close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2007174343

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 JP2007174343
PROBLEM TO BE SOLVED: To make it possible to reduce an echo feeling in a subsequent
conversation even if a tone signal is input. SOLUTION: In the echo canceller of the present
invention, it is judged whether the far-end input signal is a tone signal for call control, and if so,
the coefficients of the adaptive filter are reset. In another echo canceller of the present invention,
it is determined whether the far-end input signal is a tone signal of a predetermined type, and in
the case of a positive result, coefficient updating of the adaptive filter is performed if the echo
cancellation amount is equal to or less than the threshold. Stop it. Also, when the tone signal
ends, the echo cancellation amounts before and after the end are compared, and the coefficient
of the adaptive filter is reset and then the coefficient update is resumed or the coefficient of the
adaptive filter is not reset according to the comparison result. Resume coefficient updating.
Furthermore, at the input stage of the adaptive filter and the adder for echo removal, a band
rejection filter for removing the frequency component of the tone signal is provided. [Selected
figure] Figure 1
エコーキャンセラ
[0001]
The present invention relates to an echo canceller, and is applicable to, for example, an echo
canceller provided in a VoIP terminal accommodating a telephone terminal.
[0002]
An echo canceller, which removes echo components in the hybrid circuit, is generally provided to
15-04-2019
1
prevent the reception signal from flowing into the path of the transmission signal through the
hybrid circuit and becoming an echo component to degrade voice quality. .
By the way, although a received signal that can be an echo component is generally an audio
signal, a tone signal may be a received signal.
[0003]
Conventionally, as an echo canceller in consideration of the case where a tone signal becomes a
reception signal, there is one described in Patent Document 1. The configuration of the echo
canceller described in Patent Document 1 will be clarified in the section of the description of the
problem. Patent document 1: JP-A-2005-110307
[0004]
The echo canceller described in Patent Document 1 has the following problems 1 to 4. The echo
canceller described in Patent Document 1 is hereinafter referred to as prior art.
[0005]
(Problem 1) Since it is not known that the tone signal has been input to the echo canceller (or the
adaptive filter) only after the coefficients of the adaptive filter are destroyed by the tone signal, it
is often too late even if some measures are taken. Also, even if reconvergence, since the
coefficient value is not suitable as an initial state of convergence, it takes time until the
performance can be exhibited again.
[0006]
The prior art basically has a mechanism capable of detecting that the adaptive filter is a tone
signal after convergence to the tone component.
That is, before the tone signal is input to the adaptive filter, even if it is possible to estimate a
coefficient having a suitable wide-band frequency characteristic by speech etc., the filter
15-04-2019
2
coefficient is once made to have tonality. After being destroyed, it is often too late to use for
various processes following the determination, for example, whether or not to update the
coefficients of the adaptive filter, etc., in order to determine whether the coefficients match the
frequency characteristics of the tone. Also, even if the echo canceller is in the initial state or
when reconvergence is performed once in the convergent state, the coefficients of the adaptive
filter incorrectly converged by the tone signal are appropriate as initial values for reconvergence.
Because the reconvergence takes time, the echo quality does not disappear, and the speech
quality is degraded.
[0007]
(Problem 2) When the initial delay is small, the desired operation is not performed.
[0008]
According to the prior art, when the echo canceller estimates the echo path using a tonal signal
as a reference signal, the adaptive filter coefficient register of the echo canceller converges to the
input tone frequency.
However, when the initial delay of the echo path is small, the prior art often does not work. This
is because when the initial delay of the echo path is small, the coefficient register often does not
converge to the input tone frequency.
[0009]
Hereinafter, Problem 2 of the prior art will be described with reference to FIGS. 2 and 3. First,
cases where the prior art can be applied will be shown, and then examples of cases where the
prior art can not be applied will be illustrated.
[0010]
FIG. 2 (a) shows an echo path of an initial delay of 50 samples, and this echo path is an echo path
having a substantially flat frequency characteristic from 0 to 8 kHz as shown in FIG. 2 (c). FIG. 2
(b) shows the result of convergence of the coefficient register by inputting a tone input (in this
15-04-2019
3
example, the dial tone is set to 400 Hz) to the echo canceller. FIG. 2C shows both the frequency
characteristic of the echo path and the frequency characteristic of the coefficient register after
convergence. As can be seen from FIG. 2 (b), the adaptive filter coefficients of the echo canceller
converge in a state similar in waveform to the input tone, and from FIG. 2 (c) showing the
frequency characteristics, Among them, it can be seen that only the tone frequency portion has a
large component.
[0011]
On the other hand, FIGS. 3A to 3C show how the echo canceller operates with an echo path with
an initial delay of 10 samples. FIG. 3 (a) shows an echo path of 10 samples of initial delay, and
the frequency characteristic is an almost flat echo path from 0 to 8 kHz as shown in FIG. 3 (c).
Similar to the above, FIG. 3 (b) shows the result of convergence of the coefficient register by
inputting a tone input (dial tone 400 Hz) to the echo canceller. In FIG. 3 (b), the adaptive filter
coefficient of the echo canceller is different from that of the input tone, and it can be seen from
FIG. 3 (c) that is the frequency characteristic that there is no peak at 400 Hz of the input tone.
More specifically, in FIG. 3 (c), the maximum of the frequency characteristic is 0 Hz.
[0012]
However, both cases shown in FIGS. 2 and 3 perform well with regard to echo cancellation for
tone input. In the example of explanation, both show ERLE (power ratio of received signal to
residual) about 35 dB. In other words, in spite of the fact that the characteristics of the echo path
are not sufficiently estimated, a state occurs in which the echo removal performance is exhibited
when limiting to the input frequency. Such a phenomenon often occurs when the initial delay of
the echo path is smaller than the tone period. Specifically, when the arrangement of the echo
canceller and the hybrid circuit which is an echo generation source is close, for example, in the
same apparatus, it always occurs. Unfortunately, in VoIP terminals etc. it is more common for the
hybrid circuit and the echo canceller to be located in the same device. Therefore, in such a case,
since the frequency characteristic of the coefficient register does not have an input tone, the
tonality of the coefficient can not be detected, and the prior art is not effective.
[0013]
(Problem 3) The echo cancellation performance is severely degraded by double talk immediately
after the tone is input to the echo canceller.
15-04-2019
4
[0014]
Here, to explain double talk, a state in which both the calling and called parties are generating
voice is referred to as double talk, and in the echo canceler, the transmission signal and the
reception signal are simultaneously input. It is in the state of
As described in the prior art, in the echo canceller, in the case of double talk, ERLE is often
monitored to perform double talk detection in order to perform control to stop coefficient
updating of the adaptive filter. In the prior art, as shown in the above-mentioned Problem 1,
double-talk determination is performed using the residual based on the result of using an
adaptive filter that has misfocused on a tonal signal. Not only can the correct double-talk
determination be made, but the step gain for controlling the speed of the coefficient update is set
large for a certain period after the tone convergence because the filter residue of the incorrect
convergence is used. As a result, the adaptive filter coefficient is adapted in a state in which the
followability of the coefficient update is most agile for a predetermined period after the tone
convergence. As a result of this, the sound quality is greatly degraded because of the
inappropriate cancellation. Also, even when double-talking does not occur, re-convergence of the
adaptive filter is the same as that started with the false convergence state to the tone signal as
the initial state, and even if coefficient updating is performed, echo is stably generated. It takes a
long time to get rid of it. Naturally, the speaker hears an echo for a long time until
reconvergence, and the speech quality is degraded.
[0015]
(Problem 4) When the prior art is used as it is, the removability of echo deteriorates at the time
of telephone transfer such as extension transfer. Hereinafter, with reference to FIG. 4, the
problem at the time of extension transfer, which is the problem 4 of the prior art, will be
described.
[0016]
In FIG. 4, a case will be described in which the telephone set 100 calls the telephone set 109, and
after a while talking, the extension is transferred to the telephone set 110. As in the prior art, the
echo canceller 104 uses the output of the adder 119 and the input of the reception input
15-04-2019
5
terminal 103 to estimate the echo path by using the hybrid circuit 108 as an echo path, and the
coefficients of the adaptive filter 118 in the hybrid circuit 108. Estimate to be equal to the
impulse response. As a result of such estimation, the speaker signal from the telephone set 100
passes through the hybrid circuit 101 and is converted into a digital signal by the analog / digital
converter (A / D converter) 102 to be a digital signal, and then digital / analog. The signal is
converted into an analog signal by the converter (D / A converter) 105, passes through the
switch 106, is subjected to two-wire conversion by the hybrid circuit 108, and then reaches the
telephone set 109. A part of the speaker signal to the telephone set 109 is reflected by the
hybrid circuit 108 to become an echo signal y 1, converted to a digital signal by the A / D
converter 113 via the switch 112, and input to the adder 119. Be done. The pseudo echo signal y
'output from the adaptive filter (ADF) 118 is input to the adder 119, and if the adaptive filter 118
converges, y' y y1, so echo (echo signal y ') Is canceled.
[0017]
A known algorithm such as NLMS is used to update the coefficients of the adaptive filter 118.
Here, the state of the coefficient update of the NLMS algorithm is briefly described.
[0018]
Assuming that the voice signal input from the far-end telephone 100 to the reception input
terminal is x (n) and the filter coefficient of the adaptive filter 118 is hk (n), the filter coefficient
hk (n) is represented by equations (1) and (2) Update to follow.
[0019]
When the denominator of the second term on the right side of the equation (1) is 0, the
coefficient update amount is set to 0 or the coefficient update is stopped.
The filter coefficient hk (n) represents the tap coefficient of the k-th adaptive filter at the n-th
sample time. Also, α is a step gain, which is a constant for determining the convergence speed of
the adaptive filter 118, and 0 <α <2. If α is large, the convergence speed is high, but the
fluctuation of steady-state characteristics is large, and the influence of noise is also large. On the
other hand, if α is small, the convergence speed is slow but the fluctuation of the steady-state
characteristic is small, and the influence of noise is also small. According to the prior art, it is
shown that the step gain is reduced to a small value after the adaptive filter 118 sufficiently
15-04-2019
6
converges.
[0020]
Now, at the time of extension transfer, a call is temporarily made between the telephones 100
and 109, and the coefficient of the adaptive filter 118 is also estimated by the hybrid circuit 108
to some extent and converged. Here, if the adaptive filter 118 is completely converged, the echo
removal residual e (n) of the equation (2) becomes 0, so that the coefficient updating is the same
as stopping.
[0021]
Thereafter, the case where a call is transferred from the telephone set 109 to the telephone set
110 by the private branch exchange or the like will be described in the following four cases.
