close

Вход

Забыли?

вход по аккаунту

?

JPH11298990

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 JPH11298990
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to
audio devices, and more particularly to an audio device that controls the loudness of an audio
signal in a noisy environment to be equal to the loudness of an audio signal in a noiseless
environment.
[0002]
2. Description of the Related Art When listening to music under noise such as car audio, the noise
causes a masking phenomenon of the auditory sense (a phenomenon that interferes with the
hearing), and the music becomes difficult to hear. In order to cope with such problems, various
devices such as an auto volume device and a loudness correction device have been proposed. FIG.
13 is a block diagram showing an example of an auto volume apparatus, in which 16 is a gain
control section with variable gain to which an audio signal is input, 17 is an amplifier, and 18 is a
speaker for emitting audio sound according to the audio signal to acoustic space. , 19 is a
microphone for detecting a synthesized sound signal of audio sound and noise at a
predetermined observation point in the acoustic space, 20 is a detection unit for calculating an
average level of the audio signal, 21 is a synthesized sound signal detected by the microphone
(microphone Reference numeral 22 denotes a gain determination unit which determines a gain
based on the difference between the average levels of the microphone signal and the audio
signal. The detection units 20 and 21 calculate the average level of the audio signal and the
microphone signal, respectively, the operation unit 22 'subtracts the audio signal average level
from the microphone signal average level, and the gain determination unit 22 calculates the
08-05-2019
1
difference as the noise average level estimated value. The gain is determined based on the noise
average level, and the gain control unit 16 multiplies the audio signal by the determined gain. As
a result, if the noise increases, the gain increases and the volume increases, and if the noise
decreases, the gain decreases and the volume decreases.
[0003]
The principle of calculating the noise average level estimated value in the auto volume apparatus
is as follows. Assuming that the audio signal is S and the noise signal is N, the signal M captured
by the microphone 19 is a superposition of the audio signal S and the noise signal N. Therefore,
the average signal power E [M2] captured by the microphone is E [M2] = E [(S + N) 2] = E [S2] +
2E [S.N] + E [N2]. Here, E [•] is an expectation value operator, which is equivalent to taking a long
time average. By subtracting the average power of the audio signal from the average power of
the signal captured by the microphone, E [M2] -E [S2] = E [S2] + 2E [SN] + E [N2] -E [S2] It
becomes 2E [S * N] + E [N2]. Here, since the audio signal S and the noise signal N are
uncorrelated, by taking long time averaging, the term of E [S · N] approaches 0 and the average
power of the noise signal N can be obtained. From the above, in order to improve the calculation
accuracy of the noise average level estimated value, a long time average is required.
[0004]
FIG. 14 is a block diagram showing an example of a loudness correction apparatus, wherein 23 is
a gain control section with variable gain to which an audio signal is input, 24 is an amplifier, and
25 is a speaker for emitting audio sound according to the audio signal into acoustic space. , 26 is
a gain determination unit that determines the gain of each frequency band according to the
loudness level curve based on the level of the input audio signal. The unit of "sound loudness
(loudness)" perceived by human beings is sone, and the magnitude of a pure tone of 1 KHz, 40
dB is 1 sone. Because it is based on human perception, 2sone sounds twice as large as 1sone.
Loudness varies not only with the intensity of the sound but also with the frequency spectrum.
FIG. 15 is a graph obtained by connecting the sound pressure levels of pure tones having the
same loudness as a 1 kHz pure tone in the absence of external noise, and is called an equal
loudness level curve. That is, the equal loudness level curve is a plot of the levels of other
frequencies at which one sounds as loud as a 1 kHz sine wave. The equal loudness level curve
indicates that as the level decreases, if the level in the low frequency range and the high
frequency range is not raised, the sound may be heard smaller or less audible than the sound in
the intermediate frequency range. In the loudness correction apparatus shown in FIG. 14, the
gain determination unit 26 determines the gain for each frequency based on the equal loudness
08-05-2019
2
level curve so that the input audio signal level can be heard to be the same in magnitude. Based
on the level adjustment for each frequency.
