close

Вход

Забыли?

вход по аккаунту

?

JP2016225884

код для вставкиСкачать
Patent Translate
Powered by EPO and Google
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2016225884
Abstract: The present invention provides a sound collection device that can easily acquire a
speaker's voice with high accuracy while suppressing reduction in the number of parts and
complication of the structure, and a control method of the sound collection device. SOLUTION: It
is judged whether or not it is an operation mode (utterance mode, call mode) for acquiring a
voice of an occupant who is a utterer (step S1A), and in another operation mode (voice
reproduction mode, head position measurement mode) When there is, the directivity of the
microphone array is set to the outside of the occupant or omnidirectional (steps S4A and S5A),
the input voice of the microphone array is acquired, and the occupant's voice is acquired. In the
case of mode), the directivity of the microphone array is directed to the occupant (step S2A).
[Selected figure] Figure 6
Sound collector and control method of sound collector
[0001]
The present invention relates to a sound collection device capable of controlling the directivity of
a microphone array in which a plurality of microphones are arranged, and a control method of
the sound collection device.
[0002]
There is known an in-vehicle device that carries a microphone on a dashboard, a ceiling, a
steering wheel, etc. of a car and performs hands-free calling and voice recognition.
03-05-2019
1
A device that performs this type of sound collection (hereinafter referred to as a sound collection
device) includes a microphone array in which a plurality of microphones are arranged to control
the directivity of the microphone array by beamforming control, in order to further improve the
performance. Has been proposed (see, for example, Patent Document 1). In this patent document
1, a position detector for detecting the position of the head of a speaker in a car based on the
position of a rearview mirror or a seat provided in the car is provided, the detected head position,
and each microphone The directivity calculation processing is performed based on the position
of.
[0003]
WO 2009/157195
[0004]
However, the conventional configuration requires a position detector that physically detects the
position of the rearview mirror or the seat, which tends to increase or decrease the number of
parts and to complicate the structure.
The present invention has been made in view of the above-described circumstances, and it is
possible to control the sound collection device and the sound collection device that can easily
acquire the voice of the speaker with high accuracy while suppressing the reduction in the
number of parts and the complication of the structure. Intended to provide a method.
[0005]
In order to achieve the above object, according to the present invention, in a sound collection
device capable of controlling the directivity of a microphone array in which a plurality of
microphones are arranged, a determination unit that determines whether or not it is a first state
for acquiring the voice of a speaker And in the second state which is not the first state, the
directivity of the microphone array is set to the outside of the speaker or non-directional to
acquire the input sound of the microphone array, and in the first state. In this case, the control
unit may be configured to direct the directivity of the microphone array to the speaker.
[0006]
In the above configuration, the microphone array is disposed behind the speaker, and the control
03-05-2019
2
unit sets the directivity of the microphone array in a range from the back of the head to the neck
of the speaker in the first state. You may do so.
[0007]
In the above configuration, the control unit acquires ambient noise based on the input voice
acquired in the second state, and removes the ambient noise from the input voice of the
microphone array in the first state. Noise removal processing may be performed.
[0008]
Further, in the above configuration, the sound collection device has a voice output unit that
outputs a voice via a predetermined speaker, and an utterance mode for acquiring the voice of
the speaker as an operation mode, and the speaker The second mode may include a speech mode
for outputting a voice and acquiring the speech of the speaker, and the second state may include
a state in which the speech mode and the speech of the speaker in the speech mode are not
acquired.
[0009]
Further, in the above configuration, as the operation mode, a position specifying mode for
obtaining a reflected sound output from the speaker and reflected by the speaker and specifying
a position of the speaker based on the reflected sound is provided, The second state may include
the state of the position specifying mode, and in the first state, the directivity of the microphone
array may be controlled in accordance with the position of the speaker specified in the position
specifying mode.
[0010]
Further, in the above configuration, the audio output unit outputs an ultrasonic wave signal from
the speaker in the audio output mode for outputting audio via the speaker, acquires the reflected
sound, and outputs the reflected sound as the reflected sound. A position specifying process for
specifying the position of the speaker may be performed based on the position.
[0011]
In the above configuration, the microphone array and the speaker may be disposed in a housing
disposed behind the speaker.
