close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2006279157

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 JP2006279157
An object of the present invention is to provide a howling suppression system that reliably
suppresses howling without degrading the sound quality of the entire acoustic system.
SOLUTION: An amplification unit 101 of an acoustic microphone system which is weak in
howling but amplifies the output of an acoustic microphone P1 whose sound quality is superior
to the vibration microphone P2, and a vibration microphone P2 which is strong in howling but
inferior in sound quality to the acoustic microphone P1. The output of the amplification unit 102
of the vibration microphone system that amplifies the output of the signal is synthesized. Further,
a howling detection unit 7 connected to the acoustic microphone system to detect the occurrence
of howling, an output of the acoustic microphone system synthesized by the synthesis unit based
on the output of the howling detection unit 7 and an output of the vibration microphone system
And a combining ratio changing unit 8 that changes the combining ratio of When the howling
detection unit 7 detects the occurrence of howling by analysis by the f analysis unit 6, the
combination ratio change unit 8 decreases the amplification factor of the acoustic microphone
system and increases the amplification factor of the vibration microphone system. [Selected
figure] Figure 1
Howling suppression system
[0001]
The present invention relates to a howling suppression system with good sound quality.
[0002]
In howling, in a system consisting of a microphone, an amplifier, and a speaker, the audio output
from the speaker is fed back to the microphone to form a closed loop, and certain frequency
15-04-2019
1
components in the feedback signal diverge and grow with time It is an unstable phenomenon
caused by
[0003]
Conventionally, in order to suppress howling, an equalizer or a notch filter has been used.
[0004]
As an equalizer, there is a method of manually adjusting a parametric equalizer in a concert hall
or the like.
[0005]
Moreover, in order to suppress howling, there is a dynamic equalizer which detects or predicts
howling by frequency analysis and dynamically adjusts the equalizer (see, for example, Patent
Documents 1 to 5).
The dynamic equalizer mainly includes howling detection means, a notch filter, and the like.
[0006]
Also, bone conduction microphones (see, for example, Patent Documents 6 and 7).
) And vibration pickup microphones (hereinafter collectively referred to as “vibration
microphones”.
There is a loud-sounding system which uses and, for example, a nose is attached and detected
(for example, refer nonpatent literature 1).
These howling suppression systems take advantage of the fact that the vibration microphone is
less sensitive to air noise, and therefore the sensitivity to the howling is more stable than the
acoustic microphone. JP-A-7-147700 JP-A-8-223684 JP-A-8-149593 JP-A-5-137191 JP-A-5-
15-04-2019
2
236588 JP-A-5-199577 JP-A-5-276587 Japan Proceedings of the Acoustical Society of Japan
"Consideration on the system for recognizing internal conductive sound" March 2004 1-10-20
p539
[0007]
However, manual adjustment of the equalizer is very skillful, and if the filter frequency is always
constant during the operation of the acoustic system, there is a problem in the ability to adapt
when the howling frequency changes during the operation time. there were.
[0008]
Further, as shown in FIG. 4 of the patent document 1, by using a notch filter that lowers the gain
of a specific frequency using a dynamic equalizer, it is possible to reliably stop howling.
However, there is a drawback that the sound quality of the entire acoustic system becomes
unnatural during operation of the howling suppression system since the signal in the filtered
band is dropped. Such documents do not disclose what compensates for this missing frequency
band.
[0009]
On the other hand, the vibrating microphone has a disadvantage that the sound quality is
relatively not so good.
[0010]
Then, an object of this invention is to provide the howling suppression system which suppresses
howling reliably, without impairing the sound quality of the whole sound system.
[0011]
In the present invention, means for solving the above-mentioned problems are configured as
follows.
[0012]
(1) An acoustic microphone system for amplifying an output of an acoustic microphone, a
15-04-2019
3
vibration microphone system for amplifying an output of a vibration microphone, a synthesis
unit for synthesizing an output of the acoustic microphone system and an output of the vibration
microphone system, the acoustic microphone A howling detection unit connected to the system,
which detects or predicts howling occurrence, and changing a synthesis ratio of an output of the
acoustic microphone system synthesized by the synthesis unit based on an output of the howling
detection unit and an output of the vibration microphone system And a combining ratio changing
unit.
[0013]
With this configuration, it is possible to dynamically detect or predict the occurrence of howling.