[0005]
The conventional auto volume apparatus of FIG. 13 has the following problems. That is, (1) it
takes a long time to separate and output a noise signal with high accuracy, and it is difficult to
change the gain value at high speed. Therefore, the consonant component and the vowel
component of human voice Can not respond to a signal whose level changes rapidly, and
consonant components and aftertone components with small levels can not be heard. (2) When
the fluctuation of the audio signal is large, an error occurs in the estimated value of the average
power of the noise signal, so that accurate volume control can not be performed. (3) While the
audio signal output from the gain control unit 16 passes through the amplifier 17 and the
acoustic space and enters the microphone 19, the gain and the frequency characteristic change
variously, but there is no part that copes with such a point. (4) The gain set in the gain control
unit 16 changes from moment to moment, and thus the volume of the audio signal input to the
microphone 19 also changes from moment to moment, but there is no part to cope with such a
change. (5) In order to change the gain uniformly over the entire frequency, under low noise
such as in a car where the spectrum is biased noise, the low frequency of music can not be heard,
and the high frequency is likely to become too large.
[0006]
Also, in the loudness correction apparatus of FIG. 14, (1) attention is focused only on the audio
signal level, and the influence of noise is not taken into consideration, so the low frequency of
music can not be heard especially in the vehicle compartment. Since the audio signal level from
the music source is not the level at the actual ear point, there are problems such as that the
amount of correction is not optimal due to changes in the audio system and the car. From the
above, the object of the present invention is an audio device which solves the problems of the
prior art and has higher control accuracy than the prior art and which closely adheres to human
auditory characteristics and can enjoy music well even under noise It is to provide. Another
object of the present invention is to provide an audio device which makes it possible to listen to
music such as car audio as much as possible in noise but without noise.
[0007]
According to the present invention, the above object is achieved by: (1) a noise separation unit
08-05-2019
3
for separating a noise signal from a synthetic speech signal of audio sound and noise detected at
a predetermined observation point in an acoustic space; (2) A gain for making the loudness of an
audio signal in a noisy environment equal to the loudness of an audio signal in a noiseless
environment, a gain that makes the audio signal level correspond to a signal level, and a gain that
stores signal level characteristics for each noise level Signal level characteristic storage unit, (3)
Noise level calculation unit for calculating the noise level in each frequency band of the noise
signal output from the noise separation unit, (4) Calculation of audio signal level in each
frequency band of the audio signal Signal level calculation unit, (5) the gain / signal level
characteristic corresponding to the noise level of each frequency band, for each frequency band
A gain determination unit that determines the gain according to the Dio signal level, (6) a signal
gain control unit that causes the gain of each frequency band to act on each frequency band
component of the audio signal, (7) output from the signal gain control unit This is achieved by an
audio device provided with a speaker that emits an audio sound according to the audio signal
into the acoustic space.
[0008]
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS (a) Principle FIG. 1 shows the
correspondence between physical sound pressure levels and the "sound loudness (called
loudness)" felt by humans when the sound is heard. It is shown and called a loudness curve.
In the loudness curve, the horizontal axis is the physical sound pressure level (in units of Sound
Pressure Level SPL (dB)), and the vertical axis is the loudness (in units of sone) obtained by
digitizing the loudness of human feeling. In FIG. 1, (a) is a loudness curve in a quiet environment,
and (b) is a loudness curve under noise. However, (b) is a curve in the noise that the human
minimum audible value increases by about 35 dB, and this curve also changes variously as the
noise changes.
[0009]
The loudness curve shows that a person feels that the sound is the same magnitude if the value
of the loudness on the vertical axis is the same. Therefore, the sound that a person feels at
0.1sone may be a physical sound pressure level of 12 dB SPL in the quiet environment of (a) but
a physical sound pressure level of 37 dB SPL in the noise of (b) is necessary. In other words, in
order to let people hear the sound of 12 dB SPL out of the audio system in a quiet environment,
it is necessary to make the sound of 37 dB SPL out of the audio system under the noise of (b)
08-05-2019
4
There is. In other words, we need to add 25 dB of gain. Also, the sound that a person feels at
1sone is a physical sound pressure level of 42 dB SPL in the quiet environment of (a), but a
physical sound pressure level of 49 dB SPL is necessary under the noise of (b) So, I have to add 7
dB of gain. From the above, it is necessary to change the gain according to the level of the audio
signal even under the same noise. A graph showing the relationship between the audio signal
level and the gain under the same noise is shown in FIG. 2 (solid line). The horizontal axis in FIG.
