close

Вход

Забыли?

вход по аккаунту

JPH06327087

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 JPH06327087
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a
voice detecting apparatus provided with a noise canceling function unit using an adaptive filter.
[0002]
2. Description of the Related Art In general, a voice detection apparatus provided with a noise
cancellation function unit is used to detect voice in a noisy environment.
[0003]
FIG. 5 shows a conventional speech detection apparatus 50. As shown in FIG.
The voice detection apparatus 50 includes a microphone for detecting voice and noise and a first
detection unit 51 including an analog-digital converter, a second detection unit 52 including a
microphone and an analog-digital converter for detecting only noise, an adaptive filter Adaptive
filter 53 and subtraction processing unit 54 are configured. The voice detection apparatus 50 as
described above obtains the digital first signal Da from the first detection unit 51 and obtains the
digital second signal Db from the second detection unit 52. Then, the digital second signal Db is
filtered by the adaptive filter 53, and the subtraction processing unit 54 subtracts the output
signal Dc of the adaptive filter 53 from the digital first signal Da, whereby the output of the
subtraction processing unit 54 is output. Is a voice signal including only voice components whose
03-05-2019
1
noise components have been canceled.
[0004]
By the way, the principle of the above-mentioned conventional speech detection apparatus 50 is
to estimate the transfer function in the speech detection circuit system by the detection signal
from the second detection unit 52 which detects only the noise. It is in. However, when including
noise with a wide dynamic range such as white noise in general, the characteristic curve Mg of
the transfer function G [dB] with respect to the frequency f [Hz] exhibits a complicated change as
shown in FIG. The characteristic curve Mp of the phase P [rad] with respect to [Hz] also exhibits a
complex change as shown in FIG. For this reason, when noise with a wide dynamic range is
included, in the conventional voice detection apparatus 50, the amount of noise cancellation
becomes insufficient, and there is a problem that voice can not be detected with high quality.
[0005]
SUMMARY OF THE INVENTION The present invention solves the problems existing in such
conventional techniques, and a voice detection apparatus capable of sufficiently removing noise
and detecting high-quality voice even when noise is included in a wide dynamic range. The
purpose is to provide
[0006]
According to the present invention, a first microphone 2 capable of detecting voice and noise, a
second microphone 3 capable of detecting only noise, and detection signals from the respective
microphones 2 and 3 are respectively digitized. Analog-to-digital conversion processing unit 4 for
converting one signal Df and digital second signal Ds, filtering the digital second signal Ds by the
adaptive filter, and subtracting the output signal of the adaptive filter from the digital first signal
to filter noise In particular, when configuring the voice detection device 1 including the noise
cancellation function unit 5 for outputting the canceled digital output signal Do and the digitalto-analog conversion processing unit 6 for converting the digital output signal Do into the analog
signal Ao, First signal Df and digital second signal D Of the digital filter unit 7 that divides each
into a plurality of bands Za, Zb, and Zc, a plurality of noise cancel function units 5a, 5b, and 5c
according to each band, and digital outputs of the noise cancel function units 5a, 5b, and 5c It is
characterized by comprising an addition processing unit 8 for adding the signals Doa, Dob, Doc.
[0007]
In this case, the first microphone 2 and the second microphone 3 are integrally provided in a
03-05-2019
2
single casing 9, and the in-phase or anti-phase signal component in each detection signal from
the first microphone 2 and the second microphone 3 is It is desirable to include the same sound
signal component removal circuit 10 to be removed.
[0008]
According to the voice detection device 1 according to the present invention, while the voice
including noise is detected by the first microphone 2, only the noise is detected by the second
microphone 3, and the detection signals from the microphones 2, 3 are detected. Are converted
into a digital first signal Df and a digital second signal Ds by the analog-digital conversion
processing unit 4, respectively.
[0009]
The digital first signal Df and the digital second signal Ds are divided by the digital filter unit 7
into a plurality of bands Za, Zb, and Zc, respectively.
Then, noise cancellation processing is performed by the plurality of noise cancellation functional
units 5a, 5b, and 5c corresponding to each band.
That is, in one noise cancellation function unit 5a (5b and 5c are also the same), the digital
