close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2015050685

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 JP2015050685
An audio signal processing apparatus capable of performing more appropriate gain adjustment
and obtaining higher quality voice by calculating a gain adjustment amount in consideration of
not only a resonance band signal but also an input signal. provide. A band division unit divides an
input signal into a resonance band signal and a resonance band outside signal, and an analysis
unit extracts a resonance analysis feature value from the resonance band signal and the input
signal. The gain adjustment amount calculation unit 23 calculates the gain adjustment amount of
the resonance band based on the resonance analysis characteristic amount of the resonance
band signal and the input signal, and the gain adjustment unit 24 calculates the resonance band
of the input signal based on the gain adjustment amount. Adjust the gain of. As a result, higher
quality speech can be obtained. The present invention can be applied to a laptop personal
computer. [Selected figure] Figure 1
Audio signal processing apparatus and method, and program
[0001]
The present technology relates to an audio signal processing apparatus and method, and a
program, and more particularly, to an audio signal processing apparatus and method, and a
program capable of obtaining higher quality voice.
[0002]
In recent years, miniaturization of portable devices such as laptop personal computers has
progressed, and the situation where the speaker and parts such as a base and a keyboard are
inevitably located close to each other, and the parts themselves have become smaller and thinner
And there is a tendency for resonance to occur when playing back audio.
11-04-2019
1
[0003]
However, since hardware measures against resonance are costly, measures may be taken within
a range that can be achieved at low cost, or resonance may be regarded as meaningless, with the
result that the sound quality may be sacrificed.
[0004]
In addition, as a measure against resonance, there is a method of lowering the volume of the
entire voice so that resonance does not occur, but depending on the sound source, the volume
may become too low to be heard, or sound quality may not be satisfactory.
On the other hand, there has been proposed a technique for suppressing resonance by
suppressing the gain of the resonance frequency band with a notch filter or the like.
[0005]
By the way, it is known that when reproducing movie content and music content with a portable
device such as a laptop personal computer that incorporates a small speaker, the sound volume
of the sound is reduced as a whole.
[0006]
Therefore, there is a technique of emphasizing the reproduction level of the audio signal of the
audio to make the audio easier to hear (see, for example, Patent Document 1 and Patent
Document 2).
In this technique, the characteristics of the amplitude conversion function dynamically change
according to the analysis result of the input audio signal to emphasize the reproduction level.
[0007]
JP, 2012-60379, A JP, 2012-235310, A
11-04-2019
2
[0008]
However, with the above-described technology, sufficiently high quality voice can not be
obtained.
[0009]
For example, in the resonance phenomenon, even if the signal level in the resonance band is the
same, the magnitude of resonance may change depending on the signal level outside the
resonance band.
Therefore, in a technique in which the amount of suppression gain is determined only by a
resonance band such as a notch filter, the gain is suppressed even for a signal that does not need
to be suppressed, which adversely affects the volume and the sound quality of the sound.
[0010]
Furthermore, in the technology of emphasizing the reproduction level, if the reproduction level is
emphasized with a notebook personal computer or the like, the above-mentioned resonance
becomes larger, and the adverse effect on the sound quality of the sound also becomes large.
In addition, it has not been possible to make the amount of emphasis on the reproduction level
sufficiently large to suppress resonance.
Therefore, it was not possible to obtain a sufficiently high quality voice.
[0011]
The present technology has been made in view of such a situation, and makes it possible to
obtain higher quality speech.
[0012]
An audio signal processing apparatus according to one aspect of the present technology divides a
11-04-2019
3
voice input signal into bands to generate a resonant band signal, and an analyzer that extracts
each feature amount from the resonant band signal and the input signal. And a gain adjustment
amount calculation unit that calculates a gain adjustment amount of the resonance band of the
input signal based on the feature amount of the resonance band signal and the feature amount of
the input signal, and the gain adjustment amount. And a gain adjustment unit that adjusts the
gain of the resonance band in the input signal.