2 is the physical sound pressure level of the quiet environment (corresponding to the audio
signal level), and the vertical axis is the same size as the audio signal in the quiet environment
under the noise in FIG. This is a gain value required to hear
[0010]
In the above, as shown in FIG. 1 (b) as the noise environment, the case where the minimum
audible value of a person increases by about 35 dB is described, but as shown in FIG. The gain
and signal level characteristics also change according to the noise environment as shown in FIG.
In FIG. 3, (a) is an ideal loudness curve given by the following equation when the loudness is
represented by φ and the sound intensity is represented by I. In fact, since there are
physiological noise such as blood flow noise and extraneous noise, the loudness curve is
expressed by the following equation φ = K (I-Ith), where the intensity of the minimum audio
value under these noises is Ith. (b) is the loudness curve when there is no external noise (it
corresponds to the quiet state in FIG. 1 (a)) with only physiological noise such as blood flowing
noise, and (c) a noise that raises the minimum audio value by 15 dB (D) is the loudness curve
with the noise raising the minimum audio value by 35 dB, (e) the noise with the noise increasing
the minimum audio value by 55 dB Loudness curve.
[0011]
In the audio apparatus of the present invention, (1) gain and signal characteristics (FIG. 2, FIG. 4)
at various noise levels are stored in memory beforehand, and (2) gain according to the noise level
in the actual vehicle interior A signal level characteristic is selected, (3) an optimum gain
according to the audio signal level is calculated with reference to the gain and signal level
characteristic, and (4) the gain is added to the audio signal. This makes it possible to hear an
audio signal equivalent to a quiet environment, even in noise. When a large capacity memory can
not be provided in the audio apparatus, as shown by dotted lines in FIGS. 2 and 4, the gain and
signal level characteristics are approximated by a straight line, and only the slope and intercept
are stored. In this way, the data is only the slope and the intercept, which can be realized even if
the memory is small. Although the above describes the case of the loudness curve whose
08-05-2019
5
frequency is 1 KHz, when the frequency changes, the loudness curve also changes. FIG. 6 is a
loudness curve when the frequency is 100 Hz, (a) is a loudness curve in a quiet environment, (b)
is a loudness curve under noise. Therefore, as another configuration method of the audio
apparatus, (1) gain-signal level characteristics shown in FIG. 2 are stored for each noise level of
each frequency band, and (2) each frequency band in an actual vehicle compartment The gain
and signal level characteristics are selected according to the noise level of (3), the optimum gain
according to the audio signal level is calculated for each frequency band with reference to the
gain and signal level characteristics, and (4) the gain To the audio signal. This can further
improve the control accuracy.
[0012]
(B) Configuration FIG. 6 is a configuration diagram of the audio device of the present invention.
In the figure, 1 is an audio source such as a CD player, 2 is a volume adjustment volume, 3 is a
correction filter as a signal gain control unit, and an optimum gain according to the noise level
and input audio signal level for each frequency In addition, it outputs. 4 is an amplifier for
amplifying an audio signal, 5 is a speaker for emitting an audio sound according to the audio
signal output from the amplifier to an acoustic space (car room), 6 is an audio sound and noise at
a predetermined observation point in the acoustic space A microphone for detecting a
synthesized sound signal, 7 separates and outputs a noise signal from the synthesized sound
signal detected by the microphone in the signal separation unit, and outputs and outputs a
propagation characteristic (transfer characteristic) of an acoustic space to an audio signal It is a
thing. A loudness compensation gain calculation unit 8 (1) stores the gain and signal level
characteristics shown in FIG. 4 for each noise level in the memory MM, and (2) calculates the
level of the noise signal in each frequency band and an audio signal Calculate the level in each
frequency band of (3), determine the gain according to the audio signal level referring to the gain
and signal level characteristics according to the noise level for each frequency band, and set it in
the correction filter 3 It is a thing. A white noise source 9 generates white noise to identify the
impulse response of the acoustic system C in the passenger compartment, and a switch 10 inputs
the white noise to the signal separation unit 7 when identifying the impulse response in the
acoustic space.
