close

Вход

Забыли?

вход по аккаунту

?

JPH04178100

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 JPH04178100
[0001]
[Industrial field of application] The present invention relates to a sound receiving system for
receiving an acoustic signal having a good SN ratio in a sound field in which noise and an
acoustic signal to be received are mixed. [Prior Art] In recent years, there have been remarkable
developments of a speech communication system such as a speech communication system and a
speech recognition system. In these systems, a desired sound (such as voice or music, hereinafter
referred to as a target signal) is received by a microphone placed at a position slightly away from
a sound source (for example, the mouth of the speaker). However, at this time, if noise (a general
term for unnecessary sound such as 41 mechanical noise and disturbing voice) exists around the
target signal, the noise is also received at the same time as the target signal, resulting in
deterioration of speech quality and false recognition of voice. It is a big problem. A conventional
sound receiving device for preventing such ambient noise mixing will be described with
reference to FIG. In FIG. 5, curve 1 represents the sensitivity directivity characteristics
(hereinafter simply referred to as directivity characteristics) of various microphones, and when
the microphone is positioned at the center of circle 2, the distance from the center of circle 2 to
curve l The distance represents the magnitude of sensitivity in each direction. The case where the
maximum sensitivity direction is the direction 3 of the target signal and the direction 4 of the
noise is in a direction different from this direction is shown. FIG. 5A is a sound pressure-time
gradient type unidirectional microphone mainly used conventionally. The directional
characteristics of (hereinafter, abbreviated as a unidirectional microphone) are shown. As shown
in FIG. 5A, since the unidirectional microphone has high sensitivity in the ± 90 ′ ′ direction
toward the target signal direction 3, noise comes from the direction 4 within ± 90 ′ ′ with
respect to the target signal direction 3. It is heard that the noise is received with little
suppression. As a result, it was not possible to obtain a satisfactory SN ratio improvement effect
with the conventional unidirectional microphone. A phase synthesis microphone is known as a
03-05-2019
1
first method for obtaining a high SN ratio improvement effect. As shown in FIG. 5B, the phase
synthesis type microphone improves the SN ratio by reducing the sensitivity to the noise
direction 4 by sharpening the directivity characteristic to the target direction 3. A configuration
using a microphone array (sound receiver composed of a plurality of microphone elements) is
shown in FIG. 6 as a configuration example of the phase synthesis type microphone. In FIG. 6, M
microphone elements 11 ° to 116 are linearly arranged, and in this example, vertically arranged
to form the microphone array 12, these microphone elements 11. The respective received sound
outputs of ∼llx are added by the adder 13 and output.
Microphone element 11. The target signal (sound wave) I4 arrives from the direction
perpendicular to the arrangement of llx to llx, in this example, the horizontal direction, and the
noise (sound wave) 15 is transmitted to the microphone element 11. It arrives obliquely to the
arrangement direction of. Therefore, although the target signals 14 coming from the horizontal
direction are emphasized because they are added in the same phase, the noise 15 coming from
the diagonal direction is each microphone element] 1 ° -11. The time difference (phase
difference) is received and then added, so it is suppressed. As a result, phase synthesis is
performed such that a plurality of acoustic signals are added so as to be in phase with the target
signal and to have a phase difference with respect to sounds coming from other directions. The
sharp directivity shown in FIG. 5B can be realized. Such sharp directional characteristics are
often referred to as superdirectivity, and thus microphones using such a sound receiving system
are often referred to as superdirective microphones, but in the following, in this specification,
This system is called a phase synthesis type sound receiving system, and a sound receiving
device based on this system is called a phase synthesis type microphone. As other means for
constructing a phase synthesis type microphone, a method using a parabolic reflector, a method
using an acoustic tube, and the like are known (for example, Nakajima, Editor: "" Acoustics
Lecture 2: Applied Electro-Acoustic "Showa) 54 years, Corona company). Now, the phase
synthesis type microphone has a good SN ratio improvement effect. However, in order to sharpen
the directional characteristics, a receiver scale several times the spatial wavelength of the target
frequency, that is, the length L (FIG. 6) of the microphone array 12 is required. For example, at
170 Hz, which is near the lowest frequency of voice, the spatial wavelength of sound waves is 2
m, and to achieve sharp directivity in this frequency band, the microphone array 12 or reflection
of several times its size (length) There was a problem that a board was needed. In other words,
the directional characteristics of the phase-combining microphone are strongly frequencydependent and the SN ratio improvement effect is small at low frequencies. 0 For example, when
the receiver size is 50 CII in FIG. 340 Hz. Curves 1 and 3 indicate the directivity characteristics
and the target signal direction, as shown in FIGS. 7A, 7B, and 7C, and the directivity
characteristics at 680) 1z, 1360 Hz are shown. From this figure, it can be seen that a large SN
ratio improvement effect can not be expected below 680 Hz. Since the linear microphone array is
used, the directivity characteristics are vertically symmetrical.