Also, by dynamically combining the output of the vibrating microphone system in which the
occurrence of howling does not easily occur, while suppressing the occurrence of howling, since
the sound of the acoustic microphone system having excellent sound quality can be input while
the howling does not occur. The sound quality of the whole sound system can be improved.
[0014]
Here, the combining ratio changing unit uniformly amplifies the respective amplification factors
of the acoustic microphone system and the vibration microphone system in the entire frequency
band (hereinafter, the amplification factors are referred to as “gains”.
And the frequency band in which the howling occurs may be adjusted.
[0015]
(2) The combining ratio change unit operates when the howling is detected by the howling
detection unit, and the amplification factor of the acoustic microphone system is determined
after the detected time until the howling stops. Lower and increase the amplification factor of the
vibration microphone system.
[0016]
According to this configuration, when the howling is not detected, the output of the acoustic
15-04-2019
4
microphone system having good sound quality but weak to howling is mainly used, and when the
howling is detected, the output of the acoustic microphone system is lowered and the sound
quality is poor. By raising the output of the strong vibration microphone system, it is possible to
achieve both suppression of howling and improvement of the sound quality of the entire acoustic
system.
[0017]
(3) After the time when the howling is predicted by the howling detection unit, the combining
ratio change unit reduces the amplification factor of the acoustic microphone system while the
howling prediction state continues, and the vibration is the vibration. Increase the amplification
factor of the microphone system.
[0018]
With this configuration, it is possible to achieve both suppression of howling and improvement in
the sound quality of the entire acoustic system, as in (2).
Since howling prediction can predict howling earlier than the configuration of (2), a further
howling suppression effect can be expected.
[0019]
(4) The synthesis ratio change unit adjusts the amplification factor of each of the acoustic
microphone system and the vibration microphone system so that the output sound from the
synthesis unit has the same volume as that before adjustment of the synthesis ratio change unit.
It is characterized by being.
[0020]
With this configuration, the volume becomes equal before and after the operation of the
combining ratio changing unit, and the sound quality becomes natural.
Here, “equal volume” means that when the gains of the acoustic microphone system and the
vibration microphone system are adjusted uniformly over the entire frequency band, the sum of
15-04-2019
5
the gain of the flat part of the frequency is the operation of the combining ratio changing unit
Make the volume equal before and after.
On the other hand, in the case of adjusting at least one frequency band, the frequency
characteristic and the gain in the band including the adjusted frequency band are made equal
before and after the operation of the combining ratio changing unit.
[0021]
(5) The combining ratio changing unit is configured to obtain amplification factors of the
acoustic microphone system and the vibration microphone system only in a predetermined band
centered on the howling frequency obtained by analyzing the signal frequency of the acoustic
microphone. It is characterized in that it is to be adjusted.
[0022]
According to this structure, since the combining ratio changing unit can mix the inputs of the
vibration microphone system having poor sound quality only with respect to the howling
frequency, it is possible to minimize the deterioration of the sound quality.
Here, the howling frequency is a frequency at which howling is expected to occur, which is
detected as howling is occurring.
The constant band refers to a band of several hundreds Hz before and after the howling
frequency.
[0023]
According to the present invention, howling is reliably suppressed without impairing the sound
quality of the entire acoustic system.
[0024]
The howling suppression system according to the first embodiment will be described.
15-04-2019
6
The system is configured as shown in (A) and (B) of FIG.
[0025]
As shown in FIG. 1, the vibrating microphone P2 is used by attaching it to the ear canal, throat,
lip, head, chest and the like.
The amplification unit 101 amplifies the input signal 1 obtained from the acoustic microphone
P1.
On the other hand, the amplification unit 102 amplifies the input signal 2 obtained from the
vibration microphone P2. The audio signals amplified through the amplifying unit 101 and the
amplifying unit 102 are converted into an acoustic microphone system output signal 11 and a
vibrating microphone system output signal 12, and are synthesized as a synthesized speech input
3 by the synthesizing unit 13 and passed through a power amplifier (not shown). The sound is
output from the speaker 4.