Here, when transfer occurs, the terminal of the switch 106 is closed to a2, and the terminal of the
switch 112 is closed to b2. A signal output from the D / A converter 105 passes through the
terminal a2 of the switch 106 and is partially output to the telephone 110 via the hybrid circuit
111 and partially reflected by the hybrid circuit 111 to be an echo signal y2. The signal is input
to the A / D converter 113 via the terminal b 2 of the switch 112 and is input to the Sin terminal
114. Hereinafter, combinations of the convergence state of the echo canceller and the magnitude
of the echo path change due to the transfer will be respectively described.
[0022]
(Problem 4; Extension Transfer Case A) The extension transfer case A is when convergence of the
adaptive filter 118 is insufficient and the characteristics of the hybrid circuits 108 and 111
hardly change.
[0023]
In the case before the transfer in this case A, since the echo removal residual e (n) of the equation
(2) is not 0, the coefficient update of the equation (1) is executed while being valid.
That is, as in the prior art, if x (n) is a tonal signal, the coefficients of the adaptive filter 118 are
15-04-2019
7
updated according to the input x (n) according to equation (1). As a result, the coefficients of the
adaptive filter 118 converge to cancel the tonal echo signal y1, rather than the characteristics of
the hybrid circuit 108. Naturally, as shown in FIG. 2, the coefficients of the adaptive filter 118
converge to characteristics different from the impulse response of the hybrid circuit 108.
Thereafter, a transfer occurs, the hybrid circuit 111 is connected, and the telephone 110 is
connected to start a call.
[0024]
In this case, the hybrid circuit characteristics can be considered equal before and after transfer,
but the coefficients of the adaptive filter 118 converge according to the tone. Thereafter, when
the speech signal x (n), which is a non-tone signal, is input, the adaptive filter 118 estimates only
the component of the tone signal, so that the adaptive filter 118 The convergence state does not
reflect the response characteristics of the hybrid circuit. Therefore, the echo of the audio signal
having a frequency component wider than the tone frequency can not be canceled by almost all
components, and is output from the adder 119. The output of the adder 119 is input to a double
talk detector (DTD) 117 as shown in FIG. The double talk detector 117 calculates the magnitude
of the ratio ERLE between the adder output and the input rin (n) from the reception input
terminal, and detects the double talk based on the magnitude of the value of ERLE or the
magnitude of the change of ERLE. ERLE is calculated, for example, as shown in equation (3).
[0025]
As described in the prior art, if the value of ERLE is large, such as 30 dB, then e (n) is sufficiently
small and convergence state or convergence is possible, and if ERLE is small, such as 6 dB, e (n)
And so on) and stop updating the coefficients of the adaptive filter 118 because there is voice on
the called side because it is still large. Alternatively, as another method, abrupt deterioration of
ERLE is detected, it is determined that there is transmission voice on the called side, and
coefficient updating of the adaptive filter is stopped.
[0026]
However, as described above, since the echo component by the voice signal from the receiving
input terminal is not canceled by most components immediately after the tone erroneous
convergence, ERLE has a small value, and from the viewpoint of ERLE change, The ERLE by the
15-04-2019
8
voice immediately after the tone is rapidly degraded. The double talk detector 117 determines
that this is the presence of the transmission voice signal on the called side, thereby detecting that
the double talk state has occurred. As a result, since the double talk detector 117 stops updating
the coefficients of the adaptive filter 118, the adaptive filter 118 subsequently freezes the
coefficients, which are maintained thereafter. That is, the coefficient update of the adaptive filter
118 is frozen in a state where the echo does not disappear. Furthermore, if the response is a
hybrid circuit 108, 111 with a small initial delay, then the coefficients of the adaptive filter 118
will not be the same as the input tonality. That is, the adaptive filter 118 reflects only
characteristics different from the nature of the input signal and the nature of the hybrid circuits
108 and 111. As a result, it is not possible to recover from the state of erroneous convergence.
[0027]
Luckily, even when double-talking is not detected, the coefficients of the adaptive filter 118
converge to an inappropriate state for echo cancellation that reflects the tone signal, rather than
the characteristics of the hybrid circuits 108 and 111. Even if the characteristic of the hybrid
circuit 111 converges again, it is not suitable as an initial setting for reconvergence of the
adaptive filter 118, so it takes time to reconverge, and in the meanwhile echo is generated. It will
[0028]
(Problem 4; Extension Transfer Case B) The extension transfer case B is a case where the
convergence of the adaptive filter 118 is insufficient and the characteristics of the hybrid circuits
108 and 111 are different.
[0029]
Case B is more problematic than case A.
This is because the characteristics of the hybrid circuit 111 after the tone misconvergence are
completely different from the characteristics of the hybrid circuit 108, and the filter coefficients
of the adaptive filter 118 incorrectly misconverged with the hybrid circuit 108 and the tone are
newly added. Since the characteristic of the impulse response of the connected hybrid circuit 111
is not reflected at all, the echo of the voice of the wide frequency component after the tone signal
can not be canceled at all, and the echo becomes noticeable.
Furthermore, it is impossible to escape the false determination of the double talk detector 117.
15-04-2019
9
As a result, as described in Case A, the adaptive filter 118 freezes the coefficient update with the
echo not disappearing. Furthermore, if the response is a hybrid circuit 108, 111 with a small
initial delay, the coefficients of the adaptive filter 118 will not be the same as the input tonality,
so it can not recover from the frozen state.
[0030]
Therefore, after transmission, echo remains in the generated state.
[0031]
(Problem 4; Extension Transfer Case C) In the extension transfer case C, the convergence of the
adaptive filter 118 is sufficient, and the characteristics of the hybrid circuits 108 and 111 hardly
change.
[0032]
In the case C, since the echo removal residual e (n) = 0 of the equation (2) is originally obtained,
the coefficient update according to the equation (1) is not updated because the update amount is
0.
Furthermore, as in the prior art, if the step gain is reduced in response to convergence,
coefficient updating is prevented.
[0033]
However, in practice, it is rare that the echo removal residual e (n) becomes completely zero due
to background noise on the near-end speaker (Sin speaker) side.
In such a case, when the tone signal is input from the reception input terminal 103 for a long
period of time, the coefficients of the adaptive filter 118 gradually converge to the tonality. At
this time, once the adaptive filter 118 erroneously converges on the tone signal, the same
problem as in case A occurs. In the case C, it is desirable to stop the coefficient update before the
tone erroneous convergence of the adaptive filter 118 is not progressing, instead of detecting the
result of the erroneous convergence of the adaptive filter 118 as in the prior art. .
15-04-2019
10
[0034]
(Problem 4; Extension Transfer Case D) The extension transfer case D is when convergence of the
adaptive filter 118 is sufficient and the characteristics of the hybrid circuits 108 and 111 are
different.
[0035]
Also in case D, when a long-time tone signal is input from the telephone set 100, the coefficients
of the adaptive filter 118 may be misfocused in the tone signal.
The problem that occurs immediately after such a tone misconvergence is the same as in case B.
However, in the case D, even if the coefficient update is stopped before the erroneous tone
convergence of the adaptive filter 118 progresses, the response characteristic of the hybrid
circuit 111 after the extension transfer is the same as the characteristic of the hybrid circuit 108
before transfer. Is a separate item, so in any case echo cancellation is almost impossible after
transfer. Therefore, ERLE becomes a small value. Even if it is considered as a change, the method
of change of ERLE also becomes sharply smaller after the occurrence of transfer, so the double
talk detector 117 erroneously judges the change of ERLE as a double talk state. Moreover, even if
the coefficients of the adaptive filter 118 are stopped before the tone erroneous convergence
progress, this means that the coefficients that have converged on the audio signal are frozen as
they are. There is no tonality even when the coefficients are analyzed, and ERLE is small.
Therefore, the adaptive filter 118 of the echo canceller freezes the coefficients in a state where
the echo can not be canceled, and the echo continues to be output continuously after the transfer
occurs. That is, even if the convergence is incorrect, a defect such as Case B occurs, and even if
the coefficient updating is stopped before the convergence, the echo is continuously output after
the transfer.
[0036]
The present invention has been made in consideration of the above points, and it is an object of
the present invention to provide an echo canceller that can reduce the feeling of echo in
subsequent conversations even if a tone signal is input.
[0037]
According to a first aspect of the present invention, in an echo canceller for removing echo by a
15-04-2019
11
hybrid circuit using an adaptive filter for generation of pseudo echo, (1) determining whether the
far-end input signal is a tone signal for call control And (2) coefficient control means for resetting
the coefficient of the adaptive filter when the call control tone determination means determines
that it is a call control tone signal. It is characterized by
[0038]
According to a second aspect of the present invention, in an echo canceller for removing echo by
a hybrid circuit using an adaptive filter for generation of pseudo echo, (1) it is determined
whether the far-end input signal is a tone signal of a predetermined type Tone type
determination means, (2) Addition means for subtracting the pseudo echo signal from the above
adaptive filter from the near end input signal, and (3) Echo cancellation amount for calculating
the signal output from the addition means When the cancellation amount calculation means and
(4) the tone type determination means determine that the far-end input signal is a tone signal of
a predetermined type, the adaptation according to the establishment of the predetermined
condition for the echo cancellation amount. And tone input coefficient control means for
stopping the updating of filter coefficients.
[0039]
Here, when the tone type determination means determines that the far-end input signal is a tone
signal of a predetermined type, it also catches the end of the tone signal, and the tone input
coefficient control means determines The echo cancellation amounts before and after the timing
at which the tone type determination means determines the end of the tone signal are compared,
and according to the comparison result, the coefficients of the adaptive filter are reset and
coefficient updating is restarted, or the adaptive filter Preferably, coefficient update is forced to
resume without resetting the coefficient of.
[0040]
Further, when the tone type determination means determines that the far end input signal is a
tone signal of a predetermined type, the frequency component of the tone signal of the
predetermined type determined from the far end input signal to the adaptive filter is When the
first band rejection filter provided at the input stage of the adaptive filter to be removed, or the
tone type determination unit determines that the far-end input signal is a tone signal of a
predetermined type, to the addition unit It is preferable to have a second band rejection filter
provided at the input stage of the addition means for removing the frequency component of the
determined predetermined type of tone signal from the near-end input signal of
[0041]
15-04-2019
12
Furthermore, instead of the tone type determination means for determining whether the far-end
input signal is a predetermined type of tone signal, arbitrary tone determination means for
determining whether the far-end input signal is any tone property signal Can be applied.
[0042]
According to a third aspect of the present invention, in an echo canceller for removing an echo
by a hybrid circuit, which uses an adaptive filter for generating a pseudo echo, (1) it is
determined whether or not the far end input signal is a predetermined type of tone signal. A
predetermined type determined from the far-end input signal to the adaptive filter when the tone
type determination means and (2) the tone type determination means determine that the far-end
input signal is a predetermined type of tone signal And a first band rejection filter provided at the
input stage of the adaptive filter, for removing frequency components of the tone signal.
