close

Вход

Забыли?

вход по аккаунту

JP2007251801

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 JP2007251801
An audio signal processing apparatus, an audio signal processing method, and an audio signal
processing program capable of easily obtaining a superdirective signal in a built-in microphone
of a video camera. SOLUTION: A plurality of acoustic signals are inputted from a plurality of
acoustic channels, and signal levels in a predetermined period of the plurality of acoustic signals
are detected by a plurality of level detection means 8, and the smallest level is among the
detected level values. A signal having a level value is selected by the selection unit 6 at
predetermined intervals, and the selected signal is band-limited by the band limitation unit 7 to
output a signal having different directivity patterns obtained from a group of adjacent
microphones. Abstract: An acoustic signal processing apparatus, an acoustic signal processing
method, and an acoustic signal processing program capable of strongly extracting only an AND
region from a plurality of acoustic signals by minimum value selection processing are provided.
[Selected figure] Figure 1
Acoustic signal processing apparatus, acoustic signal processing method, and program for
acoustic signal processing
[0001]
The present invention relates to an audio signal processing apparatus, an audio signal processing
method, and an audio signal processing program, and in particular, an audio such as a zoom
microphone which is less susceptible to wind noise due to superdirective characteristics that can
be built into small electronic devices such as video cameras. The present invention relates to a
signal processing device, an acoustic signal processing method, and a program for acoustic signal
processing.
04-05-2019
1
[0002]
Conventionally, as an acoustic sound collector equivalent to an optical zoom lens or telephoto
lens such as a camera, a parabolic sound collector, an acoustic phased-tube directional
microphone, an array microphone, etc. can be mentioned. However, it is not suitable for being
installed in small electronic devices such as home video cameras.
In addition, image processing is performed by applying matrix processing to audio signals
obtained from a plurality of low-order directional microphones such as a unidirectional
microphone, and further selecting a signal of the minimum level after dividing into a plurality of
bands and recombining bands. Patent Document 1 discloses a microphone in which the
correlation coefficient of audio signals extracted from a plurality of microphones is calculated in
order to emphasize the voice from the center direction of the plane, and the input audio signal is
enhanced according to the magnitude of the correlation coefficient. ing.
[0003]
In addition, in order to suppress or enhance an acoustic signal localized near the center, the
similarity of the level phase between the channels of the stereo signal is determined, or the
stereo phase difference component between the channels of the stereo signal is determined. An
acoustic signal processing apparatus and method thereof are disclosed in Patent Document 2.
[0004]
Furthermore, in order to realize a microphone with superdirective characteristics, a method using
an acoustic phase tube or a microphone array, from a microphone of low order sound pressure
gradient to a secondary sound pressure gradient (consistent with secondary sound pressure
gradient, secondary directivity) Patent Document 3 discloses a method of generating a
microphone of the present invention by calculation, and, for example, a microphone device and a
sound collection method of extracting an acoustic signal input from a predetermined angle range
by a correlation signal extraction circuit from a phase difference between channel signals. .
[0005]
Furthermore, Patent Document 4 discloses a zoom microphone which cancels an acoustic signal
from behind or an inter-channel difference component by an adaptive filter to generate
directivity in the forward direction, and varies its characteristics.
04-05-2019
2
JP-A-2001-8285 JP-A-2002-78100 JP-A-2004-72630 JP-A-3277954
[0006]
As the microphone described in Patent Document 1, the correlation coefficient of audio signals
extracted from a plurality of microphones is calculated to emphasize the audio from the center
direction of the imaged screen, and the audio input according to the magnitude of the correlation
coefficient Although it had the problem of having to use correlation coefficients to emphasize the
signal, the present invention is high-performance, high-emphasis by further emphasizing the
target signal extracted from the input signal without using the correlation coefficient. A sensitive
acoustic signal processor, an acoustic signal processing method, and a program for acoustic
signal processing are obtained.
[0007]
In the stereo audio signal processing apparatus and processing method described in Patent
Document 2, to determine the level phase similarity between channels of the stereo signal in
order to suppress or enhance the audio signal, or the level phase between channels of the stereo
signal The present invention has the problem that the determination of the level phase similarity
between channels and the determination of the level phase difference component between
channels has been made in order to determine the difference component of It is possible to
obtain an acoustic signal processing apparatus, an acoustic signal processing method, and an
acoustic signal processing program, which can be easily calculated without the need to determine
various difference components and similarities.
[0008]
In the superdirective microphones described in Patent Document 3 and others, a method using
an acoustic phase tube or a microphone array, a method of generating a secondary sound
pressure gradient microphone from a low order sound pressure gradient microphone by
calculation, Alternatively, although a method of obtaining microphones of superdirective
characteristics from a predetermined angle range by a correlation signal extraction circuit from
phase differences between channel signals is described, the present invention uses an acoustic
phase tube, a microphone array, or a phase difference. An acoustic signal processing apparatus
and an acoustic signal processing method of the superdirective characteristic and a program for
acoustic signal processing can be further obtained from the microphone of the secondary sound
pressure gradient without the noise.
04-05-2019
3
[0009]
In a zoom microphone capable of changing the characteristics of the acoustic signal processing
device described in Patent Document 4, an adaptive filter is used to generate directivity by
canceling an acoustic signal from the rear or a difference between channels by an adaptive filter.
In the present invention, an acoustic signal processing apparatus and an acoustic signal
processing method of an acoustic sound collector and a program for acoustic signal processing
can be obtained without using an adaptive filter.
[0010]
As described above, various proposals have been made to make acoustic sound collection devices
such as microphones superdirective, but as with array microphones, too, it is necessary to
increase the number of microphones in order to improve directivity; It was not possible to obtain
a sharp directional characteristic.
Furthermore, in general, as the directional characteristics become sharper, the microphone is
more susceptible to wind noise (wind noise, wind blow noise) generated outdoors, etc., and it is
not possible to clearly collect the target audio signal such as voice. There was a problem.
Therefore, an object of the present invention is to provide an acoustic signal processing
apparatus and an acoustic signal processing method, and an acoustic signal processing program,
which obtain zoom microphones and telephoto microphones of superdirective characteristics
that can be incorporated in small electronic devices and are not susceptible to wind noise. To
propose.
