close

Вход

Забыли?

вход по аккаунту

JPH0576094

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 JPH0576094
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a
sound collecting apparatus for collecting an audio signal with a microphone.
[0002]
2. Description of the Related Art In recent years, camera-integrated video tape recorders have
become widespread and video and audio recording has become more frequent. The image is
recorded through the lens and the object is easy to zoom in, but it is often easy to use a
unidirectional microphone for sound recording, and directivity control is not easy. An example of
controlling the directivity is shown, for example, in the technical research report EA 89-92 of the
Institute of Electronics, Information and Communication Engineers.
[0003]
Below, the directivity control apparatus which is the conventional sound collection apparatus is
demonstrated. FIG. 2 is a block diagram of this conventional directivity control apparatus. In FIG.
2, 1 is a plurality (23) of microphones, which are arranged in a line at intervals of 2 cm. 2, a
plurality of analog-digital converters (hereinafter referred to as "ADC") connected to each of the
microphones 1, a plurality of digital filters 3 to which respective output signals of the ADC 2 are
inputted, and 4 respective outputs of the digital filters 3 Adder for adding signals, 5 is a digital-
03-05-2019
1
analog converter (hereinafter referred to as DAC) for converting the output signal of the adder 4
into an analog signal, 6 is an output terminal connected to the DAC 5 and from which collected
sound is output It is.
[0004]
The operation of the directivity control apparatus configured as described above will be
described below.
[0005]
The signal collected and output by the microphone 1 is converted into a digital signal by the ADC
2 and added to the digital filter 3.
The respective output signals of the digital filter 3 are added by the adder 4 and then converted
back to an analog signal by the DAC 5 and output to the output terminal 6. Here, the digital filter
3 group constitutes a two-dimensional digital filter. Then, by controlling the coefficient of each
digital filter 3 and changing the sampling frequency of the digital circuit, it is possible to change
the directivity characteristic collected by the microphone 1.
[0006]
However, in the above-described conventional configuration, a large number of microphones and
ADCs must be provided, and the digital filter 3 uses a high-order (30th-order) FIR (Finite Impulse
Response) filter. The problem is that the scale is large.
[0007]
The present invention solves the above-mentioned conventional problems, and an object of the
present invention is to provide a sound collecting device which is small in size and can collect a
target sound.
[0008]
SUMMARY OF THE INVENTION In order to achieve the object, a sound collection apparatus
according to the present invention comprises two band pass filter groups for separating an input
signal into a plurality of frequency bands, and a band pass filter group. A level detector group for
detecting a level of an output signal of a band pass filter and an output signal of a pair of level
03-05-2019
2
detector groups connected to band pass filters of the same frequency band of each of the two
band pass filter groups are stored. A signal processor group that calculates a time difference
between level change points output from the memory and each of the pair of level detectors from
the memory contents of the memory and selects and outputs a coefficient value according to the
calculation result, and one band pass filter group And an adder for adding an output signal of the
multiplier group, and a multiplier group for multiplying an output signal of each of the frequency
bands of the above and an output signal of the respective signal processor group.
[0009]
According to the present invention, according to the above configuration, the signals collected by
the two microphones are respectively input to the band pass filter group and separated into
signals of narrow frequency bands.
Then, the level of the signal separated by the band pass filter of the same frequency band is
detected by the pair of level detectors among the two band pass filters, and the signals output by
the pair of level detectors are stored in the memory The time difference between level change
points is calculated by the signal processor from the memory contents of the memory, and a
coefficient whose value decreases from 1 when the time difference becomes large is output to
the multiplier.
In the multiplier, the signal separated by the band pass filter is multiplied by the coefficient
selected as described above.
As described above, the same process as described above is performed for the output signals of
all the band pass filter pairs, and the output signals of the multiplier group are added by the
adder to form a wide frequency band signal. Here, the fact that there is no time difference
between the level change points of the pair of level detectors indicates that the sound source is
located at an equal distance from the two microphones, and that the time difference is large
means that the sound source is separated from that position Indicates that. That is, the sound of
the target sound source equidistant from the two microphones passes and the sound away from
the target is attenuated. Thus, the target sound can be collected.
[0010]
An embodiment of the present invention will be described below with reference to the drawings.
03-05-2019
3
[0011]
FIG. 1 is a block diagram of a sound collecting apparatus according to an embodiment of the
present invention.
In FIG. 1, 11, 12 are microphones placed at a certain distance (for example, 20 cm), 13, 14 are
ADCs for converting the output signals of the microphones 11, 12 into digital signals, and 15, 16
are output signals of the ADCs 13, 14 A band pass filter (hereinafter referred to as "BPF") group
for frequency band division is formed by BPFs 151 to 15n and BPFs 161 to 16n. Reference
numeral 17 is a first level detector for detecting the output signal level of the BPF 151, 18 is a
