JP2020039048A - Voice collecting device and voice collecting method - Google Patents

Abstract

To provide a voice collecting device and a voice collecting method that can select a microphone appropriately even in an environment that is easily affected by a reflected sound.SOLUTION: A voice collecting device includes a voice acquisition unit 241 that acquires voice signals from a plurality of microphones, a first microphone identification unit 242 that identifies the microphone on the basis of the strength and/or arrival time of the voice signal obtained from the microphone, a user identification unit 243 that identifies a user corresponding to the voice signal obtained from the microphone, a second microphone identification unit 244 that identifies a microphone associated with the identified user in a user table that associates one microphone among a plurality of microphones with each user, and a microphone selection unit 245 that selects the microphone identified by the second microphone identification unit as a microphone used for voice recognition processing when the microphone is identified by the second microphone identification unit and the microphone identified by the first microphone identification unit does not match the microphone identified by the second microphone identification unit.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a voice collecting device and a voice collecting method.

  2. Description of the Related Art In recent years, an information processing apparatus that collects voice spoken by a user, specifies a command based on the collected voice, and executes a process corresponding to the specified command has been used. According to such an information processing apparatus capable of voice input, a user can execute a desired process without operating an input device such as a touch panel, a keyboard, and a mouse. For example, a user driving a vehicle performs a desired process such as setting a destination without interrupting the driving operation by speaking a voice to a car navigation system capable of voice input mounted on the vehicle. Can be done.

  The voice spoken by the user is acquired by a microphone, converted into a voice signal, and input to a voice recognition device. In order for the voice recognition device to appropriately execute voice recognition processing (processing for specifying a command based on an input voice signal), it is desirable to obtain a voice signal that faithfully corresponds to the voice uttered by the user.

  In order to obtain an audio signal faithfully corresponding to the voice uttered by the user, the voice uttered by the user may be obtained by a plurality of microphones having different directional characteristics. In this case, the audio signals obtained by the plurality of microphones are selected and used for voice recognition processing.

  Patent Document 1 discloses an in-vehicle voice recognition that detects the direction of a speaker, detects the voice of the speaker, adjusts the directional characteristics to increase in the direction of the speaker, and recognizes the adjusted voice of the speaker. An apparatus is described. Such an in-vehicle voice recognition device can reduce noise other than the voice of the speaker and improve the voice recognition rate.

JP-A-11-219193

  In the speech recognition device described in Patent Literature 1, the gain of each microphone is adjusted based on the direction of the speaker calculated based on the signal levels and the delay times at the outputs of the plurality of microphones.

  Here, the microphone may acquire a sound in which a sound (direct sound) directly reaching the microphone and a sound (reflected sound) reaching the microphone by being reflected by a peripheral object such as a wall surface are mixed. In particular, in a narrow closed situation such as a vehicle cabin, the vehicle is easily affected by reflected sound. Since the reflected sound is delayed more than the direct sound, the mixed sound may hinder the voice recognition processing by the direct sound.

  The sound obtained by the microphone selected based on the signal level and the delay time may be affected by the reflected sound, and in such a case, the selected microphone is not always the best microphone for the voice recognition processing. .

  Whether or not the microphone is easily affected by the reflected sound tends to be determined by the characteristics of the voice of the uttering user. For example, an utterance by a loud user is easily affected by the reflected sound.

  Therefore, an object of the present invention is to provide a sound collection device and a sound collection method that can appropriately select a microphone even in a situation easily affected by a reflected sound.

  A sound collection device according to the present invention includes a sound obtaining unit that obtains sound signals from a plurality of microphones, a first microphone specifying unit that specifies a microphone based on the strength and / or arrival time of the sound signals obtained from the microphones, A user identification unit that identifies a user corresponding to an audio signal acquired from a microphone, and a second microphone that identifies a microphone associated with the identified user in a user table that associates one of a plurality of microphones for each user If the microphone is specified by the specifying unit and the second microphone specifying unit, and the microphone specified by the first microphone specifying unit does not match the microphone specified by the second microphone specifying unit, the second microphone specifying unit A microphone selection unit that selects the specified microphone as a microphone used for the voice recognition processing.

  Further, in the sound collection device according to the present invention, the microphone is arranged so as to be able to acquire a sound signal corresponding to the sound in the vehicle cabin of the vehicle, and the user table is configured to output guidance voice to the vehicle room in addition to the user. One microphone among the plurality of microphones is associated with each of the level, the noise level in the vehicle compartment, the speed of the vehicle, and / or the closed state of the vehicle compartment, and the second microphone specifying unit is specified in the user table. It is preferable to specify the user who has been connected and the output level of the guidance voice to the passenger compartment, the noise level in the passenger compartment, the speed of the vehicle, and / or the microphone associated with the closed state of the passenger compartment.