[0043]
Here, it is preferable to input the near-end input signal passed through the first band rejection
filter to the double talk detection means for detecting the double talk state of the far end talker
and the near end talker.
[0044]
In addition, it has addition means for subtracting the pseudo echo signal from the adaptive filter
from the near end input signal, and when the tone type determination means determines that the
far end input signal is a tone signal of a predetermined type, the above addition It is preferable to
have a second band rejection filter provided at the input stage of the adding means for removing
the frequency components of the determined predetermined type of tone signal from the near
end input signal to the means.
[0045]
A fourth aspect of the present invention is an echo canceler for removing echoes by a hybrid
circuit, which uses an adaptive filter for generating pseudo echoes, (1) detecting that the far-end
input signal is a tonal signal, and detecting the tonality signal The tone end judging means which
also catches the end of the polarity signal, (2) adding means for subtracting the pseudo echo
signal from the adaptive filter from the near end input signal, and (3) the signal outputted from
the adding means When the echo cancellation amount calculation means for calculating the echo
cancellation amount, and (4) the tone end determination means determines that the tone signal
has ended, the adaptive filter is performed according to the establishment of the predetermined
condition for the echo cancellation amount. And at the end of tone coefficient control means for
stopping the updating of the coefficients.
15-04-2019
13
[0046]
According to the present invention, it is possible to provide an echo canceller which can reduce
the feeling of echo in the subsequent speech even if a tone signal is input.
[0047]
(A) First Embodiment Hereinafter, a first embodiment in which an echo canceller according to the
present invention is applied to an echo canceller provided in a VoIP terminal will be described
with reference to the drawings.
[0048]
The echo canceller according to the first embodiment is made in view of the problems 1 and 2 of
the prior art described above, and it is based on the fact that the real harm in tone signal
misconvergence occurs most often at the start of call setting. It was done.
The echo canceller of the first embodiment is such that the initial value becomes appropriate
when the echo canceller reconverges.
[0049]
(A-1) Configuration of First Embodiment FIG. 5 is a block diagram showing the configuration of a
telephone communication system including the echo canceller of the first embodiment.
[0050]
The telephone communication system shown in FIG. 5 is, for example, a system for executing
communication by the existing telephones 1 and 9 via the IP network 5.
The telephones 1 and 9 are accommodated in the corresponding VoIP terminals 4 and 6,
respectively, and are connected to the IP network 5 via the VoIP terminals 4 and 6.
15-04-2019
14
[0051]
In FIG. 5, the base A side is the call receiving side and the base B side is the calling side, and FIG.
5 prevents the call signal arriving from the base A to the base B from flowing into the base A
from the base B as an echo. It is illustrated from the viewpoint of
[0052]
The VoIP terminal 4 has a hybrid circuit 2 connected to the telephone 1 by two wires, and a
digital / analog converter (D / A / D converter (D / A / D converter) that digital / analog converts
a digital call signal directed to the telephone 1 An A converter 11 and an analog / digital
converter (A / D converter) 3 for analog / digital conversion of an analog call signal via the
hybrid circuit 2 output from the telephone 1 are provided.
Of course, the VoIP terminal 4 may be equipped with an echo canceller.
[0053]
On the other hand, the VoIP terminal 6 includes the echo canceller 12 of the first embodiment,
the hybrid circuit 8 connected to the telephone 9 by two wires, and the digital call signal output
from the echo canceller 12 and directed to the telephone 9. A digital / analog converter (D / A
converter) 7 which converts digital / analog and outputs it to the hybrid circuit 8 and an analog
call signal which is output from the telephone 9 via the hybrid circuit 8 is analog / digital
converted and echo canceler 12 And an analog-to-digital converter (A / D converter) 10.
[0054]
FIG. 1 is a block diagram showing the detailed configuration of the echo canceller 12 of the first
embodiment provided at the position as described above, together with the surrounding
configuration, and the same or corresponding parts as in FIG. Show.
In the echo canceller 12 of the first embodiment, the input and output signals are digital signals.
[0055]
15-04-2019
15
In FIG. 1, the echo canceller 12 of the first embodiment has, as input and output terminals, a
reception input terminal Rin, a reception output terminal Rout, a transmission input terminal Sin,
and a transmission output terminal Sout.
The echo canceller 12 includes an adder 13, a double talk detector (DTD) 14, a tonality
determiner 15, a signal type determiner 16, a coefficient controller 17, and an adaptive filter
(ADF) 18.
[0056]
The adder 13 subtracts the pseudo echo signal y 'from the signal from the transmission input
terminal Sin to remove an echo component.
[0057]
The double talk detector 14 detects a talk state such as a double talk state based on the signal
from the reception input terminal Rin and the output signal of the adder 13.
The double talk detector 14 according to the first embodiment also clears the internal state by
the control signal from the coefficient controller 17.
[0058]
The tonality determiner 15 determines whether the signal input from the reception input
terminal Rin is a tonal signal.
A signal having tone characteristics refers to a repetitive waveform signal having a
predetermined cycle, such as various tones (for example, dial tones) for call control.
The tone property determination unit 15 may be of any type as long as it can determine whether
or not it has tone property, but, for example, the one described in JP-A-2000-295641 is applied.
obtain.
15-04-2019
16
[0059]
The signal type determination unit 16 determines which call control signal (a tone signal related
to call control) the signal is, when the tonality determination unit 15 determines that the signal
has tonality.
The determination method will be clarified in the operation explanation section.
[0060]
The coefficient controller 17 determines whether to update the filter coefficient in the adaptive
filter 18 according to the detection result of the double talk detector 14 or the determination
result of the signal type determination device 16, clears the filter coefficient, and the like. It is to
control.
[0061]
The adaptive filter 18 generates a pseudo echo signal y 'from an input signal from the reception
input terminal Rin and an internal filter coefficient, and supplies the pseudo echo signal y' to the
adder 13 as a subtraction input.
The adaptive filter 18 also updates the filter coefficients while also taking into account the echo
removal residual signal e.
However, the adaptive filter 18 clears the filter coefficient when the coefficient controller 17
instructs the filter coefficient to be cleared.
[0062]
(A-2) Operation of the First Embodiment Next, the operation of the echo canceller 12 of the first
embodiment will be described together with the operation of the telephone communication
system including the echo canceller 12.
15-04-2019
17
[0063]
When the calling telephone 9 lifts the handset to call the speaker (not shown) of the telephone 1,
a dial tone is output to the telephone 9.
The dial tone may be output from, for example, an IP gateway apparatus (not shown) between
the VoIP terminal 6 and the IP network 5 shown in FIG. 5 and may be input to the reception
input terminal Rin. It may be input to the reception input terminal Rin from a provided tone
generator (not shown).
In the following description, it is assumed that the signal is output from the IP gateway device
(not shown) and then input to the reception input terminal Rin.
The dial tone is a so-called "two-tone" sound that is heard when the handset is lifted, and in Japan
is a tone signal with a frequency of 400 Hz.
[0064]
The received signal (tone signal, speech signal, etc .; digital signal) input from the reception input
terminal Rin is input to the double talk detector 14, the tone property determination unit 15, the
reception output terminal Rout, and the adaptive filter 18.
[0065]
The received signal (digital signal) input from the reception input terminal Rin and output as it is
from the reception output terminal Rout is converted to an analog signal by the D / A converter
7, and the analog signal is converted to the telephone 9 via the hybrid circuit 8. Output to
However, part of the converted analog signal is reflected by the hybrid circuit 8.
The signal reflected by the hybrid circuit 8 is input to the transmission input terminal Sin via the
A / D converter 10.
15-04-2019
18
[0066]
A signal input to the transmission input terminal Sin is appropriately input to the double talk
detector 14, the adaptive filter 18, and the transmission output terminal Sout after an echo
component is appropriately removed via the adder 13.
A signal is output from the transmission output terminal Sout toward the IP network 5 side.
[0067]
It is determined by the tonality determiner 15 whether the signal input to the tonality determiner
15 is a tonal signal.
If the input signal is a tonal signal, the tonality determiner 15 outputs a tonal signal TON,
otherwise it outputs a tonal signal TOFF to the signal type determiner 16, and the input signal Is
output to the signal type determination unit 16 as it is.
If the signal input to the tonality determiner 15 is a dial tone, the tonality determiner 15 usually
determines that there is tonality.
[0068]
The signal type determination unit 16 analyzes the frequency of the input signal by, for example,
a known FFT (Fast Fourier Transform), a known zero cross method, or the like. In the following, it
is assumed that the zero crossing method, that is, the method of detecting the number of
crossings of the signal amplitude with the horizontal axis is used.
[0069]
Here, assuming that the digital sampling frequency is 16 kHz, the input signal is a 400 Hz signal
when the number T1 of samples of the period (interval) of the zero crossing in the same direction
15-04-2019
19
of the input signal waveform satisfies equation (4). It is determined that
[0070]
40-M1 ≦ T1 ≦ 40 + M1 (4) where 40 is 16 kHz / 400 Hz, and the input signal is in the same
direction (negative to positive or positive to positive) in the case of a tone signal such as a sine
wave signal It is the number of samples of the period of the zero crossing) of either negative
one).
M1 is a parameter that defines an allowable error range, and for example, 5 can be used. That is,
when the number T1 of samples from the zero crossing to the next zero crossing satisfies the
equation (4), it is determined that the input signal is a 400 Hz signal. With regard to signals of
other frequencies f, since the interval of the zero crossing becomes 16000 / f samples, an
equation for detecting the frequency may be defined similarly. The equation (5) relates to the
signal of the frequency f and corresponds to the equation (4). 16000 / f−M1 ≦ T1 ≦ 16000 / f
+ M1 (5) The signal type determination unit 16 receives the tonality presence / absence signals
TON and TOFF supplied from the tonality determination unit 15, and frequency information
obtained by frequency analysis. From this, it is determined whether the input signal is a known
call control signal, and when it is determined that it is a call control signal, the coefficient reset
signal RST is output to the coefficient controller 17 and the double talk detector 14 . For
example, the signal type determination unit 16 receives the tonal signal TON, and outputs the
coefficient reset signal RST when the frequency information obtained by the frequency analysis
has a predetermined frequency (for example, 400 Hz or 2100 Hz). For example, at the time of
dial tone input, since the predetermined frequency 400 Hz is provided, the signal type
determination unit 16 outputs the coefficient reset signal RST even if the type of dial tone is not
specified.