[0013]
The gain adjustment amount calculation unit can calculate the gain adjustment amount based on
the feature amount using a statistical analysis model generated by learning.
[0014]
The gain adjustment amount calculation unit may calculate the gain adjustment amount for each
of the plurality of resonance bands, and the gain adjustment unit may perform gain adjustment
for each of the plurality of resonance bands.
[0015]
The audio signal processing apparatus further comprises the mapping control information based
on the reproduction level emphasis feature extracted from the input signal and information for
obtaining mapping control information from the reproduction level emphasis feature generated
by learning. And a mapping processing unit for converting the amplitude of the input signal
whose gain has been adjusted based on a linear or non-linear mapping function determined by
the mapping control information. The unit may calculate the gain adjustment amount based on
the mapping control information and the feature amount extracted by the analysis unit.
[0016]
The gain adjustment amount can be made larger as the ratio of the resonance band signal
included in the input signal becomes larger.
[0017]
The band dividing unit divides the input signal into the resonant band signal and the out-ofresonance band signal, and the gain adjusting unit adjusts the gain of the resonant band signal,
and the audio signal processing apparatus An addition unit may be further provided to add the
gain-adjusted resonance band signal and the resonance band out signal.
11-04-2019
4
[0018]
The audio signal processing method or program according to one aspect of the present
technology divides an audio input signal into bands to generate a resonance band signal, extracts
each feature amount from the resonance band signal and the input signal, and the resonance
band The gain adjustment amount of the resonance band of the input signal is calculated based
on the feature amount of the signal and the feature amount of the input signal, and the gain
adjustment of the resonance band in the input signal is calculated based on the gain adjustment
amount. Including the steps of
[0019]
In one aspect of the present technology, a speech input signal is divided into bands to generate a
resonance band signal, each feature amount is extracted from the resonance band signal and the
input signal, and the feature amount of the resonance band signal, The gain adjustment amount
of the resonance band of the input signal is calculated based on the feature amount of the input
signal, and the gain adjustment of the resonance band in the input signal is performed based on
the gain adjustment amount.
[0020]
According to one aspect of the present technology, higher quality voice can be obtained.
[0021]
In addition, the effect described here is not necessarily limited, and may be any effect described
in the present disclosure.
[0022]
It is a figure which shows the structural example of an audio signal processing apparatus.
It is a flow chart explaining resonance band ingredient suppression processing.
It is a figure explaining generation | occurrence | production of resonance.
It is a figure which shows the example of a gain adjustment model.
11-04-2019
5
It is a figure which shows the structural example of an audio signal processing apparatus.
It is a flow chart explaining resonance band ingredient suppression processing.
It is a figure which shows the structural example of an audio signal processing apparatus.
It is a flow chart explaining resonance band ingredient suppression processing.
It is a figure which shows the structural example of an audio signal processing apparatus.
It is a flow chart explaining resonance band ingredient suppression processing.
It is a figure explaining a mapping control model.
It is a figure explaining a mapping function.
It is a figure showing an example of composition of a computer.
[0023]
Hereinafter, embodiments to which the present technology is applied will be described with
reference to the drawings.
[0024]
First Embodiment <Configuration Example of Audio Signal Processing Device> FIG. 1 is a diagram
showing a configuration example of an audio signal processing device to which the present
technology is applied.
11-04-2019
6
For example, the audio signal processing apparatus is a notebook personal computer or the like.
[0025]
The audio signal processing device 11 includes a band division unit 21, an analysis unit 22, a
gain adjustment amount calculation unit 23, a gain adjustment unit 24, and an addition unit 25.
In the audio signal processing device 11, an input signal that is an audio signal of speech is
supplied to the band division unit 21 and the analysis unit 22.
[0026]
The band division unit 21 is formed of, for example, a band pass filter, performs band division on
the input signal, and generates a resonant band signal and a resonant out-of-band signal from the
supplied input signal. That is, the input signal is divided into the resonance band signal and the
out-of-resonance band signal by band division with respect to the input signal. The band division
unit 21 supplies the resonance band signal to the analysis unit 22 and the gain adjustment unit
24, and supplies the out-of-resonance band signal to the addition unit 25.
