close

Вход

Забыли?

вход по аккаунту

?

JP2018019413

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 JP2018019413
Abstract: The present invention discloses a method, apparatus and hands-free calling device for
improving the call quality of a hands-free calling device. A hands-free calling device includes a
transmit end comprising one main microphone and at least one auxiliary microphone, the
method comprising the steps of: scanning within an initial first collection angle of the transmit
end After the audio feature signal is scanned within the first collection angle, determining a
second collection angle less than the first collection angle within the first collection angle
according to the direction of the audio feature signal; Calibrating the transmit end in the
direction determined by the second collection angle. The method can achieve the purpose of
significantly reducing the interference of ambient noise and improving the signal-to-noise ratio of
the transmission by performing voice pickup using a relatively small voice protection angle. The
voice during the call of the hands-free device has more directivity, and the quality of the call is
improved. [Selected figure] Figure 1
Method, device and hands-free calling device for improving call quality of hands-free calling
device
[0001]
The present invention relates to the field of hands-free calling devices, and more particularly, to a
method, apparatus and hands-free calling devices for improving call quality of hands-free calling
devices.
[0002]
03-05-2019
1
Hands-free calling devices in the prior art, for example smart arm-mounted devices such as smart
watches, smart bracelets, etc., when in hands-free calling mode, because of the relative position
uncertainty between the smart watch and the user's mouth, Usually, a relatively large voice
protection angle is used, and at the same time as picking up the voice, a large amount of
environmental noise is collected.
As a result, it affects the signal-to-noise ratio of the transmission, and at the same time, the sound
emitted from the speaker of the hands-free device such as a smart watch can not only be heard
by the caller but also other people in the vicinity It also makes it easier to leak relatively private
information. In short, due to the above-mentioned defects of the transmitting end and the
receiving end, the call quality of the hands free calling device is not high in the prior art.
[0003]
The present invention is a method, apparatus and method for improving the call quality of handsfree speech equipment in order to solve the problem that much environmental noise is collected
and the signal-to-noise ratio of transmission is low when the hands-free speech equipment talks.
Provide hands-free calling equipment.
[0004]
According to one aspect of the present invention, there is provided a method of improving the
call quality of a hands-free handset including a transmit end comprising one main microphone
and at least one auxiliary microphone, the method comprising: Performing a scan within an initial
first collection angle of the speaking end; and after the audio feature signal is scanned within the
first collection angle, the first within the first collection angle according to the direction of the
audio feature signal. Determining a second collection angle less than one collection angle, and
calibrating the transmit end in a direction determined by the second collection angle.
[0005]
Here, after the audio feature signal is scanned within the first collection angle, determining a
second collection angle smaller than the initial collection angle within the first collection angle
according to the direction of the audio feature signal: After being scanned for voice feature
signals in the γ1 direction within the initial first collection angle of the transmit end, draw a
reverse extension through the main microphone along the γ1 direction and center the main
microphone Draw a circle whose radius is the connecting line between the main microphone and
any one auxiliary microphone, and determine the position where the arc and the reverse
extension line intersect with the virtual microphone of the main microphone, the main
microphone and the main microphone The microphone virtual microphone is used as a new
03-05-2019
2
voice array, and an angle β1 smaller than the first collection angle is defined, and it is
determined in real time whether or not there is an audio feature signal within the angle β1, If
there is an audio feature signal, check the angle β1 as the second collection angle, and if there is
no audio feature signal, use the connecting line between the main microphone and any one
auxiliary microphone as the symmetry axis, Defining the mirror angle β2 with respect to the axis
of symmetry, and verifying the angle β2 with the second collection angle.
[0006]
Here, it is determined in real time whether or not there is a voice feature signal within the angle
β1 if envelope energy detection is performed on the voice feature signal in the γ1 direction,
and the energy detection value is larger than a first predetermined threshold Performing the zero
crossing rate detection on the voice feature signal in the γ1 direction, and if the voice feature
signal is within the angle β1 when the zero crossing rate of the voice feature signal in the γ1
direction reaches a second predetermined threshold Including confirming and.
[0007]
Here, performing envelope energy detection on the voice feature signal in the γ1 direction
includes performing envelope energy detection according to the following equation.