[0011]
The acoustic signal processing apparatus according to the present invention comprises: input
means for inputting a plurality of acoustic signals from a plurality of acoustic channels; a
plurality of level detection means for detecting signal levels in a predetermined period of a
plurality of acoustic signals; A selecting means for selecting a signal having the smallest level
value among the level values to be selected at predetermined intervals, and a band limiting
means for band limiting the signal from the selecting means, and the output of the band limiting
means is an output signal It is supposed to be.
[0012]
04-05-2019
4
The acoustic signal processing method according to the present invention receives a plurality of
acoustic signals from an input unit including a plurality of acoustic channels, detects signal levels
in a predetermined period of the plurality of input acoustic signals by a plurality of level
detection units, and detects the levels. The signal having the smallest level value among the level
values detected from the means is selected by the selection means at predetermined intervals,
and the signal from the selection means is subjected to band limitation by the band limitation
means and the band limited output is output signal It is supposed to be.
[0013]
The program for acoustic signal processing of the present invention comprises the steps of
inputting a plurality of acoustic signals from input means comprising a plurality of acoustic
channels, and detecting the signal levels of the inputted plurality of acoustic signals in
predetermined periods by a plurality of level detection means. Step of selecting the signal having
the lowest level value among the level values detected by the level detecting means by the
selecting means every predetermined period, and performing band limitation of the signal from
the selecting means by the band limiting means The band limited output obtained in the above is
used as an output signal.
[0014]
According to the present invention, it is possible to strongly extract only the AND region (the
same phase and the same level component) by the minimum value selection processing from a
plurality of acoustic signals having different directivity patterns, which are obtained particularly
from close microphone groups. With the built-in microphone of the video camera, it is possible to
obtain an acoustic signal processing apparatus and an acoustic signal processing method, and a
program for acoustic signal processing, which can easily obtain a superdirective signal.
[0015]
Hereinafter, an acoustic signal processing apparatus, an acoustic signal processing method, and
an acoustic signal processing program according to an embodiment of the present invention will
be described with reference to FIGS.
FIG. 1 is a system diagram for explaining an audio signal processing apparatus and an audio
signal processing method according to one embodiment of the present invention, and an audio
signal processing program, and FIGS. 2 (a) to 2 (e) are for the present invention. FIG. 3 (a) to FIG.
3 (c) are acoustic directivity characteristic diagrams for explaining various directivity
04-05-2019
5
characteristics of the acoustic signal processing apparatus and acoustic signal processing method
of one embodiment and a microphone unit used for an acoustic signal processing program. An
acoustic signal processing apparatus and an acoustic signal processing method according to one
embodiment of the present invention, and acoustic pointing for explaining an AND region when
angles of a plurality of main axes with respect to a pointing direction of a microphone unit used
in an acoustic signal processing program are changed. FIG. 4 is an acoustic signal processing
apparatus and an acoustic signal processing method according to one embodiment of the present
invention, and a single directivity of a microphone unit used for a program for acoustic signal
processing. Directivity characteristic diagram showing combination of flexibility and bidirectional characteristics, FIG. 5 shows an acoustic signal processing apparatus and an acoustic
signal processing method according to one embodiment of the present invention, and three
single directivity microphone units used in an acoustic signal processing program FIG. 6 is an
acoustic signal processing apparatus and an acoustic signal processing method according to an
embodiment of the present invention and extraction of an AND area of a microphone unit used in
an acoustic signal processing program according to an embodiment of the present invention. FIG.
7 is a waveform diagram for explaining the operation of FIG. 1 and an acoustic directivity
characteristic diagram for explaining the method and its emphasizing method.
[0016]
First, before describing the system diagram of the acoustic signal processing apparatus according
to one embodiment of the present invention shown in FIG. 1, various microphone units
(hereinafter referred to as microphone units) connected to the acoustic signal processing
apparatus will be described with reference to FIG. ) To FIG. 2 (e).
2 (a) to 2 (e) show acoustic directivity characteristic diagrams (hereinafter referred to as polar
patterns) of the microphone unit, and this polar pattern represents the sensitivity level of each
microphone unit from the entire circumferential direction. Polar coordinates are displayed.
In FIGS. 2A to 2E, the shooting direction in the video camera is 0 °, the sensitivity level in the
radial direction indicates a relative value, and the center is zero in sensitivity.
[0017]
FIG. 1 (a) is omnidirectional (all-directional) and has the same level of sensitivity characteristics
04-05-2019
6
in all directions.
FIG. 2B shows primary (single) directivity, which is often used when directivity is given in a
certain direction, and in this case, directivity is given in the photographing direction 3.
FIG. 2C is a second directivity having stronger direction selectivity than the first directivity
shown in FIG.
2 (d) and 2 (e) are referred to as bi-directional, having a maximum sensitivity in a certain
direction and in the direction of the opposite pole, and making the sensitivity in the 90 °
direction zero with that. (D) and FIG. 2 (e) have characteristics orthogonal to each other.
Further, the plus (+) characteristic and the minus (−) characteristic are opposite to each other,
and FIGS. 2 (d) and 2 (e) are 180 ° out of phase with each other. そして。 These directional
characteristics can be generated by combining a microphone unit alone or by combining a few
microphone units.
[0018]
Here, in the present invention, a microphone unit of superdirective characteristics with enhanced
selectivity in a predetermined direction is realized by a small-scale system, and further, an
acoustic signal processing device and acoustic signal processing with a large degree of freedom
in characteristic variation The purpose is to propose a method and a program for acoustic signal
processing. First, one embodiment of the microphone unit of the present invention having a
single directivity characteristic will be described with reference to FIGS. 3 (a) to 3 (c). 3 (a) to 3
(c) generate the single directivity shown in FIG. 2 (b) in two different directions, and change the
angle formed by the main axes 1 and 1 of each. a) to FIG. 3 (c) show the spindles 1 and 1 with an
arbitrary angle. Here, in this example, an overlapping area that changes according to the angle
formed by the main axes 1 of the two polar patterns, that is, an AND area (filled portion) 2 is
extracted, and this AND area 2 extracts a polar pattern having a new directivity. It is generated.
As a result, the directivity characteristic can be easily varied, and as shown in FIG. 3C, any
superdirectivity can be easily generated in the photographing direction 3.