03-05-2019
2
An adaptive microphone array (hereinafter referred to as an adaptive array) is known as a second
method of obtaining a high SN ratio improvement effect. The adaptive array, as shown in FIG. 5C,
detects the noise direction 4 and improves the SN ratio by automatically reducing the sensitivity
of the direction (eg, Monzlng). andMiller、
”IntroductiontoAdaptiveArrays、
”JohnWiley&5ons+1980)。 FIG. 8 shows an example of the configuration
of the adaptive array. In FIG. 8, the microphone array 22 is configured by the M microphone
elements 211 to 21i, and the microphone elements 21. The outputs of 2111 are respectively
filters 23. The signal is added and output by the adder 24 through 23.4. フィルタ23. ∼
23. Each filter coefficient of I is controlled by the filter control unit 25. In this configuration,
the filter control unit 25 adds the filter 23.x so that the noise power PN contained in the output
signal 26 of the addition EH 24 is minimized. To 23 are controlled. However, at that time, if noise
power PN is simply minimized, all filter coefficients become zero, and noise power PN becomes
zero, but it does not mean that the target signal is not output. Gives a constraint on the filter
coefficients to minimize the noise power PN. As examples of this constraint condition, there are
known: no distortion constraint condition (a condition that the target signal is not distorted at
all), and a second 22 constraint condition (a condition that no distortion exceeding the allowable
range in the target signal occurs) ing. Minimizing noise under such constraints is equivalent to
lowering the sensitivity of the noise direction while maintaining the sensitivity of the target
signal direction, and the directivity shown in FIG. 5C. Is realized. A specific filter coefficient
calculation means is described in, for example, Japanese Patent Application No. 58-149500.
Although the SN ratio improvement effect of the adaptive array is smaller than that of the phasecombining microphone, noise reduction can be achieved without realizing sharp directivity
characteristics, so that the SN ratio can be improved with a small microphone array. There is a
feature. Further, the frequency dependency in the improvement of the SN ratio is smaller than
that of the phase synthesis type microphone, and it is characterized that the SN ratio can be
improved over a wide frequency band. However, according to the conventional experimental
results, it is known that good performance can not be obtained even in an adaptive array in a low
frequency band of several hundreds Hz or less (hereinafter this frequency band is referred to as
an extremely low frequency band in this specification). It is done.
For example, when the distortion-free condition is used, the SN ratio improvement effect in the
very low frequency region can not be obtained, and when the second-order condition is used,
distortion of the target signal in such low frequency region It has been known that it becomes a
problem as a high quality voice and music receiving device. Furthermore, the adaptive array also
has a problem that the hardware scale becomes large because the adaptive signal processing
involved in the control of the filter is necessary. An object of the present invention is to provide a
high S / N ratio sound receiving system which solves the problems of the conventional sound
receiving system as described above. [Means for Solving the Problems] The present invention
03-05-2019
3
divides the frequency band of an acoustic signal to be received into two, and for example, divides
a received signal at a frequency fc determined based on a required receiver scale. It is
characterized in that the two high frequency bands (hereinafter referred to as the high frequency
band and the low frequency band) are divided, the high frequency band is received by the phase
synthesis type sound receiving system, and the low frequency band is received by the adaptive
array system. I assume. In the present invention, when the scale of the receiver is given, the
frequency band in which sharp directional characteristics (directional characteristics with the
sharpness required according to the purpose of use) can be formed by the phase combination
type sound reception method. A high frequency range and a lower frequency range are regarded
as a low frequency range. As a result, in the high frequency range, a desired SN ratio
improvement effect can be obtained by the phase combination type sound reception method. On
the other hand, by receiving sound in the low frequency band by the adaptive array system, a
good improvement of the SN ratio can be obtained except for the extremely low frequency band.