Here, power is the energy value of the voice feature signal, the parameter alpha is a weighting
coefficient, and the parameter N is a voice feature signal at a certain point in time.
Performing zero crossing rate detection on the voice feature signal in the γ1 direction includes
performing zero crossing detection according to the following equation.
Here, Z_rate is the zero crossing rate of the speech feature signal, n is one value in the discrete
time series, and sgn [x] satisfies the following equation.
[0008]
Here, the hands-free device further includes a receiving end including at least one speaker, and
the method includes providing a virtual speaker of one speaker at the receiving end, and a
connecting line between the speaker and the virtual speaker is γ1. The method may further
comprise the step of pointing the direction and defining a third collection angle pointing the
03-05-2019
3
direction determined by the second collection angle.
[0009]
According to another aspect of the present invention, there is provided a device for improving
the call quality of a hands-free calling device, wherein the device scans within the initial first
collection angle of the transmit end and An audio feature determination for determining a second
collection angle less than the first collection angle within the first collection angle according to
the direction of the audio feature signal after the audio feature signal is scanned within one
collection angle Means for calibrating the transmitting end in a direction determined by the
second collection angle.
[0010]
The device is scanned for voice feature signals in the γ1 direction within the initial first
collection angle of the transmit end, then draws a reverse extension through the main
microphone along the γ1 direction and A virtual circle to draw a circle whose center is the
connecting line between the main microphone and one of the auxiliary microphones with the
microphone as the center of the circle, and to determine the position where the arc and the
reverse extension intersect the virtual microphone of the main microphone The microphone
creation means and the virtual microphones of the main microphone and the main microphone
are used as a new voice array, and an angle β1 smaller than the first collection angle is defined,
and whether there is a voice feature signal in the angle β1 in real time If there is a voice feature
signal, check the angle β1 as the second collection angle, and if there is no voice feature signal,
the main microphone and any one auxiliary microphone As a symmetrical axis connecting lines,
define the mirror angle β2 with respect to the axis of symmetry of the angle .beta.1, further
comprising an angle confirmation means for confirming the angle β2 and a second collection
angle, the.
[0011]
Here, the angle check means further performs envelope energy detection on the voice feature
signal in the γ1 direction, and if the energy detection value is larger than a first predetermined
threshold, the zero crossing rate for the voice feature signal in the γ1 direction The detection is
performed, and is used to confirm that the voice feature signal is within the angle β1 when the
zero crossing rate of the voice feature signal in the γ1 direction reaches a second predetermined
threshold.
[0012]
03-05-2019
4
Here, the angle confirmation means includes an envelope detection means and a zero crossing
detection means, and the envelope detection means is used to perform envelope energy detection
on the audio feature signal in the γ1 direction by the following equation.
Here, power is the energy value of the voice feature signal, the parameter alpha is a weighting
coefficient, and N is the voice feature signal at a certain point in time.
The zero crossing detection means is used to perform zero crossing rate detection on audio
feature signals in the γ1 direction according to the following equation.
Here, Z_rate is the zero crossing rate of the speech feature signal, n is one value in the discrete
time series, and sgn [x] satisfies the following equation.
[0013]
The apparatus is provided with a virtual speaker of one speaker at the receiving end so that the
connecting line between the speaker and the virtual speaker points in the γ1 direction, and the
third collecting angle is included in it. The receiver further comprises earpiece positioning means
for defining a third collection angle including the direction determined by the second collection
angle.
[0014]
A method and apparatus for improving the speech quality of a hands-free speech device such as
the present invention performs speech calibration and positioning with respect to the speech end
to ensure that the speech pickup angle of the speech end is within a relatively precise range. By
throttling, the voice signal picked up at the transmitting end contains a lot of environmental
noise, and it is avoided that the signal-to-noise ratio of the voice is affected, and the call quality of
the hands-free handset is not high. Solve the problem and improve the call quality of the handsfree device.
[0015]
Other features and advantages of the present application are described in the following
specification, some of which are apparent from the specification, or are understood by practicing
03-05-2019
5
the present application.
The objectives and other advantages of the present application can be realized and obtained by
the structure specified in the description, claims and drawings.
