close

Вход

Забыли?

вход по аккаунту

?

JP2007181099

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 JP2007181099
PROBLEM TO BE SOLVED: To reliably detect an utterance sound from a speaker on the own
apparatus side even in a situation where an utterance sound from the partner apparatus side is
emitted and a speaker on the own apparatus side speaks. The microphone beam selection circuit
8 detects that there is an emitted voice by detecting the input detection Hi data. In this situation,
when the microphone beam selection circuit 8 detects the output detection Hi data, the
microphone beam selection circuit 8 is in the search mode, and selects and outputs the collected
sound beam signal with the highest signal strength (S5 → S3 → S4). On the other hand, in the
same situation, when the microphone beam selection circuit 8 detects the output detection low
data, the microphone beam selection circuit 8 enters the hold mode, and continuously outputs
the sound collection beam signal currently selected (S5 → S6). Here, when the output detection
Hi data is detected again, the search mode is restored, and the collected sound beam signal
having the highest signal strength is selected and output (S6 → S1 → S2 → S5 → S3 → S4).
[Selected figure] Figure 1
Sound emission device
[0001]
According to the present invention, a sound emitting and collecting apparatus is used for voice
conferences between remote places and the like, and the voice of the user of the other apparatus
is emitted and the voice of the user of the own apparatus is collected. The invention relates to a
sound emission and collection device that selects and outputs the voice of the user who is
actually speaking on the side of the user's own device from the collected sound even in the
situation where the voice sound is emitted. is there.
04-05-2019
1
[0002]
As a voice communication system for conducting a voice conference (communication conference)
at a remote place, a sound emission and collection device integrally provided with a speaker and
a microphone is often used.
The sound emission and collection device converts the input voice signal based on the utterance
sound of the other party device into a voice and emits the sound from the speaker. In addition,
the sound emission and collection device collects an utterance sound of the own device side to
generate an output sound signal, and transmits it to the other device side. At this time, the sound
emitting and collecting apparatus has a function of detecting from which direction the collected
sound is propagated when there are a plurality of speakers on the apparatus side or when the
speakers move. The apparatus of Patent Document 1 has a function of detecting the direction of
the speaker, and the sound volume from the speaker corresponding to the microphone of the
detected speaker direction is reduced to be picked up by the microphone. It suppresses the
amount of wraparound sound from the speaker. Japanese Patent Application Laid-Open No. 1155784
[0003]
However, in the device of Patent Document 1, it is difficult for the user on the own device side to
listen to the utterance sound on the opposite device side by reducing the sound volume of the
speaker.
[0004]
Also, in a situation where the voice of the other party is emitted and the speaker of the own party
speaks, the wraparound sound from the speaker detects the direction of the louder voice in the
direction of the speaker. In some cases, it may not be possible to reliably detect the voice of the
speaker speaking.
[0005]
Therefore, according to the present invention, even in a situation where the voice from the
partner apparatus is emitted and the speaker on the own apparatus speaks, the voice from the
speaker on the own apparatus is reliably detected. The purpose is
[0006]
04-05-2019
2
The sound emission and collection device of the present invention controls the sound collection
signal of the microphone device to generate sound collection beam signals corresponding to a
plurality of different collection directivity, and the sound emission from the speaker device.
Whether or not to select and output the collected sound beam signal of the highest signal
strength from the plurality of collected sound beam signals based on the combination of the
signal strength of the input sound signal to be output and the signal strength of the output sound
signal of the regression sound removal means Sound collection beam selection means for
continuously outputting a sound collection beam signal at a current time point, generating a
pseudo-regression sound signal based on the input sound signal and subtracting it from the
selected sound collection beam signal to obtain an output sound signal It is characterized by
having the regressive sound removal means to obtain.
[0007]
In this configuration, the speaker device emits the voice of the user on the other party device,
that is, the input voice signal, and the microphone device picks up the voice of the user on the
own device side.
At this time, the wraparound sound is collected by the microphone device.
The sound collection control means generates a sound collection beam signal having sound
collection directivity in different directions from the sound collected by the microphone device.