  Also, in the voice collection device according to the present invention, the user identification unit may include, in a terminal table for associating a user with each information terminal wirelessly connectable to the vehicle, a user associated with the identifier of the information terminal wirelessly connected to the vehicle. It is preferable to specify the user corresponding to the signal.

  In addition, the audio collection device according to the present invention may further include a buffer control unit that temporarily stores the obtained audio signal for each of the plurality of microphones and outputs the audio signal obtained and temporarily stored by the selected microphone. Is preferred.

  In addition, the voice collection device according to the present invention receives a result of voice recognition of a voice signal acquired by the selected microphone, and associates the specified user with the microphone in the user table based on the result. It is preferable to further include a learning unit for changing.

  A voice collecting method according to the present invention includes: a voice obtaining step of obtaining voice signals from a plurality of microphones; a first microphone specifying step of specifying a microphone based on the strength and / or arrival time of the voice signals obtained from the microphones; A user identification step of identifying a user corresponding to an audio signal acquired from the microphone, and a second microphone identifying a microphone associated with the identified user in a user table associating one of a plurality of microphones for each user If the microphone is specified in the specifying step and the second microphone specifying step, and the microphone specified in the first microphone specifying step does not match the microphone specified in the second microphone specifying step, the microphone is specified in the second microphone specifying step. Select the specified microphone as the microphone to be used for speech recognition processing. Characterized in that it comprises a microphone selection step of, a.

  ADVANTAGE OF THE INVENTION According to the audio | voice collection apparatus of this invention, a microphone can be appropriately selected according to the user who speaks even in the situation which is easily influenced by a reflected sound.

  ADVANTAGE OF THE INVENTION According to the audio | voice collection method of this invention, a microphone can be appropriately selected according to the user who speaks even in the situation easily affected by the reflected sound.

It is a figure explaining the operation | movement outline of a voice collection apparatus. It is a hardware schematic diagram of the vehicle provided with the voice collection device. (A) is an example of a terminal table, and (b) is an example of a user table. It is a schematic diagram which shows the schematic structure of a voice collection device. It is a processing flowchart of a voice collection device.

  Hereinafter, the voice collecting apparatus and the voice collecting method will be described in detail with reference to the drawings. However, it should be understood that the invention is not limited to the drawings or the embodiments described below.

  The sound collection device of the present invention specifies a microphone based on the strength and / or arrival time of a sound signal acquired from the microphone. On the other hand, as described above, whether or not a microphone is susceptible to reflection tends to be determined by characteristics of a user's voice. Therefore, the voice collecting apparatus stores a microphone in association with a microphone, reflecting the susceptibility of the user to the reflection. Then, the sound collection device specifies the user corresponding to the sound signal acquired from the microphone, and specifies the microphone associated with the specified user. The voice collection device may be configured to identify the microphone associated with the user, and if the microphone identified based on the sound intensity and / or the arrival time does not match the microphone associated with the user, Select the microphone as the microphone to be used for the voice recognition processing. Accordingly, the voice collection device can appropriately select a microphone according to the user who speaks even in a situation where the sound is easily affected by the reflected sound.

  If the user who made the utterance cannot be identified, or if the user who made the utterance is a new user and no microphone is stored in association with the user, the voice collection device will use the voice intensity and / or arrival time. Select the specified microphone based on the. Also, even in the case of a new user, the learning unit described later reflects the susceptibility of the user to the influence of the reflection, and the microphone is newly associated and stored. The microphone which was made can be specified.

  FIG. 1 is a diagram illustrating an outline of the operation of the voice collection device.

  The vehicle 1 includes a voice collection device 2, microphones 3a to 3c, and a storage device 5. The user X gets on the vehicle 1 while carrying the terminal P and speaks. The voice collection device 2 acquires the utterance of the user X from the microphones 3a to 3c.

  The graph G1 is a graph in which the vertical axis represents the input level of the audio signal acquired from the microphone 3a, and the horizontal axis represents time. Similarly, the graphs G2 and G3 show the input levels of the audio signals acquired from the microphones 3b and 3c.