[0027]
Here, the resonance band signal is a signal consisting of a frequency at which a predetermined
resonance of the input signal occurs, that is, a band component of the resonance frequency.
Therefore, the resonance band signal is a signal of a resonance band component to be
suppressed among the input signals.
[0028]
Further, the out-of-resonance band signal is a signal of a component obtained by removing the
resonance band component from the input signal. That is, the out-of-resonance band signal is a
signal of a frequency band component other than the resonance frequency in the input signal,
and is a signal of a frequency band component which does not have to be suppressed.
11-04-2019
7
[0029]
The analysis unit 22 extracts a resonance analysis feature amount from the supplied input signal
and the resonance band signal supplied from the band division unit 21, and supplies the
resonance adjustment feature amount to the gain adjustment amount calculation unit 23. The
gain adjustment amount calculation unit 23 calculates the optimum gain adjustment amount for
suppressing the resonance band component of the input signal based on the resonance analysis
feature amount supplied from the analysis unit 22, and supplies the calculated amount to the
gain adjustment unit 24.
[0030]
The gain adjustment unit 24 adjusts the gain of the resonance band signal by converting the
amplitude of the resonance band signal supplied from the band division unit 21 based on the
gain adjustment amount supplied from the gain adjustment amount calculation unit 23. The
resulting resonance band signal is supplied to the adder 25.
[0031]
The addition unit 25 is a signal in which the resonance band component in the input signal is
suppressed by adding the resonance band signal supplied from the gain adjustment unit 24 and
the resonance band outside signal supplied from the band division unit 21. Generate and output
an output signal.
[0032]
<Description of Resonant Band Component Suppression Processing> Next, the operation of the
audio signal processing device 11 will be described.
[0033]
When the input signal is supplied, the audio signal processing device 11 starts a resonance band
component suppression process for suppressing the resonance band component and generates
an output signal.
Hereinafter, the resonance band component suppression process by the audio signal processing
11-04-2019
8
device 11 will be described with reference to the flowchart of FIG.
[0034]
In step S11, the band division unit 21 divides the supplied input signal into the resonance band
signal and the out-of-resonance band signal, and supplies the resonance band signal to the
analysis unit 22 and the gain adjustment unit 24. The signal is supplied to the addition unit 25.
For example, a band pass filter performs filtering on an input signal to generate a resonant band
signal and a non-resonant band signal.
[0035]
In step S 12, the analysis unit 22 extracts a resonance analysis feature amount from the supplied
input signal and the resonance band signal supplied from the band division unit 21, and supplies
the resonance adjustment feature amount to the gain adjustment amount calculation unit 23.
[0036]
Specifically, for example, the analysis unit 22 performs an input by calculating the following
equation (1) for the section of the input signal consisting of N samples centered on the nth
sample to be processed in the input signal. The root mean square RMS_ab (n) of the signal is
calculated to be a resonance analysis feature.
[0037]
[0038]
In equation (1), x (m) represents the m-th sample of the input signal, and the input signal is such
that the value of each sample x (m) is −1.0 ≦ x (m) ≦ 1.0 Shall be normalized to
[0039]
Further, the analysis unit 22 calculates the resonance band signal by calculating the following
equation (2) for the section of the resonance band signal consisting of N samples centered on the
nth sample to be processed in the resonance band signal The root mean square RMS_rb (n) of is
calculated as a resonance analysis feature.
11-04-2019
9
[0040]
[0041]
In Equation (2), x_rb (m) represents the m-th sample of the resonance band signal, and the
resonance band signal is determined such that the value of each sample x_rb (m) is −1.0 ≦ x_rb
(m) ≦ 1.0. It is assumed to be normalized to be
[0042]
The root mean square RMS_ab (n) of the input signal thus obtained and the root mean square
RMS_rb (n) of the resonance band signal are supplied to the gain adjustment amount calculator
23 as resonance analysis feature quantities.