[0008]
The sound collection beam selection means switches the method of selecting the sound collection
beam signal based on a combination of the signal strength of the input sound signal and the
signal strength of the output sound signal from which the influence of the wraparound sound has
been removed by the regressive sound removal means.
For example, when there is no input audio signal or low signal strength, the selection method is
switched depending on whether there is an output audio signal and high signal strength or no
output audio signal or weak signal strength.
04-05-2019
3
Also, when there is an input audio signal and the signal strength is high, the selection method is
switched depending on whether the output audio signal is high and the signal strength is high or
not or the signal strength is low. Thus, the method of selecting the sound collection beam signal
is set according to each situation. Then, an optimal sound collection beam signal is selected and
output.
[0009]
The regression sound removal means subtracts the pseudo regression sound signal from the
sound collection beam signal output from the sound collection beam selection means to obtain
an output sound signal. Since this output sound signal is the result of removing the influence of
the wraparound sound on the collected sound beam signal, that is, the sound collection beam
signal from which the sound component based on the input sound signal is removed, Based on
the vocal sound, this output voice signal is transmitted to the other device. Then, the intensity
information of the output sound signal is fed back to the sound collection beam selection means.
[0010]
With such a configuration, the method of selecting the collected sound beam signal is switched
based on whether there is a voice from the other party device to be emitted from the speaker and
whether there is an utterance from the user on the own device side. A sound collection beam
signal corresponding to a vocal sound from the user on the apparatus side which is not a
wraparound sound from the speaker is appropriately selected, and the regressive sound is
removed and transmitted.
[0011]
Further, the sound collection beam selection means of the sound collection and collection device
of the present invention selects the output sound signal when the input sound signal is equal to
or more than the first threshold and selects and outputs a predetermined sound collection beam
signal. When it becomes less than 2 threshold values, the new selection process of the sound
collection beam signal is not performed, and the sound collection beam signal selected at the
time is continuously output.
[0012]
04-05-2019
4
In this configuration, specifically, in the situation where the sound collection beam signal whose
directivity axis is directed to the user on the apparatus side already in the predetermined
direction is selected and the input voice signal is strong, that is, wraparound voice When the
amount is large, even if the output sound signal based on the currently selected sound collection
beam signal becomes weak, the sound collection beam signal currently selected with the
directivity axis oriented in the direction in which the user of the local apparatus is present.
Continue to output as it is.
As a result, the user's own apparatus's voice can be continuously picked up without being
affected by the wraparound sound from the speaker.
[0013]
Further, the sound collection beam selection means of the sound collection and collection device
of the present invention stops the continuation process when detecting that the output sound
signal becomes equal to or more than the second threshold while continuously outputting a
predetermined sound collection beam. The present invention is characterized in that the collected
sound beam signal having the highest signal strength is selected from the plurality of collected
sound beam signals.
[0014]
In this configuration, when the signal strength of the output voice signal becomes greater than or
equal to the second threshold, it is determined that a new utterance sound has appeared in any
direction including the direction currently selected, and a plurality of collected sounds are
collected again. The collected sound beam signal with the highest signal strength is selected from
the beam signals.
As a result, when there are a plurality of users on the apparatus side and in different directions
with respect to the apparatus, if the speaking user is switched, a collected sound beam signal
with the directivity axis oriented in the direction of the user after this switching Is selected.
That is, even if the user switches, it is possible to reliably pick up a new user's voice and transmit
it to the other device. In addition, also when the user who is speaking moves, similarly, the sound
collection beam signal in which the axis of directivity is oriented in the new sound source
direction is selected.
04-05-2019
5
[0015]
Also, the sound emission and collection device of the present invention includes a low pass filter
that passes only the low frequency band of the output sound signal, and the sound collection
beam selection unit collects the sound based on the output sound signal consisting of only low
frequency components. It is characterized in that sound beam signal selection is performed.
[0016]
In this configuration, based on the fact that noise in the low frequency band is more reliably
eliminated in the low frequency band than in the high frequency band, the output sound signal
consisting only of low frequency components is used by the regression sound removal means.
