close

Вход

Забыли?

вход по аккаунту

?

JP2008104001

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 JP2008104001
An object of the present invention is to provide an auditory sense improvement circuit which
improves the auditory sense by suppressing the noise level of the audio signal and is adapted to
the auditory sense characteristic. SOLUTION: An auditory sense improvement device shapes the
frequency characteristic of an input audio signal into a predetermined auditory sense
characteristic and outputs the result, and the level difference of the main component signal from
the level difference between adjacent sampling points of the audio signal from the equalizer
circuit. A phase difference, an amplitude value, and a frequency are obtained, thereby performing
filter processing centering on the frequency near the main component signal to remove noise
components, and an instantaneous frequency analysis circuit that outputs the main component
signal; A correction filter, an equalizer control circuit which obtains an error between an output
signal of an instantaneous frequency analysis circuit and an output signal of an auditory sense
correction filter, and an audio signal characteristic of the equalizer circuit and an S / N ratio of
the instantaneous frequency analysis circuit Switch for selecting one of the output signal of the
equalizer circuit and the output signal of the instantaneous frequency analysis circuit Provided
with a door. [Selected figure] Figure 2
Hearing improvement circuit
[0001]
The present invention is a circuit intended for digital audio signals such as an audio circuit for
digital broadcasting, and improves the audibility characteristics of audio information by reducing
noise level, emphasizing signal level, and correcting audio spectrum. The present invention
relates to a hearing improvement circuit.
08-05-2019
1
[0002]
Conventionally, in the reception of a television broadcast or the like, in order to improve the
audibility of an audio signal, the audio signal level is measured, its average level or peak level is
detected, and the gain of the audio amplifier is always constant. By adjusting the above, the
control is performed so that the listening level does not always change to improve the hearing
(see Patent Document 1).
Unexamined-Japanese-Patent No. 6-121393
[0003]
However, in digital broadcast receivers and DVD players, it is premised that multiple speakers are
installed to output audio signals from various directions for the purpose of enhancing the sense
of reality, and that a dynamic range is secured. It was outputting from a delicate minute signal to
a loud sound signal.
[0004]
In an environment in which such an audio device is provided, it is extremely useful to be able to
watch a realistic movie or to watch an orchestra, but it is not necessarily suitable for listening to
human voice.
[0005]
Usually, the audibility characteristics of the ear are the highest in the frequency band of 2 to 4
KHz and become low in the high frequency band and the low frequency band, but the tendency
of the decrease in sensitivity is emphasized as the loudness level becomes smaller.
[0006]
In a wide range, in news programs where high accuracy is required, it is often difficult to
accurately grasp the program contents.
This is a significant problem because the sound pressure level at which the old people can hear is
lower than that of the young hearing people, and the band of the sound to be heard is narrowed.
08-05-2019
2
[0007]
Further, in the conventional system in which the output level of the audio signal is always
constant, the circuit operates at a high level or a low level of the frequency, and the level in the
audible frequency band is not necessarily constant.
[0008]
SUMMARY OF THE INVENTION It is an object of the present invention to provide an auditory
sense improvement circuit which reduces the noise level of an audio signal to improve auditory
sense and matches the audio signal to the auditory sense characteristic.
[0009]
In order to solve the above problems, according to a first technical means of the present
invention, there is provided an equalizer circuit that shapes and outputs frequency
characteristics of an input audio signal to a predetermined auditory characteristic, and adjacent
sampling of audio signals from the equalizer circuit. The phase difference, amplitude value, and
frequency of the main component signal are obtained from the level difference of the signal of
the point, and the filter processing centering on the frequency near the main component signal is
performed to remove noise components and output the main component signal. Together with an
instantaneous frequency analysis circuit for obtaining an S / N ratio, an auditory sense correction
filter having the auditory sense characteristic, and an error between an output signal of the
instantaneous frequency analysis circuit and an output signal of the auditory sense correction
filter And an equalizer control circuit that corrects the auditory sense characteristic.
[0010]
According to a second technical means, in the first technical means, the removal of the noise
component is performed by converting the audio signal from the equalizer circuit into a complex
signal by Hilbert transform and performing complex averaging processing of the complex signal.
