close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JPH0667692

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 JPH0667692
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an
adaptive noise reduction type voice input device.
[0002]
2. Description of the Related Art A circuit as shown in FIG. 5 is known as an adaptive noise
reduction device. In FIG. 5, 1 is a main input terminal, 2 is a reference input terminal, and a
signal inputted through the main input terminal 1 is supplied to the synthesis circuit 4 through
the delay circuit 3. Further, the signal input through the reference input terminal 2 is supplied to
the combining circuit 4 through the adaptive filter circuit 5 and is subtracted from the signal
from the delay circuit 3. The output of the synthesis circuit 4 is fed back to the adaptive filter
circuit 5 and is led to the output terminal 6.
[0003]
In this noise reduction device, the main input terminal 1 receives a signal obtained by adding the
desired signal s and an unnecessary voice signal (noise) no correlated with the desired signal s.
On the other hand, the noise n1 is input to the reference input terminal 2. The noise n1 of the
reference input is uncorrelated with the desired signal s, but is correlated with the noise n0.
10-04-2019
1
[0004]
The adaptive filter circuit 5 processes the reference input noise n1 and outputs a signal y
approximating the noise n0. As the output signal y of the adaptive filter circuit 5, it is also
possible to obtain a signal having the same amplitude as that of the noise n0. The delay circuit 3
is for adjusting the time of the signal to be subjected to the subtraction process in consideration
of the processing time in the adaptive filter circuit 5.
[0005]
The adaptive algorithm in the adaptive filter circuit 5 serves to minimize the subtraction output
(residual output) e which is the output of the combining circuit 4. That is, assuming that s, n0, n1,
and y are statistically stationary and the average value is 0, the residual output e is e = s + n0−y.
Since the expected value of the square of this is uncorrelated with s and n0 and y, E [e2] = E [s2]
+ E [(n0-y) 2] + 2E [s (n0-y) ] = E [s2] + E [(n0-y) 2]. Assuming that the adaptive filter circuit 5
converges, the adaptive filter circuit 5 is adjusted so as to minimize E [e2]. At this time, E [s2] is
not affected, so Emin [e2] = E [s2] + Emin [(n0-y) 2].
[0006]
That is, E [(n0-y) 2] is minimized by minimizing E [e2], and the output y of the adaptive filter
circuit 5 becomes an estimator of the noise n0. The expected value of the output from the
combining circuit 4 is only the desired signal s. That is, adjusting the adaptive filter circuit 5 to
minimize the total output power is equal to the subtraction output e being the least-squares
estimate of the desired speech signal s.
[0007]
In some cases, the synthesis circuit 4 may be acoustic synthesis means. That is, the adaptive filter
circuit 5 forms a noise cancellation voice signal -y having the same phase as noise and the
opposite phase of noise and supplies this to a speaker or the like to acoustically add to the main
voice to reduce noise. Do. The residual e in this case is picked up by the residual detection
microphone.
10-04-2019
2
[0008]
The adaptive filter circuit 5 can be realized either by an analog signal processing circuit or a
digital signal processing circuit. An example of the case where the adaptive filter circuit 5 is
realized using a digital filter is shown in FIG. In this example, a so-called LMS (Least Mean
Squares) method is used as an adaptation algorithm.
[0009]
As shown in FIG. 6, in this example, an adaptive linear combiner 300 of FIR filter type is used.
This includes a plurality of delay circuits DL1, DL2,..., DLm (m is a positive integer) each having a
delay time Z-1 of unit sampling time, an input noise n1 and respective delay circuits DL1, DL2,. A
weighting circuit (coefficient multiplier) MX0, MX1, MX2,... MXm for performing multiplication
of the output signal of and the weighting factor, and an adding circuit 301 for adding the outputs
of the weighting circuits MX0 to MXm. The output of the summing circuit 310 is y.
[0010]
Weighting coefficients to be supplied to the weighting circuits MX0 to MXm are formed based on
the residual signal e from the synthesizing circuit 4 by an LMS operation circuit 320 comprising,
for example, a microcomputer. The algorithm executed by this LMS arithmetic circuit 320 is as
follows.