Such an output voice signal is more purely due to the vocal sound from the user on the apparatus
side. As a result, the presence or absence of a vocal sound from the user on the apparatus side
can be detected more accurately.
[0017]
Further, in the sound emission and collection device according to the present invention, the
microphone device is formed of a microphone array in which a plurality of microphones are
arranged in a predetermined pattern, the speaker device is formed of a speaker array in which a
plurality of speakers are arranged in a predetermined pattern. It is characterized in that it is
integrally formed with the speaker device.
[0018]
In this configuration, as a specific example of the microphone device and the speaker device, the
microphone device is configured of a microphone array, and the speaker device is configured of a
speaker array.
[0019]
According to the present invention, in the situation where the voice of the user of the other
party's device is being emitted from the speaker device, and the volume of the voice of the user
04-05-2019
6
of the other device's side becomes lower, By holding in the direction of the user on the self
apparatus side, it is possible to reliably continue to pick up and transmit the voice of the user on
the self apparatus side.
Furthermore, in the situation where the voice of the other party's device user is being emitted
and the other user on the own device speaks or the user moves, etc., the sound source direction
is changed, this new sound source Utterances from directions can be reliably picked up and
transmitted.
That is, even under the situation where the voice from the other party's device is being emitted, it
is possible to reliably pick up and transmit the voices from each user on the own device's side.
[0020]
A sound emission and collection device according to an embodiment of the present invention will
be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration
of the main part of the sound emission and collection device of the present embodiment.
Moreover, FIG. 2 is a front view of the sound emission and collection device of this embodiment.
In the present embodiment, the number of speakers and the number of microphones are each
eight. However, the present invention is not limited to this, and the respective numbers may be
set as appropriate. Of course, the number of speakers and the number of microphones do not
necessarily match. good.
[0021]
As shown in FIG. 1, the sound emission and collection device of this embodiment includes an
input / output I / F 1, a speaker signal processing circuit 2, D / A converters 31A to 31H, output
amplifiers 32A to 32H, speakers 4A to 4H, and a microphone 5A. To 5H, input amplifiers 61A to
61H, A / D converters 62A to 62H, microphone beam generation circuit 7, microphone beam
selection circuit 8, echo canceller 9, output level detection circuit 10, input level detection circuit
11, and input / output connector 12 Prepare.
[0022]
The housing 45 has a rectangular parallelepiped shape, and the speakers 4A to 4H and the
04-05-2019
7
microphones 5A to 5H are arrayed and installed along the long side direction to form a speaker
array and a microphone array.
At this time, the speakers 4A to 4H and the microphones 5A to 5H are disposed at substantially
equal intervals, and one speaker and one microphone (for example, the speaker 4A and the
microphone 5A) are arranged along the short side direction of the arrangement surface. Are
arranged at predetermined intervals. Further, the portion excluding the input / output connector
12 of the sound emission and collection device shown in FIG. 1 is installed in the housing 45, and
the input / output connector 12 is formed on any wall of the housing 45 (shown in FIG. ). The
housing 45 is installed such that, for example, the long side direction is horizontal and the short
side direction is vertical.
[0023]
The input / output connector 12 includes a LAN connection unit, an analog audio signal input /
output unit, a digital audio signal input / output unit, and the like. When a voice communication
conference is performed between remote locations, the LAN connection unit is connected to the
other side of the sound emission and collection device (other device) via the network.
[0024]
The input / output I / F 1 receives an input audio signal from the input / output connector 12
and supplies the signal to the speaker signal processing circuit 2 via the echo canceller 9.
Further, the input / output I / F 1 transmits an output sound signal output from the echo
canceller 9 through the input / output connector 12.
[0025]
The input level detection circuit 11 monitors an input audio signal from the input / output I / F 1
and outputs input detection data to the microphone beam selection circuit 8. More specifically,
the input level detection circuit 11 includes a peak hold circuit, and “Hi” data (input detection
Hi) indicating that the input audio signal is present when the input audio signal becomes equal to
or greater than a preset first threshold. It is called data. ) Is output to the microphone beam
selection circuit 8. On the other hand, if the input audio signal is less than the first threshold, the
04-05-2019
8
input level detection circuit 11 indicates “Low” data (no input detection Low data) indicating
that there is no input audio signal. ) Is output to the microphone beam selection circuit 8.