03-05-2019
3
Further, in the above configuration, the housing may be a headrest.
[0012]
Further, according to the present invention, there is provided a control method of a sound
collection apparatus capable of controlling the directivity of a microphone array in which a
plurality of microphones are arranged, determining whether or not it is a first state for acquiring
a voice of a speaker; In the case of the second state other than the one state, the directivity of the
microphone array is set outside the speaker or in the omnidirectional manner to acquire the
input sound of the microphone array, and in the case of the first state, And B. directing the
directivity of the microphone array to the speaker.
[0013]
In the present invention, it is determined whether or not the first state for acquiring the voice of
the speaker is determined, and in the case of the second state which is not the first state, the
directivity of the microphone array is outside the speaker or not omnidirectional To set the input
sound of the microphone array to direct the directivity of the microphone array to the speaker in
the case of the first state, so that the directivity of the microphone array is outside the speaker or
no directivity It becomes possible to obtain ambient noise and position information of the
speaker, which are information useful for speech processing at the time of speech, based on the
input speech when set to.
It becomes possible to perform voice processing such as removal of ambient noise and directivity
control at the time of speech using this obtained information, and a dedicated position detector
or the like is unnecessary.
Therefore, it is easy to obtain the voice of the speaker with high accuracy while suppressing the
reduction of the number of parts and the complication of the structure.
[0014]
It is a figure showing an example of use of a headrest device concerning a 1st embodiment.
03-05-2019
4
It is a perspective view of a headrest device.
It is a sectional side view of a headrest device.
It is the figure which looked at a headrest device from the upper part with a crew member's head.
It is a block diagram which shows the electric constitution of a voice input / output unit. It is a
flowchart which shows control regarding the directivity of a microphone array. It is a flow chart
which shows control at the time of head position measurement mode. It is a flowchart which
shows control at the time of audio | voice reproduction mode. It is a flow chart which shows
control at the time of utterance mode. It is a flow chart which shows control at the time of a talk
mode.
[0015]
Hereinafter, embodiments of the present invention will be described with reference to the
drawings. (First Embodiment) FIG. 1 is a view showing a usage example of a headrest device 11
according to a first embodiment. FIG. 2 is a perspective view of the headrest device 11, and FIG.
3 is a side sectional view of the headrest device 11. In each of the drawings, the front in each
direction with reference to the occupant M is indicated by a reference F, the upper side is
indicated by a reference U, and the left is indicated by a reference L. As shown in FIG. 1, the
headrest device 11 is installed on a passenger seat 12 provided in a vehicle (in the present
embodiment, an automobile), and a passenger (in the present embodiment, a driver) seated on
the passenger seat 12 It functions as a headrest placed behind M's head MH. In addition, this
headrest device 11 can be adopted not only in a car but also in various vehicles such as an
airplane, a ship, and a train. Further, the headrest device 11 can be applied not only to a vehicle
but also to furniture used in offices and homes such as office chairs and sofas, and in general, it
can be widely applied to equipment provided with a headrest. .
[0016]
The headrest device 11 includes a hollow casing 13 forming the outer shape of the headrest
03-05-2019
5
device 11 and a pair of left and right headrest stays 14 extending downward from the casing 13.
In the housing 13, a cushioning material (not shown) for shock reduction and a voice input /
output unit 21 are accommodated. The voice input / output unit 21 includes a pair of left and
right speakers 22, a plurality of microphones (hereinafter referred to as microphones) 23, and a
circuit board 24. Under the control of the circuit board 24, voice is transmitted via the speakers
22. , And input (collect) an external voice through the microphone 23. Thus, the headrest device
11 has an audio input / output function.
[0017]
As shown in FIGS. 2 and 3, the voice input / output unit 21 is formed thin in a short length in the
front-rear direction, and is accommodated in the front space of the housing 13. By making the
compact shape thin in the front-rear direction in this manner, the voice input / output unit 21
can be easily accommodated in the headrest of various shapes (that is, the housings 13 of various
shapes). Further, the housing 13 is movable up and down via the headrest stay 14 and is tiltable
back and forth with respect to the headrest stay 14 (the tilting direction is indicated by the arrow
K in FIG. 3). Thus, the position of the housing 13 can be appropriately adjusted in accordance
with the position of the head MH of the occupant M.