[0016]
The drawings form a part of the specification and, while providing a further understanding of the
invention, are used to interpret the invention in conjunction with embodiments of the invention
and are not a limitation on the invention.
[0017]
5 is a flow chart of a method for improving the call quality of a handsfree calling device
according to one embodiment of the present invention.
FIG. 1 is a schematic diagram of a principle of a method for improving the call quality of a handsfree device according to one embodiment of the present invention.
FIG. 5 is a schematic diagram of calibrating a transmit end according to one embodiment of the
present invention.
FIG. 5 is a schematic diagram of calibrating a transmit end according to one embodiment of the
present invention.
5 is a flow chart of determining an audio feature signal according to one embodiment of the
present invention. FIG. 5 is a schematic view of calibrating a receiving end according to one
embodiment of the present invention. FIG. 5 is a block diagram of an apparatus for improving
call quality of a hands-free calling device according to one embodiment of the present invention.
[0018]
The core idea of the present invention is to track the direction of the sound source in real time by
the microphone array sound source positioning technology and to determine the sound
03-05-2019
6
protection angle smaller than the initial sound protection angle to perform sound pickup for the
ambient environment noise. The purpose of greatly reducing interference and improving the
signal-to-noise ratio of transmission can be achieved, and at the same time, by using direction
information of the sound source to perform direction compensation on the speaker array, These
sounds just point to the sound source, and the privacy of the reception improves.
[0019]
FIG. 1 is a flow chart of a method for improving the call quality of a hands-free handset including
a transmit end comprising one main microphone and at least one auxiliary microphone according
to one embodiment of the present invention.
[0020]
Referring to FIG. 1, the method comprises the steps of: S110 for scanning within an initial first
collection angle of the transmit end; and after the audio feature signal is scanned within the first
collection angle, the audio feature Step S120 of determining a second collection angle less than
the first collection angle within the first collection angle according to the direction of the signal,
and calibrating the transmit end in the direction determined by the second collection angle And
step S130.
[0021]
According to the above steps, since the voice protection angle of the transmitting end is
narrowed within a relatively small angle and the voice pickup is performed within the relatively
small angle determined again, interference of environmental noise with the voice signal is
avoided. The signal-to-noise ratio of the transmission is improved, and further, the call quality of
the hands-free device is improved.
[0022]
FIG. 2 is a diagram showing the principle of the method for improving the call quality of the
hands-free device according to one embodiment of the present invention.
Referring to FIG. 2, the hands-free communication device 1 includes a transmitting end and a
receiving end, the transmitting end includes a main microphone MIC-a and an auxiliary
microphone MIC-b, and the receiving end is a main speaker SPK. Is the initial scan angle of the
transmit end, and when communicating normally, the transmit end collects signals, first
03-05-2019
7
according to a relatively large collection angle α After a scan is performed and the audio feature
signal is scanned within the angle α, the collection angle is narrowed to the angle β, thus
achieving the goal of transmit end audio positioning.
As a process of transmitting end voice positioning, when it is determined that the transmitting
end scans the audio feature signal within the initial scan angle α and it is determined that the
audio feature signal is within β angle, the transmitting end is Position in the direction
determined by the angle, and the direction determined by the angle becomes the user's uttered
position or orientation (as shown in FIG. 2, the open range of the angle β covers the position of
the person's head just Positioning relative to the human head is more accurate than the angle α).
The role of the method according to the invention is to narrow the initial scan angle to a
relatively small second collection angle β, so that the voice protection angle is relatively small,
and the user's spoken position is more accurately located, The environmental noise in the audio
signal can be reduced, and the audio quality can be improved.
[0023]
FIG. 3a is a schematic diagram of calibrating a transmit end according to one embodiment of the
present invention.
Referring to FIG. 3a, reference numeral 1 represents a hands-free device, 2 represents an
auxiliary speaker, 3 represents an auxiliary microphone, and 4 represents a virtual microphone
of a main microphone.
[0024]
A specific implementation scheme for improving hands free handsets according to the present
invention is as follows.