[0026]
The speaker signal processing circuit 2 performs different delay processing and amplitude
processing on the input sound signal based on the sound emission environment setting from the
control unit (not shown), and gives the same to the D / A converters 31A to 31H. . Here, the
sound emission environment is set from the arrangement of the speakers 4A to 4H, the room in
which the sound emission and collection device is installed, the azimuth between the sound
emission and collection device and each user, the distance relationship, and the like. The D / A
converters 31A to 31H perform analog conversion on the input audio signals subjected to the
delay / amplitude processing and input the same to the output amplifiers 32A to 32H. The
output amplifiers 32A to 32H respectively amplify the audio signals and give them to the
speakers 4A to 4H, and the speakers 4A to 4H convert the given audio signals into audio and the
exterior, that is, the sound emitting and collecting apparatus is installed. Sounds indoors. Here,
since the delay processing and the amplitude processing based on the above-mentioned sound
emission environment setting are performed, the sound emission voices from the respective
speakers 4A to 4H of the speaker array have predetermined sound emission directivity in the
horizontal direction. Given.
[0027]
The microphones 5A to 5H constituting the microphone array pick up the voices uttered by the
user in the room and give them to the input amplifiers 61A to 61H. The input amplifiers 61A to
61H amplify the collected sound signals and apply them to the A / D converters 62A to 62H,
respectively, and the A / D converters 62A to 62H convert the collected sound signals to digital
and give the microphone beam generation circuit 7 .
[0028]
The microphone beam generation circuit 7 performs a delay-sum process or the like on the
collected sound signals collected by the microphones 2A to 2H to obtain a collected sound beam
signal having sound collection directivity in which the direction different from each other is the
directivity axial direction. Generate MB1 to MB6. In the present embodiment, an example of
04-05-2019
9
generating six sound collection beam signals MB1 to MB6 is shown, but the number of beams
may be set appropriately according to the specification.
[0029]
The microphone beam selection circuit 8 receives the sound collection beam signals MB1 to
MB6, acquires input detection data from the input level detection circuit 11, and acquires output
detection data from an output level detection circuit 10 described later.
[0030]
The microphone beam selection circuit 8 determines the selection method from the relationship
between “Hi” and “Low” of the input detection data and the output detection data by a
method described later, selects the sound collection beam signal as appropriate, and Output to
[0031]
The echo canceller 9 includes an adaptive filter 91 and a post processor 92.
The adaptive filter 91 includes a digital filter such as an FIR filter, estimates a transfer function of
an acoustic propagation path from the speaker array to the microphone array, generates a
pseudo-regression sound signal, and provides the post-processor 92.
The post processor 92 subtracts the pseudo-regression sound signal from the sound collection
beam signal output from the microphone beam selection circuit 8 to generate an output sound
signal, and outputs the output sound signal to the input / output I / F 1. Here, the signal output
from the post processor 92 is fed back to the adaptive filter 91 as a reference signal. Then, the
pseudo-regression sound signal is optimized by using an adaptive algorithm based on the input
speech signal using the feedback reference signal. This makes it possible to suppress the
influence of the wraparound sound from the speaker array to the microphone array, so that the
output sound signal is substantially controlled by the voice from the user who is present in the
room where the sound emitting and collecting apparatus is installed. Ru.
[0032]
04-05-2019
10
The output level detection circuit 10 includes a level detection unit 101 and a low pass filter 102,
monitors an output voice signal, and outputs output detection data to the microphone beam
selection circuit 8. More specifically, the low pass filter 102 of the output level detection circuit
10 attenuates high frequency components of the output sound signal and supplies the attenuated
signal to the level detection unit 101. By attenuating the high frequency band in this way, it is
possible to remove the wraparound voice component in the high frequency band that is difficult
to remove by the echo canceller 9, so that it is possible to obtain a signal with clearer voice
presence or absence. This signal is input to the level detection unit 101.