The present invention is characterized in that filtering processing is performed to remove
harmonic components whose frequency and amplitude constantly change from the audio signal.
[0011]
A third technical means is the first technical means or the second technical means, wherein the
input voice signal, the output signal of the equalizer circuit, and the instantaneous frequency
analysis circuit are based on the S / N ratio of the instantaneous frequency analysis circuit. A
voice switching switch is provided to select any one of the output signals from the above.
08-05-2019
3
[0012]
The auditory sense improvement circuit of the present invention can correct the audio signal to a
signal that has passed through the auditory sense correction filter, can remove harmonic
components that become noise by complex averaging processing, and can be used for audio
signal in digital broadcasting When viewing a program that prioritizes the ease of listening to
news, such as in the presence of realistic audio effects, it is possible to convert it into an easy-tohear speech spectrum or reduce background noise to ensure that speech contents can be heard.
[0013]
Hereinafter, embodiments of the present invention will be described in detail based on the
drawings.
The case where the auditory sense improvement circuit of the present invention is used in a
digital broadcast receiver will be described with reference to FIG.
The digital broadcast receiver 1 demodulates the received signal received from the antenna by
the demodulator 3 through the tuner (station selection) 2 and separates the demodulated signal
into the video signal and the audio signal by the demultiplexer (DMUX) 4 The video signal is
decoded into an output video signal by the video signal decoding unit 5 and output, and after the
audio signal is decoded into an audio signal by the audio signal decoding unit 6, filter processing
for noise reduction to be described later in the audibility improvement circuit 7 And output as an
audio output.
[0014]
The hearing sense improvement circuit 7 will be described in detail with reference to FIG.
The speech signal always changes with time, and can be divided into a real part and an imaginary
part, as shown in FIG.
08-05-2019
4
The audio signal includes a fundamental frequency component of the main component signal and
a harmonic component which is a noise component.
The fundamental frequency component is stable (substantially constant) in both the
instantaneous frequency and the instantaneous amplitude, but the harmonic component is
unstable (variation) in both the instantaneous frequency and the instantaneous amplitude.
[0015]
In the present invention, the harmonic component which is a noise component is canceled and
the S / N ratio is improved by superimposing the signals using the nature that the harmonic
component is unstable both in instantaneous frequency and instantaneous amplitude.
For that purpose, it is necessary to quickly detect the instantaneous frequency and the
instantaneous amplitude which change with time.
[0016]
The auditory sense improvement circuit 7 receives the signal from the audio signal decoding unit
6 as an input.
Further, the output of the audibility improving circuit 7 is switched by the voice switch 49 and
output.
The voice switching switch 49 switches the voice signal in the best state from among the three
types of signals according to the condition.
[0017]
The first one is directly connected to the terminal A of the voice selection switch 49 so that the
output signal of the voice signal decoding unit 6 is used as it is. The signal input to the terminal A
08-05-2019
5
of the voice switching switch 49 is for bypassing the auditory sense improvement circuit 7 when
viewing a voice such as a movie or orchestra with excellent sound effects.
[0018]
The second one is to output the output signal of the equalizer circuit 41 to the terminal B of the
voice changeover switch 49. This is because when the output signal of the equalizer circuit 41
has a good S / N ratio, the output signal of the equalizer circuit 41 is output to the terminal B of
the voice changeover switch 49.
[0019]
Third, the output signal y [k] of the instantaneous frequency analysis circuit 44 is output to the
terminal C of the voice switch 49. The main component signal of the audio signal from which the
harmonic component which becomes noise is removed from the output signal of the equalizer
circuit 41 is output to the terminal C of the audio switching switch 49.
[0020]
The auditory sense correction filter 42 constituting the reference characteristic of the auditory
sense improvement circuit 7 is, for example, a filter having a frequency characteristic (JIS-A
characteristic) defined as a standard as shown in FIG. 4. The filter characteristics of this auditory
sense correction filter 42 are matched to the human ear's auditory sense. Human hearing is
difficult to hear low frequency audio signals, and it is also difficult to hear high frequency audio
signals above 5 kHz, but the 2 kHz to 4 kHz frequency signals among those frequencies are the
easiest to hear . For example, the frequency characteristic of FIG. 4 is a characteristic in which a
section between 2 kHz and 4 kHz is a peak of about + 1-2 dB, and the lower band side is -37 dB
at 30 Hz and the higher band side is -7 dB lower at about 16 kHz. An output signal of the
audibility correction filter 42 is input to the equalizer control circuit 45 as a reference signal.
