close

Вход

Забыли?

вход по аккаунту

?

JPWO2017072958

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 JPWO2017072958
Abstract Adjusts for each frequency so that the noise characteristics included in the input signals
of multiple microphones are aligned with each other, and determines that the speaker has input
speech from among the input signals of multiple microphones whose noise characteristics have
been adjusted Switch the input signal of the microphone to the signal to be sent to the other
party.
Hands-free control device
[0001]
The present invention relates to a hands-free control device that determines a microphone to
which a speaker has input speech from among a plurality of microphones in a hands-free call,
and controls an input signal of the determined microphone to be transmitted to the other party .
[0002]
BACKGROUND In recent years, a hands-free call system in which a driver can make a call without
holding a mobile phone by hand has become widespread.
Also, in the hands-free communication system, it is possible to make a plurality of speakers talk
with a far-end speaker by using a plurality of microphones. In this case, in order to transmit the
uttered voice collected clearly to the other party, it is necessary to appropriately determine the
microphone where the speaker has input the uttered voice.
03-05-2019
1
[0003]
For example, Patent Document 1 describes a technique for determining, from among a plurality
of microphones, a microphone to which a speaker has input a speech sound. In the system
described in Patent Document 1, among a plurality of microphones, a microphone whose input
level exceeds a threshold continuously for a predetermined time or more as a speaker input a
speech voice is set as a microphone corresponding to the position of the speaker It is determined.
By using this determination method for a hands-free call system, it is possible to appropriately
determine a microphone from which a speaker has input a speech from among a plurality of
microphones.
[0004]
Also, in the system described in Patent Document 1, the dark noise level in the vicinity of each
microphone is measured in advance, and correction is performed to match the dark noise level of
each microphone with the dark noise level of the entire room. The microphone corresponding to
the position of the person is identified. This makes it possible to make a determination in which
the influence of the dark noise level of each microphone has been eliminated.
[0005]
JP 2007-174155 A
[0006]
In a hands-free call system mounted on a vehicle, road noise generated as the vehicle travels is
input to each of a plurality of microphones, and a voice including road noise is transmitted to the
other party.
At this time, since noise characteristics such as the gain of noise input to the microphone and the
frequency characteristics are different for each of the plurality of microphones, when the speaker
changes and the input signal of the microphone is switched, the texture of the noise heard by the
other party changes significantly And there was a problem of giving a sense of incongruity.
03-05-2019
2
[0007]
For example, when the microphone is used in combination with functions other than hands-free
calling such as voice recognition and ANC (Active Noise Control), the acoustic characteristics of
the microphone may differ depending on the function used and the installation position. In this
case, the noise characteristics also differ according to the difference in acoustic characteristics.
Therefore, when the input signal of the microphone to be transmitted to the telephone of the
other party is switched, the texture of the noise heard by the other party will greatly change.
[0008]
Further, even in the case of using a plurality of microphones for hands-free calling, if the audio
processing to be applied to the input signal is different, the characteristics of the noise included
in the input signal after the audio processing will be different accordingly. Furthermore, even if
microphones with the same acoustic characteristics are used for a plurality of microphones, if
there are individual differences in the gain, frequency characteristics, temperature
characteristics, etc. of the respective microphones, noises included in the input signal of the
microphones accordingly The characteristics of are also different. Therefore, also in these cases,
when the input signal of the microphone to be transmitted to the telephone of the other party is
switched, the texture of the noise heard by the other party will be largely changed.
[0009]
In the system described in Patent Document 1, as described above, correction is performed on
the input signal so as to match the dark noise level of each microphone with the dark noise level
of the entire room. For this reason, if the corrected signal is transmitted to the telephone of the
other party, even if the input signal of the microphone heard by the other party is switched, it is
expected that the change in the texture of noise heard by the other party will be reduced.
[0010]
However, the dark noise level of each microphone in Patent Document 1 is obtained by
03-05-2019
3
measuring in advance how much dark noise is present in the vicinity of each microphone, and it
is assumed that there is little noise variation. ing. On the other hand, in a vehicle, for example,
road noise occurs as the vehicle travels. Road noise is noise generated due to the friction between
the tire surface and the road surface, and the aspect largely varies depending on the state of the
road surface. For example, road noise mainly occurs in a low frequency band, but the dry road
surface and the wet road surface cause the noise level for each frequency to greatly fluctuate
dynamically. For this reason, in the hands-free communication system mounted in a vehicle, even
if dark noise correction processing as described in Patent Document 1 is performed, the
characteristics of the noise included in the input signals of the respective microphones can not be
made uniform. Therefore, when the input signal of the microphone to be transmitted to the
telephone of the other party is switched, the texture of the noise heard by the other party will
change significantly.