[0025]
Referring to FIG. 3a, when communicating, the transmit end first performs an audio scan at a
relatively large first collection angle α and if there is an audio feature signal in the γ1 direction
within the first collection angle α After that, the virtual microphone is a reverse extension
03-05-2019
8
passing through the γ1 direction of the main microphone MIC-a, and the connection line
between the main microphone MIC-a and the auxiliary MIC-b with the main microphone MIC-a as
a circle. A virtual microphone MIC-c located in an arc as a radius is provided, and a new array is
configured by the main microphone MIC-a and the virtual microphone MIC-c, and the directivity
angle of the new array is very small. The angle β1 points to a relatively small range in the γ1
direction, and the value of β1 is specifically selected according to the different application
situations.
[0026]
Since there is no auxiliary array element, the sound direction comes from the left side (γ1) of
the connecting line between the main microphone MIC-a and the auxiliary MIC-b by the array
consisting of two dot array elements (MIC-a and MIC-c) It is difficult to distinguish whether it is
coming from the right side (γ2).
One of the core contents of the method according to the present invention is to make a judgment
by the virtual microphone MIC-c. As a method of judgment, first, the virtual microphone MIC-c is
provided in the γ1 direction in FIG. When β1 is formed, the voice feature signal within the β1
pointing angle is determined in real time, and the voice feature signal is found, the pointing angle
β1 is determined as the second collection angle of the transmitting end, and the transmitting
end is Calibrate in the direction determined by the second collection angle (β1).
If the voice characteristic signal is not found, a virtual microphone MIC-c is newly provided on
the opposite side of the mirror of the connection line between the main microphone MIC-a and
the auxiliary MIC-b, and the voice direction is judged.
[0027]
FIG. 3b is a schematic diagram of calibrating the transmitting end according to one embodiment
of the present invention. Referring to FIG. 3b, symbol 1 represents a hands-free device, 2
represents an auxiliary speaker, 3 represents an auxiliary microphone, and 4 represents a virtual
microphone of the main microphone.
[0028]
03-05-2019
9
If no audio feature signal is found within the pointing angle of β1 described in FIG. 3a, the angle
of symmetry of β1 on the other side of the connecting line between the main microphone MIC-a
and the auxiliary microphone MIC-b described in FIG. 3b. If β2 is provided and it is determined
that there is an audio feature signal within the angle of β2, the transmit end is calibrated in the
direction determined by β2. Through the above steps, positioning and calibration of the transmit
end is completed and the directivity of the voice pickup is improved.
[0029]
Hereinafter, the determination of the audio feature signal will be combined to specifically
describe the positioning of the transmitting end.
[0030]
FIG. 4 is a flow chart of determining a voice feature signal according to one embodiment of the
present invention, and referring to FIG. 4 to specifically determine in real time whether there is a
voice feature signal within the angle β1. 1, collecting the signal, that the collected signal
becomes a scanned voice characteristic signal in γ1 direction, 2 performing envelope detection
for the voice characteristic signal, and the energy value is first If it is larger, the zero crossing
rate detection is performed on the voice feature signal, and if not, the direction of the voice
feature signal and the voice feature signal are determined again. 3. Determine the zero crossing
rate of the voice feature signal, and if the zero crossing rate is greater than a second
predetermined threshold, determine that the voice feature signal is within the collection angle
β1 and compare the directivity angle β1 In As a second collection angle of the further includes
calibrating the transmitter end by the second collection angle, the.
[0031]
In the present embodiment, performing envelope energy detection on the voice feature signal in
the γ1 direction includes performing envelope energy detection according to the following
equation.
Here, power is the energy value of the voice feature signal, the parameter alpha is a weighting
coefficient, and the parameter N is a voice feature signal at a certain point in time.
03-05-2019
10
Moreover, the detection sensitivity is controlled by adjusting two parameters of alpha and N, and
the stability of the envelope energy detection is guaranteed, and once the envelope energy power
is set to the first predetermined threshold (according to the actual situation) If it is detected that
the value is greater than zero), the zero crossing rate detection step is entered.
[0032]
Performing zero crossing rate detection on the voice feature signal in the γ1 direction includes
performing zero crossing detection according to the following equation. Here, Z_rate is the zero
crossing rate of the speech feature signal, n is one value in the discrete time series, and sgn [x]
satisfies the following equation. If the zero crossing rate Z_rate is greater than the second
predetermined threshold, it is considered that the collected signal within the β1 angle has an
audio feature signal, and it is determined that it is an audio operation.