[0021]
Next, the equalizer circuit 41 adjusts the frequency characteristic of the signal input from the
08-05-2019
6
audio signal decoding unit 6. In this equalizer circuit 41, a plurality of band pass filters having
frequency characteristics as shown in FIG. 5 are connected in parallel as shown in FIG. 3 while
changing the center frequency. For example, as shown in FIG. 5, the band pass filter has a center
frequency of 1 kHz, a peak of 0 dB between approximately 800 Hz and 1300 Hz, and a low band
of -40 dB at approximately 430 Hz and a high band The characteristic is -40 dB decrease at
2100 Hz.
[0022]
The equalizer circuit 41 is a second-order Butterworth filter having a center frequency (fc) of 1
kHz and cutoff frequencies of 707 Hz and 1414 Hz, respectively. The band pass filters 50 to 59
shown in FIG. 3 have center frequencies (fc) of 31.5 Hz, 63 Hz, 125 Hz, 250 Hz, 500 Hz, 1 kHz, 2
kHz, 4 kHz, 8 kHz and 16 kHz, respectively. Also, the cutoff frequency is determined by the ratio
to the center frequency, and all the characteristics are those of the second-order Butterworth
filter.
[0023]
As shown in FIG. 3, the equalizer circuit 41 amplifies band pass filters BPFs 50 to 50 having
different center frequencies via amplifiers 60 to 69, and then changes the output signal level by
adjustable level adjustment circuits 70 to 79. Each of them is individually input to the adder
circuit 80. The equalizer circuit 41 constitutes a set of BPFs connected in series, an amplifier and
a level adjustment circuit, and adjusts the level for each frequency so that the output sound
signal matches the auditory sense characteristic. The output of the adder circuit 80 is output to
an instantaneous frequency analysis circuit 44 or the like.
[0024]
The instantaneous frequency analysis circuit 44 performs signal processing based on the
flowchart shown in FIG. Hereinafter, step Sn is described as "Sn". However, "n" means arbitrary
positive integers.
[0025]
08-05-2019
7
First, at step (S1) of averaging (LPF, low pass filter) of input data, a signal of a frequency band
higher than the audible area is removed from the input voice signal, and the input signal is
averaged (step S1). The characteristics of the filter at this time are such that the voice can be
efficiently generated with a small number of taps so as not to lower the level of the main
component signal and prevent it from taking too much time to process the instantaneous
frequency and instantaneous amplitude. It is necessary to remove out-of-band signals. Therefore,
the frequency information of the main component signal obtained in the step of calculating the
frequency (S7) is fed back to perform signal processing including averaging around the
frequency near the main component signal.
[0026]
Next, in order to perform complex signal processing, Hilbert transform (Hilbert transform-1) is
performed on the signal from which the noise component in the frequency range higher than the
audio frequency is removed in S1 (S2). In this Hilbert transform, the signal is converted into a
complex signal to be processed immediately so as not to cause a signal delay, and a signal
component whose phase is advanced is taken out. Thereby, two signals (I1: in-phase component,
Q1: quadrature component) whose phases are shifted by 90 degrees as shown in FIG. 8 are
obtained. In addition, since this Hilbert transform is a kind of band pass filter, not only high
frequencies of the frequency but also considerably low components such as signal level
fluctuations rather than audio signals are removed.
[0027]
In this case, if noise components remain only by one Hilbert transform, Hilbert transform (Hilbert
transform-2) is performed again any number of times, for example, (S3). Thereby, as shown in
FIG. 9A, the in-phase component I11 and the quadrature component Q12 are further generated
from the in-phase component I1 of S2, and as shown in FIG. 9B from the quadrature component
Q1 of S2, Component Q21 and orthogonal component I22 are generated. Since the quadrature
component Q1 is 90 degrees ahead of the phase, it is in phase with Q21 and the quadrature
component Q12, which are 90 degrees ahead of the in-phase component I1, and I22 and I11 are
90 degrees ahead of the phase from the quadrature component Q1. Become. By this processing,
the respective in-phase component and quadrature component are added. As a result, the inphase component I11 and the quadrature component I22 are 180 degrees out of phase, so they
are canceled out by the addition. Further, the quadrature component Q12 and the in-phase
component Q21 have the same phase.