second level detector for detecting the output signal level of the BPF 161, and 19 is an output of
the first and second level detectors 17, 18 A signal processor adds a signal, stores the value in
the memory 20, determines the time difference between level change points, and outputs a
coefficient according to the time difference, 21 outputs the output signal of the BPF 1151 and
the signal processor 19. A multiplier that multiplies coefficients, an adder 22 to which the output
signal of the multiplier 21 is added, a DAC 23 that converts the output signal of the adder 22 into
an analog signal, and an output of the sound collection result obtained by the DAC 23 Output
terminal. Although not shown, the same is also applied to BPFs 152 to 15n and BPFs 162 to 16n
to detect the levels of the output signals of BPF group 15 and BPF group 16 in the same manner
as BPF 151 and BPF 161, and to perform multiplication. It has a configuration. Therefore, 1 to n
multiplication results are added to the adder 22.
[0012]
The operation of the sound collection apparatus of the present embodiment configured as
described above will be described below.
[0013]
The signal obtained by the microphone 11 is converted into a digital signal by the ADC 13.
The voice signal converted into the digital signal is divided into n frequency bands at BPFs 151
to 15 n in the BPF group 15. Similarly, the audio signal obtained by the microphone 12 passes
through the ADC 14 and is divided by the BPFs 161 to 16 n in the BPF group 16. The output
03-05-2019
4
signals of BPF 151 and BPF 161 through which signals of the same frequency band pass are
applied to first and second level detectors 17 and 18, respectively, and the average value of the
signals is detected. The detected average value is added to the signal processor 19, divided into a
certain time, and stored in the area 1 and the area 2 of the memory 20, respectively. The signal
processor 19 detects the change point of the value stored in the area 1 and searches the area 2
for the same pattern as the change point. If the same pattern is found, the time difference at
which the change point is stored is known from that location. Then, a coefficient value
corresponding to the time difference is selected from the read only memory in the signal
processor 19 and output. As the coefficient value, 1 is selected when the time difference is zero,
and a value smaller than 1 is selected when the time difference is increased. The coefficient value
thus selected is applied to the multiplier 21 and multiplied with the output signal of the BPF 151
to change the level. The signal processor 19 operates to select a coefficient value when the
output signal of the level detector 17 exceeds a set value, and when the level is less than the set
value, the coefficient value holds the previous value. The same signal processing as described
above is performed on the pass signals of BPFs 152 to 15 n and BPFs 162 to 16 n. The output
signals of the n multipliers thus obtained are added by the adder 22, and the addition result is
converted back into an analog signal by the DAC 23 and output to the output terminal 24. If
there is no time difference between the level change points of the output signals of BPF group 15
and BPF group 16, the signal source is equidistant from microphones 11 and 12, that is, the
target sound source, and the time difference becomes large because the sound source is away
from it The sound separated from the target sound source is reduced in coefficient by the
multiplier 21 because the coefficient output from the signal processor 19 becomes smaller. The
signals obtained from the microphones 11 and 12 are divided into many (for example, 32)
frequency bands from BPFs 151 to 15n and BPFs 161 to 16n, so even if the sound other than the
target sound is input, the target sound is different. The coefficient selected by the signal
processor 19 in the frequency band through which the signal passes does not change, and the
sound collection level does not change. Also, it is possible to prevent voice level fluctuation by
not selecting a coefficient smaller than a coefficient for a few seconds (for example, 5 seconds)
after a coefficient is selected.
Since the adder 22 adds the signals of all the divided frequency bands, it is possible to obtain a
collected sound signal in which the non-target signal is attenuated. When the coefficient value
corresponding to the time difference is selected by the signal processor 19, the sound collection
characteristic of the target sound can be changed by slightly modifying the coefficient value.
[0014]
In the embodiment, an example in which sound sources equidistant from the two microphones
03-05-2019
5
11 and 12 have been described as the target sound source has been described. At the time of
determination, the time difference equivalent to the time for sound to travel between the
microphones 11 and 12 is given as an offset value, and the offset value is subtracted from the
time difference so that the target sound source is on the line connecting the microphones 11 and
12 It may be configured. Also, since the BPFs 15 and 16 and ADCs connected thereto, level
detectors, signal processors, multipliers, etc. can be configured using one digital signal processor
in time division, the circuit scale Will not grow.
[0015]
As described above, according to the present invention, two sets of band pass filters, a level
detector group for detecting the signal level of the band pass filters, and a level change point
output from a pair of level detectors. A signal processor group that calculates a time difference
and selects a value according to the time difference, a multiplier group that multiplies an output
signal of a band pass filter group and a coefficient value output from the signal processor group,
and an output signal of a multiplier group By adding an adder that adds up, it is possible to
detect an arrival time difference of an unintended audio signal, select a small coefficient value,
multiply it with the input signal and attenuate the level, and the target sound is effectively
effective. Can be collected.
03-05-2019
6
1/--страниц
Пожаловаться на содержимое документа