[0033]
After determining that there is an audio feature signal within the β1 angle, the transmit end is
positioned and calibrated in the range determined by the angle β1. If there is no voice feature
signal within the β1 angle (γ1 is not the voice source direction), whether there is a voice
feature signal within the angle β2 symmetrical to β1 with respect to the connection line
between MIC-a and MIC-b Again, at this time, the audio feature signal within β 2 may be further
verified by envelope energy detection and zero crossing detection.
[0034]
From the above, accurate positioning of the transmitting end is realized by virtual array element
positioning and audio feature signal detection.
[0035]
FIG. 5 is a schematic diagram of calibrating the earpiece according to one embodiment of the
present invention.
Reference numeral 1 represents a hands-free device, 2 represents an auxiliary speaker, and 3
03-05-2019
11
represents a main microphone.
[0036]
After positioning of the transmitting end is completed, the direction information of the speaker
array is used for direction compensation using the direction information of the sound source so
that the sound emitted from the speaker array is directed to the sound source, ie, Reception
privacy is improved by adjusting as far as possible the speaking direction of the speaker array to
the speaking position of the user where the speaking end is positioned, ie, the direction
determined by the second collection angle. Specifically, virtual array element technology is
applied to virtualize the speaker SPK-b so that the connection line between the virtual speaker
SPK-c and the speaker SPK-a points in the audio direction, and A pointing angle is provided so
that the coverage area of the third collection angle (i.e. the area range of sound propagation
included in the third collection angle) covers the voice direction determined by the second
collection angle. (An open range below the SPK-a covers the position of the person's head, as
shown in FIG. 5), preferably the third collection angle is determined by the second collection
angle Direct the directed voice direction.
[0037]
Referring to FIG. 5, the hands-free device further includes a receiving end including at least one
speaker, and after the second collecting angle and direction are determined to complete the
calibration of the transmitting end, the method is performed. Calibrate the receiving end of the
hands-free device and provide a virtual speaker of one speaker at the receiving end so that the
connection line between the speaker and the virtual speaker points in the γ1 direction, and The
method further comprises the step of defining a third collection angle pointing in a direction
determined by the collection angle of the third collection angle, in fact the third collection angle
being the range of angles for reproducing the sound; By including the two collection angles and
directions, the speaker is reproduced as accurately as possible towards the direction of the sound
source, the sound reproduction range is narrowed, and the privacy of the reception is improved.
[0038]
Because the distance between the microphone and the speaker is close during use of the handsfree speech device, calibration can be performed on the receiver using the direction of the voice
feature signal determined when calibrating the transmitter. .
[0039]
03-05-2019
12
Referring to FIG. 5, main speaker SPK-a and a reverse extension passing through the γ1
direction are virtual speaker SPK-c of auxiliary speaker SPK-b, and main speaker SPK-c having
main speaker SPK-a as a circle. Virtual speaker SPK-c located in a circular arc whose radius is the
connecting line between -a and the auxiliary speaker SPK-b, and a new array is configured by the
main speaker SPK-a and virtual speaker SPK-c, and pointing of the new array Define the corner
as the third collection angle.
The positioning and calibration of the receiving end is completed, with the third collection angle
pointing in the direction determined by the second collection angle determined when calibrating
the transmit end.
[0040]
As can be understood, the angles shown in FIGS. 3a, 3b and 5 are for schematically describing
the range of angles when positioning the transmitting end and the receiving end, and the angles
shown in the drawings are actually Not the size of the angle.
Embodiments of the present invention improve the signal-to-noise ratio of hands-free speech
signals by positioning and calibrating with respect to the transmit end and improve the privacy
of the receive call by positioning the receive end.
[0041]
Summarizing the above, the method for improving the speech quality of hands-free equipment
such as the present invention performs calibration of the transmitting end and calibration of the
receiving end to reduce the interference of ambient environmental noise and to reduce the
transmission signal The objective is to improve the noise ratio and to improve the call quality of
the handsfree handset. Then, by performing direction compensation on the speaker array, the
emitted sound just points to the voice, the privacy of the reception is improved, and the user
experience is improved. The method is applicable to a smart device having hands-free calling
function such as a smart watch, and can greatly improve the calling performance of the smart
device.