[0026]
The amplification factors of the above-described amplification units 101 and 102, that is, the
gains G1 and G2 are described as “CTL” in the control unit 5 (FIG. 1A). Control signals C1 and
C2 are output, and are adjusted based on this. As illustrated in FIG. 1B, the control unit 5
performs an f analysis unit 6 that analyzes the frequency, a howling detection unit 7 that
performs howling detection or howling prediction, and a combining ratio change unit 8 that sets
gains of G1 and G2. It consists of In the present system, the howling detection unit 7 performs
howling detection and does not perform prediction thereof (howling prediction is performed by
the system of the second embodiment as described later).
[0027]
When the speech signal x (t) which is a function of time t is input, the f analysis unit 6 analyzes
15-04-2019
7
the frequency of the speech signal x (t) at predetermined time intervals ΔT, and the power
spectrum for the frequency f P (f) is output to the howling detection unit 7. This predetermined
interval is performed at an interval small enough to detect howling. As shown in FIG. 1A, the
frequency analysis by the f analysis unit 6 is performed exclusively based on the input signal 1 of
the acoustic microphone P1. This is because the acoustic microphone P1 has high sensitivity to
air noise, so the input from the acoustic microphone P1 is likely to cause howling, and it is easy
to detect howling early. On the other hand, since the vibration microphone P2 has low sensitivity
to air noise, it is difficult to detect howling early from the input signal 2 from the vibration
microphone P2.
[0028]
The power spectrum calculation is, of course, also calculated based on time domain data of a
fixed time interval before time t, and the time interval is for detecting a band in which the
howling frequency may occur. Sufficient time.
[0029]
When it is detected in the howling detection unit 7 that howling occurs, a signal to the effect that
the howling is being performed is referred to the combining ratio change unit 8 (hereinafter
referred to as “howling detection signal”).
If it is detected, F = 1, otherwise it is set to F = 0. And howling frequency is output. The
combining ratio changing unit 8 receives the signal such as howling, and outputs the control
signal C1 so as to lower the amplification factor with respect to the input 1 of the acoustic
microphone P1, that is, the amplification factor G1 of the amplification unit 101. At the same
time, the control signal C2 is output to increase the amplification factor for the input signal 2
from the vibration microphone P2, that is, the amplification factor G2 of the amplification unit
102. As described above, since the acoustic microphone P1 has a large influence of howling and
the vibration microphone system has a small influence of howling, the combination ratio of the
acoustic microphone system is decreased to increase the synthesis ratio of the vibration
microphone system, thereby preventing howling. It is. At this time, the synthesized voice input 3
amplified by G1 and G2 and synthesized by the synthesizing unit 13 is output from the speaker 4
at a constant volume. In the system of the first embodiment, when the howling detection signal is
received from the howling detection unit 7, the combining ratio of the gains of G1 and G2 is
uniformly changed for all frequencies regardless of the howling frequency. In this way, the
amplification factors of G1 and G2 are simply changed, and there is no need to manipulate the
frequency domain, so the combining ratio changing unit 8 can be configured simply.
15-04-2019
8
[0030]
FIG. 2 shows an operation flow of the howling detection unit 7 shown in FIG. 1 (B). The flow is
performed at predetermined time intervals (in the first embodiment, the time interval is ΔT). The
howling detection means is a known technique, and is described, for example, in FIG. 7 of Patent
Document 1 and FIG. 2 of Patent Document 4.
[0031]
At S21 in FIG. 2, the peak of the frequency is analyzed based on the result of analysis by the f
analysis unit 6 for every ΔT. For example, it is assumed that the f analysis unit 6 analyzes as
shown in FIG. Here, the horizontal axis f is a frequency, and the vertical axis P (f) is a power
spectrum. In this case, the peak has a power spectrum of size L1, and the frequency of this peak
is fH.
[0032]
In S22 of FIG. 2, the average level of the power spectrum is calculated. The average level may be
calculated, for example, by averaging the power spectrum squared by frequency. In the example
of FIG. 3, the following description will be given assuming that L2 is the average level of the
power spectrum.
[0033]
In S23 of FIG. 2, the difference between the peak L1 and the average level L2 is obtained, and if
the value (L1-L2) is equal to or greater than a predetermined threshold TH, Y is obtained, and the
process proceeds to S24. If it is less than the predetermined value TH, it is N, and the process
proceeds to S25.
[0034]
15-04-2019
9
In S24 of FIG. 2, it is detected that howling occurs, and the howling detection signal is sent to the
combination ratio changing unit 8 as F = 1. The flow then ends.