08-05-2019
8
[0028]
Such an operation is performed for several sampling clocks in order to remove the influence of
noise and is averaged (S4). However, there is a phase difference between the in-phase component
and the quadrature component. Therefore, in the addition of the in-phase component and the
quadrature component, it is necessary to perform the phase correction by performing signal
processing with the signal shifted from the sampling point of each component, and thereafter
performing the addition.
[0029]
Basically, the phase difference (S6), the amplitude value and the frequency (S7) of the main
component signal are obtained from the level difference between adjacent sampling points of the
audio signal, and the filter processing centered on the frequency near the main component signal
By performing (S1), the noise component can be removed and only the main component signal
can be output. Also, the noise level is calculated (S8), and the output signal is switched according
to the ratio of the main component signal to the noise level (S10).
[0030]
The averaged signal is also output to the equalizer control circuit 45 and the audio selector
switch 49. Since most of the signal to this instantaneous frequency analysis circuit 44 is a main
component signal, if there is a lot of noise in the output signal of the audio signal decoding unit
6, this signal is sent to the audio signal decoding unit 6 through the audio switch 49 Output as
voice output of
[0031]
Further, the instantaneous frequency analysis circuit 44 calculates the real part and the
imaginary part of the complex signal in which the in-phase component and the quadrature
component are averaged (S5). The calculated level (amplitude value) of the complex signal is
taken as the signal level (S) when the S / N ratio is calculated. Thereafter, the phase angle is
08-05-2019
9
calculated from the level of the real part and the level of the imaginary part of the calculated
complex signal (S6).
[0032]
As described above, since the phase angle is obtained, the phase difference can be obtained from
the difference between the phase angles of the signals at the adjacent sampling points, and the
frequency of the main component signal can be obtained (S7). Information on this frequency is
averaged and fed back to S1 to adjust the frequency characteristics of the LPF (low pass filter,
S1) in order to acquire the main component signal.
[0033]
In addition to the averaging step (S1) of the input signal, the signal input to the instantaneous
frequency analysis circuit 44 is also input to the difference detection step (S8) with respect to the
main component signal. In addition, difference detection is performed by comparing with the
main component signal generated in the averaging step (S4) of the in-phase component and the
quadrature component.
[0034]
Further, the minimum value and the maximum value of the difference signal are detected every
several samplings (S9). This makes it possible to detect the noise level (harmonic component) of
the high frequency component appearing in the main component signal.
[0035]
Further, the S / N ratio is calculated by comparing the calculation result of the main signal
obtained in the calculation step (S5) of the real part and the imaginary part with the average
(S10).
[0036]
The data of this S / N ratio is output to the voice switching switch 49, and when there is a certain
degree of S / N ratio, it can be clearly distinguished from the main signal component, so the voice
08-05-2019
10
signal from the equalizer circuit 41 is switched to the voice switching switch 49 is selected, and
when the S / N ratio is small, the main signal component and the harmonic component can not
be clearly discriminated. Therefore, in the switching control in which the audio switching switch
49 selects the audio signal from the instantaneous frequency analysis circuit 44 Used.
[0037]
At the stage where the processing of S 4 in the instantaneous frequency analysis circuit 44 is
completed, the output signal of the averaging process (S 4) of the in-phase component and the
quadrature component is also input to the equalizer control circuit 45.
The equalizer control circuit 45 compares this input signal with the signal from the audibility
correction filter circuit 42, detects an error signal thereof, and detects each of the level
adjustment circuits 70 to 79 in FIG. 3 by the LMS (Least Mean Square) method. Determine the
output level.
[0038]
That is, assuming that the output signal of the instantaneous frequency analysis circuit 44 is y
[k], the input of the equalizer circuit 41 is x [k], and the coefficients of the level adjustment
circuits 70 to 79 are am (a0 to a9) y [k] = It becomes Σam x [k-m].