[0011]
The present invention solves the above-mentioned problems, and it is an object of the present
invention to provide a hands-free control device capable of reducing a change in the texture of
noise heard by the other party when switching the input signal of the microphone to be
transmitted to the other party's telephone To aim.
[0012]
A hands-free control device according to the present invention includes a determination unit, a
noise characteristic adjustment unit, and a switching unit.
The determination unit determines, from among the plurality of microphones, the microphone to
which the speaker has input the uttered voice. The noise characteristic adjustment unit adjusts
each frequency so that the characteristics of the noise included in the input signals of the
plurality of microphones are aligned with each other. The switching unit switches the input
signal of the microphone determined by the determination unit among the input signals of the
plurality of microphones whose noise characteristics have been adjusted by the noise
characteristic adjustment unit to a signal to be transmitted to the communication partner.
[0013]
According to the present invention, the characteristics of the noise contained in the input signals
of the plurality of microphones are adjusted for each frequency so that they match each other, so
the other party hears when the input signal of the microphone to be transmitted to the other
03-05-2019
4
party's phone is switched It is possible to reduce the change in the texture of noise.
[0014]
It is a block diagram which shows the structure of the handsfree telephone call system provided
with the handsfree control apparatus which concerns on Embodiment 1 of this invention.
It is a figure which shows the outline | summary of the handsfree telephone which the several
speaker exists. FIG. 3A is a diagram showing a hardware configuration for realizing the function
of the hands-free control device according to the first embodiment, and FIG. 3B is a diagram
showing execution of software for realizing the function of the hands-free control device
according to the first embodiment It is a figure which shows the hardware constitutions to carry
out. 5 is a flowchart showing an operation of the hands-free control device according to the first
embodiment. FIG. 7 is a diagram showing an outline of adjustment processing of noise
characteristics in the first embodiment. It is a figure which shows the case where the frequency
band used as adjustment object of a noise characteristic is changed. It is a figure which shows the
case where the frequency width used as adjustment object of a noise characteristic is changed.
[0015]
Hereinafter, in order to explain the present invention in more detail, embodiments for carrying
out the present invention will be described according to the attached drawings. Embodiment 1
FIG. 1 is a block diagram showing a configuration of a handsfree call system 1 provided with a
handsfree control device 2 according to Embodiment 1 of the present invention. FIG. 2 is a
diagram showing an overview of a hands-free call in which a plurality of speakers exist. Referring
to FIG. 2, in the hands-free communication system 1 shown in FIG. 1, there are an occupant A
seated on the driver's seat SA as a speaker and an occupant B seated on the front passenger seat
SB. The sound is collected, and the voice of the occupant B is collected by the microphone 5C.
Note that FIG. 2 shows the case where two microphones are arranged on the driver's seat SA side
and one microphone on the assistant driver's seat SB side as an example, but the combination of
the number of microphones and the characteristic is arbitrary.
[0016]
03-05-2019
5
As shown in FIG. 1, the handsfree call system 1 is a system which is mounted on a vehicle 20 and
performs handsfree calls between the vehicle 20 and a communication partner outside the
vehicle 20. The configuration includes a hands-free control device 2, a communication device 3, a
portable telephone 4A on the vehicle 20 side, microphones 5A to 5C, an amplifier 6, and
speakers 7A and 7B.
[0017]
The communication device 3 is a device that performs wireless communication with the mobile
phone 4A brought into the vehicle 20. As a wireless communication method, for example, near
field wireless communication compatible with Bluetooth (registered trademark) is used. Also, the
communication device 3 performs communication by connecting to the mobile phone 4A using a
handsfree communication protocol.
[0018]
The mobile telephone 4A establishes communication connection with the far-end mobile
telephone 4B via a mobile telephone line. For example, the mobile phone 4A transmits the voice
signal of the speaker received from the communication device 3 to the mobile phone 4B via the
mobile phone line. In addition, when the portable telephone 4A receives the voice signal of the
other party from the portable telephone 4B via the portable telephone line, the portable
telephone 4A transmits the voice signal to the communication device 3 using the handsfree
communication protocol.