[0011]
Let the reference input vector Xk at time k be Xk = [x0k x1k x2k... Xmk] T, as also shown in FIG.
... m), the input-output relationship is as shown in the following equation 1,
[0012]
Then, if the weight vector Wk at time k is defined as Wk = [w0kw1kw2k... Wmk] T, the input /
output relation is given by yk = XkT.Wk (1).
10-04-2019
3
Assuming that the desired response is dk, the error ek with the output is expressed as follows. ek
= dk-yk = dk-Xk T · Wk (2) In the LMS method, updating of the weight vector is performed
according to the equation Wk + 1 = Wk + 2μ · ek · Xk (3). Here, μ is a gain factor (step gain)
that determines the speed and stability of adaptation.
[0013]
In the above equation (3), the vector for correcting the coefficient vector Wk at a certain time k is
the second term of the right side of equation (3), but the gain factor μ and the instantaneous
error ek are scalar values, both Directly influence the value. Since the reference input vector Xk
also works in the form of a product, this also influences the correction value. The average
convergence time constant τa is represented by τa = (n + 1) / 4μ · trE [Xi Xj T]. Here, n is the
order of the reference input vector (corresponding to the number of taps of the FIR filter), and
trE [Xi Xj T] is the average power of the reference input. That is, the larger the number of taps of
the FIR filter, the slower the convergence speed, and the larger the gain factor μ, the faster the
convergence speed.
[0014]
In the case of a stationary signal, if the convergence speed is fast, the final residual noise level is
large, and if the convergence is slow, the final noise level is small. However, when the target
signal fluctuates like speech, the nature of the signal changes before convergence is complete, so
the amount of cancellation becomes larger if the convergence speed is faster to some extent.
[0015]
The value of the gain factor μ satisfies the following conditions in order to converge so that the
output y of the adaptive filter circuit 5 approaches that canceling the noise n0 and the output of
the apparatus becomes equivalent to the desired signal s There is a need. If 0 <μ <(power of
signal) / (number of taps of FIR filter + 1) (4) becomes larger than the range of equation (4), the
output y of the adaptive filter circuit 5 diverges, and It produces a loud noise as an output.
10-04-2019
4
[0016]
Conventionally, the value of the gain factor μ takes into account the above equation (4) taking
into account the number of taps of the adaptive filter 300 directly affecting the quality of the
processed signal and the magnitude (power) of the reference input signal. In order for the
adaptive filter circuit 5 to operate normally (converge) satisfactorily, it is set to a fixed value.
[0017]
By the way, for example, there are telephones which are provided with a microphone and a
speaker separately from the receiver, and are provided with a so-called speakerphone function
capable of making a call while the receiver is left.
In a telephone apparatus equipped with this function, the microphone catches the voice of the
talker, while the other party's voice can be heard from the speaker, so that so-called hands-free
communication can be performed and the hand is blocked, It is useful when you want to talk with
the other party.
[0018]
Also, even in a teleconferencing system between remote points, a plurality of voice terminals
consisting of microphones and speakers are provided for each of a plurality of attendees in each
conference room, and each attendee is likewise hands-free, While being able to hear the other
party's remark, it is possible to speak at any time.
[0019]
However, in the phone equipped with the speaker phone function as described above, ambient
noise enters the microphone and interferes with the call, and in each voice terminal of the
television conference system, the voices of other attendees in the same conference room It is
picked up unnecessarily.
[0020]
Also, for example, it is desirable that the sound pickup microphone of the camera integrated VTR
generally picks up only the sound from the subject side and does not pick up the sound from the
rear, for example, the sound of the photographer.
10-04-2019
5
[0021]
In order to achieve the above object, it is conceivable to use a noise reduction device using the
above-mentioned adaptive processing.
That is, in the adaptive processing as described above, it is conceivable to reduce unnecessary
signals (noise) in the main input signal using the unnecessary signal as a reference input signal.