[0013]
In the signal separation unit 7, 7a is an adaptive control unit for identifying the impulse response
of the acoustic system C in the vehicle compartment, 71 is an adaptive control unit, and 72 is an
adaptive filter. The adaptive control unit 71 performs, for example, adaptive control using an
08-05-2019
6
LMS (Least Mean Square) adaptive algorithm to identify an impulse response (transfer
characteristic) C ^ of the acoustic system C, and a coefficient corresponding to the impulse
response C ^ is an FIR digital filter It sets to the adaptive filter 72 of a structure. A filter 7b
simulates the impulse response C ^ of the acoustic system C from the speaker 5 to the
microphone 6, generates a signal corresponding to the audio signal at the observation point, and
inputs it to the loudness compensation gain calculation unit 8. 7c is a filter in which the same
characteristic as the correction filter 3 is set, which is rewritten whenever the coefficient of the
correction filter 3 changes, 7d is a signal synthesis unit, and 7e is observed from the synthesized
sound signal output from the microphone 6 It is an operation unit that subtracts an audio signal
at a point and outputs a noise signal.
[0014]
(C) Identification of Impulse Response of Acoustic System C Before starting normal operation, it
is necessary to identify the impulse response of the acoustic system C and set it in the filter 7b.
FIG. 7 is an explanatory diagram of the identification of the impulse response of the acoustic
system C, and the signal path relating to the identification of the impulse response is indicated by
a thick line. When identifying the impulse response of the acoustic system C, the audio source 1
is turned off to stop the output of the audio signal. As a result, the audio sound correction filters
3 and 7c and the loudness compensation gain calculation unit 8 have no influence on the
identification control. In the stopped state of the audio signal, the white noise source 9 is
activated and the switch 10 is turned on. As a result, the white noise passes through the amplifier
4 to reach the speaker 5, is radiated to the vehicle interior sound space, and is detected by the
microphone 6. The white noise is also input to the adaptive controller 7a and subjected to
filtering processing by the adaptive filter 72. The arithmetic unit 7e subtracts the output signal of
the adaptive filter 72 from the detection signal of the microphone 6, and feeds back the
difference as an error signal e to the adaptive controller 7a. The adaptive control unit 71
performs adaptive control according to the LMS algorithm so as to minimize the power of the
error signal e, and determines and sets the coefficients of the adaptive filter 72. Thereafter, the
above-mentioned adaptive control is repeated, and finally the output of the adaptive filter 72 and
the microphone output become equal, whereby the impulse response C ^ of the acoustic system
C is set in the adaptive filter 72. Strictly speaking, the impulse response characteristic (transfer
characteristic) from the input of the amplifier 4 to the microphone 6 is set in the adaptive filter
72.
[0015]
08-05-2019
7
After the identification of the impulse response of the acoustic system C is completed, the
coefficients of the adaptive filter 72 are copied to the filter 7b, and the switch 10 is turned off to
enable normal operation. Such identification control of the impulse response of the acoustic
system C is performed, for example, when the user attaches the product to a vehicle. By
performing this identification control, it is possible to consider the gain characteristics of the
amplifier 4 and the frequency characteristics of the acoustic system which are different for each
user.
[0016]
(D) Normal operation In normal operation, the filter 7b (FIG. 6) applies the impulse response
characteristic C ^ of the acoustic system C to the audio signal input from the volume 2 to
generate an audio signal at the observation point (microphone position) , Loudness compensation
gain calculation unit 8. The arithmetic unit 7 e subtracts the audio signal from the synthesized
sound signal at the observation point output from the microphone 6 to generate a noise signal,
and inputs the noise signal to the loudness compensation gain calculation unit 8.