[0019]
04-05-2019
7
In the present invention, the AND region 2 can also be extracted from two polar patterns having
different directivity characteristics. FIG. 4 shows an example of the uni-directionality shown in
FIG. 2 (b) and the polar pattern according to bi-directionality shown in FIG. 2 (d). 2 can be
extracted. Furthermore, in the present invention, the AND region 2 composed of three or more
directional characteristics can also be extracted. FIG. 5 is a polar pattern with a single directivity
shown in FIG. 2 (b) directed to the center channel 4 and the left and right channels 5L and 5R in
three different directions. In this manner, it is possible to easily extract the AND region 2
overlapping all polar patterns, and to realize the superdirective characteristic with enhanced
selectivity.
[0020]
A system diagram of the acoustic signal processing device for obtaining the microphone unit
having the superdirective characteristic described above will be described with reference to one
embodiment (Example 1) of FIG. FIG. 1 is a system diagram showing the most basic configuration
of the present invention, and FIG. 1 extracts an AND area 2 from an input signal having two
directional characteristics. In FIG. 1, an A channel (hereinafter referred to as Ach) signal Sa is
input to the input terminal 10 from a microphone unit of any directivity characteristic, and a
microphone unit of any directivity characteristic is input to the input terminal 11 Each of the
signals Sb is input as Bch. The input Ach signal Sa and Bch signal Sb are input to the fixed
contacts a and b of the changeover switch 6 and to the level value detection / determination
means 8, and the level value detection / determination means 8 detects the level of each signal in
a predetermined period. A value is calculated, and the channel signal (referred to as a ch signal)
on the smaller side of the level is selected by the movable contact c of the changeover switch 6,
and the low pass filter (hereinafter referred to as LPF) 7 is selected. An output signal is output
from the output terminal 12.
[0021]
The operation of the above-described acoustic signal processing apparatus of FIG. 1 will be
described with reference to FIGS. 6 (a) to 6 (c). FIG. 6 (a) is a case where the AND region 2 is
extracted from the unidirectional signal having two different main axes 1 and 1 shown in FIG. 3
(b), and one Ach signal Sa of the unidirectionality is extracted. Level value detection /
determination means 8 when an input signal consisting of a sound source of an arbitrary
predetermined level is input from direction A at a certain moment by inputting the other Bch
04-05-2019
8
signal Sb to the input terminal 10 and inputting it to the input terminal 10. The level on the point
1a indicated by the x mark and the level on the point 1b indicated by the circle are detected, and
the changeover switch 6 is selected so as to select the input signal on the side of the small level
1b. Similarly, in the input signal of the sound source from direction B, direction C, and direction D
in FIG. 6A, the AND region 2 can be obtained by always selecting and outputting the signal 1b on
the ○ side with a small level. It is extracted.
[0022]
Here, an acoustic signal processing method in the case where actual input signal waveforms of
the Ach signal Sa and the Bch signal Sb are input to FIGS. 7A to 7D will be described. In this
example, since a small microphone system mounted on a video camera or the like is premised,
the distance between the plurality of microphones installed with respect to the distance to the
sound source is sufficiently small and can be treated as a plane wave The plurality of input
signals input to the microphones have phase differences depending on the sound source
direction and the distance between the microphones. In FIGS. 7 (a) to 7 (d) showing one example,
the Ach signal Sa shown in FIG. 7 (a) is inputted with a predetermined phase delay than the Bch
signal Sb shown in FIG. 7 (b). In the case where the minimum signal in both predetermined
periods is selected by the level value detection / determination means 8 and the changeover
switch 6 is selected by the determination output signal Sd, as shown by the thick line in FIG. 7C.
The selection output signal Ss is output. Then, when the low-pass component is extracted by the
LPF 7 in the subsequent stage, the LPF output signal Sf shown in FIG. 7D is output. Here, the
output signal Sf is an in-phase component of the Ach signal Sa and the Bch signal Sb, and can be
a component of the above-described AND region 2.
[0023]
Further, the sampling interval for selecting the minimum value (minimum) in this example and
the band-limited frequency by the LPF 7 after that will be described. In the selection output
signal Ss of FIG. 7C described above, the level is compared and the minimum value is selected.
Although the time unit has been described as the sampling interval which is the minimum time
unit of the digital signal, the cutoff frequency of the band limitation in the LPF 7 at this time is
set to Fs / 2 or less according to the sampling theorem, assuming that the sampling frequency is
Fs. Just do it. Therefore, if the band to be used is only the lower band, the sampling interval of
the minimum value selection can be further lengthened conversely.
04-05-2019
9
[0024]
Next, another embodiment (Example 2) of the present invention will be described with reference
to FIG. Hereinafter, in the description of the present invention, parts corresponding to those in
FIG. FIG. 8 shows the case of extraction / emphasis by adding the components of the AND region
2 shown in FIG. 6A extracted in FIG. 1 to the input signal again by the adders 15 and 16. In FIG.
8, Ach signal Sa and Bch signal Sb from input terminals 10 and 11 are fixed contacts a and b of
switch 6, level value detection / determination means 8, one input terminal of adder 15 and
adder 16, and And the LPF output signal Sf of the LPF 7 connected to the movable contact c of
the changeover switch 6 is supplied to the other input terminal of the adder 15 and the adder 16,
respectively, and the respective adders 15 and 16 The Ach signal Sa and the Bch signal Sb are
output via the output terminals 13 and 14.
[0025]
By adding such an adder 15 and the adder 16, for example, when the Ach signal Sa and the Bch
signal Sb constitute stereo 2ch, the AND region 2 extracted as shown in FIG. 6 (b) to obtain a
plain pattern (polar pattern) having an AND area 2a in which the sound in the shooting direction
3 is emphasized while maintaining the sense of stereo sound field (independence) with the
synthesized directional characteristics as shown in 6 (b) it can.
[0026]
FIG. 9 shows an acoustic signal processing apparatus according to still another embodiment
(third embodiment) of the present invention.