[0071]
FIG. 6 is an explanatory view showing the relationship between the call control signal (tone
signal) and the frequency. Dial tone (DT), ring back tone (RBT) busy tone and the like are 400 Hz
tone signals, and FAX communication start tone is 2100 Hz tone signals. Therefore, for example,
the signal type determination unit 16 stores 400 and 2100 in the internal memory, and reads
out the stored 400 or 2100 as the frequency f when applying the equation (5), (5) Make a
judgment according to the formula. Although two types of predetermined frequencies are shown
in FIG. 6, three or more types of predetermined frequencies may be described.
15-04-2019
20
[0072]
The coefficient controller 17 outputs a signal CL_T that clears the coefficient of the adaptive
filter 18 to the adaptive filter 18 when the coefficient reset signal RST is input. The coefficient
controller 17 does not output the clear instruction signal CL_T when the coefficient reset signal
RST is not input. The adaptive filter 18 receiving the clear instruction signal CL_T clears the filter
coefficient. The adaptive filter 18 re-executes the updating of the filter coefficients from the
initial state when the clear instruction signal CL_T does not arrive.
[0073]
Further, the clear instruction signal CL_T signal is also given to the double talk detector 14 as
described above. The double talk detector 14 detects double talk using ERLE, but clears the
internal state when the clear instruction signal CL_T is given.
[0074]
As described above, when the signal input to the reception input terminal Rin is a call control
signal such as DT, the coefficient of the adaptive filter 18 is cleared. In fact, when the calling
party speaker on the telephone 9 side dials and the human voice signal starts to flow through the
call path, the signal type determination unit 16 does not output the coefficient reset signal RST,
which causes adaptation The filter 18 can start estimation of the echo path from the desired
initial state of coefficient clearing and can perform echo path estimation quickly. Also, the double
talk detector 14 calculates ERLE by using the cancellation addition output of the adaptive filter
18 which has started to converge from the cleared desired initial state and the input signal from
the reception input terminal Rin, and correct double talk detection It can be implemented.
[0075]
(A-3) Effects of the First Embodiment According to the first embodiment, a call control signal
which is inevitably generated at the start of a call or the like is detected, the coefficient of the
adaptive filter is cleared, and double talk detection is performed. By clearing the internal state of
15-04-2019
21
the device, false convergence due to a tone signal (call control tone) can be prevented, and
updating of the echo canceller adaptive filter can be started from the desired initial state, and
echo path estimation can be performed quickly. It is possible to provide an echo canceller that
can be implemented and has less performance degradation due to double talk. Here, even if the
initial delay of the echo path is short, a good call without echo can be realized.
[0076]
(B) Second Embodiment Next, a second embodiment in which an echo canceller according to the
present invention is applied to an echo canceller provided in a VoIP terminal will be described
with reference to the drawings.
[0077]
The echo canceller of the second embodiment is made in view of the above-described problems
1, 2 and 3 of the prior art.
[0078]
In addition, in the case where the first embodiment is applied, it is made in view of the problem
that occurs when the tone detector (not shown in FIG. 1) is provided at the following position in
the VoIP terminal 6.
The first embodiment functions effectively if no tone detector is provided at the following
positions.
[0079]
When the echo canceller 12 is disposed in the VoIP terminal 6 as shown in FIG. 1 (the first
embodiment) described above, a tone not shown is shown on the output side of the transmission
output terminal Sout regardless of the tone property determination unit 19. Often a detector is
provided.
The reason is that a push button signal (DTMF signal: PB signal) or the like input from the
telephone set 9 is detected. However, if the output signal from the reception input terminal Rin is
15-04-2019
22
detected as a tonal signal as in the first embodiment, and the echo canceller is reset so as not to
remove the echo, the following inconvenience occurs. Do.
[0080]
For example, part of the dial tone signal input from the reception input terminal Rin passes
through the reception output terminal Rout, the D / A converter 7, the hybrid circuit 8 and the A
/ D converter 10 to the transmission input terminal Sin. It is input as an echo y1. As described
above, when the DTMF signal s1 which is a push button signal is output from the telephone set 9,
the echo y1 and the DTMF signal s1 are added together in the hybrid circuit 8, and the echo
canceler 12 is added via the A / D converter 10. , And output to the transmission output terminal
Sout without echo removal. As a result, the above-described tone detector is disturbed by the
extra echo signal y1 other than the DTMF signal, which causes inconveniences such as failure to
detect the push button number. Such problems occur frequently with VoIP terminals that do not
assume the above-mentioned problems caused by echo, and while picking up the handset and
listening to the dial tone (two tone), it is quite easy to recognize the phone number even if you
press the dial button Cause a phenomenon that can not be In the worst case, it may appear as a
serious problem such as being unable to make a call.
[0081]
In the second embodiment, in view of such a problem, in addition to the same effect as the first
embodiment, the telephone 9 can be used even when a signal such as a dial tone is input from
the reception input terminal Rin. It is also an object of the present invention to provide an echo
canceller which can detect a push button signal from.
[0082]
(B-1) Configuration of Second Embodiment FIG. 7 is a block diagram showing the detailed
configuration of the echo canceller 12A of the second embodiment together with the
surrounding configuration, and FIG. 1 described above according to the first embodiment. The
same or corresponding parts are denoted by the same reference numerals.
[0083]
In FIG. 7, an echo canceller 12A according to the second embodiment includes an adder 13, a
double talk detector 14, a tonality determiner 15, a signal type determiner 16A, and an adaptive
15-04-2019
23
filter 18 as in the first embodiment. Have.
On the other hand, the echo canceller 12A of the second embodiment does not include the
coefficient controller 17, and instead, band rejection filters (NF) 21 and 22 are provided, and the
signal type judgment unit 16A is also the first one. It performs processing somewhat different
from that of the embodiment of.
[0084]
The signal type determination unit 16A of the second embodiment determines not only the call
control signal but also the specific call control signal type of the input signal.
For example, it distinguishes and determines the same 400 Hz dial tone and busy tone. The
signal type determination unit 16A of the second embodiment collates the spectrum pattern
obtained by FFT or the like with the spectrum pattern of the reference, or collates the waveform
pattern of the input signal with the reference waveform pattern after aligning the dynamic range.
The specific type of call control signal is also determined, for example. When the input signal is a
tone signal and a specific type is obtained, the signal type determination unit 16A provides type
information to the band rejection filters 21 and 22.
[0085]
The band rejection filter 21 is provided between the transmission input terminal Sin and the
adder 16. On the other hand, the band rejection filter 22 is provided between the reception input
terminal Rin and the double talk detector 14 and between the reception input terminal Rin and
the adaptive filter 18.
[0086]
Each of the band rejection filters 21 and 22 blocks passage of frequency components determined
by the type information when the type information is given from the signal type determination
unit 16A. That is, the band rejection filters 21 and 22 are band rejection filters that can change
the pass rejection band and the like.
15-04-2019
24
[0087]
(B-2) Operation of Second Embodiment Next, the operation of the echo canceller 12A of the
second embodiment will be described focusing on differences from the first embodiment.
[0088]
When the signal type determination unit 16A of the second embodiment determines the specific
type of the input signal, the signal type determination unit 16A outputs the determination result,
that is, the signal KIND_TONE indicating the type of the signal to both band rejection filters 21
and 22.
The signal type determination unit 16A does not output anything when the input signal is a nontone signal.
[0089]
FIG. 8 is an explanatory diagram of the relationship between the determination result of the
signal type determination unit 16A and the output signal KIND_TONE. The signal type
determination unit 16A internally stores information representing the relationship as shown in
FIG. 8, and outputs a signal KIND_TONE having a value according to the determination result. In
the example of FIG. 8, KIND_TONE = 1 represents a dial tone. The value of the signal KIND_TONE
may follow any system as long as it represents the frequency and the type of signal, and is
different from FIG. 8, but for example, the numerical value 400 directly representing the
frequency is directly signal KIND_TONE It may be made an output value of Note that, even when
the input signal is a non-tone signal, the signal KIND_TONE having a value (for example, 0)
indicating that the signal is a non-tone signal may be output.
[0090]
Each of the bandstop filters 21 and 22 to which the signal type determination signal KIND_TONE
is input performs known bandstop filter processing on the input signal so as to block the
frequency corresponding to the signal KIND_TONE. Each band rejection filter 21, 22 respectively
15-04-2019
25
passes the input signal in the absence of the signal KIND_TONE.
[0091]
The band rejection filters 21 and 22 select one or more of a plurality of filters prepared in
advance, that is, band rejection filters that block a predetermined frequency (for example, 400
Hz) or other frequencies, and indicate them by the signal KIND_TONE A filter that blocks the
frequency may be implemented, or a frequency blocking filter may be adaptively configured each
time from the input frequency. Describing a method for realizing the band rejection filters 21 and
22 adaptively, for example, assuming that the blocking frequency indicated by the signal
KIND_TONE is 400 Hz and the sampling frequency is 16 kHz, the input signal is 40 (= 16 kHz /
400 Hz). 2.) A filter that removes a 400 Hz tone can be implemented by inverting the signal
delayed by samples (corresponding to a 400 Hz period) and adding it with the undelayed input
signal. The point is that any method may be used as long as a filter that removes a predetermined
frequency can be realized according to the information on the signal type and the frequency. As
described above, the band rejection filters 21 and 22 block passing of the call control signal to
the input signal type.
[0092]
An output signal from the band rejection filter 21 is input to the adder 13 and echo removal is
appropriately performed. On the other hand, the output signal from the band rejection filter 22 is
input to the double talk detector 14 and the adaptive filter 14 and appropriately used for
detection of the double talk state and formation of the pseudo echo signal y '.
[0093]
If the input signal from the reception input terminal Rin is determined to be a tone signal by the
signal type determination unit 16A, the tone signal is removed by the band rejection filter 22, so
the adaptive filter 18 and the double talk detector 14 It is the same if nothing is input. At this
time, even if the DTMF signal s1 is output from the telephone set 9 to the hybrid circuit 8, the
signal after being band-blocked by the band rejection filter 21 is input to the adder 13 as only
the signal s1. On the other hand, the tone signal from the reception input terminal Rin is blocked
by the band rejection filter 22, and nothing is input to the adaptive filter 18 and the double talk
detector 14. That is, at this time, the input signal x (n) to the adaptive filter 18 becomes 0, and
15-04-2019
26
the above equation (1) can be expressed as equation (6), and the coefficient update is
substantially stopped.
[0094]
hk (n + 1) = hk (n + 1) +0 (6) Further, in addition to such control of coefficient updating, the
denominator of the second term of the right side of equation (1) becomes 0 so that division does
not diverge The update is stopped when the sum of squares of the input signal which is the
denominator MIN MIN (dBm0) × product sum number (0 dBm0 reference value of dBm0 = 0
dBm0 is known ITU (International Telecommunication Union)-T ( Communications department)
As described in standard recommendation G.711). For example, although MIN = −50 dBm0 can
be applied, it is not limited to this, and may be set as appropriate.