[0017]
When the noise signal is input, the loudness compensation gain calculation unit 8 calculates the
level in each frequency band of the noise signal, and when the audio signal is input, the level in
the frequency band of the audio signal is calculated. Next, the loudness compensation gain
calculation unit 8 selects a gain / signal level characteristic according to the noise level for each
frequency band, and refers to the gain / signal level characteristic to obtain a gain according to
the audio signal level for each frequency band. It determines and sets to the filter 3 for correction
| amendment, and copies the said gain to the filter 7c simultaneously. That is, the loudness
compensation gain calculation unit 8 determines the gain for making the loudness of the audio
signal in the environment where noise is generated equal to the loudness of the audio signal in
the noiseless environment, and sets the correction filter 3 And the filter 7c. The correction filters
3 and 7c apply the gain of each frequency band to each frequency band component of the input
audio signal and output the result. Thereafter, the same control as described above is repeated,
and even in the presence of noise, music such as car audio can be enjoyed as much as possible
without noise.
[0018]
08-05-2019
8
(E) Loudness compensation gain calculation unit The loudness compensation gain calculation
unit 8 stores in advance the gain and signal level characteristics at various noise levels in the
memory MM, and the gain and signal level according to the noise level in the actual vehicle
interior The characteristic is selected, and the optimum gain according to the audio signal level is
calculated and output for each frequency with reference to the gain and signal level
characteristics. FIG. 8 is a block diagram of such a loudness compensation gain calculation unit.
81 is a first FFT operation unit that outputs the average value of noise signals in each frequency
band by FFT operation, and 82 is an audio signal in each frequency band by FFT operation The
second FFT operation unit that outputs the average value of the noise level, and 83 is a noise
level adjustment unit, which uses the well-known Zwicker loudness calculation method and the
Stevens loudness calculation method to add each auditory characteristic to each frequency The
noise signal level is adjusted, 84 is a gain / signal level characteristic selection unit that selects
the gain / signal level characteristic according to the noise level for each frequency band, 85 is
the gain / signal level characteristic at various noise levels ( FIG. 4) is stored in memories MM1
to MMm, and the audio signal level is referenced with reference to the gain and signal level
characteristics corresponding to the noise level. It is a frequency band gain determination unit
that outputs optimum gains G1 to Gn according to frequency for each frequency.
[0019]
The power spectrum of the noise is not flat and the audio sound is not masked uniformly over
the entire frequency band. That is, since the degree of receiving masking varies depending on the
noise level of each frequency band, it is necessary to calculate and respond to each frequency
(for example, every 1/3 octave band) for each gain. In this case, noise of a certain frequency
causes masking not only to the audio sound of the same frequency but also to the audio sound of
higher frequencies. For this reason, the noise level adjustment unit 83 adjusts the noise signal
level for each frequency in consideration of human auditory characteristics using the loudness
calculation method (ISO 532B) of Zwicker and the loudness calculation method (ISO 532A) of
Stevens. Do.
[0020]
The first FFT operation unit 81 performs FFT operation (Fast Fourier Transform) on the noise
signal for each short time block, and calculates an average level of the noise signal for each
predetermined frequency band. The noise level adjustment unit 83 receives the average level for
each frequency band, and adjusts and outputs the noise signal level of each frequency band
08-05-2019
9
based on the Zwicker loudness calculation method or the Stevens loudness calculation method
described above. The gain / signal level characteristic selection unit 84 selects the optimum gain
/ signal level characteristic according to the noise level of each frequency band for each
frequency band. The second FFT operation unit 82 performs an FFT operation on the audio
signal for each short-time block and calculates an average level of the audio signal for each
predetermined frequency band. Frequency band gain determination unit 85 determines optimum
gains G1 to Gn corresponding to the audio signal level of the corresponding frequency band with
reference to gain / signal level characteristics corresponding to the noise level of each frequency
band, for each frequency. Output. Although the gain and signal level characteristics are provided
commonly in the frequency band for each noise level in the above, the gain and signal level
characteristics may be provided for each noise level in each frequency band. In such a case, (1)
gain / signal level characteristics are stored in the memory MM of the frequency band gain
determination unit 85 for each noise level of each frequency band, and (2) the gain / signal level
characteristic selection unit 84 Select gain and signal level characteristics according to the noise
level of each frequency band in the actual vehicle compartment, and (3) Refer to the gain and
signal level characteristics in the frequency band gain determination unit 85 and select the
optimum according to the audio signal level Gains G1 to Gn are calculated and output for each
frequency band. According to this modification, assuming that the number of noise levels is m
and the number of frequency bands is n, it is necessary to store m × n gain / signal level
characteristics in the memory MM.