second signal Dsa is filtered by the adaptive filter 11a, and the output signal Dma of the adaptive
filter 11a is subtracted from the digital first signal Dfa. .
As a result, digital output signals Doa, Dob, Doc in which noise is canceled for each band are
obtained at the outputs of the noise cancellation functional units 5a, 5b, 5c.
Then, the digital output signals Doa, Dob, Doc are added by the addition processing unit 8, and
the digital output signal Do output from the addition processing unit 8 is converted into an
analog signal Ao by the digital-analog conversion processing unit 7. Therefore, this analog signal
Ao becomes an audio signal in which noise is canceled.
[0010]
Next, preferred embodiments according to the present invention will be described in detail with
03-05-2019
3
reference to the drawings.
[0011]
First, the configuration of the voice detection device 1 according to the present embodiment will
be described.
[0012]
FIG. 1 shows a block circuit of the voice detection apparatus 1, wherein 2 is a first microphone
capable of detecting voice and noise, and 3 is a second microphone capable of detecting only
noise.
Each of the microphones 2 and 3 is integrally provided in a single casing 9 as shown in FIG.
In this case, the microphones 2 and 3 use directional microphones, and the sound receiving units
2f and 3f in the microphones 2 and 3 are disposed with the 180 ° opposite directions. The
microphones 2 and 3 may be separately disposed. Reference numeral 10 denotes a homonymous
signal component removal circuit for removing signal components of reverse phase (or in phase)
in detection signals from the first microphone 2 and the second microphone 3.
[0013]
An analog-digital conversion processing unit 4 converts the detection signal from the first
microphone 2 output from the same sound signal component removal circuit 10 into a digital
first signal Df, and the same sound signal component removal circuit 10 And an analog-to-digital
converter 4s that converts a detection signal from the second microphone 3 output from the
second microphone 3 into a digital second signal Ds.
[0014]
A digital filter unit 7 divides the digital first signal Df and the digital second signal Ds into a
plurality of bands Za, Zb, and Zc, respectively.
03-05-2019
4
In the embodiment, as shown in FIG. 2, the divided bands Za, Zb and Zc are respectively set to the
low band, the middle band and the high band. Therefore, the digital filter unit 7 filters the low
band filter 12a, 13a having the same property to filter the low band, and the mid band filter 12b,
13b having the same property to filter the mid band. It is comprised by a pair of high pass filters
12c and 13c which have the same characteristic. In this case, it is desirable that each of the
filters 12a to 13c use a digital FIR filter having linear phase characteristics, which can prevent
phase shift in each band.
[0015]
A noise cancellation function unit 5 includes three noise cancellation function units 5a, 5b, and
5c corresponding to each band, that is, low, middle, and high frequencies. The noise cancellation
function unit 5a for processing the low band is the adaptive filter 11a for filtering the digital
second signal Dsa obtained from the low band filter 13a, the output signal Dma of the adaptive
filter 11a and the low band filter on the digital first signal Df side A subtraction processing unit
14a that performs subtraction processing on the output signal Dfa from 12a is configured. On
the other hand, the noise cancellation functional units 5b and 5c for processing the middle and
high frequencies are also configured in the same manner as the noise cancellation functional unit
5a. Reference numerals 11b and 11c denote adaptive filters, and reference numerals 14b and
14c denote subtraction processing units. Each adaptive filter 11 a, 11 b, 11 c is connected to the
state signal generator 15.
[0016]
An addition processing unit 8 has a function of adding the digital output signals Doa, Dob, and
Doc output from the noise cancellation function units 5a, 5b, and 5c. A digital-to-analog
conversion processing unit (digital-to-analog converter) 6 converts the digital output signal Do
output from the addition processing unit 8 into an analog signal Ao.
[0017]
Next, the operation of the voice detection device 1 according to the present embodiment will be
described.
[0018]
First, at the time of voice non-generation, the filter coefficients of the adaptive filters 11a, 11b,
03-05-2019
5
and 11c included in the noise cancellation function units 5a, 5b, and 5c are set (changed).
[0019]
That is, in this case, the first microphone 2 detects a voice including noise, and the second
microphone 3 detects only the noise.
A detection signal (analog signal) from each of the microphones 2 and 3 is applied to the same
sound signal component removal circuit 10, and an opposite phase (or in phase) signal
component in each detection signal, for example, a collision sound when an object etc. Etc. are
removed.
Then, the detection signals output from the same sound signal component removal circuit 10 are