[0035]
In S25 of FIG. 2, it is detected that howling has not occurred, and the howling detection signal is
set to F = 0 and is sent to the combining ratio changing unit 8. The flow then ends.
[0036]
FIG. 4 shows an operation flow of the combining ratio changing unit 8. As described above, the f
analysis unit 6 described in FIG. 1B analyzes the frequency of the audio signal x (t) at
predetermined time intervals ΔT, and the howling detection unit 7 analyzes the frequency as
shown in FIG. The howling detection signal F is sent to the combining ratio changing unit 8 at
each time interval ΔT. The combining ratio changing unit 8 also performs the flow of FIG. 4
described below for each time interval ΔT.
[0037]
In S41 of FIG. 4, the combining ratio change unit 8 determines whether the howling detection
signal sent from the howling detection unit 7 is F = 1. As described above, if howling is detected,
F = 1 is input to the combining ratio changing unit 8, the determination in S41 is Y, and the
process proceeds to S42. If it is not F = 1, it becomes N, and it progresses to S45.
[0038]
In S42 of FIG. 4, it is determined whether the howling detection signal before ΔT from the
current time is F = 1. If F is not 1, in S43, the combining ratio changing unit 8 sends the control
signal C1 so that the gain of G1 becomes G1-ΔG with respect to the gain G1 before howling
occurs (FIGS. )reference.). Next, in S44 of FIG. 4, the combining ratio change unit 8 outputs the
control signal C2 so that the gain G2 before howling occurs becomes G2 + ΔG (see FIGS. 1A and
1B). However, ΔG indicates a fluctuation gain of a predetermined magnitude. If F = 1 in S42 of
15-04-2019
10
FIG. 4, the flow is Y, and the flow ends. That is, when the howling detection signal is F = 1 from
the present time to the present time by ΔT, the gains determined in S43 and S44 are maintained
before ΔT.
[0039]
The order of S43 and S44 may be earlier, or may be performed simultaneously.
[0040]
In S41 of FIG. 4, when the howling detection signal sent from the howling detection unit 7 is not
F = 1, that is, F = 0, it is determined in S45 whether the howling detection signal before ΔT from
the current time is F = 1. .
If F = 1 before ΔT from the current time, it becomes Y, and in S46, the gains of G1 and G2 are
restored. That is, howling occurred at F = 1 before ΔT from the current time, but when the
occurrence of howling stops at F = 0 now, the gains of G1 and G2 are restored and the flow
thereafter ends. Do. In S45, if the howling detection signal before ΔT from the current time is
not F = 1, it is originally the case that the howling does not occur, so the original gain is
maintained.
[0041]
In summary, the gain increase / decrease in FIG. 4 changes the gain when the howling detection
signal is changed, but maintains the gain when the howling detection signal F is not changed.
Specifically, at the time when F = 1 of the howling detection signal is detected for the first time at
a specific frequency, the determination in S41 is Y and the determination in S42 is N, so the gain
is adjusted in S43 and S44. If F = 1 also at the time when ΔT has elapsed from that time, the
determinations in S41 and S42 are both Y, so the gains set in S43 and S44 are maintained. The
maintenance of this gain continues as long as F = 1. After that, if the howling disappears, the
determination in S41 is N, the determination in S45 is Y, and the gains of G1 and G2 are restored
in S46. Furthermore, when howling does not occur originally, both S41 and S45 become N, and
the original gain before the occurrence of howling is maintained.
[0042]
15-04-2019
11
In this way, the combining ratio of the gain of the acoustic microphone system and the vibration
microphone system can be varied only when howling occurs, so that the input of the vibration
microphone system, which is inferior in sound quality, is not excessively mixed. . Therefore, by
dynamically combining the output of the vibrating microphone system in which occurrence of
howling does not easily occur, the generation of howling can be suppressed, and while the
occurrence of howling does not occur, sound of the acoustic microphone system with excellent
sound quality can be input. The sound quality of the entire sound system can be improved.
[0043]
In S46, the increase or decrease of the gain ΔG performed in S43 and S44 is not immediately
returned to the original state, but as the time interval ΔT elapses, the gain increase or decrease
ΔG of the gains G1 and G2 gradually becomes the original state. A return process (not shown)
may be performed.
