can be mixed. Furthermore, it is possible to mix information types. Hereinafter, a case where a plurality of connected devices 20 and the speech recognition apparatus 10 according to the present embodiment are connected will be described with reference to FIG. 8 as an example.

  In the present embodiment, the speech recognition apparatus 10 includes the microphone 107 as a means for capturing sound from outside, but this is not a limitation. The voice recognition device 10 can be used in place of the microphone 107 as long as it is a device that acquires voice from the outside.

(Modification)
FIG. 8 is a block diagram showing a modification of the present embodiment. Hereinafter, the plurality of connection devices 20 illustrated in FIG. 8 are referred to as connection device 20A, connection device 20B, and connection device 20C, respectively, and are distinguished from each other. The configurations of these connection devices 20A, 20B, and 20C are storage units 204A and 204B. , 204C other than the connection device 20 shown in FIG.

  As shown in FIG. 8, the connected device 20A includes a voice recognition unit 201, a connection unit 202, and a storage unit 204A. Similarly, the connected device 20B includes a voice recognition unit 201, a connection unit 202, and a storage unit 204B, and the connected device 20C includes a voice recognition unit 201, a connection unit 202, and a storage unit 204C. Since the operations of the voice recognition unit 201 and the connection unit 202 of the connection devices 20A, 20B, and 20C are the same as those of the voice recognition unit 201 and the connection unit 202 of the connection device 20 described above, detailed description thereof is omitted.

  With reference to FIG. 9, the structure of the memory | storage part of each connection apparatus 20A, 20B, 20C is demonstrated. FIG. 9 is a diagram for explaining the configuration of the storage unit of each of the connected devices 20A, 20B, and 20C. In FIG. 9, for the sake of explanation, some of the configurations of the connected devices 20A, 20B, and 20C are omitted, and only the configurations of the storage units 204A, 204B, and 204C are shown.

  The storage unit 204A of the connected device 20A holds “music information data A” which is information that becomes “character data to be recognized by voice” in addition to the music data. The “music information data A” includes a music title list (music title 1, music title 2,..., Music title N) based on the music information data A.

  In addition to the music data, the storage unit 204B of the connected device 20B holds “music information data B” that is information that becomes “character data to be recognized by voice”. The “music information data B” includes a music name list (song name A, music name B,..., Music name Z) based on the music information data B.

  The storage unit 204C of the connected device 20C holds “phone book data C”, which is information that becomes “voice recognition target character data”, in addition to the telephone number data. The “phone book data C” includes a list of person names (person name α, person name β,..., Person name ω) based on the phone book data C.

  The connection unit 102 receives information (“music information data A” and “phone book data C”) that is “character data for speech recognition” from each of the connected devices 20A, 20B, and 20C, and receives an instruction from the control unit 103. Is transmitted to each connected device 20A, 20B, 20C.

  Hereinafter, it is assumed that the speech recognition apparatus 10 has already acquired “music information data A” from the connected device 20A and has already acquired “phone book data C” from the connected device 20C. Furthermore, it is assumed that the speech recognition apparatus 10 has not acquired “music information data B” from the connected device 20B. Therefore, the storage unit 104 of the speech recognition apparatus 10 holds “music information data A” and “phone book data C” (see FIG. 9). In other words, the connection devices 20A and 20C correspond to the connection device 20 of <Case 1> described above, and the connection device 20B corresponds to the connection device 20 of <Case 2> described above. The storage unit 104 holds “music information data A” and the connected device 20A from which the “music information data A” is acquired in association with each other. Similarly, the storage unit 104 holds “phone book data C” and the connected device 20C from which the “phone book data C” is acquired in association with each other.

  The G2P conversion unit 105 performs conversion from “character data to be recognized by speech recognition” stored in the storage unit 104 to phonemes for speech recognition processing.

  The speech recognition dictionary 106 holds a vocabulary “call” related to the telephone function in addition to the vocabulary “Play Music”, “Stop”, and “Pause” as fixed commands. Further, the speech recognition dictionary 106 of the speech recognition apparatus 10 uses, as variable commands, a song name list (song name 1, song name 2,..., Song name N) based on the song information data A of the connected device 20A and the phone book data C of the connected device 20C. A personal name list (person name α, person name β,..., Person name ω) based on is stored.

  When the speech recognition unit 101 determines, based on an instruction from the control unit 103 described later, a vocabulary that matches well with the speech captured from the microphone 107 among the vocabulary described in the speech recognition dictionary 106, is a speech uttered by the user. Speech recognition processing is performed assuming that this vocabulary is spoken.

