close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JPS60176313

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 JPS60176313
[0001]
Industrial Application Field of the Invention The present invention relates to a howling
suppressor that prevents howling that occurs in loudspeakers used in conference houses and
lecture halls and the like, which is configured of Nubica or the like. The configuration of the
conventional example and its problems The conventional howlink suppressing apparatus will be
described with reference to FIG. (1) is an input terminal, (2) to (4) are band pass filters, (5) to (7)
are measuring circuits that periodically measure the magnitude of a signal, and (8) is a frequency
with a certain bandwidth Divided (comparison circuit that compares the size of No. 8; (9) is a
control circuit that controls the attenuator; (IG-soda is an attenuator that attenuates the
frequency-divided signal, cao is an adder, Q? is an attenuator that attenuates the entire signal,
and ?? is an output terminal. First, an input signal is divided into a plurality of frequency bands
by band pass filters (2) to (4), and the output is branched into two. One of the branched signals is
input to measurement circuits (5) to (7), the magnitude of each signal is periodically measured,
and input to the comparison circuit (8). In this comparison circuit (8), with respect to the divided
signals of each band at each point of time, the comparison of the signal magnitudes with the
bands adjacent to both of that band is performed, and the difference in the magnitude of the
signal is a threshold The detection of the band which is over is performed. When this band
detection is performed, the control circuit (9) controls an attenuator corresponding to the band
in which the detection is performed among the attenuators ati-ua, and applies a certain
attenuation to the attenuator. Be Then, the outputs of the respective attenuators ?1 to (b) are
added by an adder (1 ?) and inputted to the attenuator Q41. When the number of bands
detected by the comparison circuit (8) reaches a certain number, this attenuator Q is controlled
by the control circuit (9) and given a certain amount of attenuation. With such a configuration,
detection and suppression of occurrence of a link are performed. FIG. 2 is a flow chart showing a
hack link processing algorithm of the conventional hacking suppressor. In this flowchart, i is the
15-04-2019
1
number of a plurality of frequency division bands, Ei is the signal size in l band, gI is an
attenuator corresponding to l band, G2 is the sum of attenuations applied to the attenuators, Si,
gi is the difference between the signal size of 111 and the adjacent bands of that band, and TH is
the how-link detection threshold considered to generate a hack in that band if G is a value
greater than G; Amount of attenuation applied to the attenuator gi when it is detected that H has
been generated, Hn is the number of bands in which the total amount of attenuation applied to
the attenuator is 02, m is the number thereof, gt is the total signal An attenuator that attenuates
the magnitude 3 indicates G3.
For each of Ei (i =] to n: n is the number of bands) taken in periodically, check the number Hn of
bands for which the attenuation provided to the attenuator gi is 62 as far as possible. If the
enemy has reached a certain number m, the attenuation amount G3 is given to the attenuator gt
for attenuating the magnitude of the entire signal, and the processing is completely terminated. If
Hn is less than m, start processing from i @, and check whether attenuation of attenuator g1
corresponding to l band has reached 62, and if G2 has been reached, processing in l band is
Move processing to the next band without doing. When the attenuation amount of the attenuator
g1 does not reach 62, the signal of the l band is compared with the amplitude and the signal
magnitudes Ei-x and Ei + ? of both adjacent bands of the band, If a difference 5iji is detected and
either value exceeds a threshold 1H, it is detected that howlink has occurred in that band i, and a
certain amount of attenuation G in the attenuator gi corresponding to that band is detected. Add
? to move to the processing of the next band i 11. Then, when this process is performed in all
bands i = 1 to n, the same process is applied to Ei (i = I = n) taken in at the next time point. And
processing is continued until the number of Hn becomes m. The hacking suppression process is
performed by using such a hacking detection and suppression algorithm. As described above,
although two conventional examples are shown, in the conventional configuration and algorithm,
since the continuous signal is observed in the conventional configuration and algorithm,
particularly when there is no voice, there is no one in the room for howling prevention
processing. When it is done (it is very effective to call such a processing method 1-preset J in the
following, it is very effective, but if there is voice (if the speaker actually speaks)) There is + jJ