[0044]
Next, the howling suppression system according to the second embodiment will be described.
In the system of the second embodiment, FIG. 1 and its description are generally applicable to
the second embodiment. However, the system of the second embodiment is different only in
predicting howling. Therefore, the howling detection unit 7 is referred to as the howling
prediction unit 7 and will be described below.
[0045]
Although the howling detection unit 7 of the system of the first embodiment outputs the howling
detection signal F indicating whether or not howling occurs, the howling prediction unit 7 of the
system of the second embodiment predicts howling and howling. The prediction signal F is
output. If the growth of howling is confirmed, F = 2, if the decrease of howling is confirmed, F =
1, and if the howling does not grow nor attenuates and is stable, then F = 0.
[0046]
15-04-2019
12
As described above, when the howling has already grown, the system of the first embodiment
detects howling using the predetermined threshold TH. On the other hand, the howling
prediction according to the second embodiment predicts howling from the time variation of the
level of the peak frequency. Then, a howling frequency is defined from this time variation, and
howling is predicted based on whether or not the howling frequency increases with time.
[0047]
FIG. 5 shows an operation flow of howling prediction in the present system according to the
second embodiment. The howling prediction unit 7 operates the flow at each time interval ΔT.
However, ΔT is extremely short, approximately several microseconds, with respect to ΔT
according to the system of the first embodiment.
[0048]
In S51 of FIG. 5, the howling frequency is calculated from the time variation of the level of the
peak frequency, and the howling frequency is stored in the storage unit (not shown).
[0049]
In S52 of FIG. 5, the howling frequency before time ΔT stored in the storage unit is compared
with the howling frequency calculated in S51 to determine whether the howling frequency is
increased.
If it is increasing, it becomes Y and the process proceeds to S53. If not, the result is N, and the
process proceeds to S54.
[0050]
In S53 of FIG. 5, the howling prediction unit 7 predicts howling is growing, and outputs F = 2 to
the combining ratio changing unit 8.
[0051]
15-04-2019
13
In S54 of FIG. 5, it is determined whether the howling frequency is reduced by the same
comparison as in S52.
If it is decreasing, it becomes Y and proceeds to S55. If not decreased, the result is N, and the
process proceeds to S56.
[0052]
In S55 of FIG. 5, the howling prediction unit 7 outputs F = 1 to the combining ratio change unit 8.
This indicates that the howling frequency is decreasing.
[0053]
In S56 of FIG. 5, the howling prediction unit 7 outputs F = 0 to the combining ratio change unit 8.
This indicates that the howling frequency of a particular frequency is neither increasing nor
decreasing, and is stable.
[0054]
After any of S53, S55, and S56 of FIG. 5, the operation flow of FIG. 5 ends.
[0055]
FIG. 6 shows an operation flow of the combining ratio changing unit 8 according to the present
system.
The combining ratio changing unit 8 performs the operation of the flow for each time interval
ΔT.
[0056]
15-04-2019
14
In S61 of FIG. 6, it is determined whether the howling prediction signal F output from the
howling prediction unit 7 is F = 2. If F = 2, the result is Y, and the process proceeds to S62. If it is
not F = 2, it becomes N, and it progresses to S64.
[0057]
In S62 of FIG. 6, the combining ratio changing unit 8 reduces the control signal C1 (see FIG. 1) so
that the gain G1 of the acoustic microphone system is reduced by ΔG from the gain before it is
predicted that howling will occur. Output).
[0058]
In S63 of FIG. 6, the combining ratio changing unit 8 increases the control signal C2 (see FIG. 1)
so that the gain G2 of the vibration microphone system is increased by ΔG from the gain before
it is predicted that howling will occur. Output).
[0059]
In S64 of FIG. 6, the combining ratio changing unit 8 determines whether the howling prediction
signal F output from the howling prediction unit 7 is F = 1, that is, the howling frequency
decreases. If F = 1, the result is Y and the process proceeds to S65. If F is not 1, it becomes N, and
the operation flow of FIG. 6 ends.
[0060]
In S65 of FIG. 6, the combining ratio changing unit 8 outputs the control signal C1 to increase
the gain G1 of the acoustic microphone system by ΔG. At S66 in FIG. 6, the control signal C2 is
output so as to reduce the gain G2 of the vibration microphone system by ΔG.