  In the graph G1, a series G11 represents an input level of a direct sound acquired from the microphone 3a, and a series G12 represents an input level of a reflected sound acquired from the microphone 3a. In the graph G1, for the sake of simplicity, the direct sound and the reflected sound are represented in different series, but in practice, an audio signal is obtained from the microphone 3a in a state where these sounds are combined.

Direct sound of the voice uttered the user arrives at the microphone 3a at time t _0. Input level of the direct sound of the user's uttered voice is L _1. Reflected sound of the sound uttered the user arrives at the microphone 3a at time t _3. Input level of the reflected sound of the user's uttered voice is L _5.

Direct sound acquired from the microphone 3b represented by series G21 in the graph G2 reaches the microphone 3b at time t _1, the input level is L _2. Further, the reflected sound acquired from the microphone 3b represented by series G22 in the graph G2 reaches the microphone 3b at time t _4, the input level is L _6.

Direct sound acquired from the microphone 3c represented by series G31 in graph G3 reaches the microphone 3c to time t _2, the the input level is L _3. Further, the reflected sound acquired from the microphone 3c represented by series G32 in graph G3 reaches the microphone 3c at time t _5, the input level is L _4.

  Thus, the sound collection device 2 specifies the microphone based on the strength and / or the arrival time of the sound signal acquired from each microphone. In the example of FIG. 1, the voice collection device 2 specifies the microphone 3a with the highest voice signal intensity and the earliest voice signal arrival based on the obtained voice signal strength and arrival time.

  Further, the sound collection device 2 may specify the microphone based on the strength or arrival time of the obtained sound signal. For example, the sound collection device 2 may specify the microphone having the largest sound signal strength, or may specify the microphone to which the sound signal reaches earliest.

  On the other hand, the sound collection device 2 specifies the user corresponding to the sound signal acquired from the microphone. Specifically, the voice collection device 2 first detects the identifier of the terminal P carried by the user.

  Then, the voice collection device 2 refers to the terminal table 51 stored in the storage device 5 and specifies the user X who is the user corresponding to the terminal P.

  Further, the voice collection device 2 refers to the user table 52 stored in the storage device 5 and specifies the microphone 3b that is the microphone associated with the user X.

  Next, the voice collection device 2 determines whether the microphone specified based on the strength and / or the arrival time of the voice signal is the same as the microphone associated with the user. In the example of FIG. 1, the microphone specified based on the strength and / or arrival time of the audio signal is the microphone 3a, and the microphone corresponding to the specified user X is the microphone 3b. Are determined not to match.

  The voice collection device 2 that determines that the microphones do not match selects the microphone corresponding to the specified user as the microphone used for the voice recognition processing. In the example of FIG. 1, the voice collection device 2 determines that the microphones do not match, and selects the microphone 3b, which is the microphone corresponding to the specified user, as the microphone used for the voice recognition processing.

  By selecting a microphone in this way, the voice collecting apparatus can appropriately select a microphone to be used for the voice recognition process, depending on the user who speaks, even in a situation that is easily affected by the reflected sound.

  FIG. 2 is a schematic hardware diagram of a vehicle including the voice collecting device.

  The vehicle 1 includes a voice collection device 2, a microphone 3, a wireless interface 4, a storage device 5, a voice recognition device 6, and a car navigation system 7.

  The voice collecting device 2 is an information processing device that obtains a voice signal from a plurality of microphones 3 and outputs a voice signal of one selected microphone to the voice recognition device 6. The detailed configuration of the voice collection device 2 will be described later.

  The microphone 3 is an input device that outputs a sound signal that is an electric signal corresponding to the sound. The microphone 3 is connected to the sound collection device 2 and provides a sound signal to the sound collection device 2. The microphone 3 includes a plurality of microphones, and each of the plurality of microphones included in the microphone 3 is attached to each location in the passenger compartment of the vehicle.

  For example, the microphone 3a is mounted near the center of the upper end of the windshield, the microphone 3b is mounted above the right B pillar, and the microphone 3c is mounted above the left B pillar.

  The wireless interface 4 is an interface used for performing wireless communication with an external device. The wireless interface 4 is a device that provides a Bluetooth (registered trademark) interface, and is connected to the voice collection device 2 to enable communication with a Bluetooth-compatible external device such as a smartphone or a tablet computer. The wireless interface 4 may be a wireless local area network such as IEEE802.11ac, a short-range wireless communication standard such as ISO / IEC 18092, or the like.

  The storage device 5 is a device that stores data. The storage device 5 is an SSD (Solid State Drive) using a semiconductor element memory as a storage element, and is connected to the audio collection device 2 to transmit and receive data. The storage device 5 may be a hard disk drive.