ability to misrecognize occurrence of hacking, and in fact, it is used as a preset type, and there is
a condition that processing is not performed when voice is present. SUMMARY OF THE
INVENTION The present invention solves the above-mentioned conventional drawbacks, and is
not limited to only Prescent, but by detecting a silent interval of voice even when a speaker
actually exists and an axial voice is input, An object of the present invention is to provide a noise
reduction device capable of detecting and suppressing the occurrence of hacking. In order to
achieve the above-mentioned object of the invention, the Klink suppression apparatus of the
present invention comprises band dividing means for dividing an incoming signal into signals of
a plurality of bands and O whole bands having a certain frequency bandwidth, and this band
Splitting means for splitting a plurality of signals frequency-divided by a bandwidth which is in
the splitting means into two, and one divided ? + J signal split by this branching means and the
15-04-2019
2
size of the double of the entire band] Measuring means for measuring in a static manner,
comparing means for comparing the magnitudes of the signals measured by the measuring
means, and first attenuation means for respectively attenuating the other divided signal branched
by the 111 branching means A valving means for summing the output signals of the first
damping means, a second damping means for damping the output signal of the summing means,
and a ratio soft result by the ratio soft means; And the second #, ?A configuration in which
example ? and control means for controlling the.
Description of Embodiments Hereinafter, an embodiment of the present invention will be
described based on the drawings. FIG. 3 is a circuit diagram of a howlink suppressor according to
an embodiment of the present invention, in which 0Q is an input terminal, 0.eta..about.01 is a
band pass filter, and (7) .about. Is a periodic signal magnitude. (C) is a comparison circuit that
compares the magnitude of the signal to detect the silent section of the voice and the band in
which the howling occurred, (B) controls the entire control of ~ to ?Circuit, (A) Attenuator
Attenuates the size of the frequency bandwidth 7'l 14 Attenuator (2) is an adder, Attenuator
Attenuates the overall signal magnitude A 0 Is an output terminal. Signals input from the input
terminal (IQ are frequency-divided by a plurality of band pass filters ?? to OI. Then, the outputs
of the respective band pass filters a? to 4 are branched into two, and one signal is attenuated to
+ a (e) to (d), and the other (+ # 4 j ? is added to the measuring circuit 4 to 4 respectively) It is
input. In addition, the input signals of all bands can be threatened to the measuring circuit. The
signals input to the measuring circuit 4 to (i) are measured for the magnitudes of the respective
signals at predetermined intervals, and are taken into the comparing circuit (iii). In this
comparison circuit, the magnitudes of the signals of all bands taken at periodic intervals within
the time are compared at regular time intervals to detect the moment when the magnitude of the
signal becomes the smallest. By doing this, it is possible to detect a silent section of voice. Then,
the magnitudes of the frequency-divided signals at the detected time point are compared, and for
each band, the difference between the magnitudes of the signals of both adjacent bands is
calculated, and the value exceeds a certain threshold If it is detected, it detects that snowfall has
occurred in that band. Then, the control circuit controls the attenuator corresponding to the band
detected as being generated by the noise among the attenuators (7) to ?, and the signal passing
through the band is attenuated by a fixed amount. Such processing is repeated at regular time
intervals, and even when meat and hack links occur in the same zone, hacking detection is
performed in the same manner as described above, and a certain amount of attenuation is added
to the attenuator corresponding to that zone. , The signal passing through that band is
attenuated. Then, when the total number of attenuators added to this attenuator reaches a
certain value, hacking suppression is performed, and when the number of bands for which the
amount of attenuation has reached a certain value is hacking, After the suppression process is
completed, the attenuator (7) is controlled to attenuate the entire signal level by a fixed amount,
and the hacking suppression process is completely finished.
15-04-2019
3
FIG. 4 shows a flowchart of a hacking processing algorithm of the hack-link suppressing
apparatus in the present embodiment. In this flowchart, Et represents the magnitude of the signal
of the entire band measured at dl- at one time point, t represents a certain time point within a
fixed time, and is from 1 to X time. E ? ti was divided into multiple frequency bands at one time!