[0018]
The pair of left and right speakers 22 is disposed on the front plate portion 13A of the housing
13 at an interval to the left and right, and outputs sound forward. For these speakers 22,
vertically long speakers are used. As a result, even if the position of the head MH of the occupant
M fluctuates up and down, voice can be output toward the head MH of the occupant M. Further,
by moving the headrest device 11 up and down, even if the position of the head MH is different
due to the difference in the physical size of the occupant M, voice can be output toward the head
MH of the occupant M. Also, the plurality of microphones 23 are disposed between the left and
right speakers 22, and are provided to be able to collect at least the sound in front of the
headrest. A circuit board 24 is disposed behind the microphones 23 and between the left and
right speakers 22. In this manner, the microphone 23, the circuit board 24, and the speaker 22
are disposed compactly, and the overall miniaturization is achieved.
[0019]
03-05-2019
6
Although the microphone 23 is attached to the circuit board 24 in the present embodiment, the
microphone 23 may be provided separately from the circuit board 24. Further, since the voice
input / output unit 21 is accommodated in the front space of the housing 13, the layout in which
the speakers 22 and the microphones 23 are directed forward can be maintained even if the
angle of the housing 13 changes.
[0020]
FIG. 4 is a top view of the headrest device 11 together with the head MH of the occupant M. As
shown in FIG. As shown in FIG. 4, since the head MH of the occupant M is located in front of the
housing 13 of the headrest device 11, the sounds of the pair of left and right speakers 22 are
efficiently output toward the left and right ears of the occupant M. be able to. A plurality of (two
in the present configuration) microphones 23 are disposed at the same height with an interval on
the left and right, respectively, to collect sounds from the front of the headrest. Non-directional
microphones are used as the microphones 23 themselves, and can widely collect surrounding
sounds including the voice of the occupant M.
[0021]
In addition, when the occupant M speaks, the voice of the vocal cords comes out rearward
through the skin in the range from the back of the occupant M to the neck (cervical spine), and
the voice of the occupant M is accurately determined by collecting this voice. I can collect sound.
In addition, since the microphones 23 of the present embodiment are arranged at intervals in the
left and right, the sounds from the left and right can be collected at different phases, and the left
and right sounds can be easily determined.
[0022]
As shown in FIG. 3, the opening 23K in front of the microphone 23 (microphone opening) 23K is
such that it is not affected even if the installation angle of the voice input / output unit 21 or the
angle adjustment position of the headrest changes. It is formed in the mortar-shaped opening
shape which diameter-expands. As a result, it becomes easy to pick up the voice of the occupant
M emitted backward between the back of the occupant M and the neck, regardless of the layout
and adjustment angle of the headrest of the audio input / output unit 21 and various vehicle
types Is easy to deploy.
03-05-2019
7
[0023]
As shown in FIG. 4, the left and right end portions of the headrest device 11 have a sound cut
structure that does not allow sound to pass through. As a result, the sound leaking from the left
and right speakers 22 to the rear and the like is cut, and the influence on the sound entering the
microphone 23 can be suppressed. In addition, what is necessary is just to apply the well-known
sound cut structure of providing a sound absorption material in the speaker 22 back, etc. as a
sound cut structure.
[0024]
FIG. 5 is a block diagram showing the electrical configuration of the voice input / output unit 21.
As shown in FIG. The circuit board 24 is configured by mounting the connectors 31 and 32, the
DC-DC converter 33, the communication unit 34, the audio processing unit 35, the amplifier 36,
and the remote control 37. The connector 31 receives the vehicle power supply 41 and receives
an operation signal of the operation unit 42 operated by the occupant M. Further, the
microphones 23 are connected to the other connectors 32 respectively. The DC-DC converter 33
DC-converts the power from the vehicle power supply 41 into a predetermined power, and
supplies the power to each part of the audio input / output unit 21.
[0025]
The communication unit 34 functions as a communication device that communicates with
another device (not shown), and in the present embodiment, performs near field communication
according to a communication standard such as Bluetooth (registered trademark). The
communication unit 34 outputs the reception result to the voice processing unit 35, and
transmits the microphone voice input through the voice processing unit 35 to another device.