[0022]
However, since it is difficult to completely change the directional characteristics of the
microphone for reference input sound collection so as not to collect the desired sound signal, the
desired sound signal is mixed in the reference input signal at a certain level. It will
[0023]
This situation deviates from the precondition of adaptive processing that both input signals of
the desired speech and the reference input speech are uncorrelated, especially when the level of
unwanted signal (noise) in the reference signal is quite low. In the above-mentioned normal
adaptive processing, there is a problem that the desired voice signal itself becomes a target of
reduction.
[0024]
SUMMARY OF THE INVENTION In view of the above, it is an object of the present invention to
provide a noise reduction device capable of preventing the reduction of desired voice itself.
[0025]
SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, an adaptive
noise reduction type voice input device according to the present invention collects desired voice
and corresponds to the reference symbols of the embodiments described later. A first
microphone 11 from which an output signal of polarity is obtained, and an audio input from
which an output signal different in polarity according to the audio input direction is obtained,
and an output signal of reverse phase to the polarity of the output signal of the first microphone
is obtained Adaptive filter means 24 including a second microphone 21 whose direction is
oriented in the direction of arrival of the desired voice, and a coefficient multiplier, to which an
audio signal from the second microphone is supplied, and an output signal of the adaptive filter
means And subtracting means 15 for subtracting from the voice signal of one microphone,
10-04-2019
6
adjusting the adaptive filter means so that the output power of the subtracting means is
minimized, and supplying it to the coefficient multiplier Weight coefficient Wk is always those to
limit the operation of the adaptive filtering means such that Wk ≧ 0.
[0026]
When the second microphone 21 is disposed in the direction of arrival of the desired voice so as
to obtain an audio input direction in which an output signal in phase with the output signal of the
first microphone 11 is obtained, addition means instead of the subtraction means 15 The
weighting factor Wk supplied to the coefficient multiplier may be set so as to limit the operation
of the adaptive filter means so that Wk ≦ 0.
[0027]
In the present invention of the above configuration, the main input speech signal from the first
microphone and the desired speech component in the output speech signal of the second
microphone are in reverse phase.
Further, the main input speech signal from the first microphone and the unnecessary voice
(noise) in the output speech signal of the second microphone are in phase.
Therefore, if the weighting coefficient Wk supplied to the coefficient multiplier in the adaptive
filter means is Wk ≧ 0, the main input speech signal is subtracted from the main input speech
signal by the subtraction means if Unwanted voices in the media are eliminated and only desired
voices can be obtained.
[0028]
If Wk <0, the subtracting means subtracts the desired voice in the reference input signal from the
main input voice signal, resulting in a reduction in the desired voice.
In the present invention, the adaptive processing is limited so that the coefficient Wk00 at all
times, so that the situation where the desired voice is reduced can be prevented.
10-04-2019
7
[0029]
When an audio input direction in which an output signal in phase with the output signal of the
first microphone 11 is obtained is disposed in the arrival direction of the desired voice and the
addition means is provided instead of the subtraction means 15 Is the operation of the adaptive
filter means such that the phase relationship between the desired voice component in the
reference input voice signal and the unnecessary voice component is reversed with respect to the
main input voice signal, and the weighting coefficient Wk is always Wk ≦ 0. Is limited.
[0030]
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an adaptive
noise reduction type speech input apparatus according to the present invention will be described
below with reference to FIGS.
[0031]
In FIG. 1, reference numeral 11 denotes a main input microphone for picking up desired voice,
and 21 denotes a reference input microphone for picking up unnecessary voice and ambient
noise in a direction to be removed as noise.
In this example, the arrival direction of the desired voice is a direction (hereinafter referred to as
a front direction) going from the top to the bottom on the figure mainly as shown by an arrow AR
in FIG. This is an example of realizing a noise reduction type voice input device that prevents
unwanted sounds from being collected as noise.
[0032]
In this example, as shown in FIG. 2, the main input microphone 11 is configured by a
unidirectional microphone having a high sensitivity in the desired voice incoming direction and a
low sensitivity in the back direction.
On the other hand, as shown in FIG. 2, the reference input microphone 21 is composed of a bidirectional (eight-character directivity) microphone having high sensitivity in the desired voice
arrival direction and the back direction.