In addition, if the application frequency bandwidth of the adaptive array system is limited to the
low frequency range, the sampling frequency can be reduced, and as a result, the signal
processing amount in the adaptive array system is advantageously reduced. This is explained
below. Filter shown in the configuration example of FIG. Consider the case of implementing ~ 23
with a digital FIR filter. Here, for example, in the case of an adaptive array having a two-element
microphone array, it is necessary to realize a delay time of at least d / c (d: distance between
microphone elements, C: sound velocity) in each filter It is known. The number of taps required
to realize a certain delay time with the FIR filter is proportional to the sampling frequency.
Furthermore, since the amount of processing of the adaptive array system is proportional to the
number of taps of the filter, reducing the sampling frequency leads to a reduction in the amount
of arithmetic processing. Next, it will be described below that by limiting the application
frequency bandwidth of the adaptive array system, the great advantage of improving the SN ratio
improvement effect in the extremely low frequency range is produced.
First, simulation results conducted to investigate the cause of deterioration of the SN ratio
improvement effect in the extremely low frequency range will be described. In the simulation, it
was assumed that a target signal arrived from a linear direction connecting two microphone
elements using an array of two elements of microphones, and that noise came from a direction
perpendicular to the linear. The microphone element interval d was 4.25 CI. The sampling
frequency is 8 kHz (application frequency band 4 kHz), and the filter 23 shown in the
configuration example of FIG. The ~ 23P + used a 16 tamp FIR digital filter. The filter coefficients
were calculated using distortion free constraints. The experimental results are shown in FIG. In
FIG. 9, the curve m27 represents the frequency response of the adaptive array to the target
signal, and the curve 28 represents the frequency response of the adaptive array to noise. From
this figure, it can be seen that in the frequency band of 500 Hz or more, there is a difference of
about 20 dB or more in the magnitude of the response to the target signal and the response to
noise, and the SN ratio is improved by 20 dB or more. However, it can be seen that the
03-05-2019
4
improvement effect of the SN ratio is small in the extremely low frequency region of 500 Hz or
less. Also, the following can be considered from FIG. Since the directivity characteristic is in
principle non-directed to the DC component, the value of the response to the target signal and
the value of the response to the noise coincide at the point where the frequency is zero. b. The
performance deterioration in the extremely low frequency range is largely due to the influence of
the mountain starting from the frequency spread among the response 28 to noise, and if the
width of this mountain is reduced, the performance of the extremely low frequency range
reduction is improved . The number of peaks of the response 28 to C0 noise is 8, which
corresponds to the frequency resolution of the FIR filter in the human power-output system, that
is, 1/2 of the number of filter tamps. From the above consideration, if the frequency resolution is
improved by increasing the number of taps of the FIR filter, the enemy of the mountain of the
response 28 to noise is increased, and the width of the mountain becomes smaller accordingly
(the low band characteristic is improved) It is expected to be possible. Therefore, the simulation
was performed with the number of filter taps doubled (32). The results are shown in FIG. 10.
From this figure, it is apparent that the frequency bandwidth in which the SN ratio improvement
effect is degraded is approximately 1⁄2, and is approximately inversely proportional to the
increase in the number of taps of the filter. Became. Therefore, it has been found that the number
of taps of the FIR filter should be increased in order to improve the low-pass characteristics of
the adaptive array, but the amount of operation processing increases in proportion to the
number of taps, and the calculation accuracy is also improved. In many cases, it is difficult for
practical use because it decreases.
However, since the frequency resolution of the filter is (applied frequency band × 2) ÷ (number
of taps of the filter), the applied frequency band of the adaptive array is not increased as in the
present invention without increasing the number of taps. If reduced, the frequency resolution is
improved, and the performance improvement in the extremely low frequency range is expected.
In order to confirm this, a simulation similar to the above was performed with a sampling
frequency of 1.6 k lz (application frequency band of 800 Hz) and a tap number of 16. The results
are shown in FIG. 11. According to this figure, the bandwidth for improving the SN ratio is about
5001 (z or less) by reducing the common frequency band from 4 kHz (in the case of FIG. 9) to
800 Hz. It can be seen that the frequency is reduced to a band of about 100 Hz or less. From the
foregoing, it has become apparent that the present invention has a great improvement effect on
the performance degradation of the adaptive array in the extremely low frequency range.
Embodiment J FIG. 1 shows a first embodiment of the present invention. A phase matching
receiver 41 and an adaptive array receiver 42 are provided. The output of the phase matching
receiver 41 is supplied to the high pass filter 43, and the output of the adaptive array receiver 42
is The signal is supplied to the low pass filter 44, and the output of the high pass filter 43 and
the output of the low pass filter 44 are added by the adder 45 and output as an output signal 46.