Communication is not limited to wireless communication, and may be wired communication.
Other devices are, for example, a mobile phone (including a smart phone), a tablet terminal, a car
audio, and a car navigation device.
[0026]
03-05-2019
8
The audio processing unit 35 is configured by an audio DSP (Digital Signal Pocessor) having a
microphone amplifier. The voice processing unit 35 executes a control program stored in
advance to determine a current operation state (such as an operation mode), a control unit that
controls each unit of the voice input / output unit 21, and It functions as an arithmetic
processing unit that performs various arithmetic processing, an audio amplification unit that
amplifies input sound from each microphone 23, and the like. The voice processing unit 35 and
the communication unit 34 realize linkage processing between another device and the voice
input / output unit 21 as described later. The arithmetic processing performed by the audio
processing unit 35 includes beam forming control (arithmetic) processing of microphone input,
reproduction output control (including arithmetic processing for thinning speaker output)
corresponding to noise level of microphone input, and head position measurement mode. There
are test tone output, distance calculation processing by microphone input, and voice recognition
processing of microphone voice.
[0027]
The beam forming control process is a process of controlling the directivity of the microphone
array 23A having a plurality of microphones 23. More specifically, based on the fact that sound
propagation from the sound source to each of the microphones 23 is different, the sound from a
specific direction is emphasized or reduced. For example, it has a plurality of directivity patterns,
and switching the pattern or changing parameters of each pattern switches the voice of the
occupant M to a directivity pattern that collects the voice with high accuracy, or surrounding
voice (such as noise) Switch to a directivity pattern that collects sound with high accuracy. The
reproduction output control according to the noise level of the microphone input is a process of
controlling the volume or the frequency so as to reduce the influence of ambient noise.
Hereinafter, when it is not necessary to distinguish the microphones 23 in particular, the
microphones 23A will be described.
[0028]
The head position measurement mode is a mode in which a test signal is output from the speaker
22 and the position of the head MH of the occupant M is measured based on the reflected sound
collected by the microphone array 23A. In this mode, position identification processing is
performed to calculate the separation distance L1 (see FIG. 4) between the head MH of the
occupant M and the microphone array 23A. This separation distance L1 is used when controlling
the directivity of the microphone array 23A to the head MH by beam forming control. In
03-05-2019
9
addition, these processes can apply a well-known process widely. As a result, it is possible to
suppress the deterioration in the quality of the collected sound of the occupant M and the
deterioration in the speech recognition rate due to the influence of the seat position, the physical
size of the occupant M, and the separation distance L1 with the microphone array 23A.
[0029]
By the way, when the position of an ear or a head is specified by a sensor in order to increase the
accuracy of directivity etc., it becomes a very complicated system, resulting in a significant cost
increase. Also, considering that the occupant M moves the face, when the microphone array 23A
is installed by an arm or the like around the face, the microphone array 23A or the like may get
in the way, so it is necessary to consider the influence of the air bag operation. It also happens. In
this embodiment, the microphone array 23A is disposed behind the head MH of the occupant M,
and the voice of the occupant M is collected by controlling the directivity of the microphone
array 23A, thereby reducing the number of parts and complexity of the structure. In this way,
voice recognition and hands-free calling can be realized without obstructing the movement of the
face of the occupant M.
[0030]
The amplifier 36 drives the speaker 22 based on the audio data acquired through the
communication unit 34 under the control of the audio processing unit 35, and causes the
speaker 22 to output an audio corresponding to the audio data. By applying a digital amplifier to
the amplifier 36, the amplifier 36 can be miniaturized. The remote control 37 controls the
operation of the audio processing unit 35 based on the operation signal of the operation unit 42.
The operation unit 42 receives the operation of the occupant M, and switches the operation
mode of the voice input / output unit 21 and the like.
[0031]
In this operation mode, there is a head position measurement mode for measuring the position of
the head MH of the occupant M (the separation distance L1 shown in FIG. 4), and an audio for
reproducing audio (such as music and navigation audio) sent from other devices. There are a
reproduction mode, an utterance mode for recognizing the voice of the occupant M, and a call
mode for realizing a hands-free call using a mobile phone.