[0019]
The microphones 5A to 5C are microphones installed in the cabin of the vehicle 20, and are used
for hands-free calling, for example, to collect speech voice of the speaker. In the following, the
microphones 5A to 5C will be described as having the following functions. The microphones 5A
and 5B are installed on the driver's seat SA side in the vehicle cabin, and constitute stereophonic
microphones using both the microphones 5A and 5B. The microphones 5A and 5B are directional
microphones that collect speech sounds of the occupant A (for example, a driver who operates
the steering wheel H) sitting on the driver's seat SA. In FIG. 2, the sound collection area of the
microphones 5A and 5B is indicated by an area C. Further, the microphone 5C is a nondirectional
microphone which is installed on the side of the front passenger seat SB and collects speech
03-05-2019
6
voice of the occupant B seated on the front passenger seat SB. In FIG. 2, the sound collection area
of the microphone 5 </ b> C is indicated by an area D. The microphones 5A to 5C may be
nondirectional microphones.
[0020]
The amplifier 6 is an amplifier that amplifies the signal of the voice of the other party of the
calling party received by the mobile phone 4A and input through the communication device 3.
The speakers 7A and 7B are set in the vehicle compartment of the vehicle 20, and output the
voice of the other party's speech signal amplified by the amplifier 6 as voice.
[0021]
The hands-free control device 2 determines, of the microphones 5A to 5C, a microphone to which
a speaker on the vehicle 20 side has input a speech voice, and outputs an input signal of the
determined microphone to the communication device 3. The configuration includes audio
processing units 10A to 10C, a determination unit 11, a noise characteristic adjustment unit 12,
and a switching unit 13.
[0022]
The voice processing units 10A to 10C are processing units that perform voice processing on
input signals from the microphones 5A to 5C. Further, the audio processing unit 10A performs
audio processing on the audio signal collected by the microphones 5A and 5B on the driver's seat
SA side, and the audio processing unit 10B processes the audio signal collected by the
microphone 5C on the passenger seat SB side. Voice processing. The audio processing units 10A
and 10B perform, for example, a beam forming process, an echo cancellation process, and a level
correction process on an input audio signal.
[0023]
The beam forming process is a process of emphasizing only the voice arriving from the position
of the speaker (driver) with respect to the voice signal collected by the microphones 5A and 5B.
03-05-2019
7
For example, the voice signals collected by the microphones 5A and 5B are delayed and
synthesized, and the level of the synthesized signal is detected based on the delay time prepared
in advance for each position. Then, the position corresponding to the delay time at which the
level becomes the largest is determined as the position of the occupant A. By thus identifying the
position of the occupant A and emphasizing the audio signal from that direction, it is possible to
improve the SN (Signal to Noise) ratio of the uttered voice. In FIG. 1, the beam forming process is
performed only by the audio processing unit 10A that processes the input audio from the
microphones 5A and 5B.
[0024]
The echo cancellation process is a process for canceling the echo of the voice of the other party
of the call collected by the microphones 5A to 5C. For example, an adaptive filter is used to learn
impulse responses of the speakers 7A and 7B and the microphones 5A to 5C, and a pseudo echo
obtained by convoluting the learned impulse response is generated for the audio signal output
from the speakers 7A and 7B. Generate Then, the pseudo echo is regarded as the sound from the
speakers 7A and 7B added to the microphones 5A to 5C, and the echo is canceled by subtracting
it from the input sound of the microphones 5A to 5C.
[0025]
The level correction process is a process of correcting a difference in signal level caused by each
characteristic of the microphones 5A to 5C. The characteristics of the microphones 5A to 5C are
errors due to gain with respect to input signals, frequency characteristics, directivity, individual
differences among the microphones. The level correction processing is, for example, correction
that compares the sound levels when assuming that the dark noise levels of the microphones 5A
to 5C are 0, and makes them constant. By performing these processes, the voice signals a1 and
b1 from the voice processing units 10A and 10B become signals in which the difference in level
due to the microphone characteristic is canceled from the voice signal of the occupant A or the
occupant B.
[0026]
In addition, in the audio processing unit 10C, for example, noise cancellation processing,
frequency characteristic correction processing, automatic gain control processing (hereinafter
03-05-2019
8
referred to as AGC processing), limiter processing on signals input from the audio processing
units 10A and 10B. Conduct. The noise cancellation process is a process of reducing the level of
noise included in the input signals of the microphones 5A to 5C by a predetermined level. For
example, the noise is reduced by generating a signal obtained by inverting the phase of the noise
contained in the input signal of the microphones 5A to 5C and combining it with the input signal.