[0061]
15-04-2019
15
After S63 and S66 of FIG. 6, the flow of FIG. 6 ends.
[0062]
The order of S62 and S63 may be earlier, or may be simultaneously performed.
In addition, either S65 or S66 may be preceded by one, or may be performed simultaneously.
[0063]
Further, although not shown in FIG. 6 for the sake of simplicity, in the second embodiment, as in
FIG. 4 of the system of the first embodiment, the howling prediction signal output by the howling
prediction unit 7 is F = 0 or F When changing from = 1 to F = 2, the combining ratio changing
unit 8 increases or decreases the gain at S62 and S63, and when the howling prediction signal
changes from F = 0 or F = 2 to F = 1. By the steps S65 and S66, the increase or decrease of the
gain is restored. Therefore, when there is no change in the howling prediction signal, the gain is
not changed and maintained in S62, S63, S65, and S66. An embodiment in which the gain is
increased or decreased as the time interval ΔT elapses will be described in the third embodiment
below.
[0064]
Next, a howling suppression system according to the third embodiment will be described. The
system is an application of the second embodiment, and differs from the second embodiment
only in the operation of the combining ratio changing unit 8. In the system of the third
embodiment, the sound quality is improved by managing the gain more finely than the system of
the second embodiment. In the present system, as shown in FIG. 7, the combining ratio changing
unit 8 manages the gains of G1 and G2 in several stages. That is, in FIG. 7, when howling is
growing, the gain G1 of the acoustic microphone system is decreased by ΔG at the first time t. At
the same time, the gain G2 of the vibrating microphone system is increased by ΔG. Thereafter, if
howling growth is observed even after a predetermined time interval Δt elapses, the gain G1 of
the acoustic microphone system is decreased by ΔG, and the gain G2 of the vibration
microphone system is increased by ΔG. Furthermore, the gain is changed as time passes, and
nΔt (n is a positive integer). If the growth of howling is observed even after lapse of time, the
15-04-2019
16
gain G1 of the acoustic microphone system is reduced by nΔG, and at the same time, the gain G2
of the vibration microphone system is increased by nΔG. When the howling frequency turns to
decrease at time t + (n + 1) Δt due to such change in the gains of G1 and G2, an increase or
decrease of ΔG is added to the gains G1 and G2, contrary to the case of growth. How to restore
G1 and G2 to their original condition before howling growth. In the process, when the howling is
stable without growth or decrease, as shown at time t + mΔt (m is an integer) in FIG. 8, the gain
set at t + (m−1) Δt is maintained. After that, when the state in which the howling frequency is
decreasing continues, the gains G1 and G2 are returned to the original state before the howling
growth, and if howling growth is not predicted in the process, FIG. As shown at time t + kΔt (k is
an integer), G1 and G2 are returned to the state of the original gains G1 and G2 before
occurrence of howling is predicted.
[0065]
By changing the gains of G1 and G2 like this, the howling frequency can be managed more finely.
That is, it is possible to increase the ratio of the gain of the acoustic microphone system with
high sound quality while minimizing the howling tendency, and to minimize the ratio of the gain
of the vibration microphone system with poor sound quality.
[0066]
The fluctuation gain ΔG can be made variable according to the growth degree of howling. For
example, it is conceivable to increase the fluctuation gain with the passage of time, or to make it
variable according to the slope of the rise of the howling frequency and the speed of the rise.
[0067]
Next, a howling suppression system according to a fourth embodiment will be described. The
fourth embodiment is different from the first to third embodiments in that a notch filter for
reducing a specific frequency as shown in FIG. 9 is used. On the other hand, in the first to third
embodiments, when howling occurs or is predicted, the gains G1 and G2 are uniformly increased
or decreased.
[0068]
15-04-2019
17
The howling suppression system according to the fourth embodiment is shown in FIG. As shown
in FIG. 8A, the input signal from the acoustic microphone P1 is amplified by the amplification
unit 101, but a notch filter 9 whose specific frequency is reduced (see symbol 9 in FIG. 9
described later). ) To generate an acoustic microphone system output signal 11. On the other
hand, the input signal from the vibration microphone P2 is amplified by the amplification unit
102, but a filter in which the specific frequency of the spectrum opposite to that of the notch
filter 9 of G1 is emphasized (see reference numeral 10 in FIG. Hereinafter, it is called "inverse
filter". ) To generate a vibrating microphone system output signal 12. Then, the signals 11 and
12 are combined by the combining unit 13 and output to the speaker 4.