[0021]
FIG. 9 is another configuration diagram of the loudness compensation gain calculation unit. In
the figure, 86 is a filter bank provided with a large number of band pass filters BPF1 to BPFn for
separating and outputting each frequency band component of the noise signal, and 87 is a block
for averaging and outputting noise signal components in each frequency band every
predetermined time block Averaging unit 88 includes a plurality of band pass filters BPF1 to
BPFn and a filter bank for separating and outputting each frequency band component of the
audio signal, and 89 is a block for averaging and outputting audio signal components in each
frequency band for each arbitrary time block It is an average part. In the configuration of FIG. 8,
the average level for each frequency band of the noise signal and the audio signal is calculated by
the FFT operation, but in the configuration of FIG. 9, filter banks 86 and 88 are used instead of
the first and second FFT operation units. And block averaging units 87 and 89 to calculate and
output an average level for each frequency band of the noise signal and the audio signal.
[0022]
08-05-2019
10
(F) Correction Filter The correction filter 3 adds the gains G1 to Gn output from the loudness
compensation gain calculation unit 8 to the audio signal, and the following three configurations
can be considered. FIG. 10 is a block diagram of a first embodiment of the correction filter, which
uses a band pass filter bank and a gain variable amplifier, and 31 has a large number of band
pass filters BPF1 to BPFn, and each frequency of the audio signal A filter bank for separating and
outputting band components, 32 has a large number of variable gain amplifiers AMP1 to AMPn,
and multiplies each frequency band component of the audio signal by the gains G1 to Gn of each
frequency band output from the loudness compensation gain calculating unit 8 The variable gain
unit 33 outputs the output signal of each gain variable amplifier and outputs the added signal.
[0023]
The filter bank 31 separates and outputs an audio signal for each frequency band. The gain
varying unit 32 multiplies the output signal of each frequency band by the gain of the
corresponding frequency band output from the loudness compensation gain calculating unit 8.
The addition unit 33 combines the signals of the frequency bands whose gains have been
adjusted and outputs the combined signal. As described above, gain control is performed so as to
compensate masking for each frequency band component of the input audio signal, and it is
possible to approach as close to a noise free state as possible even in a noise environment. The
merit of the first configuration is that it can be realized at lower cost with an analog circuit.
[0024]
Second Configuration FIG. 11 is a configuration diagram of a second embodiment of the
correction filter, which is an example using a frequency sampling filter. The spline function
interpolation unit 41 regards the gains G1 to Gn of each frequency band calculated by the
loudness compensation gain calculation unit 8 as the gains at the center frequency of the band,
and smoothly interpolates between the gains using a known spline function. (Spline
interpolation) The IFFT operation unit 42 performs IFFT (Inverse Fast Fourier Transform from)
on smooth gain characteristics in the frequency domain obtained by spline interpolation, and
converts the gain characteristics into an impulse response in the time domain (impulse response
for every predetermined sampling time) . The FIR type digital filter 43 sets the impulse response
value for each predetermined sampling time as the coefficient of the corresponding tap. In such a
state, the FIR type digital filter 43 applies filtering processing according to the impulse response
characteristic to the input audio signal and outputs it, thereby realizing desired masking
compensation. The merit of this second configuration is that a linear phase filter which can not
08-05-2019
11
be achieved by the first configuration can be realized.