Alternatively, noise may be reduced by providing transfer function characteristics to the
microphones 5A to 5C input signals. Although the noise level in the input signals of the
microphones 5A to 5C is reduced by this process, the sound quality of the uttered voice is also
degraded if the degree of cancellation is excessively high. For this reason, even in the input signal
after the noise cancellation processing, noise that the other party can hear can remain.
[0027]
The correction process of the frequency characteristic is a process of correcting the level and
phase of the input signals of the microphones 5A to 5C according to the frequency according to
the space near the microphones 5A and 5B as the reproduction space and the space near the
microphone 5C. As described above, since the microphones 5A to 5C are installed at different
positions in the vehicle compartment, the transfer characteristics of the sound due to the
reflection or absorption of the sound are different in the space near the microphones 5A and 5B
and the space near the microphone 5C. As described above, when the transfer characteristic is
different in each reproduction space, the level and phase of each frequency of the speech to be
transmitted to the other party is changed, and the sound quality is deteriorated. Therefore, the
voice processing unit 10C performs the correction process of the frequency characteristic on the
signal of the frequency band corresponding to the speech in the input signal of the microphones
5A to 5C, thereby preventing the deterioration of the sound quality due to the reproduction
space. .
[0028]
The AGC process is a process of performing gain adjustment on input signals of the microphones
5A to 5C. For example, in the audio processing unit 10C, when the reference input level is set in
advance and the levels of the input signals of the microphones 5A to 5C are less than the
reference input level, adjustment is made to lower the gain of the input signal as a silent section.
Automatically. When the input signal level is higher than the reference input level, the gain is
automatically adjusted so that the level of the input signal is slightly reduced so that the volume
of the input signal does not become too high.
03-05-2019
9
[0029]
The limiter process is a process of limiting the input signal whose signal level is higher than the
threshold level to the threshold level or less among the input signals of the microphones 5A to
5C. For example, by limiting the signal level of the input signal whose signal level exceeds the
threshold level to the threshold level, each signal level of the audio signals a2 and b2 from the
audio processing unit 10C becomes equal to or lower than the threshold level. As a result, it is
possible to prevent the deterioration of the sound quality caused by the input signal having a
high signal level that has suddenly occurred. The voice processing units 10A to 10C may be
functions of a voice processing device provided separately from the hands-free control device 2.
In this case, the hands-free control device 2 includes the determination unit 11, the noise
characteristic adjustment unit 12, and the switching unit 13, and the determination unit 11 and
the noise characteristic adjustment unit 12 appropriately input an audio signal from the audio
processing device to perform processing. It will be done.
[0030]
The determination unit 11 determines, from the microphones 5A to 5C, the microphone to which
the speaker has input the uttered voice. For example, the determination unit 11 inputs the audio
signal a1 from the audio processing unit 10A and the audio signal b1 from the audio processing
unit 10B. Then, the determination unit 11 compares the threshold level prepared in advance for
the microphones 5A and 5B with the signal level of the audio signal a1, and the threshold level
prepared in advance for the microphone 5C with the signal level of the audio signal b1. Compare.
Based on this comparison, the determination unit 11 selects one of the audio signal a1 and the
audio signal b1 having a signal level higher than the threshold level and a higher signal level, and
the speaker utters the microphone from which the selected output signal is obtained. Distinguish
from the microphone which input the voice.
[0031]
The noise characteristic adjustment unit 12 adjusts each frequency so that the characteristics of
the noise included in the input signals of the microphones 5A to 5C are aligned with each other.
For example, the noise characteristic adjustment unit 12 performs fast Fourier transform (FFT)
on the audio signals a2 and b2 from the audio processing unit 10C to obtain frequency spectra
of the audio signals a2 and b2. Then, the noise characteristic adjustment unit 12 sets a low
03-05-2019
10
frequency band corresponding to road noise in the frequency spectrum of the audio signals a2
and b2 as a frequency band to be adjusted. After that, the noise characteristic adjustment unit 12
adjusts the noise amplitude level for each frequency width predetermined in the frequency band
to be adjusted so that the characteristics of the noise included in the audio signals a2 and b2 are
aligned with each other.
[0032]
When the signal to be transmitted to the other party is switched from the input signal (audio
signal a3) of the microphones 5A and 5B to the input signal (audio signal b3) of the microphone
5C, the noise characteristic included in the audio signal b3 is an audio signal Make adjustments
to match the characteristics of the noise contained in a3. By doing this, it is possible to reduce
the change in the texture of noise heard by the calling party when the input signal of the
microphone to be transmitted to the mobile phone 4B of the calling party is switched.