03-05-2019
10
[0032]
Next, the operation of the voice input / output unit 21 will be described.
FIG. 6 is a flowchart showing control regarding the directivity of the microphone array 23A. As
shown in this figure, in the voice input / output unit 21, the voice processing unit 35 determines
whether or not it is a predetermined operation mode for acquiring the voice of the occupant
(speaker) M (step S1A). Control is performed to change the directivity of the microphone array
23A according to the result (steps S2A, S4A, S5A).
[0033]
More specifically, the voice processing unit 35 directs the directivity of the microphone array
23A to the occupant M (step S2A) in the operation mode (a speech mode, a call mode) for
acquiring the voice of the occupant M (step S1A; YES). . On the other hand, if the operation mode
is not the above-mentioned operation mode (Step S1A; NO), it judges whether or not it is the
sound reproduction mode (Step S3A). If it is the sound reproduction mode (Step S3A; YES) The
directivity control is canceled, or the directivity of the microphone array 23A is directed to the
left and right outside the occupant M (step S4A). It should be noted that whether the directivity
control of the microphone array 23A is canceled or directed to the left and right outside of the
occupant M may be in accordance with an initial setting or the like set by the user (the occupant
M or the like). If the sound reproduction mode is not set (step S3A; NO), the sound processing
unit 35 cancels the directivity control of the microphone array 23A (step S5A). When directivity
control is canceled, it becomes omnidirectional.
[0034]
Next, control regarding the speaker 22 and the microphone array 23A in each operation mode
will be described. FIG. 7 is a flowchart showing control (position specifying processing) in the
head position measurement mode. In the case of the head position measurement mode, the voice
processing unit 35 cancels the directivity control of the microphone array 23A (step S5A) and
then outputs a test signal (for example, test tone) from the speaker 22 (step S2B). The reflected
sound of the test signal is acquired by the array 23A (step S3B). In this case, the sound of the
speaker 22 is reflected by the head MH of the occupant M and collected by the microphone array
23A. More specifically, the sound of the test signal reflected in any range from the back of the
03-05-2019
11
occupant M to the neck is collected by the microphone array 23A. The voice processing unit 35
acquires the reflected sound of the test signal by performing a process of identifying the
reflected sound from the input sound of the microphone 23.
[0035]
Next, based on the delay time of the reflected sound (the time from the output of the test signal
to the time when the sound is collected by the microphone 23), the sound processing unit 35
measures the distance to the point where the sound of the speaker 22 is reflected. That is, the
separation distance L1 which is the distance from the back of the occupant M to the neck is
calculated (step S4B). Then, the voice processing unit 35 stores the information of the calculated
separation distance L1 in the memory (not shown) (step S5B), and the forming control
information for setting the directivity of the beam forming control to the information of the
separation distance L1. (Step S6B). That is, when controlling the directivity of the microphone
array 23A, the voice processing unit 35 can efficiently collect sound from the range from the
back of the occupant M to the neck based on the separation distance L1. Set the directivity of
23A. The setting of the directivity may be performed by applying known control such as
processing of selecting an optimum directivity pattern from among a plurality of directivity
patterns.
[0036]
The head position measurement mode is performed before the transition to an operation mode
(for example, a speech mode or a speech mode) in which directivity control of the microphone
array 23A is performed. For example, it may be performed at the time of calling or before the
ringing and ringing.
[0037]
FIG. 8 is a flowchart showing control in the sound reproduction mode. Since the sound
reproduction mode is an operation mode for reproducing sounds (such as music and navigation
sound) sent from other devices, the corresponding sound is output from the speaker 22. In the
case of the sound reproduction mode, the sound processing unit 35 cancels the directivity
control of the microphone array 23A, or with the directivity of the microphone array 23A
directed to the left and right outside the occupant M (step S4A), The input speech of the array
03-05-2019
12
23A is acquired (step S2C).
[0038]
This input voice is a voice including ambient noise and a voice of the speaker output (including a
reflected sound), and the voice processing unit 35 performs an arithmetic process to thin out the
voice for the speaker output from the acquired voice. Noise is extracted, and the average sound
pressure of this noise level is detected as an actual noise level (step S3C). The voice processing
unit 35 stores the detected noise information (noise level in the present embodiment) in a
memory (not shown) (step S4C), and automatically adjusts the volume and acoustic
characteristics of the speaker output based on the noise information. (Step S5C).