[0095]
As described above, when the band rejection filters 21 and 22 are used, the band rejection filters
21 and 22 block the tone signal and the echo due to the tone from the reception input terminal
Rin. Even if time is executed, the actual update amount becomes 0, and the coefficient value itself
is not updated and does not change. Further, since the adaptive filter 18 whose reference input is
0 does not output anything, it does not output an extra pseudo echo y 'to the adder 13.
Therefore, the DTMF signal s1 input to the adder 13 is not subjected to any deterioration, only
the echo y1 is removed and output to the transmission output terminal Sout. The output signal
(DTMF signal) from the transmission output terminal Sout is correctly detected by a tone
detector (not shown).
[0096]
In addition, since there is no erroneous convergence in the tone signal, the internal state of the
double talk detector 14 is not disturbed, and double talk detection in normal voice following the
tone can be performed accurately to remove the echo.
[0097]
(B-3) Effects of Second Embodiment As described above, according to the second embodiment,
the following effects can be obtained in addition to the effects similar to the first embodiment.
15-04-2019
27
Even if there is a tone signal from the reception input terminal Rin, and the telephone set in the
receiver outputs a tone signal such as a DTMF signal in the opposite direction, the tone detector
using the echo canceller output correctly identifies the DTMF signal and makes a call. Since there
is no false convergence in the tone signal regardless of the magnitude of the initial delay of the
echo path, the internal state of the double talk detector is not disturbed, and the double talk
detection in the normal voice following the tone is also accurate. You can do it well and remove
the echo.
[0098]
(C) Third Embodiment Next, a third embodiment in which the echo canceller according to the
present invention is applied to an echo canceller provided in a private branch exchange
apparatus will be described with reference to the drawings.
[0099]
The echo canceller according to the third embodiment automatically escapes from the echo
cancellation impossible state when applied to a large scale system with echo path switching such
as PBX connection in which internal line switching occurs, and there is no echo. The present
invention is intended to achieve the realization of a call, and is made in view of the abovedescribed problems 1 to 4 of the prior art.
[0100]
(C-1) Configuration of Third Embodiment FIG. 9 is a block diagram showing a detailed
configuration of the echo canceller of the third embodiment, together with its surrounding
configuration, which is the first and second embodiments described above. The same or
corresponding parts as in FIGS. 1 and 7 according to FIG.
[0101]
In FIG. 9, an echo canceller 12B of the third embodiment is provided in a private branch
exchange (PBX) 6B having an extension transfer function.
Private branch exchange apparatus 6B accommodates, for example, telephones 9-1 and 9-2
before and after transfer, and goes to hybrid circuits 8-1 and 8-2 that perform 4-line 2-line
15-04-2019
28
conversion, and telephone 9-1 or 9-2. Digital / analog converter (D / A converter) 7 for digital /
analog conversion of digital signals, analog / digital converter (A / D conversion) for analog /
digital conversion of signals output from the telephone set 9-1 or 9-2. Device 10, the switch 30-1
for supplying the output signal of the D / A converter 7 to the hybrid circuit 8-1 or 8-2, and the
signal output from the hybrid circuit 8-1 or 8-2 for A / D. It has switch 30-2 given to the
converter 10, and the echo canceller 12B of 3rd Embodiment.
The switches 30-1 and 30-2 perform switching operations interlockingly.
[0102]
The echo canceller 12B according to the third embodiment includes an adder 13, a double talk
detector 14B, a tonality determiner 15B, a signal type determiner 16B, a coefficient controller
17B, an adaptive filter 18, band rejection filters 21 and 22, and a filter It has a state determiner
31 and an echo cancellation amount calculator (ACANC) 32.
[0103]
The adder 13, the adaptive filter 18 and the band rejection filters 21 and 22 are similar to those
of the second embodiment.
[0104]
The tonality determiner 15B of the third embodiment can also determine two types of composite
tone signals such as DTMF signals, in addition to the functions of the embodiments described
above.
[0105]
The signal type determination unit 16B of the third embodiment can also determine the types of
two types of synthetic tone signals such as DTMF signals, using the determination result of the
tone quality determination unit 15B.
The signal type determination unit 16B also determines the end of the tone signal.
[0106]
15-04-2019
29
The echo cancellation amount calculator 32 calculates the echo cancellation amount from the
input and output signals of the adder 13, that is, from the input signal having an echo component
and the output signal after the echo cancellation operation is performed. .
[0107]
The filter state determination unit 31 determines the state of the adaptive filter 18 from the
determination result of the signal type determination unit 16B and the calculation result of the
echo cancellation amount calculator 32.
[0108]
The coefficient controller 17B of the third embodiment determines control contents such as
stopping of coefficient updating of the adaptive filter 18 and clearing of the coefficient from the
determination result by the filter state determination unit 31.
[0109]
The double talk detector 14B according to the third embodiment controls the coefficient update
for the adaptive filter 18 in accordance with the control contents by the coefficient controller
17B and the detection result of the double talk in itself.
[0110]
(C-2) Operation of Third Embodiment Next, the operation of the echo canceller 12B of the third
embodiment will be described.
In the following, a situation will be described in which a transfer to switch from the called
telephone 9-1 to another called telephone 9-2 occurs.
Here, since the explanation is made before and after transfer occurs, it is assumed that the calling
telephone (not shown) and the called telephone 9-1 are already in a state of having already made
a call.
[0111]
15-04-2019
30
First, an audio signal output from a calling telephone set (not shown) is converted into a digital
signal by an A / D converter (not shown) at the front stage of the echo canceller 12B and input to
the reception input terminal Rin.
The signal input to the reception input terminal Rin is supplied to the band rejection filter 22 and
the tone characteristic determination unit 15B, and also output from the reception output
terminal Rout as it is to the D / A converter 7.
[0112]
The signal converted into an analog signal by the D / A converter 7 is inputted to the hybrid
circuit 8-1 via the switch 30-1 and is given to the telephone set 9-1 via the hybrid circuit 8-1. ,
Part of the signal is reflected by the hybrid circuit 8-1 to become an echo signal y1.
The echo signal y1 is supplied to the A / D converter 10 via the switch 30-2, converted again into
a digital signal, and input to the transmission input terminal Sin of the echo canceller 12B.
[0113]
The signal input to the transmission input terminal Sin is input to the band rejection filter 21.
Each band rejection filter 21 and 22 removes the tone characteristic signal according to the
output of the signal type determination unit 16B, and outputs the signal Sin_AC after tone
removal to the echo cancellation amount calculator 32, as described later. Output to the output
unit 13.
The adder 13 adds the signal Sin_AC after tone removal and the pseudo echo y 'to remove an
echo.
The output signal res from the adder 13 is input to the double talk detector 14B and the echo
cancellation amount calculator 32, and is sent out from the transmission output terminal Sout to
15-04-2019
31
a distant calling side telephone set (not shown).
[0114]
Here, the operation of the echo cancellation amount calculator 32 will be described. The echo
cancellation amount calculator 32 calculates the echo cancellation amount ACANC (n), for
example, according to equation (4). In the equation (4), n represents the nth calculated value.
[0115]
In equation (4), the square ratio of Sin_AC (n) and res (n) is directly calculated logarithmically,
but if emphasis is placed on the rough behavior of the signal, to monitor gradual changes in
characteristics As shown in the equation (5), Sin_AC and res may be collected by a predetermined
number of samples (for example, 160 samples), summed, and then the ratio of both may be
calculated logarithmically. Also, an absolute value may be used instead of a square. In the
equation (5), M is the number of samples to be averaged, and may be 160, for example, but it is
not limited to this number of samples.
[0116]
The point is that as long as the level ratio and power ratio of the signals before and after the
adder 13 are calculated, the calculation method by the echo cancellation amount calculator 32
may be any method. The echo cancellation amount ACANC (n) obtained by the equation (4) is
obtained by calculating the square of the signal before and after the adder 13 by logarithmic
ratio, and the output signal res from the adder 13 according to the convergence of the echo
canceler 12B. Becomes smaller, the value of the echo cancellation amount ACANC (n) becomes
larger. Although 0 may be applied as an initial value of the echo cancellation amount ACANC (n),
it is not limited to this. The echo cancellation amount calculator 32 inputs the calculated echo
cancellation amount ACANC (n) to the filter state determination unit 31.
[0117]
In this third embodiment, in order to calculate the amount of echo removal, the echo cancellation
15-04-2019
32
amount calculator 32 is used to calculate the echo cancellation amount using the signals before
and after the adder 13. An example is as follows. The signal may be taken from any one point
between the reception input terminal Rin and the reception output terminal Rout and the point
immediately after the adder 13 to calculate an echo attenuation amount including the
attenuation of the echo path.
[0118]
The filter state determiner 31 receives the echo cancellation amount ACANC (n) from the echo
cancellation amount calculator 32, and receives the bandstop filter control signal KIND_TONE
and the tone end signal TEND described later from the signal type determination device 16B. .
[0119]
Hereinafter, operations of the tone characteristic determination unit 15B and the signal type
determination unit 16B according to the third embodiment will be described.
[0120]
An output signal of the tone characteristic judging unit 15B is inputted to the signal type judging
unit 16B.
The tone quality determiner 15B of the third embodiment is configured to be able to determine a
DTMF signal, that is, two types of composite tone signals, in addition to the function of the tone
property determiner of the second embodiment.
The tonality determiner 15B is configured to divide the signal into two broad bands in advance
so that two types of composite tone signals can be determined. This is based on the fact that the
DTMF signal is generated by combining two kinds of tones, a "high group" composed of high
frequencies and a "low group" composed of low frequencies.
[0121]
In the third embodiment, the tonality determiner 15B separates the “low group” from the band
pass filter having a passband of 1000 Hz to 1700 Hz to separate the “high group”. A band
pass filter having a passband of 600 Hz to 980 Hz and a pass band of 0 Hz to 500 Hz for
separating "call control signal groups" near 400 Hz as described in the first and second
15-04-2019
33
embodiments The tonality determiner 15B determines the tone frequency based on the zero
crossing similar to that described in the first embodiment, for the signals after the band division
by the respective band pass filters. It is made to do.
[0122]
The signal type determination unit 16B of the third embodiment determines the signal type as in
the second embodiment described above with respect to the signals in the frequency band of the
“call control signal group” (see FIG. 8), “high As for signals having both "group" and "low
group" frequency bands, the type of signal is determined based on the relationship information
between the input, the determination result, and the output as shown in FIG. 10 stored therein.