[0069]
Here, FIG. 9 is a diagram showing the relationship between the frequency f and the power
spectrum P in the system when filtering is performed by the notch filter 9 and the inverse filter
10. The horizontal axis represents frequency f, and the vertical axis represents power spectrum
P. The notch filter 9 is a filter in which the power spectrum P becomes small in the vicinity of a
specific frequency fH as shown in FIG. The inverse filter 10 is a filter in which the power
spectrum P becomes large in the vicinity of the specific frequency fH.
[0070]
With such filters 9 and 10, the output of the acoustic microphone system can be lowered only at
a specific frequency causing howling, so that howling can be properly suppressed. In addition,
since the acoustic microphone system is used for frequencies other than the specific frequency,
there is no need to increase the combined distribution of the vibration microphone system whose
sound quality is inferior to the necessary quality, so that the deterioration of the sound quality
can be reduced. Furthermore, for a specific frequency that causes howling, since the volume is
reinforced with a vibrating microphone system that is strong in howling, it is possible to prevent
the unnaturalness of the sound due to the lack of the signal in that band.
[0071]
FIG. 8B shows the configuration of control unit 5 (indicated as "CTL" in the figure). Although the
15-04-2019
18
control unit 5 controls the operation of the notch filter 9 by giving control signals C1 and C2 as
shown in FIG. 8A, the combining ratio changing unit 8 in the control unit 5 merely changes the
gain. Instead, the parameters of the notch filter 9 and the inverse filter 10 are set.
[0072]
A plurality of frequencies may be applied to apply the notch filter. It is not limited to the notch
filter as long as it is a filter that lowers a specific frequency.
[0073]
Further, the fourth embodiment does not contradict the first to third embodiments except that
the notch filter 9 is used. Therefore, the description of the first to third embodiments can be
applied as it is to the operations other than the operation of the combining ratio changing unit 8
and the control signals C1 and C2. In particular, the howling detection means, the calculation
method for howling prediction, FIG. 2, FIG. 3, FIG. 5 and their description are applicable to the
fourth embodiment. Further, with regard to the control signal of the combining ratio changing
unit 8, particularly, FIG. 4 according to the first embodiment, FIG. 6 according to the second
embodiment, FIG. 7 according to the third embodiment, and descriptions thereof The fluctuation
of the gain of the frequency fH at the center of a specific frequency band that raises and lowers
the gain by the notch filter 9 is applied as ΔG.
[0074]
It is a block diagram of the howling suppression system concerning a 1st embodiment. It is an
operation | movement flowchart of the howling detection part of the system which concerns on
1st Embodiment. It is a figure which shows an example of the analysis result of f analysis part,
and the howling detection method based on the system of 1st Embodiment. It is an operation |
movement flowchart of the synthetic | combination ratio change part of the system which
concerns on 1st Embodiment. It is an operation | movement flowchart of the howling estimation
part of the system which concerns on 2nd Embodiment. It is an operation | movement flowchart
of the synthetic | combination ratio change part of the system which concerns on 2nd
Embodiment. It is a figure which shows the state of the howling of the system concerning 3rd
Embodiment, and the gain which a synthetic | combination ratio change part should designate. It
is a block diagram of the system concerning a 4th embodiment. It is a related figure of frequency
15-04-2019
19
f and power spectrum P of a notch filter and an inverse filter.
Explanation of sign
[0075]
1, 2-input signal 3- synthetic speech input 4- speaker 5- control section 6-f analysis section 7howling detection section 8-synthesis ratio change section 9-notch filter 10-inverse filter 11acoustic microphone system output signal 12 -Vibration microphone system output signal 13Synthesis unit 101-Amplification unit 102-Amplification unit G1-Gain G2-Gain C1-Control signal
C2-Control signal P1-Acoustic microphone P2-Vibration microphone
15-04-2019
20
Документ
Категория
Без категории
Просмотров
0
Размер файла
31 Кб
Теги
jp2006279157, description
1/--страниц
Пожаловаться на содержимое документа