[0039]
As the contents of the automatic adjustment, for example, when the noise level changes by a
predetermined threshold or more, the volume level is adjusted gently with a fixed slope. Also, a
simple control may be used to increase the volume as the noise level increases and to lower the
volume as the noise level decreases, or as the control of the acoustic characteristics, the higher
the noise level, the lower and higher frequencies are enhanced to make it easier to hear
Equalizing control may be performed.
[0040]
By appropriately repeating the processing of steps S2C to S5C, it is possible to accurately detect
the level of ambient noise during the audio playback mode, and to perform audio playback
according to the ambient noise to make it easy for the occupant M or the like to hear Can. Note
that not only the noise level is stored as the noise information, but information such as the noise
frequency may be stored. In that case, it is preferable to utilize the frequency of noise for volume
control and control of acoustic characteristics to suppress the influence of the noise.
[0041]
FIG. 9 is a flowchart showing control in the speech mode. As a premise, when the speech
03-05-2019
13
processing unit 35 detects that the speech switch for instructing the start of speech recognition
(or speech operation) is operated by the occupant M or the like, the speech processing unit 35
shifts to the speech mode. As shown in FIG. 9, when the voice processing unit 35 detects the
operation of the speech switch (step S1D), the voice processing unit 35 detects the directivity of
the microphone array 23A based on the separation distance L1 acquired in the head position
measurement mode. Aim to M (any range from the back of the head to the neck) (step S2A). Next,
the voice processing unit 35 obtains input voice of the microphone array 23A, and obtains voice
from the occupant M by beam forming control (step S3D).
[0042]
Since the voice acquired in this manner directs the directivity of the microphone array 23A to the
occupant M, it is voice data that surely includes the voice of the occupant M, but it is highly likely
to include ambient noise such as engine noise. . If the influence of ambient noise is large, voice
recognition can not be performed with high accuracy. Therefore, the voice processing unit 35
performs noise reduction processing for removing ambient noise from the obtained voice (voice
data) based on the information of ambient noise obtained in the voice reproduction mode (step
S4D), and then performs voice recognition processing. (Step S5D).
[0043]
As described above, the directivity of the microphone array 23A is directed to the occupant M,
and ambient noise is removed to perform voice recognition. Therefore, the voice of the occupant
M can be obtained with high accuracy, and voice can be recognized with high accuracy. After the
speech recognition process, the speech processing unit 35 transmits the speech recognition
result to another device via the communication unit 34 (step S6D). Thus, other devices can be
voice-operated on the basis of the voice of the occupant M. Thereafter, the voice processing unit
35 cancels the directivity control of the microphone array 23A, or directs the directivity of the
microphone array 23A to the left and right outside the occupant M (step S7D).
[0044]
Here, even in the speech mode, after step S7D, the voice processing unit 35 performs detection
and storage of ambient noise as described in steps S3C and S3D described above. Thereby, the
information of the ambient noise can be efficiently acquired by using the timing at which the
03-05-2019
14
voice of the occupant M is not acquired. Thus, the noise information stored in the memory can be
updated to the latest information. The above is control regarding the speaker 22 and the
microphone array 23A in the speech mode.
[0045]
FIG. 10 is a flowchart showing control in the call mode. As a premise, when the voice processing
unit 35 detects that a mobile phone connected by Bluetooth or the like is in a call state (a call
origination or a call reception state), the voice processing unit 35 shifts to a call mode. As shown
in FIG. 10, when the voice processing unit 35 detects that the cellular phone is in a call state
(step S1E), the directivity of the microphone array 23A is determined based on the stored
separation distance L1 as the occupant M ( It is directed to any range from the back of the head
to the neck) (step S2E). The process of step S2E is the same as the process of step S2A described
above.
[0046]
Next, the voice processing unit 35 obtains input voice of the microphone array 23A, and
efficiently obtains voice from the occupant M by beam forming control (step S3E). The process of
step S3E is also the same as the process of step S3D described above. In the case of the call
mode, under the control of the voice processing unit 35, the process of acquiring the voice from
the other party of the call via the communication unit 34 and outputting it from the speaker 22
is also executed in parallel.