Note that "call control signal group" and "high group" are detected, "call control signal group" and
"low group" are detected, "call control signal group" and "high group" In the case where “the
low group” and the “low group” are detected, the signal may not be a tonal signal, but it is
not limited thereto.
[0123]
When the signal type determination unit 16B determines the signal type, the signal type
determination unit 16B outputs the determination result, that is, the signal KIND_TONE
indicating the type of the signal to the band rejection filters 21 and 22 and the filter state
determination unit 31. The signal KIND_TONE is, for example, any value (number) for identifying
the signal and the signal type associated with each other as shown in FIGS. It may have a system
of values.
[0124]
In the example of FIG. 8 and FIG. 10, if KIND_TONE = 1, it indicates dial tone (DT), and if
KIND_TONE = 5, it indicates “1” of the DTMF signal.
[0125]
Each of the bandstop filters 21 and 22 to which the signal KIND_TONE is input from the signal
15-04-2019
34
type determination unit 16B applies known bandstop filter processing to the input signal so as to
block the frequency determined according to the signal KIND_TONE.
The behavior of the band rejection filters 21 and 22 is similar to that of the second embodiment,
but in the case of this third embodiment, two types of signals are removed when the signal
KIND_TONE indicates a DFMF signal. The points are different from those of the second
embodiment. For example, a filter for "high group" and "low group" determined according to the
signal KIND_TONE is selected from a plurality of band rejection filters prepared in advance for
"high group" and "low group", respectively. (These are cascaded) to perform band removal
according to the DTMF signal. The configuration method for blocking the band of the DTMF
signal is not limited to this method.
[0126]
In the case of the third embodiment, the signal type determination unit 16B also outputs the
signal KIND_TONE to the filter state determination unit 31. Further, the signal type
determination unit 16B outputs the tone end signal TEND to the filter state determination unit 31
when the signal from the tonality determination unit 15B changes from the tonality signal TON
to the no tonality signal TOFF.
[0127]
The filter state determination unit 31 operates as follows according to the output signals
KIND_TONE and TEND from the signal type determination unit 16B and the output signal from
the echo cancellation amount calculator 32.
[0128]
When the signal KIND_TONE is not input from the signal type determination unit 16B, that is,
when the input signal from the reception input terminal Rin does not have tone characteristics,
the filter state determination unit 31 receives the signal from the echo cancellation amount
calculator 32. The output signal is held while being updated at predetermined time intervals.
For example, update retention is performed every 20 ms, but the update interval is not limited to
this.
15-04-2019
35
[0129]
On the other hand, when the signal KIND_TONE is input from the signal type determination unit
16B, the filter state determination unit 31 operates by classifying the output signal of the echo
cancellation amount calculator 32 into conditions 1 to 4 as follows. .
[0130]
Condition 1: ACANC (n) ≧ E_ACANC dB Condition 2: ACANC (n) <E_ACANC dB Note that, for
example, 20 dB is applied as a threshold parameter E_ACANC that defines the conditions 1 and 2.
However, the value of the threshold parameter E_ACANC is not limited to this value.
[0131]
When the condition 1 is satisfied, in other words, when the amount of echo cancellation is large,
the filter state determination unit 31 outputs a signal ADP_STP prompting the coefficient
controller 17B to stop updating the coefficient of the adaptive filter 18. The coefficient controller
17B outputs the signal ADP_STP to the adaptive filter 18 via the double talk detector 14B to stop
the coefficient update of the adaptive filter 18.
[0132]
When the period of the tone signal ends and the signal TEND from the signal type determination
unit 16B is input to the filter state determination unit 31, the filter state determination unit 31
temporarily stops updating and holding the echo cancellation amount ACANC (n). , Wait for the
next echo cancellation amount ACANC update period to come. Then, when the echo cancellation
amount ACANC (n + 1) is newly calculated by the echo cancellation amount calculator 32 and is
output to the filter state determination device 31, the filter state determination unit 31 calculates
the echo cancellation amount ACANC (n + 1) and Compare with previously held ACANC (n).
15-04-2019
36
[0133]
That is, echo cancellation amounts ACANC (. Compare). (. ) Represents any n. Thereafter,
following Condition 1, determinations of Conditions 3 and 4 to be described later are executed,
and operations according to the respective results are executed.
[0134]
On the other hand, when the condition 2 is satisfied, in other words, when the amount of echo
cancellation is small, the filter state determination unit 31 does not output ADP_STP as a signal
prompting the coefficient controller 17B to stop updating the coefficient of the adaptive filter.
[0135]
The filter state determination unit 31 determines the following conditions 3 and 4 and operates
according to the results.
Although 3 dB can be applied as Δ1 in the conditions 3 and 4 and 0 dB as Δ2, conditions are
not limited to these.
[0136]
Condition 3: Δ2 <ACANC (n + 1) <ACANC (n) −Δ1 Condition 4: When ACANC (n + 1) <Δ2
Condition 3 is satisfied, in other words, although the tone property signal is input, the echo
cancellation amount is input When it is considerably smaller than before, the filter state
determiner 31 outputs the coefficient update promotion signal ADP_F to the coefficient
controller 17B.
[0137]
When the condition 4 is satisfied, in other words, when the echo cancellation amount becomes
negative by performing the echo removal operation, the coefficient reset signal ADP_RST is
output to the coefficient controller 17B.
[0138]
When neither Condition 3 nor Condition 4 is satisfied, the filter state determination unit 31 does
15-04-2019
37
not output anything to the coefficient controller 17B.
[0139]
Instead of Condition 3 above, a modified example of Condition 3 below may be applied.
The modification of condition 3 sets margin width to the upper and lower sides using ACANC (n)
instead of lower limit fixed value delta2.
[0140]
Modification of Condition 3: ACANC (n) −Δ3 <ACANC (n + 1) <ACANC (n) −Δ1 When the
coefficient reset signal ADP_RST is input from the filter state determiner 31, the coefficient
controller 17B detects double talk The reset signal RST is output to the device 14B.
The reset signal RST is applied to the adaptive filter 18 via the double talk detector 14B, and the
adaptive filter 18 resets the filter coefficients and then resumes coefficient updating.
Although both the double talk detector 14B and the adaptive filter 18 are reset here, only the
adaptive filter 18 may be reset.
[0141]
In addition, when the coefficient update promotion signal ADP_F is input from the filter state
determination unit 31, the coefficient controller 17B resets the double talk detector 14B but does
not reset the adaptive filter 18, and the coefficient update operation is forcibly performed.
Control to run.
[0142]
In the third embodiment, the coefficient controller 17B controls the adaptive filter 18 via the
double talk detector 14B. However, the coefficient controller 17B includes the adaptive filter 18
and the double talk detector 14B. Each of them may be controlled.
15-04-2019
38
[0143]
It will be described how the operation of the third embodiment solves the problem 4 in the prior
art.
Here, as described in the section of the problem, the description will be made according to the
combination of the convergence condition of the adaptive filter of the echo canceller at the
transfer time and the magnitude of the change in echo path characteristics before and after
transfer.
[0144]
In the most typical example where extension transfer occurs, the far end speaker (the calling
party speaker in FIG. 9) (not shown) performs an operation for transfer on the push button
(DTMF signal is output) After that, it is when the callee side telephone changes from the
telephone 9-1 to the telephone 9-2 by the transfer function of the private branch exchange (PBX)
6B.
That is, in this case, the push button for designating the telephone set 9-2 different from the
telephone set 9-1 connected first is operated to actually re-connect to another telephone set 9-2.
[0145]
(Extension Transfer Case A) The extension transfer case A is a case in which the convergence of
the adaptive filter 18 is insufficient and the response characteristics of the hybrid circuits 8-1
and 8-2 before and after transfer hardly change.
[0146]
Before transfer, some calls are made, and by voice the adaptive filter 18 is in the process of
convergence.
15-04-2019
39
Or, this applies to case A if some coefficient disturbance is received before the tone detection.
[0147]
When the convergence of the adaptive filter 18 is insufficient, the far-end speaker requests a
connection to a different telephone 9-2, for example, by operating a push button (inputting a
DTMF signal).
[0148]
The DTMF signal inputted to the reception input terminal Rin is inputted to the band rejection
filter 22 and the tone characteristic judgment unit 15B, and is outputted from the reception
output terminal Rout as it is and is given to the D / A converter 7.
[0149]
The tonality determiner 15B detects tonality.
In response to the output of the tonality determiner 15B, the signal type determiner 16B outputs
the signal KIND_TONE, and the band rejection filters 21 and 22 block the passage of the
frequency according to the signal KIND_TONE.
[0150]
As a result, since the DTMF signal is blocked by the band rejection filter 21 while the DTMF
signal is generated, the adaptive filter 18 and the double talk detector 14B are in the same state
as when the DTMF signal is not input. ing.
Therefore, even if the DTMF signal is input to the reception input terminal Rin for a long time,
the coefficients of the adaptive filter 18 are not disturbed, but the convergence of the coefficients
does not progress to the optimum state.
[0151]
15-04-2019
40
Case A is when the convergence of the adaptive filter 18 is insufficient, and the amount of echo
cancellation is small. Therefore, the condition 2 is satisfied (when the condition 1 is satisfied, it
will be described later). When the DTMF signal ends, the tone type end signal TEND is output
from the signal type determination unit 16B. The filter state determination unit 31 having
received this signal TEND compares the echo consumption amount ACANC (n) held last time with
the echo cancellation amount ACANC (n + 1) newly output from the echo cancellation amount
calculator 32. In the case A, since the characteristics of the hybrid circuit 8-1 and the
characteristics of the hybrid circuit 8-2 are almost the same, if the adaptive filter 18 performs
coefficient updating, ACANC (n + 1) 適 応 ACANC (n). . Therefore, the coefficient controller 17B
does not output anything.
[0152]
When the condition 1 is satisfied, the filter state determination unit 31 temporarily stops
updating the coefficients of the adaptive filter 18 via the coefficient controller 17B. However, the
fact that the condition 1 is satisfied is in front of the DTMF signal. It means that the
characteristics of the hybrid circuit could be estimated from the signal. In the case A, the
characteristics of the hybrid circuit 8-1 before and after transfer and the characteristics of the
hybrid circuit 8-2 are equal, so it is desirable that the adaptive filter 18 proceed with the
updating of the coefficient as it is. Does not output. As a result, the adaptive filter 18 and the
double talk detector 14B can proceed with the normal echo removal operation as desired, and
can proceed with the echo removal without deterioration of the echo before and after the
generation of the tone signal.
[0153]
(Internal Transfer Case B) In the internal transfer case B, the convergence of the adaptive filter
18 is insufficient, and the characteristics of the hybrid circuits 8-1 and 8-2 change before and
after the transfer.