10-04-2019
8
[0033]
The bi-directional reference input microphone 21 is, in this example, a bipolar one in which twophase output signals of positive and negative phases can be obtained according to the voice input
direction. The sensitivities of the phase and the antiphase need not necessarily be equal.
Then, in this embodiment, the reference input microphone 21 is disposed so that the voice input
direction in which the output signal of the reverse phase is obtained is directed to the arrival
direction of the desired voice.
Further, the main input microphone 11 is capable of obtaining an output signal of positive phase.
[0034]
An audio signal picked up by the main input microphone 11 and converted into an electric signal
is obtained through the amplifier 12 and supplied to the A-D converter 13 where it is converted
into a digital signal, and the delay circuit 14 is used. Is supplied to the subtraction circuit 15.
Further, an audio signal obtained by being picked up by the reference input microphone 21 and
converted into an electric signal is supplied to the A-D converter 23 through the amplifier 22 and
converted into a digital signal to be an adaptive filter. It is supplied to the circuit 24.
[0035]
In this embodiment, the adaptive filter circuit 24 comprises an FIR filter type adaptive linear
combiner 300 and an arithmetic circuit (microcomputer) 320 for adaptively controlling the
linear combiner 300 as shown in FIG. The digital signal from the A-D converter 23 is supplied to
the arithmetic circuit 320 and to the subtractor circuit 15 via the linear combiner 300.
10-04-2019
9
The output signal of the subtraction circuit 15 is fed back to the arithmetic circuit 320 and is
converted back to an analog signal by the D-A converter 16 and is led out to the output terminal
17.
[0036]
The D-A converter 16 may be omitted, and the output signal of the subtraction circuit 15 may be
derived as it is as a digital signal to the output terminal 17. The delay circuit 14 is for
compensating for a time delay such as a propagation time in the adaptive filter circuit 24 or a
time delay required for an operation for the adaptive processing.
[0037]
Basically, in the adaptive filter circuit 24, control is performed so that the reference input speech
signal approximates to the noise contained in the main input speech signal. As a result, assuming
that the desired voice in the voice collected by the main input microphone 11 and the noise are
uncorrelated, the subtraction circuit 15 determines the reference input microphone 21 from the
voice signal of the main input microphone 11. The speech signal (noise) is subtracted and
removed, and only the desired speech signal is obtained from the subtraction circuit 15. That is,
the basic configuration of this embodiment is adaptive noise in which the output sound signal of
the main input microphone 11 is supplied as the main input, and the output sound signal of the
reference input microphone 21 is supplied as the noise as the reference input. It is the
composition of the reduction system. Then, noise, which is an unnecessary voice signal from the
rear direction in this example, is selectively removed to the output terminal 17, and as a result,
only the desired voice signal is output with the required quality. In other words, superdirectivity
is realized as an adaptive system.
[0038]
Next, with reference to FIG. 2 and FIG. 3, the operation limitation of the adaptive filter circuit of
the embodiment of FIG. 1 will be described. In a normal case, while the desired voice comes from
the front direction indicated by the arrow AR in FIG. 2, the unnecessary sound to be reduced
comes from the rear direction, and the unidirectional microphones for main input 11 as shown in
FIG. From this, both the high level desired voice signal s11 and the low level noise n11 are
output in the positive phase. On the other hand, from the bi-directional reference input
10-04-2019
10
microphone 21 as shown in FIG. 2, a high-level desired voice signal s21 of opposite phase and a
high-level noise n21 of positive phase are outputted.
[0039]
The output of the microphone 11 is supplied to the subtraction circuit 15 as a main input audio
signal, and the output of the microphone 21 is supplied to the adaptive filter circuit 24 as a
reference input audio signal. The noise component n21 in the reference input speech signal is
uncorrelated with the desired speech signal s11 in the main input speech signal, but is correlated
with the noise component n11. The reference input speech signal is controlled to approximate
noise contained in the main input speech signal, and the output of the adaptive filter circuit 24 is
subtracted from the main input speech signal in the subtraction circuit 15 to obtain a positive
phase main input. The noise component n11 in the speech signal and the positive-phase noise
component n21 in the adaptively processed reference input speech signal are canceled out.