[0047]
After the process of step S3E, the audio processing unit 35 performs an echo cancellation
process and a noise reduction process (step S4E). The echo cancellation process is a process of
canceling an echo caused by the microphone array 23A collecting a sound reproduced from the
speaker 22, and a widely known process can be applied. The noise reduction process is a process
of removing the ambient noise from the acquired voice (voice data) based on the stored ambient
noise information, and is the same as the process of step 4D described above. Thus, the voice of
the occupant M from which the ambient noise has been removed can be obtained.
03-05-2019
15
[0048]
Then, the voice processing unit 35 transmits voice data after the echo cancellation processing
and the noise reduction processing to the mobile phone via the communication unit 34 (step
S5E). This makes it possible to send the voice of the passenger M without echo and noise
removed to the other end of the call. Thereafter, when the voice processing unit 35 detects the
release of the call state of the mobile phone connected by Bluetooth or the like (step S6E), the
voice processing unit 35 releases the directivity control of the microphone array 23A or the
directivity of the microphone array 23A. It is turned to the left and right which is the outside of
the occupant M (step S7E).
[0049]
Here, even in the call mode, as in the case of the speech mode, after the step S7E in which the
directivity of the microphone array 23A is directed to the outside of the occupant M, the voice
processing unit 35 proceeds to the above-described steps S3C and S3D. Detect and store ambient
noise as described. Thereby, the information of the ambient noise can be efficiently acquired by
using the timing at which the voice of the occupant M is not acquired. Thus, the noise
information stored in the memory can be updated to the latest information. The above is control
regarding the speaker 22 and the microphone array 23A in the call mode.
[0050]
As described above, in the headrest device 11 according to the present embodiment, the voice
processing unit 35 obtains the voice of the occupant M as a state in which the voice of the
passenger M who is the utterer is obtained (referred to as “first state”). Function as a
determination unit (see step S1A in FIG. 6) for determining whether or not the operation mode
(speech mode, call mode) to be performed (refer to step S1A in FIG. 6). When the head position
measurement mode is set, the directivity of the microphone array 23A is set to the outside of the
occupant M or non-directional to acquire the input sound of the microphone array 23A, and in
the case of the operation mode in the first state Function as a control unit that directs the
directivity of the microphone array 23A to the occupant M.
[0051]
Thus, based on the input voice when the directivity of the microphone array 23A is set outside
03-05-2019
16
the occupant M or in a non-directional manner, ambient noise and position information of the
occupant M, which are information useful for speech processing at the time of speech, are
obtained. Becomes possible.
It becomes possible to perform voice processing such as removal of ambient noise and directivity
control at the time of speech using this obtained information, and a dedicated position detector
or the like is unnecessary. Therefore, it is easy to obtain the voice of the occupant M with high
accuracy while suppressing the reduction of the number of parts and the complication of the
structure.
[0052]
In addition, the microphone array 23A is disposed behind the passenger M who is a speaker, and
the voice processing unit 35 detects the directivity of the microphone array 23A in the operation
mode (the speech mode, the communication mode) in the first state. Since the range from the
back of the head to the neck of M is set, the voice of the occupant M can be efficiently acquired
at the rear of the occupant M. In addition, the voice processing unit 35 acquires ambient noise
based on the input voice acquired in the operation mode in the second state (audio reproduction
mode) (see FIG. 8), and the operation mode in the first state (utterance mode, call mode) In the
above case, noise reduction processing (noise removal processing) for removing ambient noise
from the input voice of the microphone array 23A is performed, so ambient noise can be easily
and easily obtained with high accuracy, and sufficient noise removal can be facilitated.
[0053]
Further, the voice processing unit 35 and the amplifier 36 function as a voice output unit that
outputs voice via the speaker 22 and outputs a voice via the speaker 22 and an utterance mode
for acquiring the voice of the occupant M and the passenger M. The directivity of the microphone
array 23A is set to the outside of the occupant M or omnidirectional when the operator M does
not acquire the voice of the occupant M (corresponding to the second state) even in the call
mode for acquiring the voice of Then, the input sound of the microphone array 23A is acquired
(step S7D in FIG. 9, step S7E in FIG. 10, etc.), and ambient noise is detected and stored. As a
result, ambient noise can be acquired even in the operation mode for acquiring the voice of the
occupant M, and the latest ambient noise can be easily obtained.