In FIG. 9, the Ach signal Sa and the Bch signal Sb input from the input terminals 10 and 11 are
respectively high-pass components by high-pass filters (hereinafter referred to as HPF) 22, LPF
23, HPF 25, and LPF 24. And the low-pass component is subjected to region extraction
processing by the level value detection / determination means 8, and band synthesis is
performed again by the adders 15 and 16, and it is here The extraction processing of the AND
region 2 may be performed in the same manner as described with reference to FIGS. 1 and 8
described above, but in the present example, the (A + B) ch signal Sa obtained by adding low-pass
components in the adder 26. Make b also selectable. As described above, in the small microphone
system of the present invention, although the correlation between the level and phase of a
plurality of signals obtained from each microphone is high, when the frequency of use outdoors
is high as in a video camera, Due to the vortex-like air flow generated in the vicinity of the
04-05-2019
10
microphone, a wind noise is generated due to wind. This wind noise signal is characterized by
approximating a random signal having no correlation as compared with the voice signal, and the
generated frequency band also exists in the low frequency band of 1 kHz or less.
[0027]
Therefore, in FIG. 9, the respective Ach signals Sa and Bch signals Sb supplied to the input
terminals 10 and 11 are supplied to the HPF 22, HPF 25 and LPF 23 and LPF 24 respectively,
and the outputs of the HPF 22 and HPF 25 are the adder 15 and The signal is supplied to one
terminal of the adder 16. The outputs of the LPF 23 and the LPF 24 are supplied to the adder 26,
the level value detection / determination means 8, and the fixed contacts a and c of the
changeover switch 6, respectively. The (A + B) ch signals Sa and b of the Ach signal Sa and Bch
signal Sb added by the adder 26 are attenuated in half via the 1⁄2 attenuator 27 and supplied to
the fixed contact b of the changeover switch 6. The movable contact d of the changeover switch
6 is supplied to the other input terminals of the adder 15 and the adder 16 through the LPF 7
and the Ach and Bch signals are output from the output terminals 13 and 14 of the adder 15 and
the adder 16 Be done.
[0028]
As described above, when the Ach signal Sa and the Bch signal Sb are two signals including noncorrelated components as shown in FIGS. 10A and 10B, the outputs from the LPFs 23 and 24 are
the adder 26. The two are added together, and the addition average signal through the 1⁄2
attenuator 27 is (A + B) / 2 signals Sa and b shown in FIG. 10C, and the Ach signal Sa and the Bch
signal Sb have their level values detected / The determination means 8 and fixed contacts a and c
of the changeover switch 6 are input. Then, if the switching switch 6 is switched so as to always
output the signal with the smallest level within a predetermined period, the selection output
signal Ss1 is output as shown by the solid line in FIG. When the low frequency component is
extracted, the LPF output signal Sf1 is output as shown in FIG. 10 (f). Furthermore, if the same
signal is processed only with the Ach signal Sa and the Bch signal Sb as shown in FIGS. 1 and 8,
the selection output signal Ss2 is output as shown by the broken line in FIG. When the lowfrequency component is extracted at S, the signal Sf2 is output as shown in FIG. Here, if the
signal waveforms in FIG. 10 (e) and FIG. 10 (f) are compared, in the input signal including the
non-correlation component, the effect of the non-correlation component is smaller in FIG. It is
understood that it is extracting.
04-05-2019
11
[0029]
As an application example of the acoustic signal processing apparatus shown in FIG. 9, by setting
the passbands of the LPFs 23 and 24 to the generation band of the wind blowing noise, there is
an effect that the AND region 2 can be extracted without being affected by the noise.
[0030]
Here, the level value detection / determination means 8 shown in FIG. 9 will be described with
reference to FIG.
While the level value detection / determination means 8 of FIGS. 1 and 8 selects two channels,
the configuration of FIG. 11 is different from FIGS. 1 and 8 in that the number of channels is
three. Therefore, the level value detection / determination means 8 of FIG. 9 will be described on
behalf of FIG. In FIG. 11, Ach signals Sa, (A + B) ch signals Sab, and Bch signals Sb from the input
terminals 10, 17, 11 are respectively input and supplied to the absolute value processing means
26, 27, 28 and positive values. Is absolute value. Further, the level detection units 29, 30, 31
detect the respective levels. The outputs of the level detectors 29, 30, 31 are supplied to level
value judging means 32, where the respective levels are compared, and the judgment result is
outputted from the terminal 18 as a judgment output Sd.
[0031]
Further, the operation of the level value determination means 32 described in FIG. 11 described
above will be described with reference to the flow chart of FIG. First, the Ach level value of the
Ach signal Sa is input in the first step ST1, the Bch level value of the Bch signal Sb is input in the
second step ST2, and the level value of the (A + B) ch signal Sab is input in the third step. In the
fourth step ST4, (A + B) ch ≦ Bch is determined, and if YES , the process proceeds to the fifth
step ST5, (A + B) ch ≦ Ach is determined, and if YES , the sixth step The process proceeds to
ST6, and (A + B) ch is set as the determination output. If the fourth step ST4 is "NO", the process
proceeds to the seventh step ST7, AchchBch is determined, and if "YES", the process proceeds to
the eighth step ST8 to set Ach as the determination output Sd, and If "NO" in step ST7, the Bch is
set to the determination output Sd in the ninth step ST9, and the set determination output Sd is
output to the output terminal 18 in the tenth step ST10. Therefore, the signal with the smallest
level is always selected as the determination output Sd.
04-05-2019
12
[0032]
FIG. 13 is a system diagram of an acoustic signal processing apparatus showing still another
embodiment (example 4) of the present invention, and the acoustic signal processing apparatus
of FIG. 13 is the same as the embodiment 1 shown in FIG. An object of the present invention is to
generate a microphone unit with superdirective characteristics more than the second
embodiment shown in and the third embodiment shown in FIG. In FIG. 13, the Ach signal Sa
input from the input terminal 10 and the Bch signal Sb input from the input terminal 11
respectively include level change means 34 and 35, level value detection / determination means
8, and changeover switch 6. The fixed contacts A and b are input to the control coefficient
generation unit 33 through the changeover switch 6 and the LPF 7 to generate the level control
coefficient Scf according to the signal level to obtain the level change units 34 and 35. The level
is controlled, and the Ach signal Sa and the Bch signal Sb are output from the output terminals
13 and 14.