[0033]
Although the low pass filter 102 may or may not be installed, the following effects are also
obtained.
[0034]
FIG. 3 is a diagram showing the directivity width at each frequency, and a real thin line shows the
directivity width of an audio signal of about 100 Hz, and a broken line shows the directivity
width of an audio signal of about 1000 Hz (1 kHz). A thick line indicates the directivity width of
an audio signal of about 8000 Hz (8 kHz).
As shown in FIG. 3, since the directivity width is wider and the change in attenuation of the
sound volume at the time of sound collection is smaller as the audio signal in the low frequency
band is smaller, even when the sound source direction to be described later is switched The
output sound signal from the sound source direction can be uniformly detected without
attenuation.
[0035]
The level detection unit 101 includes a peak hold circuit, and “Hi” data (referred to as output
detection Hi data) indicating that there is an output sound signal when the reference signal based
on the output sound signal becomes equal to or greater than a preset second threshold. . ) Is
output to the microphone beam selection circuit 8. On the other hand, if the reference signal
based on the output sound signal is less than the second threshold, the level detection unit 101
indicates “Low” data (output detection low data) indicating that there is no output sound
signal. ) Is output to the microphone beam selection circuit 8.
04-05-2019
11
[0036]
Next, the method of selecting the collected sound beam signal by the microphone beam selection
circuit 8 will be described more specifically.
[0037]
FIG. 4 shows a sound collection beam signal selection condition table stored in the microphone
beam selection circuit 8.
As shown in FIG. 4, the microphone beam selection circuit 8 stores in advance a method of
selecting a sound collection beam signal according to a combination of input detection data and
output detection data. The selection method includes a search mode and a hold mode. The search
mode compares the signal strengths of the respective collected beam signals MB1 to MB6 to
select a collected beam signal having the highest signal strength. In the hold mode, when the
hold mode is switched to, the sound collection beam signal which has already been selected is
continuously output as it is. (1) In the case of a combination of input detection low data and
output detection low data This is a case where there is neither an input sound signal nor an
output sound signal or a low signal strength, and there is no sound emission by the speakers 4A
to 4H Is low and the microphones 5A to 5H do not pick up sound or the sound volume is low. In
this case, the search mode is executed.
[0038]
(2) In the case of a combination of input detection Low data and output detection Hi data This is
a case where there is no input audio signal or low signal strength, and there is an output audio
signal and high signal strength. This corresponds to the case where there is no sound or the
discharge volume is low and the sounds collected by the microphones 5A to 5H are equal to or
higher than the predetermined volume, and the search mode is also executed in this case.
[0039]
(3) In the case of a combination of input detection Hi data and output detection Low data This is
a case where there is an input audio signal, a high signal strength, and no output audio signal or
a low signal strength. This corresponds to the case where the sound volume is equal to or higher
than the predetermined volume and the microphones 5A to 5H do not pick up sound or the voice
collection volume is low. In this case, the hold mode is executed.
04-05-2019
12
[0040]
(4) In the case of the combination of the input detection Hi data and the output detection Hi data
This is a case where both the input audio signal and the output audio signal are present and the
signal strength is high. This corresponds to the case where the sound collection by ~ 5H has a
predetermined sound volume or more, and in this case, the search mode is executed.
[0041]
Next, processing of the microphone beam selection circuit 8 in various situations will be
described with reference to a flowchart shown in FIG. 5 and an explanatory view shown in FIG.
[0042]
FIG. 5 shows a flow chart showing selection processing of the microphone beam selection circuit
8.
[0043]
FIG. 6 is an explanatory view showing a selected state of the collected sound beam signal.
In FIG. 6, reference numerals 501 and 502 denote users (speakers) who are present in the front
direction of the case 45 of the sound emission and collection apparatus on the side of the own
apparatus, 511 denotes the voice of the user 501, 512 denotes the voice of the user 502, MB1
denotes a sound collection beam whose directivity axis is in the direction of the user 501, MB2
denotes a sound collection beam whose directivity axis is in the direction of the user 502, and
700 denotes emitted sound.