  The storage device 5 stores a terminal table 51 and a user table 52. The details of the terminal table 51 and the user table 52 will be described later.

  The voice recognition device 6 is a device that converts a voice signal output from the voice collection device 2 into a command to a peripheral device. The speech recognition device 6 is configured by a processor executing a speech recognition program developed in a main memory. The speech recognition device 6 is connected to the speech collection device 2 and receives a speech signal from the speech collection device 2.

  The car navigation system 7 is a device that provides route guidance to a destination based on the current position of the vehicle 1. The car navigation system 7 is connected to the voice recognition device 6 via the local area network, and performs operations such as setting a destination or changing a display based on a command received by the voice recognition device 6.

  FIG. 3A is an example of a terminal table, and FIG. 3B is an example of a user table.

  The terminal table 51 stored in the storage device 5 associates a user with each information terminal wirelessly connectable to the vehicle 1. For example, in the terminal table 51, the user X is associated with the terminal P.

  The storage device 5 stores the association between the information terminal and the user in the terminal table 51 according to the setting by the user.

  In the user table 52 stored in the storage device 5, one microphone among the plurality of microphones 3 is associated with each user. For example, in the user table 52, the microphone 3b is associated with the user X.

  In addition to the user, the user table 52 has a plurality of output levels for the guidance sound to the vehicle interior of the vehicle 1, a noise level in the vehicle interior, a speed of the vehicle 1, and / or a closed state of the vehicle interior. One of the microphones 3 may be associated. In this case, for example, a different microphone can be assigned to the user X depending on whether or not the output level of the guidance voice is equal to or higher than a predetermined threshold.

  The output level of the guidance voice to the cabin is the level of the voice that the car navigation system 7 outputs as guidance for the occupant. It is considered that the user increases the output level of the guidance voice when the guidance voice is difficult to hear. Therefore, when the user raises the output level of the guidance sound more than usual, it is considered that the guidance sound is difficult to hear in the vehicle cabin. By associating a microphone with each output level of the guidance voice, the voice collection device 2 can select a microphone suitable for a vehicle cabin where the guidance voice is difficult to hear.

  Similarly, the noise level in the vehicle compartment, the speed of the vehicle 1, and the closed state of the vehicle compartment are also factors that may affect the acquisition of the uttered voice of the user by the microphone 3, and by associating these with the microphone, the voice collection device 2 can select a microphone suitable for the situation in the cabin.

  Note that the output level of the guidance voice to the cabin can be acquired from the car navigation system 7. The noise level in the vehicle compartment can be calculated based on the audio signal acquired from the microphone 3. The speed of the vehicle 1 can be obtained from the traveling control ECU that governs the traveling of the vehicle 1. The closed state of the vehicle compartment can be obtained based on the opening degree of the door window.

  The storage device 5 stores the association between the user and the microphone 3 in the user table 52 according to the setting by the user. The storage device 5 changes the association between the user and the microphone 3 by a learning process described later.

  FIG. 4 is a schematic diagram showing a schematic configuration of the voice collecting device.

  The voice collection device 2 includes an input unit 21, an output unit 22, a storage unit 23, and a calculation unit 24. The voice collection device 2 is mounted on the vehicle 1 as an ECU (Electronic Control Unit).

  The input unit 21 is a circuit that receives signals from the microphone 3, the wireless interface 4, the storage device 5, and the like. The input unit 21 supplies the received signal to the calculation unit 24.

  The output unit 22 is a circuit that transmits a signal to the storage device 5, the voice recognition device 6, and the like. The output unit 22 transmits the signal supplied from the calculation unit 24 to the storage device 5, the voice recognition device 6, and the like.

  The storage unit 23 is a nonvolatile memory that stores information. The storage unit 23 stores an instruction, data, and a threshold used for the operation by the operation unit 24.

  The storage unit 23 has a buffer 231. The buffer 231 temporarily stores the acquired audio signal for each of the plurality of microphones 3g.

  The arithmetic unit 24 performs an arithmetic operation based on the signal supplied from the input unit 21 and outputs a signal to the output unit 22. The calculation unit 24 has a processor that performs a calculation by executing a predetermined program.

  The calculation unit 24 includes a voice acquisition unit 241, a first microphone identification unit 242, a user identification unit 243, a second microphone identification unit 244, a microphone selection unit 245, a buffer control unit 246, and a learning unit 247. Have. Each of these units included in the arithmetic unit 24 is a functional module executed on a processor included in the arithmetic unit 24. Alternatively, these units included in the arithmetic unit 24 may be realized by a dedicated circuit.