[0033]
Among the input signals of the microphones 5A to 5C whose noise characteristics have been
adjusted by the noise characteristic adjustment unit 12, the switching unit 13 has an input signal
of the microphone determined by the determination unit 11 as shown by a double arrow in FIG. ,
Switch to the signal to send to the other party. For example, the switching unit 13 includes a
switch that switches to any one of the audio signals a3 and b3 from the noise characteristic
adjusting unit 12 and outputs the signal to the communication device 3. The switching unit 13
outputs the audio signal a3 to the communication device 3 when the discrimination unit 11
discriminates the microphones 5A and 5B, and communicates the audio signal b3 when the
discrimination unit 11 discriminates the microphone 5C. Output to device 3
[0034]
FIG. 3A is a diagram showing a hardware configuration for realizing the function of the handsfree
control device 2 according to the first embodiment, and FIG. 3B shows software for realizing the
function of the handsfree control device 2 according to the first embodiment. It is a figure which
shows the hardware constitutions to implement. Each function of the audio processing units 10A
to 10C, the determination unit 11, the noise characteristic adjustment unit 12, and the switching
unit 13 in the hands-free control device 2 is realized by a processing circuit. That is, the hands-
03-05-2019
11
free control device 2 includes a processing circuit for performing the processing from step ST1
to step ST3 shown in FIG. 3 described later. The processing circuit may be dedicated hardware or
a CPU (Central Processing Unit) that reads and executes a program stored in a memory.
[0035]
When the processing circuit is a dedicated hardware processing circuit 100 as shown in FIG. 3A,
the processing circuit 100 may be, for example, a single circuit, a composite circuit, a
programmed processor, a parallel programmed processor, an ASIC (Application Specific) An
integrated circuit), an FPGA (field-programmable gate array), or a combination thereof is
applicable. In addition, the functions of the audio processing units 10A to 10C, the determination
unit 11, the noise characteristic adjustment unit 12, and the switching unit 13 may be realized by
processing circuits, or the functions of the units may be realized by one processing circuit. You
may
[0036]
As shown in FIG. 3B, when the processing circuit is the CPU 101, the functions of the audio
processing units 10A to 10C, the determination unit 11, the noise characteristic adjustment unit
12, and the switching unit 13 are realized by software, firmware or a combination of software
and firmware. Be done. The software and the firmware are described as a program and stored in
the memory 102. The CPU 101 implements the functions of the respective units by reading and
executing the program stored in the memory 102. That is, the hands-free control device 2
includes the memory 102 for storing a program that is to be executed as a result of the process
of each step shown in FIG. 4 when executed by the CPU 101. Further, these programs cause the
computer to execute the procedures or methods of the audio processing units 10A to 10C, the
determination unit 11, the noise characteristic adjustment unit 12, and the switching unit 13.
[0037]
Here, the memory means, for example, nonvolatile or volatile semiconductor memory such as
RAM (Random Access Memory), ROM, flash memory, EPROM (Erasable Programmable ROM),
EEPROM (Electrically EPROM), magnetic disk, flexible disk, etc. An optical disk, a compact disk, a
mini disk, a DVD (Digital Versatile Disk), etc. correspond.
[0038]
03-05-2019
12
The audio processing units 10A to 10C, the determination unit 11, the noise characteristic
adjustment unit 12, and the switching unit 13 may be partially realized by dedicated hardware
and partially realized by software or firmware. .
For example, the audio processing units 10A to 10C realize their functions with a dedicated
hardware processing circuit 100, and the determination unit 11, the noise characteristic
adjustment unit 12 and the switching unit 13 are programs stored in the memory 102 by the
CPU 101. Implement the function by executing. Thus, the processing circuit can implement the
above-described functions by hardware, software, firmware, or a combination thereof.
[0039]
Next, the operation will be described. FIG. 4 is a flowchart showing the operation of the handsfree control device 2 according to the first embodiment, and a series of processes from the
determination of the microphone to which the speaker has input the uttered voice to the
switching of the signal to be transmitted to the other party Is shown. In addition, the process
from step ST1 to step ST3 shown in FIG. 4 shall be implemented repeatedly. First, the
determination unit 11 compares the signal level of the audio signal a1 with the threshold level
corresponding to the microphones 5A and 5B, and compares the signal level of the audio signal
b1 with the threshold level corresponding to the microphone 5C. Then, the determination unit 11
selects one of the audio signal a1 and the audio signal b1 whose signal level is higher than the
threshold level and has a higher signal level, and the speaker speaks the microphone whose
microphone has obtained the selected output signal. The input microphone is determined (step
ST1).