03-05-2019
17
[0054]
Further, even in the head position specifying mode (corresponding to the second state) in which
the position of the occupant M is specified using the speaker 22 and the microphone array 23A,
the microphone array 23A is set to non-directional to input the microphone array 23A. Since the
voice is acquired, it is possible to acquire the reflected sound reflected by the occupant M and
acquire the position information of the occupant M with high accuracy. By using this position
information, the directivity of the microphone array 23A can be controlled so that the voice of
the occupant M can be acquired with high accuracy.
[0055]
In addition, since the microphone array 23A and the speaker 22 are disposed in the housing 13
disposed behind the occupant M, a compact device with excellent sound collection performance
and audio output function is required without a dedicated position detector or the like. Devices
can be provided. Moreover, since the housing 13 is a headrest, the microphone array 23A and
the speaker 22 can be arranged using the space of the existing headrest.
[0056]
Second Embodiment The headrest device 11 according to the second embodiment receives an
ultrasonic signal under the control of the audio processing unit 35 in the operation mode (audio
output mode) in which audio is output through the speaker 22. A reflected signal of the test
signal is output through the microphone array 23A as a test signal, and a position specifying
process for specifying the position of the occupant M is performed based on the reflected sound.
The second embodiment is the same as the first embodiment except that the position specifying
process is performed. The voice output mode may be, for example, a voice playback mode for
playing voice (music, navigation voice, etc.) sent from another device, and a call mode, or any one
mode.
[0057]
Since the ultrasonic signal is out of the audible range, it is not recognized by the occupant M, and
the directivity is sharp and it is easy to accurately measure the reflected sound. As a result, the
03-05-2019
18
position of the occupant M can be specified with high accuracy while performing music
reproduction, hands-free calling and the like, and the occupant M does not feel uncomfortable. As
a result, in addition to the various effects of the first embodiment, it is possible to obtain the
position of the occupant M with high accuracy while outputting a sound to be heard by the
occupant M. Furthermore, the position of the occupant M can be specified at any timing, which
makes it easy to obtain the latest information. Therefore, directivity control of the speaker 22
based on the position of the occupant M can be performed with high accuracy.
[0058]
It should be noted that although it is sufficient to use an ultrasonic wave band that can be output
by the speaker 22 as the ultrasonic wave signal, since recent music sometimes includes sounds
outside the audible band, the frequency band outside of the music band is used. It is preferable to
use the frequency of Further, since this position specifying process may be performed instead of
the head position measurement mode of the first embodiment, it is possible to omit the head
position measurement mode.
[0059]
The embodiment described above merely illustrates one embodiment of the present invention,
and any modification and application are possible without departing from the spirit of the
present invention. For example, the number of microphones 23 in the microphone array 23A is
not limited to two, and a speaker for bass or high tone may be added to the speakers 22.
Although the headrest device 11 functioning as a headrest and the case where the present
invention is applied to a control method thereof have been described, the directivity of a
microphone array 23A in which a plurality of microphones 23 are arranged is controlled without
being limited to a device functioning as a headrest. The present invention is broadly applicable to
possible sound collectors and control methods therefor.
[0060]
Moreover, although the case where the control program for performing the said control was
stored beforehand in the headrest apparatus 11 was demonstrated by each embodiment
mentioned above, it is not restricted to this, A magnetic recording medium, an optical recording
medium The program may be stored in a computer-readable recording medium such as a
03-05-2019
19
semiconductor recording medium, and the computer may read this control program from the
recording medium and execute it. Also, the control program may be downloaded from a
distribution server or the like via a communication network (a telecommunication line).
[0061]
11 headrest device (sound collection device) 13 case 21 voice input / output unit 22 speaker 23
microphone 23A microphone array 34 communication unit 35 voice processing unit
(determination unit, control unit) 36 amplifier
03-05-2019
20
Документ
Категория
Без категории
Просмотров
0
Размер файла
35 Кб
Теги
jp2016225884
1/--страниц
Пожаловаться на содержимое документа