  The sound acquisition unit 241 acquires sound signals from the plurality of microphones 3.

  The first microphone identification unit 242 identifies the microphone based on the intensity and / or arrival time of the audio signal acquired from the microphone 3.

  The user specifying unit 243 specifies a user corresponding to the audio signal acquired from the microphone 3.

  The second microphone identification unit 244 identifies the microphone associated with the identified user in the user table 52 that associates one microphone among the plurality of microphones 3 for each of the plurality of users.

  In addition, in addition to the user, a plurality of output levels of the guidance sound to the cabin of the vehicle 1, the noise level in the cabin, the speed of the vehicle 1, and / or the closed state of the cabin are displayed in the user table 52. When one of the microphones 3 is associated, the second microphone identification unit 244 may identify the microphone associated with these microphones.

  When the microphone specified by the first microphone specifying unit 242 and the microphone specified by the second microphone specifying unit 244 do not match, the microphone selecting unit 245 recognizes the microphone specified by the second microphone specifying unit 244 by voice recognition. Select the microphone to be used for processing.

  The buffer control unit 246 causes the buffer 231 to temporarily store the acquired audio signal for each of the plurality of microphones 3. Further, the buffer control unit 246 outputs an audio signal acquired by the selected microphone and temporarily stored in the buffer 231.

  By outputting the audio signal temporarily stored by the buffer control unit 246, the audio collection device 2 can output the content uttered by the user before the microphone is selected to the audio recognition device 6 as an object of audio recognition. it can.

  The learning unit 247 receives the result of the voice recognition of the voice signal acquired by the selected microphone, and changes the association between the user specified in the user table 52 and the microphone based on the result.

  By the learning unit 247 changing the association in the user table 52 based on the result of the voice recognition, the voice collection device 2 can select a microphone suitable for voice recognition for each user.

  With the configuration including the above-described units, the voice collection device 2 of the present embodiment can appropriately select a microphone to be used for voice recognition processing, depending on the user who speaks, even in a situation that is easily affected by the reflected sound. .

  FIG. 5 is a processing flowchart of the voice collection device.

  First, the sound acquisition unit 241 acquires sound signals from the plurality of microphones 3 (Step S11). At this time, the buffer control unit 246 temporarily stores the acquired audio signal in the buffer 231 for each of the plurality of microphones.

  Next, the first microphone specifying unit 242 specifies the microphone 3 based on the intensity and / or arrival time of the audio signal acquired from the microphone 3 (Step S12).

  Next, the user identification unit 243 identifies a user corresponding to the audio signal acquired from the microphone 3 (step S13). If the user cannot be identified (step S13: N), the processing of the audio collection device 2 will be described later. Transit to S18.

  In step S13, the user identification unit 243 first detects an identifier of a terminal carried by the user. The user identification unit 243 can detect the identifier of the terminal using the wireless interface 4. Then, the user specifying unit 243 specifies a user associated with the identifier detected in the terminal table 51 as a user corresponding to the audio signal.

  Since the voice collecting apparatus 2 specifies the user based on the identifier of the terminal usually carried by the user, the voice collecting apparatus 2 can reliably specify the user without requesting the user to perform any particular operation.

  Further, the user specifying unit 243 may specify the user based on the output of the weight sensor incorporated in the seat, the output of the fingerprint sensor incorporated in the steering wheel, and the like. In this case, a table that associates these outputs with the users is stored in the storage device 5 instead of the terminal table 51.

  Next, the second microphone identification unit 244 identifies a microphone associated with the identified user in the user table 52 that associates one microphone among the plurality of microphones 3 for each user (Step S14), and identifies the microphone. If the microphone associated with the user cannot be specified (step S14: N), the process of the sound collection device 2 transitions to step S18 described below.

  In addition, in addition to the user, a plurality of output levels of the guidance sound to the cabin of the vehicle 1, the noise level in the cabin, the speed of the vehicle 1, and / or the closed state of the cabin are displayed in the user table 52. If one of the microphones 3 is associated, the second microphone identification unit 244 may identify the microphone associated with them in step S14.

  Next, the microphone selection unit 245 determines whether the microphone specified in Step S12 matches the microphone specified in Step S14 (Step S15).