[0040]
In the above description, the discrimination unit 11 compares the signal levels (absolute voice
amplitudes) of the input signals of the microphones 5A to 5C to discriminate the microphone, but
the input signals of the microphones 5A to 5C are The determination may be made based on the
relative comparison result with the dark noise level. For example, the determination unit 11 may
compare the signal levels obtained by subtracting the dark noise level from the signal levels of
the input signals of the microphones 5A to 5C to determine the microphone to which the speaker
has input the uttered voice.
03-05-2019
13
[0041]
Even when microphones having the same characteristics are used as the microphones 5A to 5C,
errors due to individual differences actually occur. For this reason, although the occupant A
shown in FIG. 2 has a voice level larger than that of the occupant B, the above-mentioned error
causes the absolute voice amplitude of the input signal to be reversed, and the microphone
receives the speech voice as the speaker As the microphone 5C may be determined. Therefore,
the determination unit 11 compares signal levels obtained by subtracting the dark noise level
from the signal levels of the input signals of the microphones 5A to 5C. By thus comparing the
input signals of the microphones 5A to 5C relatively with the dark noise level, it is possible to
prevent an erroneous determination caused by individual differences of the microphones 5A to
5C.
[0042]
Note that, when the signal level obtained by subtracting the dark noise level from the signal level
of the input signal is compared and the difference is separated by a predetermined value or
more, the microphone from which the input signal with the larger signal level is obtained is
obtained. It may be determined that the microphone is a speaker who has input a speech.
[0043]
For example, the signal level obtained by subtracting the dark noise level from the signal level of
the input signal of the microphones 5A and 5B is 60 dB, and the signal level obtained by
subtracting the dark noise level from the signal level of the input signal of the microphone 5C is
59 dB. explain.
Here, if the predetermined value is 10 dB, even if the microphone 5C is selected as the
microphone to which the speaker has input the speech, the signal level difference is 1 dB, so the
inputs of the microphones 5A and 5B are obtained. Do not switch the signal to be sent. On the
other hand, if the signal level obtained by subtracting the dark noise level from the signal level of
the input signal of the microphone 5C is 50 dB, the input signals of the microphones 5A and 5B
are switched to transmission targets. By doing this, the switching is not implemented due to the
difference in the error level, and the reliability of the process of the determination unit 11 can be
improved.
03-05-2019
14
[0044]
Alternatively, the microphone may be determined by comparing the SN ratio instead of the signal
level. For example, the same comparison processing as in the case of the signal level is performed
using the SN ratio and the threshold for the SN ratio in the microphones 5A and 5B and the
threshold for the SN ratio and the SN ratio in the microphone 5C. As described above, the audio
signals a1 and b1 have their level differences due to the microphone characteristics canceled. By
comparing such voice signals a1 and b1, it becomes possible to appropriately determine the
microphone to which the speaker has input the uttered voice.
[0045]
The determination unit 11 may be capable of changing the threshold of the signal level or the
threshold of the SN ratio as described above. For example, a value input by the user may be set
as the threshold. Also, as an example, the time average of the difference between the signal level
and the SN level is observed, and a value obtained by adding a specific offset from this time
average is used as a threshold. If the difference between the SN ratio of the microphone 5A and
the microphone 5B when the occupant A speaks and the SN ratio of the microphone 5C is 7 dB
on average, 4 dB obtained by subtracting the offset value 3 dB therefrom is taken as the
threshold value. The threshold may be dynamically changed in this manner.
[0046]
In addition, the determination unit 11 determines that the signal level of the input signal of the
microphone determined to have the speaker's voice input is lower than the signal level of the
input signal of another microphone, and a predetermined holding time elapses. The other
microphones may not be determined as microphones for which the speaker has input speech. For
example, after the microphones 5A and 5B are determined as the microphones to which the
speaker has input the uttered voice, the signal level of the audio signal a1 becomes equal to or
lower than the signal level of the audio signal b1 caused by the input signal of the microphone
5C. The signal output to the communication device 3 is not switched from the audio signal a3 to
the audio signal b3 until the holding time has elapsed. In this way, the signal output to the
communication device 3, that is, the signal to be transmitted to the communication partner, is
prevented from being switched frequently, and switching can be performed when it is
03-05-2019
15
determined that the speaker has surely changed. As the holding time, for example, a time of
about several hundred milliseconds can be considered.