The high pass filter 43 and the low pass filter 44 have substantially the same cutoff frequency fc.
As a result, the high frequency range of the output signal 46 is synthesized from the output of
03-05-2019
5
the phase matching receiver 41, and the low frequency range is synthesized from the output of
the adaptive array receiver 42. In this embodiment, the filters 43 and 44 can be incorporated as
parts of the sound receiving devices 41 and 42 respectively. In this way, the high frequency
components are received by the phase combination reception method, and the low frequency
components are received by the adaptive array method. FIG. 2 shows a second embodiment of
the present invention. The output of the receiver 51 is high band i! The output of the high pass
filter 52 is supplied to the pass filter 52 and the low pass filter 53, the output of the high pass
filter 52 is supplied to the processing circuit 54 for realizing phase synthesis reception, and the
output of the low pass filter 53 is adaptive array reception And the outputs of the processing
circuits 54 and 55 are added by the adder 56 and output as the output signal 57. In this
embodiment, an acoustic signal is first received by one receiver 51, for example a microphone
array. Next, the received signal is divided into two frequency bands by a high pass filter 52 and a
low pass filter 53 having substantially the same cutoff frequency fc.
Thereafter, signal processing for performing phase synthesis reception on a high frequency
region is performed by the processing circuit 544. The processing circuit 55 performs signal
processing for performing adaptive array sound reception in a low frequency band. Finally, the
outputs of processing circuits 54 and 55 are added by adder 56 to combine output 57. Now,
when the receiver size is given, the frequency f c that divides the low frequency range and the
high frequency range is determined so as to satisfy the following conditional expression when C
is the speed of sound. f c> c / L (1) The minimum frequency fcO satisfying the second condition
is 6 BOHz, for example, when the receiver scale is 50 cym, and the phase synthesis reception is
performed as shown in FIG. 7B. It will have the directional characteristics shown. When the value
of fc becomes smaller than c / L, the directivity characteristic in phase-combined reception
becomes close to non-directionality, and improvement of the SN ratio can not be expected. On
the other hand, if the fcO value is made sufficiently large, the directional characteristics in phasecombined sound reception become sharp, but the bandwidth of the low frequency band where
adaptive array sound reception is increased and the amount of signal processing is increased. It
also suffers from the disadvantage that the performance of the area also deteriorates. Therefore,
in practical use t, the bond of fc is selected in the range satisfying at least the condition of the
expression (1) in consideration of the required sharpness of the directional characteristics, the
allowable signal processing amount and the like. Next, the processing circuit 54 for realizing
phase synthesis reception and the processing circuit 55 for realizing adaptive array reception
will be described. A microphone array in which the microphone elements are arranged on a
straight line as in the microphone array 12 shown in FIG. 6 is used as the pedestal 51, and the
microphone elements are arranged in the same direction as the case of FIG. Perpendicular to the
straight line. From the sound receiver 51, a plurality of signals received by a plurality of
microphone elements are output, and are supplied to the processing circuits 54 and 55L through
the high-pass ii1 over-filter 52 and the low-pass filter 53, respectively. FIG. 3 shows an example
of the configuration of the processing surface 154 for realizing phase synthesis reception. 61. A
03-05-2019
6
signal corresponding to each sound reception output of M microphone elements supplied from
the high pass filter 52, 61. 61. l represents the processing filters 621 to 62, respectively. Are
added by the adder 63 and output as the output signal 64. The processing circuit 54 includes
processing filters 621 to 62. In the case where all the characteristics are 1 (the same
characteristics as in the case of not passing through the filter), the configuration is the same as
that shown in FIG. 6, and superdirective characteristics are realized.
Further, by giving appropriate characteristics to the processing filters 62 to 62Hc, it is possible
to correct variations in the characteristics of the individual microphone elements and to reduce
variations in the directivity characteristic beam width depending on the frequency. Processing
filter 62. more appropriate delay characteristics. ∼62. If given to I, control of the target
direction (direction of directional characteristic beam) can also be performed. FIG. 4 shows a
configuration example of the processing circuit 55 for realizing adaptive array reception. The
signals ul ′ ′ ′-uH corresponding to the respective sound receiving outputs of the M
microphone elements supplied from the low pass filter 53 have their respective adjacent
subtractors 72. ∼72.. The subtractor outputs 1 to 1 are subtracted from each other by an
adaptive filter 74. The difference between the signal ul and the signal ul is supplied to the
subtracter 75 through. In this configuration example, first, M-1 signals 1∼1 to V from which the
in-phase components included in M signals u1 to u8 are removed using M-1 subtractors 72 ° to
72 □ 1 Synthesize% -1. From the sound receiving condition, the target signal components
included in the M signals u1 to LIM are in phase, and the noise components have a phase
difference. Therefore, the signal 1∼1 to VM-1 is a signal that does not include the target signal
but only noise. These signals 1 to 1. l-1 are adaptive filters 741 to 74. It is input to-. Adaptive
filter 74. ∼74.. The output of is input to the subtractor 75 and is subtracted from the signal
u1. Adaptive filter 74 □ to 74. The output cuff 6 of the subtractor 75 is fed back to l−, and the
filter coefficient is corrected so as to minimize the root mean square value of the output cuff 6.