03-05-2019
13
[0042]
According to another aspect of the present invention, there is provided an apparatus for
improving the call quality of a hands-free calling device, and FIG. 6 improves the calling quality
of the hands-free calling device according to one embodiment of the present invention. It is a
block diagram of an apparatus. The device 600 scans within the initial first collection angle of
the transmit end, and after the audio feature signal is scanned within the first collection angle,
the first collection is performed according to the direction of the audio feature signal. Voice
feature determination means 601 for determining a second collection angle smaller than the first
collection angle within an angle, and direction calibration for calibrating the transmitting end in
the direction determined by the second collection angle And means 602.
[0043]
The device is scanned for an audio feature signal in the γ1 direction within the initial first
collection angle of the transmit end, then draws a reverse extension through the main
microphone along the γ1 direction and A virtual circle to draw a circle whose center is the
connecting line between the main microphone and one of the auxiliary microphones with the
microphone as the center of the circle, and to determine the position where the arc and the
reverse extension intersect the virtual microphone of the main microphone The microphone
creation means and the virtual microphones of the main microphone and the main microphone
are used as a new voice array, and an angle β1 smaller than the first collection angle is defined,
and whether there is a voice feature signal in the angle β1 in real time If there is a voice feature
signal, check the angle β1 as the second collection angle, and if there is no voice feature signal,
the main microphone and any one auxiliary microphone As a symmetrical axis connecting lines,
define the mirror angle β2 with respect to the axis of symmetry of the angle .beta.1, further
comprising an angle confirmation means for confirming the angle β2 and a second collection
angle, the.
[0044]
Here, the angle check means further performs envelope energy detection on the voice feature
signal in the γ1 direction, and if the energy detection value is larger than a first predetermined
threshold, the zero crossing rate for the voice feature signal in the γ1 direction The detection is
performed, and is used to confirm that the voice feature signal is within the angle β1 when the
zero crossing rate of the voice feature signal in the γ1 direction reaches a second predetermined
threshold.
03-05-2019
14
[0045]
The angle confirmation means includes an envelope detection means and a zero crossing
detection means, and the envelope detection means is used to perform envelope energy detection
on the audio feature signal in the γ1 direction by the following equation.
Here, power is the energy value of the voice feature signal, the parameter alpha is a weighting
coefficient, and N is the voice feature signal at a certain point in time.
The zero crossing detection means is used to perform zero crossing rate detection on audio
feature signals in the γ1 direction according to the following equation. Here, Z_rate is the zero
crossing rate of the speech feature signal, n is one value in the discrete time series, and sgn [x]
satisfies the following equation.
[0046]
The apparatus is provided with a virtual speaker of one speaker at the receiving end so that the
connecting line between the speaker and the virtual speaker points in the γ1 direction, and the
third collecting angle is included in it. The receiver further comprises earpiece positioning means
for defining a third collection angle including the direction determined by the second collection
angle.
[0047]
According to another aspect of the present invention, the speech quality of the above-mentioned
hands-free speech device is improved, with a transmitting end comprising one main microphone
and at least one auxiliary microphone, a receiving end comprising at least one speaker, and And a
hands-free calling device including the device.
[0048]
It should be described that the hands-free device of the present invention has a transmitting end
and a receiving end, so that improvements can be made at the receiving end and the transmitting
end according to the method for improving the hands-free calling quality described above.
However, in some applications, the hands-free device has only a talk end or a receive end,
according to the method for improving the hands free call quality at the receive end in a specific
03-05-2019
15
embodiment of the present invention, It is possible to improve the transmit end according to how
to improve the receiver end or to improve the hands-free speech quality at the transmit end.
That is, the method for improving the handsfree speech quality at the reception end and the
transmission end of the present invention may be implemented separately and separately, and
the description thereof will not be repeated here.
[0049]
What has been described above is only a preferred embodiment of the present invention and
does not limit the protection scope of the present invention.
All amendments, equivalent replacements, improvements and the like made within the spirit and
principle of the present invention shall fall within the protection scope of the present invention.
03-05-2019
16
Документ
Категория
Без категории
Просмотров
0
Размер файла
28 Кб
Теги
jp2018019413
1/--страниц
Пожаловаться на содержимое документа