Indicate the magnitude of the signal in the band, and let i be from 1 to n band. Etin indicates the
minimum value of the signal size of the entire band in a fixed time 1 =] to X, and the subscript
theory indicates the point in time. g i is an attenuator corresponding to i * # :, G 2 is the sum of
attenuations given to the attenuators, Si, S ? i is the difference in the size of the sign with the
band of both adjacent p to that of i band , TH is a detection threshold, G1 is the amount of
attenuation added to the attenuation Wgi when it is detected that howling has occurred, Hn is the
number of bands in which the sum of the amounts of attenuation applied to the attenuator has
become 02. m is the number thereof, g j is an attenuator for attenuating the magnitude of the
entire signal, and G 3 is the amount of attenuation. Compare the magnitude Et of the signal in the
entire band, which is periodically captured 1 to t = 1 to X within a fixed time, and determine
when the magnitude of the signal is minimum E, Y, voice Detect silent sections of Then,
comparison of the magnitude E ? ? (i = I ~n) of the frequency-divided signal at that time is
started. The number of bands In for which the total of the attenuations given to the attenuator g1
reaches 62 is examined, and if the number is a certain number m, the amount of attenuation G3
is reduced to the attenuator that attenuates the magnitude of the entire signal. To complete the
process completely. If Hn is less than m, start processing from i band, check whether attenuation
of attenuator gi corresponding to 1 band has reached 02, and if G2 has been reached, processing
in i band is not performed, Process the next band. Also, if gl does not reach 62, the ratio softness
of the signal size E ? ?-?, I and IiI + 1 of both adjacent bands of the band is performed with
respect to the signal size of the i band. If the difference between them Sit and Sl exceeds a
threshold value "1" H, it is considered that a hack link has occurred in that band, a certain
amount of attenuation G1 is added to the attenuator gi corresponding to that band, and Move to
processing of the band. When the above processing is performed in the entire band (i = 1 to n),
again, the size Et (t = 1 to X) of the size No. 16 of the entire band in the next predetermined time
and each frequency-divided signal The same process is performed using the size Eti (i =] to n) of.
Thus, the process is continued until the HnO number becomes In. As described above, the silent
interval of the voice can be detected by observing the size of 1g in the whole band for a certain
period of time and detecting the time when the magnitude of the signal becomes minimum. And
by comparing the magnitude of the signal in each frequency band only in the silent section of the
voice, it becomes possible to prevent misrecognition between the voice and the noise even in the
presence of the voice, and In the case where a human being is P1, it is possible to suppress the
occurrence of the hacking, and also when it is used as a fret, it is a phenomenon in which the
howlink increases monotonically. The algorithm fully enables howlink processing. As described in
detail in the invention, according to the present invention, the occurrence of hacking can be
15-04-2019
4
detected and suppressed in any case, even in the presence of voice and when used as a preset
type. Its industrial utility value is extremely large.
[0002]
Brief description of the drawings
[0003]
1 is a circuit diagram of the conventional hacking suppressor, FIG. 2 is a flow chart of a hacking
processing algorithm of the conventional hacking suppressor, and FIG. 3 is a circuit block of the
hacking suppressor in one embodiment of the present invention FIG. 4 is a flow chart of a hack
link processing algorithm of the same hack-link suppressor.
Q Q и и и Input terminal, ?? ~ o I и и и Band pass filter, (4) ~ и и и и и и и и и и и и и и и и и comparison circuit,
(C) и и и control circuit, (E) ~ threat иии Attenuator, (4) иии Adder, иииии Attenuator, G (1) иии Output
terminal agent Yoshiru Morimoto Hiroshi Fig. 2 Fig. 4
15-04-2019
5
Документ
Категория
Без категории
Просмотров
0
Размер файла
15 Кб
Теги
description, jps60176313
1/--страниц
Пожаловаться на содержимое документа