  If it is determined that the microphone specified in step S12 matches the microphone specified in step S14 (step S15: Y), the microphone selection unit 245 selects the matched microphone as the microphone to be used for the voice recognition process (step S16). ).

  If it is determined that the microphone specified in Step S12 does not match the microphone specified in Step S14 (Step S15: N), the microphone selection unit 245 determines that the microphone specified in Step S14 is used as the microphone used for the voice recognition process. Select (step S17).

  If the user cannot be specified in step S13 (step S13: N), and if the microphone associated with the user specified in step S14 cannot be specified (step S14: N), the microphone selection unit 245 determines in step S12. The selected microphone is selected as a microphone to be used for the voice recognition processing (step S18).

  Next, the buffer control unit 246 causes the voice signal acquired by the selected microphone and temporarily stored in the buffer 231 to be output to the voice recognition device 6 for voice recognition processing (step S19).

  Next, the learning unit 247 receives, from the speech recognition device 6, a speech recognition result of the speech signal acquired by the selected microphone. Then, the learning unit 247 changes the association between the specified user and the microphone in the user table 52 based on the result (Step S20).

  The voice collection device 2 executes the above-described series of processes each time a user's utterance is detected. Further, the voice collection device 2 may execute step S20 every time the number of times of processing from steps S11 to S19 reaches a predetermined number.

  The voice collection method according to the present embodiment executes the processing including the above-described steps to appropriately select a microphone to be used for the voice recognition processing according to the uttering user even in a situation where the reflected sound is susceptible. To make things possible.

  It will be understood by those skilled in the art that various changes, substitutions and modifications can be made without departing from the spirit and scope of the invention.

Reference Signs List 1 vehicle 2 sound collecting device 241 sound obtaining unit 242 first microphone specifying unit 243 user specifying unit 244 second microphone specifying unit 245 microphone selecting unit 246 buffer control unit 247 learning unit 3, 3a to 3c microphone

Claims (6)

An audio acquisition unit that acquires audio signals from a plurality of microphones;
A first microphone identification unit that identifies the microphone based on the intensity and / or arrival time of the audio signal acquired from the microphone;
A user identification unit that identifies a user corresponding to the audio signal acquired from the microphone,
A second microphone identification unit that identifies a microphone associated with the identified user in a user table that associates one microphone among the plurality of microphones for each user;
When the microphone is specified by the second microphone specifying unit and the microphone specified by the first microphone specifying unit does not match the microphone specified by the second microphone specifying unit, the microphone is specified by the second microphone specifying unit. A microphone selection unit that selects the identified microphone as a microphone to be used for voice recognition processing;
An audio collection device comprising: The microphone is arranged so as to be able to acquire an audio signal according to the audio in the vehicle compartment of the vehicle,
The user table includes, in addition to the user, the plurality of microphones for each of an output level of a guidance voice to the cabin, a noise level in the cabin, a speed of the vehicle, and / or a closed state of the cabin. One of the microphones,
The second microphone identification unit is configured to output, in the user table, the identified user and an output level of a guidance voice to the cabin, a noise level in the cabin, a speed of the vehicle, and / or The sound collection device according to claim 1, wherein a microphone associated with the closed state of the room is specified.   The user identification unit identifies a user associated with an identifier of an information terminal wirelessly connected to the vehicle as a user corresponding to the audio signal in a terminal table that associates a user with each information terminal wirelessly connectable to the vehicle. The voice collecting apparatus according to claim 2, which performs the processing.   The apparatus according to claim 1, further comprising: a buffer control unit configured to temporarily store an acquired audio signal for each of the plurality of microphones and output the audio signal acquired and temporarily stored by the selected microphone for audio recognition. 4. The voice collecting device according to any one of 3.   A learning unit configured to receive a result of voice recognition of the voice signal acquired by the selected microphone and change an association between the specified user and the microphone in the user table based on the result. The voice collection device according to claim 4. An audio acquisition step of acquiring audio signals from a plurality of microphones;
A first microphone identification step of identifying a microphone based on the strength and / or arrival time of an audio signal acquired from the microphone;
A user identification step of identifying a user corresponding to the audio signal obtained from the microphone,
A second microphone identification step of identifying a microphone associated with the identified user in a user table that associates one microphone among the plurality of microphones for each user;
When the microphone is specified in the second microphone specifying step, and the microphone specified in the first microphone specifying step does not match the microphone specified in the second microphone specifying step, the microphone is specified in the second microphone specifying step. A microphone selection step of selecting the identified microphone as a microphone to be used for speech recognition processing;
Voice collection method including.

Images