[0025]
Third Configuration FIG. 12 is a diagram showing the configuration of a third embodiment of the
correction filter, which is an example using a frequency domain filter. The spline function
interpolation unit 51 regards the gains G1 to Gn of each frequency band calculated by the
loudness compensation gain calculation unit 8 as the gains at the center frequency of the band,
and interpolates between respective gains using a known spline function. Make the gain
characteristic smooth in the frequency domain. On the other hand, the FFT operation unit 52
performs an FFT operation on the audio signal to convert it into the frequency domain. The
frequency domain filtering unit 53 filters the audio sound signal in this frequency domain with
the smooth frequency / gain characteristic determined by the spline function interpolation unit
51 and outputs it. That is, the frequency domain filtering unit 53 multiplies the gain of the
frequency band corresponding to the audio signal component of each frequency band and
outputs the result. The IFFT operation unit 54 performs IFFT operation on the frequency domain
signal (audio signal component of each frequency band) output from the frequency domain
filtering unit 53, and returns it to a time domain signal. This achieves the desired masking
compensation. In order to realize the linear filtering operation required in the third configuration,
it is preferable to use the well-known overlap-add method or the oversave-add method. The merit
of the third configuration is that computational efficiency is better than that of the second
configuration when the number of filter taps is large.
[0026]
As described above, although the correction filter 3 is realized by the first to third configurations,
when the gain changes suddenly in any configuration, the output waveform becomes
discontinuous. Therefore, the gain is gradually updated by controlling the gain according to the
following equation G (nT) = αG ((n−1) T) + βGm so that a sudden change in gain does not
occur. However, in the above equation, G (n) is the gain characteristic at time nT, G ((n-1) T) is the
gain characteristic at time (n-1) T, and Gm is the loudness compensation gain calculation unit or
spline function interpolation unit The gains α and β calculated by are constants satisfying the
following equation. Although the present invention has been described above by way of
examples, the present invention can be variously modified in accordance with the spirit of the
present invention described in the claims, and the present invention does not exclude these.
08-05-2019
12
[0027]
According to the present invention, according to the present invention, the gain and signal level
characteristics at various noise levels are stored in memory in advance, and the gain and signal
level characteristics corresponding to the noise level in the actual vehicle interior are selected.
Since the optimum gain according to the audio signal level is calculated with reference to the
gain and signal level characteristics, and the gain is added to the audio signal, audio equivalent to
a quiet environment even in noise I can hear the signal. Also, the gain and signal level
characteristics are stored for each noise level of each frequency band, and the gain and signal
level characteristics are selected according to the noise level of each frequency band in the actual
vehicle interior, and the gain and signal level are selected. The control accuracy can be further
improved by calculating the optimum gain according to the audio signal level with reference to
the characteristics for each frequency band and adding the gain to the audio signal.
[0028]
Further, according to the present invention, (1) a filter bank for separating an audio signal into
each frequency band, and (2) a gain of each frequency band calculated to compensate masking
for each frequency band component of the audio signal. Since the correction filter is configured
by the gain multiplication unit to be multiplied and (3) the addition unit to add the output of each
gain multiplication unit, the correction filter can be realized by the analog circuit more
inexpensively. Further, according to the present invention, (1) an interpolation unit which
performs processing to interpolate between gains of each frequency band calculated to
compensate for masking, (2) frequency domain obtained by interpolation processing Since the
correction filter is configured by the conversion unit that converts the gain into an impulse
response in the time domain, and (3) a digital filter in which the impulse response in the time
domain is set, a linear phase filter can be realized.
[0029]
Further, according to the present invention, (1) an interpolation unit that interpolates the gain of
each frequency band calculated to compensate for masking, and (2) a first method of converting
an audio signal from a time domain to a frequency domain. A conversion unit, (3) a frequency
domain filtering unit that applies the gain characteristic of the frequency domain obtained by the
interpolation processing to the audio signal in the frequency domain, Since the correction filter is
configured by the second conversion unit that converts the signal into an audio signal and
08-05-2019
13
outputs the result, the calculation efficiency can be improved even when the number of taps of
the filter is large. Further, according to the present invention, an impulse response of an acoustic
system from a speaker to an observation point is measured, a coefficient corresponding to the
impulse response is set in a filter to simulate the acoustic system, and an audio signal is input to
the filter. Since the filter output is subtracted from the synthesized speech signal detected at the
observation point and the noise signal is output, the car changes and the gain of the amplifier or
the acoustic space, Even when the frequency characteristic changes, this can be coped with.
08-05-2019
14
Документ
Категория
Без категории
Просмотров
0
Размер файла
28 Кб
Теги
jph11298990
1/--страниц
Пожаловаться на содержимое документа