[0154]
Even in the case B, the DTMF signal ends, the tone type end signal TEND is output from the
signal type determination unit 16B, and the filter state determination unit 31 receiving this signal
TEND newly adds the echo cancellation amount ACANC (n) previously held. The operation of
each part until the echo cancellation amount ACANC (n + 1) output from the echo cancellation
amount calculator 32 is compared is the same as in the case A.
15-04-2019
41
[0155]
When the signal KIND_TONE is input from the signal type determination unit 16B, the filter state
determination unit 31 determines that the condition 1 or 2 is satisfied based on the output from
the echo cancellation amount calculator 32.
Since convergence of the adaptive filter 18 is insufficient, in this case B, in most cases, the
condition 2 is satisfied (when the condition 1 is satisfied, it will be described later).
[0156]
Moreover, in this case B, the response characteristic of the hybrid circuit changes from the
characteristic of the hybrid circuit 8-1 which does not coincide with the characteristic of the
hybrid circuit 8-2 at the boundary of the transfer.
Even if the adaptive filter 18 tries to perform coefficient updating, the target of echo path
estimation has changed from the hybrid circuit 8-1 to the hybrid circuit 8-2. Naturally, the
pseudo echo generated by the adaptive filter 18 is inappropriate, and ACANC (n + 1) <ACANC (n).
In many cases, ACANC (n + 1) <0.
[0157]
Therefore, whether the determination result of the filter state determination unit 31 is the
condition 1 or the condition 2, the subsequent condition determination always becomes either
the condition 3 or the condition 4. Further, the double talk detector 14 B erroneously determines
the echo after the path change as the near end speaker signal according to the rapid
deterioration of the echo consumption, and judges the state as the double talk, and the coefficient
of the adaptive filter 18 Stop updating once.
[0158]
15-04-2019
42
As described above, the filter state determination unit 31 determines whether the change in the
echo cancellation amount ACANC before and after receiving the output TEND from the signal
type determination unit 16B matches the condition 3 or 4 and, first, the double talk The detector
14B is reset to cancel the erroneous determination of the double talk detector 14B. Next, the
filter state determination unit 31 controls the adaptive filter 18, but changes the control
according to the conditions 3 and 4 as follows.
[0159]
a) When the condition 3 is satisfied The fact that the condition 3 is satisfied indicates that the
echo cancellation amount ACANC is deteriorated within a predetermined allowable range.
Therefore, it is desirable to perform further convergence for optimal echo path estimation,
although some echo cancellation is possible. Therefore, the filter state determination unit 31
outputs the coefficient update promotion signal ADP_F to the coefficient controller 17B to update
the filter coefficient of the adaptive filter 18.
[0160]
b) When the condition 4 is satisfied If the parameter Δ2 in the condition 4 is, for example, 0 dB,
that the condition 4 is satisfied means that the echo cancellation amount ACANC is negative, that
is, the echo cancellation fails, rather the echo is emphasized. It shows that you are doing. In this
case, for optimal echo path estimation, it is desirable to discard the coefficients of the adaptive
filter 18 and execute the refocusing operation. Therefore, the coefficient controller 17B outputs
the coefficient reset signal ADP_RST to the adaptive filter 18 via the double talk detector 14B,
resets the filter coefficient, and restarts coefficient updating. In the third embodiment, although
the adaptive filter 18 is reset via the double talk detector 14B, the coefficient controller 17B may
directly reset the adaptive filter 18.
[0161]
Since the control of the adaptive filter 18 according to the conditions 3 and 4 is performed, even
if the characteristics of the hybrid circuit change before and after transfer, the echo canceller
12B promptly detects the characteristics of the newly connected hybrid circuit 8-2. Since it
follows again, it becomes possible to make a call with echo removed immediately.
[0162]
15-04-2019
43
(Extension Transfer Case C) The extension transfer case C is a case in which the convergence of
the adaptive filter 18 is sufficient and the characteristics of the hybrid circuit hardly change
before and after transfer.
[0163]
In this case C, the convergence of the adaptive filter 18 is sufficient, so the condition 1 is satisfied
in the filter state determination unit 31.
The estimated echo path can estimate the speech frequency band evenly, and can naturally
remove not only the speech signal but also the echo of the tonal signal.
On the other hand, since the echo removal residual res is also approximately 0, the coefficient
update amount in equation (1) is also 0, and even if a tonal signal is input, it is almost the same
as when the coefficient update has almost stopped.
[0164]
Exceptionally, if a DTMF signal is input for a long time (for example, one minute), the coefficients
of the adaptive filter 18 may be gradually destroyed, and in the case C, this may be prevented.
[0165]
As described above, the combination of the filter state determination unit 31, the signal type
determination unit 16B, and the band rejection filters 21 and 22 plays this role.
When condition 1 is satisfied in the state where the signal KIND_TONE is output from the signal
type determination unit 16B, the filter state determination unit 31 outputs the signal ADP_STP
prompting the coefficient controller 17B to stop updating the coefficient of the adaptive filter 18
Therefore, the coefficient update of the adaptive filter 18 is stopped, and after the transfer, since
the coefficient update is resumed, the echo can be removed without any problem.
[0166]
15-04-2019
44
Also in the transfer case C, the same operation is performed according to the conditions 3 and 4.
Thereafter, in the case C, it can be considered that the characteristics of the hybrid circuit before
and after the transfer do not change. Therefore, from the aspect of response characteristics, since
the hybrid circuit 8-1 and the hybrid circuit 8-2 are almost equal, the coefficients of the adaptive
filter 18 are not disturbed unless the DTMF signal input is so long, and before and after transfer
There is no deterioration of the echo cancellation amount ACANC. Therefore, there is no problem
in continuing the operation of the echo canceller 12B.
[0167]
Further, even if the DTMF signal is input for a long time, the filter state determination unit 31
receiving the output KIND_TONE of the signal type determination unit 16B prevents the
subsequent coefficient disturbance of the adaptive filter 18. If, by any chance, the determination
of the tone type is delayed, and the coefficient update stop of the adaptive filter 18 by the filter
state determiner 31 due to the input of the signal KIND_TONE is delayed, the coefficients of the
adaptive filter 18 are disturbed. Since the echo cancellation amount ACANC degrades at the
boundary, the filter state judgment unit 31 to which the signal TEND is input from the signal type
judgment unit 16B after the termination of the DTMF signal judges again on the conditions 3 and
4 of the echo cancellation amount ACANC. Thereafter, the same operation as in the cases A and B
is performed. That is, even in the worst case, after transfer, as in the cases A and B described
above, the adaptive filter 18 can be promptly reconverged to remove the echo.
[0168]
(Extension Transfer Case D) The extension transfer case D is a case in which the convergence of
the adaptive filter 18 is sufficient and the characteristics of the hybrid circuit change at the
transfer boundary.
[0169]
In this case, since the echo canceller 12B has converged sufficiently before transfer, Condition 1
is satisfied.
[0170]
Therefore, as described above, the coefficient controller 17B temporarily stops updating the
coefficient of the adaptive filter 18 by the output of the filter state determination unit 31, but
15-04-2019
45
after passing the DTMF signal, the condition 3 or 4 holds. The convergence of the adaptive filter
18 is resumed, and immediately after the transfer, the adaptive filter 18 can start reconvergence
to remove the echo.
[0171]
(C-3) Effects of Third Embodiment As described in detail above, according to the third
embodiment, there is a tone signal such as a dial tone from the reception input terminal, and the
accommodated telephone set is in the opposite direction. Even if the DTMF signal is output to
the, the tone detector using the output of the echo canceller can correctly identify the DTMF
signal and determine the telephone number to make a call, and even if there is a push button
operation, Even if the echo path is re-estimated quickly, there is little performance degradation
due to double talk, and the initial delay of the echo path is small, it is possible to quickly
reconverge to realize a good call without echo. In addition to such effects, even when applied to a
large-scale system in which an extension transfer or the like occurs, the echo removal
performance does not deteriorate due to the extension transfer, so excellent voice quality is
achieved. An echo canceller that can be provided can be realized.
[0172]
(D) Fourth Embodiment Next, a fourth embodiment in which an echo canceller according to the
present invention is applied to an echo canceller provided in a private branch exchange
apparatus will be described with reference to the drawings.
[0173]
The fourth embodiment is made in view of the fact that in human-to-human calls, human voice
may occasionally exhibit tonality.
[0174]
(D-1) Configuration of Fourth Embodiment FIG. 11 is a block diagram showing a detailed
configuration of the echo canceller of the fourth embodiment, together with its surrounding
configuration, compared with FIG. 9 according to the third embodiment. The same or
corresponding parts are indicated by the same reference numerals.
[0175]
In FIG. 11, an echo canceller 12C of the fourth embodiment is different from the third
embodiment in that the components of the tone property determiner 15B and the signal type
determiner 16B are replaced with a signal component determiner 40. Are the same as in the
15-04-2019
46
third embodiment.
[0176]
The signal component determination unit 40 forms the signals KIND_TONE and TEND based on
the input signal from the reception input terminal Rin, and the formation method will be clarified
in the operation explanation section.
[0177]
(D-2) Operation of Fourth Embodiment The fourth embodiment is different from the third
embodiment in only the operation of the signal component determination unit 40, so it will be
described. The explanation of is omitted.
[0178]
The signal component determination unit 40 determines whether the input signal is a wide-band
input signal (the input signal component is in a wide band, regardless of whether the input signal
is a known call control signal (for example, DT, DTMF, etc.) type or other. ) Or not.
The signal component determination unit 40 operates to determine, for example, a tone signal
such as 1500 Hz that is not a call control signal as a tone signal.
[0179]
Hereinafter, the determination method in the signal component determination unit 40 will be
described.
[0180]
The signal component determination unit 40 performs, for example, a fast discrete Fourier
transform (FFT) on the input signal to decompose it into individual single frequency components.
For example, an input signal is converted into frequency components using a 256-point FFT.
15-04-2019
47
Assuming that the sampling frequency is 16 kHz and an FFT of 256 points is used, each
frequency component can be calculated with 128 resolutions for the 0 to 8 kHz band of the input
signal to obtain the power spectrum P_f (k).
In the fourth embodiment, the power P_min_f of the smallest frequency min_f in each power
spectrum is regarded as the noise floor frequency power level, and when the following condition
5 is satisfied, it is regarded as “with frequency component”.
[0181]
Condition 5: P_f (k)> P_min_f + δf Here, δf is a determination offset value, and although 15 dB
can be applied, it is not limited to this value.
k represents the first frequency among the resolved frequencies and can take 128 values of 0 ≦
k ≦ 127.