[0040]
In the case of such normal adaptive processing, as described above, the weighting coefficients W
jk (j = 0, 1, 2,... M) of the coefficient multipliers MX 0 to MX m (see FIG. 6) of the adaptive linear
combiner 300 are The weighting coefficient is updated so that the positive phase noise
components n11 and n21 in both the main and reference input speech signals are canceled out
by subtraction. It does not take a value.
[0041]
In addition, since the weighting factor does not take a negative value, the desired phase signal
component s21 of the opposite phase in the reference input sound signal is added to the desired
phase signal component s11 of the positive phase in the main input sound signal by subtraction.
It will be.
In other words, in the case of using the bipolar and bi-directional microphones 21 as shown in
FIG. 2, the sensitivity as the adaptive system to the desired voice is improved.
[0042]
Then, in the case of ordinary adaptive processing, the directivity as an adaptive system is as
10-04-2019
11
shown in FIG. 3A, and as expected, unnecessary noise coming from the rear direction is a target
of reduction of the system.
[0043]
By the way, when the desired voice at the normal level comes from the front direction and the
unnecessary sound coming from the back direction is weak, the bi-directional reference input
microphone 21 as shown in FIG. Only the high-level desired voice signal s21 is output in the
opposite phase.
On the other hand, substantially only the desired voice signal s11 of positive phase and high
level is output from the unidirectional main input microphone 11 as shown in FIG.
[0044]
In this state, since the noise reduction system works to minimize the output power of the
subtraction circuit 15, the characteristics of the adaptive filter circuit 24 largely deviate from the
originally desired one, so as to reduce the desired voice signal itself. It will be in a working state.
That is, in this state, the weighting coefficients Wjk (j = 0, 1, 2,... M) of the adaptive linear
combiner 300 have negative values, and the adaptively processed reference input speech signal
is inverted in polarity, If the desired audio signal component s21 is in the normal phase and is
subtracted from the main input audio signal, that is, the desired audio signal component s11 in
the positive phase in the subtraction circuit 15, the desired audio signal itself is reduced.
[0045]
And when the weighting factor becomes negative in this way, the directivity as an adaptive
system becomes as shown by a broken line in FIG. 3B, and the reduction target direction of the
system moves in the front direction, and the desired voice Is the target of system reduction.
[0046]
In this embodiment, in order to avoid the reduction of the desired audio signal itself as described
above, the operation of the adaptive filter circuit is limited as follows.
10-04-2019
12
That is, when the adaptive processing starts, in the arithmetic circuit 320, the weighting
coefficients of the adaptive linear combiner 300 at the next time point are sequentially calculated
according to the equation (3), and the weighting coefficients are sequentially updated. At this
time, when the weighting factor at the next time point obtained by the operation becomes a
negative value, the weighting factor of the adaptive linear combiner 300 is forced without using
the negative weighting factor as the weighting factor of the next time point. Update to zero.
[0047]
Due to this operation restriction, as described above, the unnecessary sound to be reduced is
weak, and the main and reference input audio signals become substantially only the desired
audio signals s11 and s21 of the positive phase and the negative phase. However, since the
weighting factor does not take a negative value, the desired audio signal itself is not reduced by
the subtraction process. When the weighting factor is set to zero, the directivity as an adaptive
system is as shown by a solid line in FIG. 3B, and the desired voice coming from the front
direction is not a reduction target of the system.
[0048]
In the above embodiment, the direction of the reference input microphone 21 is opposite to that
in the example of FIG. 2, and the voice input direction in which an output signal in phase with the
output voice signal of the main input microphone 11 is obtained is the arrival of the desired
voice. It is also possible to limit the operation of the adaptive filter circuit so as to avoid reduction
of the desired audio signal itself, so that it is oriented toward the direction so that the weighting
factor Wk always takes only zero or a negative value (Wk 0 0) it can. Instead of reversing the
direction of the reference input microphone 21 to the example shown in FIG. 2, an inverter is
inserted into the two-phase output of the reference input microphone 21 and a signal obtained
by inverting the phase is used as the reference input microphone 21. The output signal of the
phase may be used.