[0044]
FIG. 6A shows a situation in which there is no emitted voice and the user 501 is speaking, and
FIG. 6B shows a situation in which there is no emitted voice and the user 502 is speaking.
Further, FIG. 6 (C) shows a situation where the user 501 is speaking and there is an emitted
04-05-2019
13
voice, and (D) shows a situation where the user 501 is speaking at a low volume and there is an
emitted voice and E) shows a situation where the user 502 has an utterance and has a voice to be
emitted.
[0045]
The microphone beam selection circuit 8 acquires input detection data and output detection data
at predetermined intervals, for example, at sampling timings (S1).
Here, when the input detection low data is detected, the search mode is executed.
That is, the microphone beam selection circuit 8 compares the signal strengths of the collected
sound beam signals MB1 to MB6, and selects and outputs the collected sound beam signal
having the highest signal strength (S2 → S3 → S4).
For example, as shown in FIG. 6A, when the user 501 utters the vocalization sound 511 in a
situation where there is no speech from the counterpart apparatus side user and there is no
emitted voice, the sound collection beam corresponding to the vocalization sound 511 The signal
strength of MB1 will be higher than other sound collection beams. The microphone beam
selection circuit 8 acquires the respective collected sound beam signals MB1 to MB6, selects the
collected sound beam signal MB1 having the highest signal strength, and outputs the selected
sound collection beam signal MB1 to the echo canceller 9. Then, as shown in FIG. 6B, when the
user 501 ends the speech and the user 502 utters the vocal sound 512 instead, the signal
strength of the sound collection beam MB2 corresponding to the vocal sound 512 is another
collection. It is higher than the sound beam. The microphone beam selection circuit 8 acquires
each of the sound collection beam signals MB1 to MB6, selects the sound collection beam signal
MB2 having the highest signal strength, and outputs the selected sound collection beam signal
MB2 to the echo canceller 9. Even when there is no such an emitted voice and no collected voice,
the microphone beam selection circuit 8 executes the search mode at predetermined intervals,
for example, at each sampling timing.
[0046]
When sound is emitted from the speaker array, the microphone beam selection circuit 8 detects
04-05-2019
14
input detection Hi data. Then, when there is an utterance by the user on the own apparatus side
and the output detection Hi data is detected, the microphone beam selection circuit 8 executes
the search mode and the highest signal strength as in the case where the input detection Low
data is detected. The selected sound collecting beam signal is selected and output (S 2 → S 5 → S
3 → S 4). For example, as shown in FIG. 6C, when the user 501 utters the utterance 511 in the
presence of the emitted sound 700, the echo canceller 9 removes the wraparound component
due to the emitted sound 700, so the output The speech signal is mainly composed of
components based on the vocal sound 511. The output level detection circuit 10 outputs the
output detection Hi data to the microphone beam selection circuit 8 when detecting that the
signal based on the output audio signal is equal to or more than the second threshold value. The
microphone beam selection circuit 8 acquires the respective collected beam signals MB1 to MB6,
selects the collected beam signal MB1 having the highest signal strength, with the direction of
the user 501 as the axial direction of directivity, and outputs it to the echo canceller 9 Do.
[0047]
On the other hand, when there is sound emission from the speaker array and the voiced volume
of the user on the apparatus side decreases, that is, mainly when the user on the partner
apparatus side speaks, the microphone beam selection circuit 8 outputs output detection low
data To detect and execute the hold mode. That is, the microphone beam selection circuit 8
continues to output the sound collection beam signal MB already selected at that time (S5 → S6).
For example, as shown in FIG. 6 (C) to FIG. 6 (D), when the amount of generation of the user 501
on the own apparatus changes little under the situation where the emitted sound 700 is present,
the vocal sound 521 is less than the second threshold. It becomes. Then, the output level
detection circuit 10 detects the output detection low data and outputs it to the microphone beam
selection circuit 8. The microphone beam selection circuit 8 continuously outputs the sound
collection beam signal MB1 selected at present to the echo cancellation unit 9.