[0033]
For example, in the above configuration, if the level control coefficient Scf is generated so that
the output levels from the level changing means 34 and 35 increase as the signal level from the
LPF 7 increases, the signal level in the extracted AND region 2 increases. Then, since the
feedback loop is applied to further increase the output level, for example, in the case of FIG. 6A, it
is emphasized as shown in FIG. 6C, and FIG. 6A and FIG. A synthesized directional characteristic
is generated.
[0034]
FIG. 14 is a system diagram of an acoustic signal processing apparatus showing still another
embodiment (fifth embodiment) of the present invention.
FIG. 14 is a diagram in which the AND region 2 is extracted / emphasized only in a
predetermined band, as compared with the fourth embodiment of FIG. In FIG. 14, the Ach signal
Sa input from the input terminal 10 and the Bch signal Sb input from the input terminal 11 each
have a predetermined band pass filter (hereinafter referred to as BPF) 37 and 38, and only this
band. Band division filters (hereinafter referred to as "BEF") 36 and 39 to block out the
predetermined band, level value detection / determination means 8, fixed contacts a and b of
changeover switch 6, LPF 7, control coefficient generation means 33, the AND area 2 is
emphasized by level changing means 34 and 35, and band synthesis is performed again by the
04-05-2019
13
adders 15 and 16, and Ach signals and Bch signals are output from the output terminals 13 and
14, for example, video camera In the above, if the predetermined band is a voice band, the voice
of the object to be photographed is clearly collected by making it superdirective by the level
control coefficient Scf only when the person in the photographing direction 3 is emitting a sound.
There are effects that can be done. In FIGS. 13 and 14, the LPF 7 is not necessarily required, and
is often included in the control coefficient generation unit 33 in the subsequent stage.
[0035]
Here, arrangement examples of microphones in the present invention will be described with
reference to FIGS. 15 (a) to 15 (c). In the present invention, the directivity characteristics of the
microphones and the arrangement of the microphones are not limited because the AND region 2
can be extracted from a plurality of directivity characteristics, but in FIGS. 15 (a) to 15 (c) As
shown in FIGS. 3 (a) to 3 (c), this is an embodiment in which continuous variation is possible,
whereby a zoom microphone can be easily configured. FIG. 15A shows an omnidirectional
microphone MIC1 and two bidirectional microphones MIC2 and MIC3. The directivity changing
means using the plurality of microphones MIC1, MIC2 and MIC3 will be described in detail below
with reference to FIGS. 16 (a) to 16 (c) and 17 (a) to 17 (c).
[0036]
In FIG. 16A, the nondirectional signal 60a similar to that shown in FIG. 2A is input from the input
terminal 60 by the nondirectional microphone MIC1, and the first bidirectional microphone
MIC2 is input from the input terminal 61. The second bi-directional signal 61a similar to that
shown in FIG. 2 (d) is input from the input terminal 62, and the second bi-directional signal
similar to that shown in FIG. The directivity signal 62a is input. Then, when the first bi-directional
signal 61a and the second bi-directional signal 62a are subjected to the addition and averaging
process at the same level by the first level variable addition / subtraction combining means 40,
the bi-directional in FIG. 16 (b) As shown in the sex pattern B, vector synthesis is performed.
[0037]
Furthermore, when the synthesized signal and the nondirectional signal 60a are subjected to
addition averaging processing by the second level variable addition / subtraction / synthesis
means 41, the negative phase side is canceled and the unidirectional pattern shown by the solid
04-05-2019
14
line in FIG. Are generated and output from the output terminal 41a. Similarly, when the first level
variable addition / subtraction combining means 40 performs subtraction averaging processing
on the first bi-directional signal 61a and the second bi-directional signal 42a at the same level,
the broken line in FIG. It is vector-synthesized as bi-directional pattern A. Further, when the
synthesized signal and the nondirectional signal 60a are subjected to addition averaging
processing by the second level variable addition / subtraction / synthesis means 41, the negative
phase side is canceled and the unidirectional pattern shown by the broken line in FIG. Is
generated. Since bi-directional patterns A and B can be vector synthesized in all directions by
changing the combination ratio of the two in the first level variable addition / subtraction
synthesis means 40, as shown in FIG. The directivity can be varied continuously as shown in FIG.
3 (c).
[0038]
Also, as shown in FIG. 15 (b), directivity can be similarly changed by using nondirectional
microphones MIC4 to MIC7. That is, subtracting the nondirectional microphone MIC5 from the
nondirectional microphone MIC7 to adjust the frequency characteristics generates the first
bidirectional signal 61a, and subtracting the nondirectional microphone MIC4 from the
nondirectional microphone MIC6 generates a frequency. When the characteristics are adjusted, a
second bi-directional signal 62a is generated. Furthermore, as the nondirectional signal 60a is
generated by adding all the outputs of the nondirectional microphones MIC4 to MIC7, as shown
in FIG. 16A, as shown in FIGS. 3A to 3C. The directivity can be varied continuously.
[0039]
Furthermore, as shown in FIG. 15 (c), using a unidirectional microphone MIC8 and a bidirectional
microphone MIC9, as shown in FIG. 17 (b), to generate a primary directional pattern and a
bidirectional pattern C. Can. That is, in FIG. 17 (a), the primary directivity signal 61b shown in
FIG. 2 (b) is inputted from the input terminal 61 by the unidirectional microphone MIC8, and the
diagram by the bidirectional microphone MIC9 from the input terminal 62. A bi-directional signal
62a shown in 2 (e) is input. When the first directional signal 61b and the bidirectional signal 62a
are subjected to addition averaging processing at the same level by the level variable addition /
subtraction / synthesis means 40, the negative phase side is canceled and the single indicated by
the solid line in FIG. A directional pattern is generated. Similarly, when the level variable type
addition / subtraction / synthesis means 40 performs subtraction averaging processing at the
same level, the uni-directional pattern shown by the broken line in FIG. 17C is generated. Then,
bi-directional patterns A and B can be vector synthesized in all directions by changing the
04-05-2019
15
combination ratio of the two in level variable addition / subtraction synthesis means 40, as
shown in FIG. 3 (a) to FIG. The directivity can be varied continuously as in (c).