[0049]
In the above embodiment, the adaptive system uses the subtractor circuit 15 as means for
synthesizing the main input speech and the output of the adaptive filter circuit 24. However, the
signal to be reduced in the output of the adaptive filter circuit 24 (main input In the case where
10-04-2019
13
the noise (unwanted signal) is in antiphase with the phase of the output signal of the main input
microphone 11, an adder circuit can be used as a combining means of the adaptive system.
[0050]
When the addition circuit is used in this way, the weighting factor Wk always takes only zero or a
negative value when the direction of the reference input microphone 21 is the same as in the
example of FIG. 2 (Wk ≦ 0) Thus, the operation of the adaptive filter circuit can be limited to
avoid the reduction of the desired speech signal itself.
Also, when the reference input microphones 21 are arranged reverse to the example of FIG. 2,
the operation of the adaptive filter circuit is limited so that the weighting factor Wk always takes
only zero or a positive value (Wk 0 0). Thus, the reduction of the desired voice signal itself can be
avoided.
[0051]
In the above embodiment, as shown in FIG. 2 described above, a unidirectional microphone with
high sensitivity in the front direction is used for the main input, and sensitivity in the front and
back directions for the reference input. Use a pair of uni-directional microphones 11m and 21m
arranged in opposite directions to each other as main and reference inputs as shown in FIG. It
can also be done.
[0052]
When using a pair of unidirectional microphones for main and reference input, the system
configuration is exactly the same as in FIG. 1, and the output of both microphones has the same
polarity. Limit the operation of the adaptive filter circuit 24 so that the weighting coefficient of
the adaptive linear combiner 300 takes only zero or a positive value (W k 00), and avoid
reduction of the desired voice signal itself. be able to.
Also, when the system includes an addition circuit instead of the subtraction circuit 15, the
operation of the adaptive filter circuit is limited so that the weighting factor Wk always takes
only zero or a negative value (Wk k 0), and the desired voice The reduction of the signal itself
can be avoided.
10-04-2019
14
[0053]
When the output signal of the main input microphone 11m and the output signal of the
reference input microphone 21m are in reverse phase (including the case where the microphone
output is in reverse phase by the inverter), the adaptive system is a subtraction circuit. When 15
is included, the operation of the adaptive filter circuit 24 is limited so that the weighting
coefficient of the adaptive linear combiner 300 takes only zero or a negative value (W k ≦ 0),
and the reduction of the desired voice signal itself is avoided. When the system includes an adder
circuit instead of the subtractor circuit 15, the operation of the adaptive filter circuit is limited so
that the weighting factor Wk always takes only zero or a positive value (Wk .gtoreq.0). Thus, the
reduction of the desired voice signal itself can be avoided.
[0054]
As described above, according to the present invention, in the noise reduction type voice input
device, adaptation is performed according to the phase relationship between the output signal of
the adaptive filter means and the voice signal of the first microphone. Since the weighting factor
of the coefficient multiplier of the filter means is limited to take only zero or a positive value, or
only zero or a negative value, reduction of the desired voice itself can be prevented, and the
target of reduction can be reduced. Only unwanted sound from the voice input direction can be
reduced.
[0055]
Brief description of the drawings
[0056]
1 is a block diagram showing a configuration of an embodiment of the noise reduction type voice
input device according to the present invention.
[0057]
2 is a diagram showing an example of the directivity of the first and second microphones.
[0058]
3 is a diagram for explaining the operation of an embodiment of the present invention.
10-04-2019
15
[0059]
4 is a diagram showing another example of the directivity of the first and second microphones.
[0060]
5 is a block diagram showing an outline of an adaptive noise reduction apparatus for explaining
the present invention.
[0061]
6 is a diagram showing a configuration example of the adaptive filter circuit.
[0062]
Explanation of sign
[0063]
11 and 11 m 1st (main input) microphone 21 and 21 m 2nd (reference input) microphone 15
subtraction circuit 24 adaptive filter circuit
10-04-2019
16
Документ
Категория
Без категории
Просмотров
0
Размер файла
27 Кб
Теги
description, jph0667692
1/--страниц
Пожаловаться на содержимое документа