respectively applied to the analog-digital converters 4f and 4s, and the detection signals from the
first microphone 2 are converted into the digital first signal Df, and The detection signal from the
microphone 3 is converted into a digital second signal Ds.
[0020]
On the other hand, the digital first signal Df and the digital second signal Ds are divided by the
digital filter unit 7 into three bands Za, Zb, and Zc of low, middle, and high, respectively. That is,
from the digital first signal Df and the digital second signal Ds applied to the low pass filters 12a
and 13a, the digital first signal Dfa and the digital second signal Dsa including only the low pass
component are obtained, respectively, The digital first signal Df and the digital second signal Dsb
containing only the mid-range component are respectively obtained from the digital first signal
Df and the digital second signal Ds applied to the for-use filters 12b and 13b, and the high-pass
filter 12c , 13c from the digital first signal Df and the digital second signal Ds, respectively obtain
a digital first signal Dfc and a digital second signal Dsc including only high frequency
components.
[0021]
Then, the filter coefficients of the adaptive filters 11a, 11b, and 11c included in the noise
03-05-2019
6
cancellation functional units 5a, 5b, and 5c are changed by the noise cancellation functional units
5a, 5b, and 5c according to the respective bands. That is, in the noise cancel function unit 5a in
the low band, a difference obtained by filtering the digital second signal Dsa by the adaptive filter
11a and further subtracting the output signal Dma of the adaptive filter 11a from the digital first
signal Dfa by the subtraction processing unit 14a. While the signal Dea is obtained, the filter
coefficients of the adaptive filter 11a are changed so that the differential signal Dea becomes
zero. At this time, the state signal generator 15 applies the state signal Sc for changing the filter
coefficient to the adaptive filter 11a. On the other hand, the filter coefficients of the noise
cancellation functional units 5b and 5c are also changed in the same manner as the noise
cancellation functional unit 5a.
[0022]
On the other hand, at the time of actual voice detection, the state signal generator 15 applies the
state signal Ss for fixing the filter coefficient to each of the adaptive filters 11a, 11b, and 11c. In
this case, the filter coefficient is fixed by the state signal Ss. In addition, the change width (step
gain) of the filter coefficient may be smaller than the change width by the state signal Sc by the
state signal Ss. Therefore, the method of changing the filter coefficient can be changed by the
state signals Sc and Ss.
[0023]
Then, by similarly picking up voice and noise from each of the microphones 2 and 3, each of the
subtraction processing units 14a, 14b and 14c outputs the digital first signals Dfa, Dfb and Dfc to
the output signals of the adaptive filters 11a, 11b and 11c. Doa, Dob, Doc are subtracted, and the
noise-canceled digital output signals Doa, Dab, Dac are output. Then, the digital output signals
Doa, Dob, Doc are added by the addition processing unit 8 and output as a digital output signal
Do, and the digital output signal Do is converted into an analog signal Ao by the digital-analog
conversion processing unit 7 Get the voice signal of the noise canceled target.
[0024]
As mentioned above, although an Example was described in detail, this invention is not limited to
such an Example. For example, the voice is a concept including all the target sounds to be
detected. Also, although the embodiment exemplifies the case where the band is divided into
03-05-2019
7
three, it may be divided into two or four or more. On the other hand, it is not necessary to
provide the same sound signal component removal circuit. In addition, if necessary, a band pass
filter (digital FIR filter) may be inserted before the digital-to-analog conversion processing unit. In
addition, the configuration, method, and the like of the details can be arbitrarily changed without
departing from the scope of the present invention.
[0025]
As described above, the voice detection device according to the present invention comprises: a
digital filter portion for dividing the first digital signal containing voice and noise and the second
digital signal containing only noise into a plurality of bands; A plurality of noise cancellation
function units corresponding to respective bands for filtering out the second signal by the
adaptive filter and subtracting the output signal of the adaptive filter from the digital first signal
to cancel the noise, and outputting a digital output signal, and each noise Since the addition
processing unit for adding the digital output signals of the cancellation function unit is provided,
the remarkable effect of sufficiently removing noise and detecting high-quality voice even when
the noise includes a wide dynamic range Play.
03-05-2019
8
1/--страниц
Пожаловаться на содержимое документа