[0040]
The arrangement examples of the various microphones shown in FIGS. 15 (a) to 15 (c) are merely
examples, and can be changed within the scope of the object of the present invention. In a video
camera, by changing the directivity continuously as shown in FIG. 3 (a) to FIG. 3 (c) according to
the zoom operation by an optical lens etc., an acoustic collection more suitable for a
photographed image Sound is possible. Although illustration is omitted, for example, the AND
area 2 shown in FIG. 3A is output when the zoom is on the wide angle side, and the AND area
shown in FIG. 3C is output when the zoom is on the telephoto side. When 2 is output, a
directivity signal matched to the angle of view can be output.
[0041]
Furthermore, still another embodiment (Example 6) of the present invention will be described in
detail with reference to FIG. As shown in FIG. 18, the Ach signal Sa and the Bch signal Sb from
the input terminals 10 and 11 are the same as in FIG. 1 through level value detection /
determination means 8 and fixed contacts a and b of changeover switch 6 and LPF 7
respectively. And the AND area 2 is extracted, the AND area 2 is emphasized by the level varying
means 34, and further, the Ach signal Sa and the Bch signal Sb are again added by the adders 15
and 16 and output from the output terminals 13 and 14 . Here, for example, the zoom is reduced
on the wide-angle side by the zoom position information Sz input to the input terminal 19 based
on the zoom operation so that the variable level of the level variable unit 34 is not emphasized,
and conversely If the zoom is enhanced on the telephoto (telescope) side, the directivity can be
changed according to the zoom position, and an acoustic zoom can be realized according to the
imaging screen.
[0042]
In FIG. 18, the zoom position information Sz is used as level variable information of the level
variable means 34, but the present invention is not limited to this, and feature parameters based
on image information related to the input audio signal are used. It can be used. The zoom
position information described above is a feature parameter indicating the angle of view of a
04-05-2019
16
photographed image, but as an example, in a video camera, the distance information to the
subject can be obtained from the lens focus information and the level variable according to the
distance Becomes possible. Also, it is possible to perform level variation when a specific image is
obtained from metadata included in the image or image recognition information.
[0043]
Next, an acoustic signal processing apparatus in the case of performing 3-channel (3-ch) input
processing and 3-band band division processing will be described in detail as still another
embodiment (seventh embodiment) of the present invention in FIG. The 3-ch input processing
and the 3-band band division processing are independent events and may be performed
individually or may be processed by performing more channel inputs and band divisions. FIG. 19
shows, for example, the case where it is assumed that three channels of left and right channels
Lch5L and Rch5R and center channel Cch4 shown in FIG. 5 are input, and the extraction band of
AND region 2 is shown in Band 1 and Band 2 for each input signal. FIG. 20 shows a frequency
characteristic curve showing an example of band division which is an embodiment in the case
where the band is divided and the other bands are band 3.
[0044]
In FIG. 19, the Ach signal Sa, Cch signal Sc and Bch signal Sb input from the input terminals 10,
21 and 11 are band-divided by BPF1 to BPF3 (42 to 49) of groups I to 3 respectively. Be done.
The extraction band of the AND area 2 is extracted for each of the band 1 and band 2 shown in
FIG. 20 by the BPF1 (43-45) of the first group and the BPF2 (46-48) of the second group. First,
the Ach signal Sa, Cch signal Sc, and Bch signal Sb from the BPF 1 (43 to 45) of the first group I
are sent to the fixed contacts a, b and c of the level value detection / determination means 8a and
the changeover switch 6a. It is supplied, and the minimum value is selected as in FIG. 8 and input
to the adder 50 via the LPF 7a. Similarly, the Ach signal Sa, Cch signal Sc, and Bch signal Sb from
each second group BPF 2 (46 to 48) are level value detection / determination means 8b and fixed
contacts a, b, c of the changeover switch 6b. , And the minimum value is selected and input to the
adder 50 via the LPF 7b. Then, the output of the adder 50 is supplied to the other terminal of the
adders 13 and 14 whose output of the BPF 3 of the third group is connected to one input
terminal, and band 1 and band 2 shown in FIG. Furthermore, the Ach signal Sa, the Cch signal Sc,
and the Bch signal Sb from the respective bands 3 are band-combined and the Ach signal is
output from the output terminal 13 and the Bch signal is output from the output terminal 14. By
performing the minimum value selection processing for each band after performing band
division in this manner, it is possible to improve the extraction performance of the signal of the
04-05-2019
17
AND area 2 that is the same level and the same phase component.
[0045]
Furthermore, still another embodiment (Embodiment 8) of the present invention will be
described with reference to FIG. FIG. 21 improves the extraction performance of the signal of the
AND area 2 more than FIG. 19 by performing fast Fourier transform (FFT). Here, the Ach signals
Sa and Bch signals Sb input from the input terminals 10 and 11 convert the time axis signals to
be input into m frequency axis signals of frequencies f1 to fm by the FFT means 51 and 52,
respectively. Here, in each of the FFT means 51 and 52, the frequency axis signal of the
frequencies f1 to fm is divided into the frequencies f1 to fn of the extraction band of the AND
area 2 and the frequencies f (n + 1) to fm of the other bands. The Ach signal Sa and the Bch
signal Sb of to fn are input to the level comparing / selecting means 53, and the level comparison
is performed for each of the frequencies f1 to fn to select the ch signal with the lowest level at all
frequencies f1. It implements about-fn.
[0046]
Then, the signals of the selected frequencies f1 to fn are input to the band combining means 54,
55, band combined again with the signals of the frequencies f (n + 1) to fm, and inverse fast
Fourier transform (signals of frequencies f1 to fm) The frequency axis signal is converted to a
time axis signal and output from the output terminals 13 and 14 as an Ach signal Sa and a Bch
signal Sb. Also, in the fifth embodiment of FIG. 14 described above, the divided band may be
increased as in the seventh and eighth embodiments of FIGS. 19 and 21, and the level variable
processing may be performed for each divided band.
[0047]
The above series of extraction processing of the AND region 2 in all the embodiments of the
present invention is applied to the input signal from the microphone and may be configured as a
sound collection system or a recording system, but the present invention is not limited to this.
The present invention is not limited to this, and may be implemented in a reproduction system. It
is also apparent that the program may be implemented as a program of application software in a
computer, and may be implemented as non-real time processing at the time of editing video /
audio files, at file conversion, and at the time of DVD disc writing.