As a correction algorithm, an adaptive algorithm such as a learning identification method or LMS
method is used. In this adaptive filter operation, minimizing the output cuff 6 is equivalent to
minimizing the noise component contained in the output cuff 6, and also, only the noise in the
signal V, ~ VM-1 As the target signal component included in the signal u1 is the signal V because
it is included. Not affected by subtracting ~ V 1. Therefore, the configuration example of FIG. 4 is
a configuration example of an adaptive array that minimizes noise under the constraint of not
distorting the target signal at all. Although the configuration example of the processing circuit 55
for realizing the adaptive array sound reception has been described above with reference to FIG.
4, it is also possible to use another adaptive array sound reception system. For example, a
method for minimizing noise while allowing a certain range of distortion to a target signal (gold
circle: “AMNOR (Adaptive Noise Suppression Microphone Array)” technical report of the
Institute of Electronics and Communication Engineers, EA 85-36.1985) Compared with the
method of FIG. 4 based on the constraint condition that distortion is not allowed, the amount of
calculation increases, but it is an effective method as a method with a large SN ratio improvement
03-05-2019
7
effect.
[Effects of the Invention] As described above, the present invention divides the sound reception
signal into two frequency bands, receives the high frequency band in the phase combination type
sound reception system, and receives the low frequency area in the adaptive array system. It is
characterized by sound. According to the present invention, the problem of performance
degradation in the low frequency range, which is a drawback of the conventional phasecombination receiving system, and performance degradation in the very low frequency range,
which is a disadvantage of the adaptive array system, is improved Be done. As a result, for
example, high SN ratio sound reception over a wide band such as several tens of lz to several tens
of kHz can be realized. In addition, since the sampling frequency can be lowered by using the
adaptive array ancient only in the low frequency region, the effect of significantly reducing the
amount of calculation compared to the conventional adaptive array covering the entire frequency
band is realized. Furthermore, the size of the receiver can be made smaller than that of the
conventional phase-combination sound-receiving system. Therefore, the present invention can
provide a very widely used sound receiving system including general use such as high-quality
speech communication system and speech recognition system.
[0002]
Brief description of the drawings
[0003]
FIG. 1 is a block diagram showing a first embodiment of the present invention, FIG. 2 is a block
diagram showing a second embodiment of the present invention, and FIG. 3 realizes the phase
synthesis sound reception in FIG. FIG. 4 is a block diagram showing an example of the
configuration of the processing circuit 54, FIG. 4 is a block diagram showing an example of the
configuration of the processing circuit 53 for realizing the adaptive array reception in FIG. 2, and
FIG. Fig. 6 shows an example of the configuration of the phase synthesis microphone, and Fig. 7
shows an example of the configuration of the directivity of the phase synthesis microphone using
a linear array microphone array, Fig. 8 The figure is a block diagram showing an example of the
configuration of the adaptive microphone array, FIG. 9 shows an example of the frequency
response characteristic to the target signal and noise of the adaptive microphone array (16 task,
band 4 k) Iz), FIG. Is the target signal of the adaptive microphone array Fig. 11 shows an example
of frequency response characteristics to noise (32 knops, band 4 kHz), and Fig. 11 shows an
example of frequency response characteristics against the target signal of the adaptive
microphone array and noise (16 taps, band 0.8 kHz) FIG.
03-05-2019
8
Patent Applicant Nippon Denshin Telephone Co., Ltd. Attorney Attorney Kusano Takashi, + 12
tree 2 Figure 3 Figure 4, '175 Figure 876 Figure 70 f-680 Hz 1 = + 360) (2-8)
03-05-2019
9
Документ
Категория
Без категории
Просмотров
0
Размер файла
22 Кб
Теги
jph04178100
1/--страниц
Пожаловаться на содержимое документа