[0048]
By performing such processing, when the generation amount of the user on the apparatus side is
small and the emitted sound 700 is large, the search mode is performed to select the collected
beam signal that is strongly influenced by the emitted sound 700. It is possible to prevent the
problem that the sound collection directivity is enhanced in a direction completely unrelated to
the position of the user on the local apparatus side. As a result, even if the speech volume of the
user on the own apparatus side is smaller than the sound emission sound volume, the vocal
sound of the user on the own apparatus side can be reliably collected. Then, by performing such
04-05-2019
15
processing, it is possible to immediately perform sound collection with the optimum sound
collection directivity, when the same user on the own apparatus side speaks again after this.
[0049]
Next, when there is sound emission from the speaker array and another user on the own device
cuts in and speaks at a predetermined volume or higher from the situation where the user's own
utterance volume is small, the output detection Hi data is detected Ru. The microphone beam
selection circuit 8 stops the hold mode and executes the search mode because the output
detection data has shifted from low to high. Thereby, the microphone beam selection circuit 8
selects and outputs the collected sound beam signal of the highest signal strength (S6 → S1 →
S2 → S5 → S3 → S4). For example, as shown in FIG. 6 (D) to FIG. 6 (E), when the user 502
interrupts and speaks from a situation where the generation amount of the user 501 on the own
apparatus side is small in a situation where there is a sound emission 700 The voice sound 512
becomes equal to or higher than the second threshold, and the output level detection circuit 10
detects the output detection Hi data and outputs it to the microphone beam selection circuit 8.
The microphone beam selection circuit 8 acquires each of the sound collection beam signals MB1
to MB6, selects the sound collection beam signal MB2 having the highest signal strength, and
outputs the selected sound collection beam signal MB2 to the echo canceller 9.
[0050]
By performing such processing, it is possible to select a sound collection beam signal having an
axial direction of directivity in the direction of the user who is speaking anew, even when the
user on the apparatus side who speaks is switched. By switching, it is possible to prevent the
problem that voices from new users can not be collected efficiently.
[0051]
As described above, by using the sound emission and collection device according to the present
embodiment, it is possible to reliably collect a message from the user on the own device side
without being affected by the sound emission situation, and the user is switched. Even in this
case, the axial direction of directivity can be immediately set to a new user direction, and sound
can be collected effectively.
[0052]
In the present embodiment, the case where two users are present in front of the own apparatus
and speak without moving is described, but the above configuration may be applied even when
04-05-2019
16
the user, that is, the sound source moves. it can.
[0053]
Further, in the present embodiment, a speaker array in which a plurality of speakers are
arranged is shown as an example, but the above configuration can be applied even if a single
speaker device is used.
[0054]
It is a block diagram which shows the structure of the principal part of the sound emission and
collection apparatus of this invention.
It is a front view of the sound emission and collection device of the present invention.
It is a figure which shows the directivity width | variety in every frequency.
It is a sound collection beam signal selection condition table stored in the microphone beam
selection circuit 8.
It is a flowchart which shows the selection process of a microphone beam selection circuit. It is
explanatory drawing which shows the selection state of a sound collection beam signal.
Explanation of sign
[0055]
1-input / output I / F, 2-speaker signal processing circuit, 31A-31H-D / A converter, 32A-32Houtput amplifier, 4A-4H-speaker, 5A-5H-microphone, 61A-61H-input amplifier, 62A to 62H-A / D
converter, 7-microphone beam generation circuit, 8-microphone beam selection circuit, 9-echo
canceller, 91-adaptive filter, 92-post processor, 10-output level detection circuit, 101-level
detection Part, 102-low pass filter, 11-input level detection circuit, 12-input / output connector,
501, 502-user, 511, 512-vocal sound, 700-voice emission voice
04-05-2019
17
04-05-2019
18
Документ
Категория
Без категории
Просмотров
0
Размер файла
30 Кб
Теги
jp2007181099
1/--страниц
Пожаловаться на содержимое документа