04-05-2019
18
[0048]
The above-described present invention has input means for inputting a plurality of acoustic
signals from a plurality of acoustic channels as shown in FIG. 1, a plurality of level detection
means for detecting signal levels in a predetermined period of a plurality of acoustic signals, and
the level detection Means for selecting a signal having the smallest level value among the level
values detected by the means at predetermined intervals, and band limiting means for band
limiting the signal from the selecting means, the band limiting means Since the acoustic signal
processing apparatus is characterized in that the output signal is an output signal of the plurality
of acoustic signals having different directivity patterns, which are obtained from a group of
closely spaced microphones, an AND area Since only the phase and the same level component)
can be extracted strongly, the superdirective signal can be easily obtained in the built-in
microphone of the video camera.
[0049]
According to the configuration of FIG. 8 of the present invention, in the configuration shown in
FIG. 1, there is provided combining means for combining the signal from the band limiting means
and the plurality of acoustic signals for each acoustic channel. Since the acoustic signal
processing apparatus uses the output of each means as the output signal of each acoustic
channel, it becomes a monaural signal at the time of extraction of the AND region, but the sound
field of each acoustic channel is synthesized by combining the extracted signal with the input
acoustic signal. A superdirective signal can be obtained while maintaining a sense
(independence).
[0050]
According to the configuration of FIG. 9 of the present invention, input means for inputting a
plurality of acoustic signals from a plurality of acoustic channels, a plurality of band extraction
means for extracting a predetermined band from a plurality of acoustic signals, and a plurality of
these band extraction means A plurality of first level detection means for detecting signal levels
in predetermined periods of the signals from the plurality of band extraction means, and signals
in predetermined periods of the signals from the arithmetic means The second level detection
means for detecting the level, the level values from the plurality of first level detection means and
the level value from the second level detection means are the signals having the lowest level
value at predetermined intervals. Selection means for selecting, band limiting means for band
limiting the signal from the selecting means, and a signal from the band limiting means and a
band other than the extraction band in the plurality of band extraction means Since the acoustic
signal processing apparatus is characterized in that it has an audio signal processing apparatus
characterized in that it has band synthesizing means for synthesizing the band of each audio
04-05-2019
19
channel in each band and the output of each band synthesizing means is output signal of each
acoustic channel. In the case of an input signal with a small AND region, for example, in a
frequency band that includes a large amount of wind noise, the wind signal is extracted by
performing minimum value selection processing on the signal that has been added and averaged
in addition to the input acoustic signal. While reducing the sound, it is possible to maintain the
sense of sound field (independence) of each audio channel.
[0051]
Since the band extraction means in FIG. 9 is an acoustic signal processing apparatus configured
by a plurality of filter means or FFT means, when extracting an AND region, a plurality of BPFs or
FFT means are used for each acoustic channel, By performing the minimum value selection
processing for each band by dividing into bands and recombining bands thereafter, the
reproducibility of the signal of each acoustic channel can be further improved.
[0052]
The minimum unit constituting the predetermined period of the acoustic signal processing
apparatus of FIGS. 1, 8 and 9 is the sampling period, and the minimum time unit for performing
the minimum value selection process is the audio sampling time if it is a digital signal. The
extractable band in this case can be up to Fs / 2 if the sampling frequency is Fs according to the
sampling theorem (Nyquist theorem).
Further, when control coefficients are generated in the latter stage as shown in FIGS. 13 and 14,
the extraction band can be lowered, and if the extraction band frequency is Fx, the time length
for performing the minimum value selection process is 1 / Fs. It becomes ˜ 1 / Fx.
[0053]
According to the configuration of FIG. 13 of the present invention, control coefficient generation
means for generating a control coefficient according to the signal level from the selection means
of FIG. 1 and the levels of a plurality of acoustic signals by the control coefficient from this
control coefficient generation means The acoustic signal processing apparatus has a plurality of
level variable means for making variable and the output of each level variable means is an
acoustic channel output signal. In the configuration of FIG. 14, a plurality of acoustic signals are
outputted from a plurality of acoustic channels. Input means for inputting; a plurality of band
extraction means for extracting a predetermined band from the plurality of acoustic signals; a
plurality of level detection means for detecting a signal level of a signal from the plurality of
04-05-2019
20
band extraction means in a predetermined period; Selection means for selecting a signal having
the lowest level value among the level values from the plurality of level detection means at
predetermined intervals, and signal level of the signal from the selection means Control
coefficient generation means for generating a control coefficient corresponding to a plurality of
level change means for changing the level of the signals from the plurality of band limiting
means by the control coefficient from the control coefficient generation means, and the level
change means An acoustic signal, wherein band combining means is provided for band
combining the signal from the band and band signals other than the extraction band in the
plurality of band extracting means for each acoustic channel, and the output of each band
combining means is output signal of each acoustic channel Since the processing device is used,
the control coefficient is generated according to the level of the AND region signal extracted by
the minimum value selection processing, and the coefficient of the AND region signal or the
acoustic output signal extracted by feedback using that coefficient is modulated. Region signals
can be further enhanced or attenuated, and new directivity patterns such as superdirective
patterns can be generated.
[0054]
According to the configuration shown in FIG. 18 of the present invention, there are provided
input means for inputting a plurality of acoustic signals from a plurality of acoustic channels, a
plurality of level detection means for detecting signal levels in a predetermined period of a
plurality of acoustic signals, Selection means for selecting a signal having the lowest level value
among the level values detected by the detection means at predetermined intervals, band
limitation means for performing band limitation of the signal from the selection means, and the
band limitation means The level varying means is an acoustic signal processing apparatus which
varies the level according to a separately set control coefficient. Therefore, the AND region signal
extracted by the minimum value selection process is selected. Modulate the level of the extracted
AND domain signal or the sound output signal with the control coefficient generated separately,
and create the sound effect such as the sound zoom It is thing.
[0055]
The control coefficient generation means of FIGS. 13 and 14 of the present invention generates
the control coefficient so as to increase the variable level in the level variable means when the
signal level from the selection means is large. Can be easily generated.
[0056]
Since the separately set control coefficients in FIG. 18 of the present invention are generated
04-05-2019
21
according to the feature parameters in the image related to the input acoustic signal, for
example, even if generated as a user interface that can be directly controlled by the user Also, it
may be generated from information in the system such as image zoom, focus information, etc.