[0182]
The signal component determination unit 40 counts the number of frequencies f (k) satisfying
the condition 5 with a built-in counter, and when the count result C_F is C_F <TH_VOICE, the
input signal is regarded as an arbitrary tone characteristic signal. The signal KIND_TONE is
output to the filter state determination unit 31.
For example, 4 can be applied as the threshold parameter TH_VOICE, but is not limited thereto.
[0183]
In the fourth embodiment, since it is not necessary to determine the type of tone signal, as in the
first to third embodiments, information for forming the signal KIND_TONE is stored in advance,
and the information is referred to. There is no need to do this, only the presence or absence of
the output of the signal KIND_TONE is a problem.
15-04-2019
48
[0184]
When it is necessary to assign a number to the signal KIND_TONE in relation to the
configurations of the band rejection filters 21 and 22 and the filter state determination unit 31,
for example, the numbers not applied in FIG. 8 or 10 described above are used. .
[0185]
In the case of the fourth embodiment, since the frequency of the tone signal to be detected is not
determined in advance, the band rejection filters 21 and 22 may be subjected to a filter operation
by providing a signal KIND_TONE including frequency information.
[0186]
When the signal component determination unit 40 changes from the state in which the tone
signal is detected according to the condition 5 to the state in which the tone signal can not be
detected as described above, the filter state determination unit 31 Output.
[0187]
(D-3) Effects of Fourth Embodiment According to the fourth embodiment, an error for an
arbitrary tone signal (for example, accidental human tone voice in a human-to-human call) not
classified as a call control signal Convergence and subsequent malfunction of the echo canceller
can be prevented.
[0188]
A phenomenon in which an arbitrary tone signal is input is, for example, when music is applied at
the same level as the speaker's voice as background sound of one or both speakers, or when the
speaker sings during a call, It occurs in very rare cases, such as when the individual's voice
frequency characteristics have a specific tonality.
And since the frequency of the tonal property which generate | occur | produced is not only the
frequency of a call control tone but is arbitrary, it can not refer to a known frequency table.
In the fourth embodiment, a signal component determination unit 40 is provided to detect the
tonality of the input signal even if the frequency is arbitrary, and after the detection result, as in
15-04-2019
49
the third embodiment, an adaptive filter and Since the double talk detector is controlled, the false
convergence of the adaptive filter is prevented, and even if the initial delay of the echo path is
small, the false talk state is restored promptly, and the double talk detector is also falsely judged
As a result, it is possible to realize an echo canceller that can return and continue to remove echo
without deteriorating it.
[0189]
(E) Fifth Embodiment Next, the fifth embodiment in which the echo canceller according to the
present invention is applied to an echo canceller provided in a private branch exchange
apparatus will be described with reference to the drawings.
[0190]
The echo canceller of the fifth embodiment is intended to make the hardware scale smaller than
that of the third and fourth embodiments.
[0191]
(E-1) Configuration of Fifth Embodiment FIG. 12 is a block diagram showing a detailed
configuration of the echo canceller of the fifth embodiment, together with its surrounding
configuration, with FIG. 9 according to the third embodiment. The same or corresponding parts
are indicated by the same reference numerals.
[0192]
12, the echo canceller 12D of the fifth embodiment is different from the third embodiment in
that the band rejection filters 21 and 22 are not provided, and the function of the signal type
determination unit 16D is the third embodiment. It's different from something that is somewhat
different.
[0193]
(E-2) Operation of Fifth Embodiment Next, an operation of the echo canceller 12D of the fifth
embodiment which is different from that of the third embodiment will be described.
[0194]
A state in which transfer occurs between the called telephones 9-1 and 9-2 will be described.
15-04-2019
50
Since the explanation is made before and after the transfer occurs, it is assumed that the calling
side telephone and the called side telephone 9-1 which are not shown are already in a state of
talking.
[0195]
When the output of the tonality determiner 15B changes from the tonality signal TON to the no
tonality signal TOFF, the signal type determination unit 16D outputs a signal TEND indicating the
end of the tonal signal to the filter state determination unit 31.
[0196]
When the output TEND from the signal type determination unit 16D is received, the filter state
determination unit 31 determines the above-described conditions 1 to 4 as in the third
embodiment, and the coefficient controller according to the determination result The signal
ADP_F, ADP_STP or ADP_RST for coefficient control is output to 17 B to control the coefficient
update of the adaptive filter 18.
During such control, the adaptive filter 18 follows the coefficient update control of the double
talk detector 14B.
[0197]
Hereinafter, although the operation at the time of transfer will be described, it is not necessary to
finely divide the variation of the response characteristic of the hybrid circuit and the good or bad
of the convergence state of the adaptive filter in this fifth embodiment. So I will explain together.
[0198]
Before transfer, some calls are made, and by voice the adaptive filter 18 is in the process of
convergence.
At this time, a far-end speaker who is not shown requests a connection to a different telephone
15-04-2019
51
by inputting a DTMF signal or the like.
[0199]
The DTMF signal input to the reception input terminal Rin is input to the tonality determiner
15B, passes through the reception output terminal Rout as it is, and is input to the D / A
converter 7.
[0200]
At this time, the tonality determiner 15B detects tonality for the input signal, but the signal type
determiner 16D does not output anything at this time.
Thus, with no filter state determiner 31, adaptive filter 18 continues coefficient updating and
removes echoes from the tone.
At this time, although the coefficients of the adaptive filter 18 are different from the
characteristics of the echo path as described above, the tones are updated to coefficients that can
be erased well.
Therefore, the output of the adder 13 becomes small, and ERLE becomes to have a large value.
[0201]
Next, when the tone signal ends, the signal type determination unit 16D outputs the tone end
signal TEND to the filter state determination unit 31.
[0202]
The filter state determination unit 31 having received this signal TEND compares the echo
cancellation amount ACANC (n) held last time with the echo cancellation amount ACANC (n + 1)
newly output from the echo cancellation amount calculator 32.
15-04-2019
52
In this case, assuming that the characteristic of the hybrid circuit 8-1 is the same as the
characteristic of the hybrid circuit 8-2, regardless of the convergence state before and after the
tone end signal TEND, adaptation is performed at the time the signal TEND is input. The
coefficients of the filter 18 are conveniently disturbed to remove only the tone signal, and
condition 2, condition 3 or condition 4 is satisfied depending on the amount of echo cancellation
in the subsequent normal audio signal.
At this time, re-update of the double talk detector 14B and the adaptive filter 18 is executed
based on the condition 3 or 4 as in the third embodiment.
[0203]
The condition 1 is satisfied that the coefficient is not disturbed so much by the DTMF signal
before and after the input timing of the signal TEND, and the echo can be sufficiently removed
even by the subsequent voice signal. This occurs only in the extension transfer case C where the
convergence of the adaptive filter 18 in advance is sufficient and the characteristics of the hybrid
circuit do not change.
At this time, it is desirable that the adaptive filter 18 proceed with the updating of the coefficient
as it is, and since the coefficient controller 17B outputs anything, the adaptive filter 18 and the
double talk detector 14B perform normal echo removing operation as desired. , And before and
after the tone signal generation can proceed with echo cancellation without deterioration of the
echo.
[0204]
(E-3) Effects of Fifth Embodiment According to the fifth embodiment, although the disturbance of
the tone signal can not be prevented in advance, the adaptive filter coefficient is reset in most
cases when the transfer occurs. Reset the double talk detector so as not to disturb the coefficient
update of restart, and reconvergence is possible regardless of the magnitude of the initial delay
of the echo path. Since it is not necessary to return to start reconvergence and prepare a plurality
of band rejection filters in advance, it is possible to provide an echo canceller with a small hard
scale.
[0205]
(F) Other Embodiments Of the technical ideas of the above-described embodiments, those which
15-04-2019
53
can be combined may be combined to constitute an echo canceller.
[0206]
In the third and fourth embodiments, the echo cancellation amount calculator for calculating the
square ratio of the input before and after the adder is used to calculate the echo cancellation
amount, but the output signal of the adder and the reception input terminal Rin are used.
Alternatively, the output signal of the adder and the input signal of the reception output terminal
Rout may be used.
In this case, the amount of attenuation of the echo path itself may be added to Δ1 and Δ2
described in the third and fourth embodiments.
[0207]
In the fourth embodiment, the signal component determination unit 40 uses the method of using
the FFT to detect an arbitrary tone frequency, but any method other than the FFT may be used as
long as it can detect an arbitrary single tone. You may use it and it is not limited to the method of
using FFT.
[0208]
Furthermore, although the second to fourth embodiments show two band stop filters, at least an
input stage of the adaptive filter may be provided with a band stop filter.
[0209]
Furthermore, in the description of each of the above-described embodiments, each component
has been described as an image configured by hardware, but of course some of the components
may be realized by software.
[0210]
In the first and second embodiments, the echo canceller is mounted on the VoIP terminal, and in
the third to fifth embodiments, the echo canceller is mounted on the private branch exchange. Of
course, the device on which the canceller is mounted is not limited to these.
15-04-2019
54
[0211]
It is a block diagram which shows the detailed structure of the echo canceller of 1st Embodiment.
It is explanatory drawing (1) of the subject 2 of a prior art.
It is explanatory drawing (2) of the subject 2 of a prior art.
FIG. 16 is a block diagram for explaining Problem 4 of the prior art.
FIG. 1 is a block diagram showing a configuration of a telephone communication system
including an echo canceller of a first embodiment.
It is an explanatory view of judgment operation of a signal classification judging device of a 1st
embodiment.
It is a block diagram which shows the detailed structure of the echo canceller of 2nd
Embodiment.
It is explanatory drawing of the relationship between the determination result of the signal
classification determination device of 2nd Embodiment, and an output.
It is a block diagram which shows the detailed structure of the echo canceller of 3rd
Embodiment.
It is explanatory drawing of the relationship between the determination result of the signal
classification determination device of 3rd Embodiment, and an output.
It is a block diagram which shows the detailed structure of the echo canceller of 4th
15-04-2019
55
Embodiment.
It is a block diagram which shows the detailed structure of the echo canceller of 5th
Embodiment.
Explanation of sign
[0212]
12, 12A, 12B, 12C, 12D: echo canceler, 13: adder, 14, 14B: double talk detector, 15, 15B: tone
property determiner 16, 16, 16A, 16B, 16D: signal type determiner, 17 , 17B: coefficient
controller, 18: adaptive filter, 21, 22: band rejection filter, 31: filter state determiner, 32: echo
cancellation amount calculator, 40: signal component determiner.
15-04-2019
56
Документ
Категория
Без категории
Просмотров
0
Размер файла
81 Кб
Теги
description, jp2007174343
1/--страниц
Пожаловаться на содержимое документа