The coefficient am is determined by the LMS method. Assuming that the output of the auditory
sense correction filter 42 is d [k], the error signal e [k] of the output of the equalizer control
circuit 45 becomes e [k] = y [k] −d [k]. Thus, the coefficient am of the m-th level adjustment
circuit (one of 70 to 79) at a certain time k + 1 is set to a fixed value with a step size of α am [k +
1] = am [k] -2α · e It can be determined as [k] x [k-m].
[0039]
Thus, by adjusting the level, the frequency band to be emphasized and the frequency band to be
attenuated are selected. Then, the adder circuit 80 synthesizes the signals of the respective
frequency components.
08-05-2019
11
[0040]
When the frequency component of the input signal is flat, the equalizer circuit 41 outputs
approximately
[0041]
[0042]
となる。
[0043]
The equalizer control circuit 45 feeds back an error signal e [k] corresponding to the difference
between the input signals y [k] and d [k] to the equalizer circuit 41 to determine the output level
of each level adjustment circuit.
[0044]
In the auditory sense improvement circuit configured by the above elements, the input audio
signal is input as it is to the terminal A of the audio switching switch 49 as in the case of
performance sound of an orchestra whose auditory characteristic is good.
[0045]
The input audio signal is level-adjusted by the equalizer circuit 41 and input to the terminal B of
the audio switch 49.
[0046]
Also, since most of the signals from which harmonic components have been removed by
averaging the complex components of the audio signal are main component signals, if there is a
lot of noise in the output signal of the audio signal decoding unit 6, this signal is switched to
audio It is output as an audio output in place of the audio signal decoding unit 6 through the
switch 49.
[0047]
Since the present invention is configured as described above, it is possible to convert the
08-05-2019
12
spectrum of the audio signal into a frequency band easy for human beings to hear, or to detect
and output the main component signal of the audio signal.
[0048]
Further, in the present invention, the difference between the spectrum of the audio signal and the
signal passed through the auditory sense correction filter is taken to always correct the spectrum
so that the sense of hearing is good, and the harmonic component whose frequency and
amplitude constantly change Is converted to a complex signal by performing Hilbert transform,
removed by performing complex averaging, and further detects the level of the main component
signal and the level of the harmonic component signal, and outputs the equalizer output and the
main component signal. By switching, the hearing characteristic can be improved.
[0049]
In addition, the Hilbert transformer is a filter whose amplitude characteristic is constant
regardless of frequency, and whose phase characteristic is π / 2 delayed in the positive
frequency domain and π / 2 advanced in the negative frequency domain. Using this filter,
harmonic components of constantly changing frequency and amplitude can be removed by
performing complex averaging of Hilbert-transformed complex signals.
[0050]
Furthermore, the level of the main component signal and the level of the harmonic component
signal are detected, and the equalizer output and the main component signal output are switched.
The above solutions are used in combination to achieve the purpose.
[0051]
FIG. 5 is a simplified diagram of the decoder circuit of the present invention.
It is a block diagram of the audibility improvement circuit of this invention.
08-05-2019
13
It is a block diagram of the equalizer circuit of this invention.
It is a frequency characteristic of the audibility correction filter of this invention.
It is an example of a frequency characteristic of the band pass filter of the present invention.
FIG. 5 is a schematic view showing a phase relationship between a main component signal and a
harmonic component on a complex plane.
It is a flowchart of the signal processing in the instantaneous frequency analysis circuit of this
invention.
It is a schematic diagram of complex signaling by Hilbert transform in the case of the present
invention.
It is a schematic diagram of the phase relationship of the signal by performing Hilbert
transformation twice in the case of this invention.
Explanation of sign
[0052]
7. Audibility improvement circuit, 41: Equalizer circuit, 42: Audibility correction filter, 44:
Instantaneous frequency analysis circuit, 45: Equalizer control circuit, 50 to 59: BPF (fc differs),
60 to 69: Amplifier, 70 to 79 ... level adjustment circuit, 80 ... addition circuit.
08-05-2019
14
Документ
Категория
Без категории
Просмотров
0
Размер файла
25 Кб
Теги
jp2008104001
1/--страниц
Пожаловаться на содержимое документа