[0047]
Furthermore, when the signal level of one of the audio signals a1 and b1 is less than the lower
limit threshold, the determination unit 11 determines that the audio signal is silent and speaks
the microphone from which the other audio signal is obtained. It may be determined that the
microphone has input a speech voice. For example, when it is determined that the signal level of
the audio signal a1 is a silent level, the microphone 5C from which the audio signal b1 is
obtained is determined as the microphone to which the speaker has input the uttered voice.
Thereby, the signal to be transmitted to the communication party is automatically switched from
the audio signal a3 to the audio signal b3. Even in this case, it is possible to appropriately
determine the microphone to which the speaker has input the uttered voice.
[0048]
Next, the noise characteristic adjustment unit 12 adjusts each frequency so that the
characteristics of the noise included in the input signals of the microphones 5A to 5C are aligned
with each other (step ST2). For example, the noise characteristic adjustment unit 12 performs
fast Fourier transform (FFT) on the audio signals a2 and b2 to obtain frequency spectra of the
audio signals a2 and b2. Then, as shown in FIG. 5, the noise characteristic adjustment unit 12
sets a low frequency band corresponding to road noise in the frequency spectrum of the audio
signals a2 and b2 to the frequency band BW to be adjusted. After that, the noise characteristic
adjustment unit 12 sets a predetermined frequency width RB so that the characteristics of the
noise included in the audio signals a2 and b2 are aligned with each other in the frequency band
BW to be adjusted as indicated by arrows in FIG. Make adjustments to match the amplitude level
of the noise each time. As a result, the characteristics of the noise included in the audio signals
a3 and b3 from the noise characteristic adjustment unit 12 match each other, so even if one of
the audio signals a3 and b3 is switched to the signal to be transmitted to the other party, It is
possible to reduce the change in the texture of noise heard by the other party.
[0049]
Although only the frequency band corresponding to the noise is set to the frequency band BW to
03-05-2019
16
be adjusted, the frequency band of the speech may be included in the adjustment target. For
example, when the difference between the noise and the speech is not clear, the characteristics of
the speech and the noise are aligned with each other. This also makes it possible to reduce the
change in the texture of noise heard by the other party before and after switching.
[0050]
Further, the noise characteristic adjustment unit 12 may be capable of changing the setting of
the frequency band BW to be adjusted. For example, as shown in FIG. 6, after setting the
frequency band BW1 to be adjusted, only the frequency band BW2 in which the noise level
included in any of the audio signals a2 and b2 exceeds the threshold level Th Change to the
frequency band of By setting the above threshold level Th as the lower limit level that can be
heard by the other party, it is possible to select only the frequency band that requires adjustment
of the noise characteristic as the adjustment target, and the noise characteristic can be
appropriately aligned.
[0051]
In addition to road noise, there are pattern noise, cavity resonance noise, and the like as noise
generated as the vehicle travels. Pattern noise is noise generated by compression and release of
air in tire grooves, and is noise in a high frequency range more than road noise. In addition,
cavity resonance noise is noise generated by vibration of air filled in the inside of the tire, which
also becomes a dry sound with a high frequency range higher than road noise. Therefore, the
noise characteristic adjustment unit 12 sets the component of the higher frequency band
corresponding to the pattern noise and the cavity resonance sound in the frequency band BW to
be adjusted when the characteristics of noise other than these road noises are also equalized. By
doing this, it is possible to properly adjust the characteristics of noise according to the actual
noise environment.
[0052]
Furthermore, the noise characteristic adjustment unit 12 may be able to change the setting of the
frequency width RB. For example, in a traveling environment in which the level of road noise is
equal to or higher than the threshold, the frequency width RB is uniformly narrowed in order to
finely adjust the characteristics of the noise. On the other hand, in a traveling environment where
03-05-2019
17
the road noise level is less than the threshold, the frequency width RB is uniformly widened in
order to reduce the calculation load of the adjustment processing. By doing this, it is possible to
properly adjust the characteristics of the noise.
[0053]
The frequency width RB may be changed to a different width as well as to a fixed width. For
example, as shown in FIG. 7, the frequency width RB2 is changed from the frequency width RB2
to the frequency width RB1 to narrow the frequency width in the frequency band where the
variation of the noise level is larger than the predetermined threshold among the frequency
bands BW to be adjusted. . By doing this, the characteristics can be finely aligned in the portion
where the noise level fluctuates sharply, and the difference between the noise characteristics of
the audio signals a2 and b2 can be appropriately eliminated.