The control unit 103 determines whether “music information data” or “phone book data” held in each of the connected devices 20A, 20B, and 20C has been transferred to the storage unit 104 via the connection unit 102. When the control unit 103 determines that the voice recognition device 10 has acquired “music information data” or “phone book data” from any one of the connected devices 20A, 20B, and 20C, the microphone 107 The captured voice is routed or transferred to the voice recognition unit 101. The control unit 103 instructs the voice recognition unit 101 to perform voice recognition processing based on the transferred or routed voice.
When the voice recognition apparatus 10 acquires “music information data A” from the connected device 20A, the control unit 103 acquires “music information data A” and the connected device 20A from which the “music information data A” is acquired. Are stored in the storage unit 104. Similarly, when the speech recognition apparatus 10 acquires “phone book data C” from the connected device 20C, the control unit 103 connects “phone book data C” and the connected device from which the “phone book data C” is acquired. And stored in the storage unit 104.

  In this modification, the storage unit 104 holds “text data for speech recognition” relating to “music information data A”, that is, a song title list (song title 1, song title 2,..., Song title N) (see FIG. 9). ). Therefore, the control unit 103 determines that the voice recognition device 10 has acquired “music information data” from the connected device 20 </ b> A, and routes or transfers the voice captured from the microphone 107 to the voice recognition unit 101. Then, the control unit 103 instructs the voice recognition unit 101 to perform a voice recognition process based on the transferred or routed voice.

  Further, in this modification, the storage unit 104 holds “character data for speech recognition” relating to “phonebook data C”, that is, a personal name list (person name α, person name β,..., Person name ω) (FIG. 9). Therefore, the control unit 103 determines that the voice recognition apparatus 10 has acquired “phone book data C” from the connected device 20 </ b> C, and routes or transfers the voice captured from the microphone 107 to the voice recognition unit 101. Then, the control unit 103 instructs the voice recognition unit 101 to perform a voice recognition process based on the transferred or routed voice.

  In this modification, the storage unit 104 does not hold “character data for speech recognition” regarding “music information data B”, that is, a music title list (music title A, music title B,..., Music title Z) (see FIG. 9). ). Therefore, the control unit 103 determines whether or not the “music information data B” stored in the connected device 20B is transferred to and stored in the storage unit 104 via the connection unit 102. The control unit 103 determines that the voice recognition device 10 has not acquired the “music information data B” from the connected device 20B, and the control unit 103 transfers the voice captured from the microphone 107 to the voice recognition unit 201 of the connected device 20B. Routing or transferring is performed, and control is performed so that the voice recognition unit 201 of the connected device 20B performs voice recognition processing.

  In the above-described embodiment and its modification, when the voice captured from the microphone 107 is routed or transferred to the connection device 20, the voice recognition is actually performed by the remote device from the connection device 20 to a server or the like. Although the form to implement is also considered, it cannot be overemphasized that all are in the range of the fundamental idea of this invention.

  The speech recognition apparatus of the present invention has an effect that even if it is used for the same purpose (for example, selection of music to be played back), it is possible to properly use a speech recognition unit to be used according to the arrangement of candidate character strings for speech recognition. It is useful as an in-vehicle terminal such as a navigation device or an in-vehicle audio device.

DESCRIPTION OF SYMBOLS 10 Voice recognition apparatus 101 Voice recognition part 102 Connection part 103 Control part 104 Storage part 105 G2P conversion part 106 Voice recognition dictionary 107 Microphone 20 Connection apparatus 201 Voice recognition part 202 Connection part 204 Storage part 20A, 20B, 20C Connection apparatus

Claims (4)

A speech recognition device that performs speech recognition connected to one or more connected devices,
A microphone that captures audio,
A connection part for connecting to the one or more connection devices;
A storage unit for storing voice recognition target character data acquired from the one or more connected devices;
A conversion unit that performs conversion from character data to be recognized in the storage unit into phonemes for speech recognition processing;
A speech recognition dictionary used for speech recognition including the phonemes converted by the conversion unit;
A speech recognition unit that performs speech recognition using speech captured from the speech recognition dictionary and the microphone;
It is determined whether or not the character data subject to speech recognition is acquired from the one or more connected devices, and whether or not the speech recognition unit recognizes the speech captured from the microphone based on the determination result. A control unit for controlling,
Voice recognition device. The controller is
When the character data for speech recognition is acquired from the one or more connected devices, the speech recognition unit of its own device recognizes the speech captured from the microphone,
When the character data for speech recognition is not acquired from the one or more connected devices, the voice captured from the microphone is transmitted to the one or more connected devices via the connection unit, Make recognition
The speech recognition apparatus according to claim 1. The character data for speech recognition is music information data attached to music data and including at least the music title of the music,
The speech recognition apparatus according to claim 1 or 2. The speech recognition target character data is telephone number information data attached to name data and including attribute information of at least one telephone number associated with the name data.
The speech recognition apparatus according to claim 1 or 2.

Images