Furthermore, metadata such as information attached to image data at the time of shooting etc.
(information for specifying a certain person or thing) It may be generated from
Furthermore, since characteristic parameters that identify a certain person or object obtained
from image recognition, speech recognition, etc. are used, there is an effect that an application
that associates an image with sound can be configured.
[0057]
According to the present invention, as shown in FIG. 1, a plurality of acoustic signals are input
from input means comprising a plurality of acoustic channels, and signal levels in predetermined
periods of the plurality of input acoustic signals are detected by a plurality of level detection
means. The signal having the lowest level value among the level values detected from the
detection means is selected by the selection means at predetermined intervals, and the signal
from the selection means is subjected to band restriction by the band restriction means. Since the
sound signal processing method is characterized in that the output signal is used, an AND area
(same phase, same phase) is selected from a plurality of sound signals having different directivity
patterns obtained from the microphone group in close proximity, in particular. Since it is possible
to extract only the level component) strongly, it is possible to obtain an acoustic signal
processing method capable of easily extracting the superdirective signal in the built-in
microphone of the video camera. Can.
[0058]
In the present invention, as shown in FIG. 1, a step of inputting a plurality of acoustic signals
from input means comprising a plurality of acoustic channels, and detecting a signal level in a
predetermined period of a plurality of inputted acoustic signals by a plurality of level detection
means Level detection step, a selection step of selecting the signal having the lowest level value
among the level values detected from the level detection step by the selection means at
predetermined intervals, and band limiting the signal from the selection step Since an audio
signal processing program is characterized in that the band limited output obtained in the band
limiting step of band limiting by means is used as an output signal, different directivity patterns
obtained particularly from a group of microphones in close proximity are provided. Extracts only
AND domain (same phase, same level component) from multiple acoustic signals with Kill
Therefore, in built-in microphone of the video camera, it is possible to easily obtain a superdirectional signal can be extracted acoustic signal processing program.
04-05-2019
22
[0059]
It is a systematic diagram for explaining an acoustic signal processing device and an acoustic
signal processing method of one mode example (example 1) of the present invention, and an
acoustic signal processing program.
It is an acoustic directivity characteristic figure for demonstrating the various directivity
characteristics of the microphone unit used for the acoustic signal processing apparatus of 1
form of this invention.
FIG. 7 is an acoustic directivity characteristic diagram showing an AND region when angles of a
plurality of main axes with respect to the directivity direction of the microphone unit used in the
acoustic signal processing device of one embodiment of the present invention are changed.
It is an acoustic directivity characteristic view which shows the combination of the single
directivity characteristic of the microphone unit used for the acoustic signal processing
apparatus of 1 form example of this invention, bi-directionality, and a characteristic.
FIG. 6 is an acoustic directivity characteristic diagram showing an AND region in the case where
three single directivity characteristics of the microphone unit used in the acoustic signal
processing device of one embodiment of the present invention are combined.
It is an acoustic directivity characteristic figure for demonstrating the AND area ¦ region
extraction method of the microphone unit used for the acoustic signal processing apparatus of 1
form example of this invention, and its emphasis method.
It is waveform explanatory drawing for operation ¦ movement description of the acoustic signal
processing apparatus of 1 form of this invention.
It is another systematic diagram for demonstrating the acoustic signal processing apparatus of 1
form example (Example 2) of this invention, an acoustic signal processing method, and an
acoustic signal processing program. FIG. 18 is another system diagram for describing an acoustic
04-05-2019
23
signal processing device, an acoustic signal processing method, and an acoustic signal processing
program according to an embodiment (third embodiment) of the present invention. It is
waveform explanatory drawing for operation ¦ movement description of FIG. FIG. 6 is a system
diagram of an acoustic signal processing device, an acoustic signal processing method, and a
level value detection / determination means used in an acoustic signal processing program
according to one embodiment of the present invention. It is a flowchart for operation ¦ movement
description of the level value determination means of the level value detection / determination
means used in FIG. It is a systematic diagram of the acoustic signal processing apparatus which
shows the other example of a form (Example 4) of this invention. It is a systematic diagram of the
acoustic signal processing apparatus which shows the other further 1 embodiment (Example 5)
of this invention. It is arrangement ¦ positioning explanatory drawing of the microphone
applicable to this invention. It is a systematic diagram and polar pattern of the directivity change
means used for the acoustic signal processing apparatus of 1 form of this invention. It is a
systematic diagram and a polar pattern of another directivity variable means used for the sound
signal processing apparatus of one form example of this invention. It is a systematic diagram of
the acoustic signal processing apparatus which shows the further another example (Example 6)
of this invention. It is a systematic diagram of the acoustic signal processing apparatus which
shows the other example of a form (Example 7) of this invention. It is a frequency characteristic
figure which shows the band division zone of the input signal of the acoustic signal processing
apparatus of this invention. It is a systematic diagram of the acoustic signal processing apparatus
which shows the other example of a form (Example 8) of this invention.
Explanation of sign
[0060]
DESCRIPTION OF SYMBOLS 1 ... main axis ¦ shaft 2, 2a ... and area ¦ region, 3 ... imaging ¦
photography direction, 4 ... center channel, 5L, 5R ... left and right channel, 6 ... changeover
switch, 7 ... LPF 8 level value detection / determination means 10, 11, 17, 19, 21, 21, 60, 61, 62
... input terminals 12, 13, 14 ... output terminals 15, 16, 26, 50 · · · · Adder · 18 · · · terminal, 22,
35 · · · HPF, 23, 24 · · · · · · · · · · · · · · attenuator, 26, 27, 28 · · · absolute value processing means,
29, 30, 31 ... Level detection means, 32 ... Level value judgment means, 33 ... Control coefficient
generation means, 34, 35 ... Level variable means, 36, 39 ... ... BEF, 37, 38 ... BPF, 42, 49 ... BPF 3,
43, 44, 45 ... BP 1,46,47,48 ··· BPF2,51,52 ··· FFT unit, 53 ... band combining means, 56, 57 ...
IFFT unit
04-05-2019
24
1/--страниц
Пожаловаться на содержимое документа