[0054]
Next, among the input signals of the microphones 5A to 5C whose noise characteristics have
been adjusted by the noise characteristic adjustment unit 12, the switching unit 13 transmits to
the other party the input signal of the microphone determined by the determination unit 11.
Switch to (step ST3). For example, when the microphones 5A and 5B are determined by the
determination unit 11, the audio signal a3 is output to the communication device 3, and when
the microphone 5C is determined by the determination unit 11, the audio signal b3 is output to
the communication device 3. Do. The communication device 3 transmits the audio signal input
from the switching unit 13 to the mobile phone 4A by near field communication. The mobile
phone 4A transmits an audio signal to the mobile phone 4B via the mobile phone line. Thus, the
other party can hear the voice of the speaker of the vehicle 20.
[0055]
As described above, the hands-free control device 2 according to the first embodiment has the
configuration shown in FIG. 1, and the characteristics of the noises included in the input signals
of the microphones 5A to 5C are matched for each frequency I'm adjusting. With such a
configuration, it is possible to reduce the change in the texture of noise heard by the other party
when the input signals of the microphones 5A to 5C to be transmitted to the other party's mobile
phone 4B are switched.
03-05-2019
18
[0056]
Further, in the hands-free control device 2 according to the first embodiment, the determination
unit 11 compares the signal levels obtained by subtracting the dark noise level from the signal
levels of the input signals of the microphones 5A to 5C to compare the speaker with the uttered
voice. Determine the input microphone. By comprising in this way, the misjudgment resulting
from the individual difference of microphones 5A-5C can be prevented.
[0057]
Furthermore, in the hands-free control device 2 according to the first embodiment, the noise
characteristic adjustment unit 12 adjusts the characteristic of the noise included in the input
signal for each frequency width RB predetermined in the frequency band BW to be adjusted. . By
doing this, even if the voice signal a3 or b3 is switched to a signal to be transmitted to the other
party, it is possible to reduce the change in the texture of noise heard by the other party before
and after the switching.
[0058]
Furthermore, in the hands-free control device 2 according to the first embodiment, the setting
target of the frequency band BW to be adjusted by the noise characteristic adjusting unit 12 can
be changed. By doing this, it is possible to properly adjust the characteristics of the noise.
[0059]
Furthermore, in the hands-free control device 2 according to the first embodiment, the setting of
the frequency width RB can be changed. By doing this, it is possible to properly adjust the
characteristics of the noise.
[0060]
03-05-2019
19
Furthermore, in the hands-free control device 2 according to the first embodiment, the noise
characteristic adjustment unit 12 sets the frequency width in the frequency band where the
variation of the noise level is larger than the predetermined threshold among the frequency
bands BW to be adjusted. Narrow RB. By doing this, the characteristics can be finely aligned in
the portion where the fluctuation of the noise level is large, and the difference between the noise
characteristics of the audio signals a2 and b2 can be appropriately eliminated.
[0061]
Furthermore, in the hands-free control device 2 according to the first embodiment, the
determination unit 11 determines that the signal level of the input signal of the microphone
determined that the speaker has input the speech voice is lower than the signal level of the input
signal of another microphone. After that, the other microphones are not distinguished as
microphones for which the speaker has input speech voice until a predetermined holding time
has passed. In this way, frequent switching of the signal to be transmitted to the other party is
prevented, and switching can be performed when it is determined that the speaker has changed.
[0062]
In the present invention, within the scope of the invention, modification of any component of the
embodiment or omission of any component of the embodiment is possible.
[0063]
The hands-free control device according to the present invention can reduce the change in the
texture of noise due to the switching of the input signal of the microphone, and thus is suitable
for a hands-free call system for vehicle use.
[0064]
DESCRIPTION OF SYMBOLS 1 hands-free call system, 2 hands-free control apparatus, 3
communication apparatus, 4A, 4B mobile telephone, 5A-5C microphone, 6 amplifier, 7A, 7B
speaker, 10A-10C audio processing unit, 11 determination unit, 12 noise characteristic
adjustment Unit, 13 switching units, 20 vehicles, 100 processing circuits, 101 CPU, 102
memories.
03-05-2019
20
Документ
Категория
Без категории
Просмотров
0
Размер файла
34 Кб
Теги
jpwo2017072958
1/--страниц
Пожаловаться на содержимое документа