[0043]
Here, although an example in which the root mean square is obtained as the resonance analysis
feature amount has been described, the resonance analysis feature amount may be any other.
For example, in addition to the root mean square, the t root value of the root mean square (where
t 2 2), the strength of tone characteristics, the true peak obtained by dividing the resonance band
further finely, etc. are added as resonance analysis feature quantities Alternatively, any
combination thereof may be used as a resonance analysis feature.
In addition, logarithmic values such as root mean square, t-th power of root mean square,
intensity of tonality, and true peak may be used as the resonance analysis feature value.
[0044]
Also, although an example using a band pass filter has been described for analysis of the input
signal and the resonance band signal, other filters may be used, and the input signal is divided in
the frequency domain, and the feature quantity is extracted and analyzed. You may do so.
[0045]
11-04-2019
10
In step S 13, the gain adjustment amount calculation unit 23 calculates an optimal gain
adjustment amount for suppressing the resonance band component of the input signal based on
the resonance analysis feature amount supplied from the analysis unit 22, and outputs the
calculated amount to the gain adjustment unit 24. Supply.
[0046]
For example, when the resonance phenomenon is analyzed in detail, even if the resonance band
signal level, that is, the root mean square RMS_rb (n) value of the resonance band signal is the
same, the resonance level is determined according to the input signal level, that is, the root mean
square RMS_ab (n). It turns out that it changes.
[0047]
Here, an example of a simple signal is shown in FIG.
In FIG. 3, the vertical axis and the horizontal axis indicate the amplitude and the frequency, and F
on the frequency axis indicates the resonance frequency.
[0048]
In FIG. 3, the broken line C11 and the broken line C12 indicate the frequency characteristics of
different signals.
For example, a broken line C11 represents the frequency characteristic of a signal that tends to
resonate as an input signal, and a broken line C12 represents a frequency characteristic of a
signal that does not easily resonate as an input signal.
[0049]
When the broken line C11 and the broken line C12 are compared, the resonance frequency band
in each signal, that is, the signal level (amplitude) of the resonance frequency F is LV11, which is
the same.
11-04-2019
11
However, the signal indicated by the broken line C11 tends to easily resonate because the
component of the resonance frequency F is dominant.
That is, the level of the component of the resonance frequency F of the signal indicated by the
broken line C11 is significantly larger than the levels of the other frequency components.
[0050]
On the other hand, the signal indicated by the broken line C12 has a tendency that it is difficult
to resonate because the component of the resonant frequency F is not dominant, or that the
resonance is not a concern.
[0051]
When not only the level of the component of the resonant frequency F of the input signal but
also the levels of other frequency components are large as in the signal shown by the broken line
C12, the gain adjustment amount of the resonant band signal is about the signal shown by the
broken line C11. It may often be smaller than the gain adjustment amount.
Therefore, if the gain adjustment amount of the resonance band signal is made the same value
for the signal shown on the broken line C11 and the signal shown on the broken line C12, the
gain of the signal shown on the broken line C12 is suppressed more than necessary, and the
volume and sound pressure Adversely affect the
[0052]
Therefore, when calculating the gain adjustment amount of the resonance band signal, the audio
signal processing apparatus 11 considers not only the level of the resonance band signal but also
the balance with the level of the input signal, that is, the band of the entire input signal. , To
prevent the gain from being suppressed more than necessary.
[0053]
Specifically, for example, the gain adjustment amount calculation unit 23 obtains a gain
adjustment amount using a gain adjustment model learned in advance based on statistical
11-04-2019
12
analysis.
[0054]
In the learning of the gain adjustment model, the gain adjustment amount F (X, Y) is taken as an
explanatory variable, and root mean square RMS_ab (n) and root mean square RMS rb (n)
extracted from the audio signal for learning The square value of the root mean square, and the
offset term irrelevant to the root mean square or its square value are used as explanatory
variables.
Then, in the gain adjustment amount calculation unit 23, the statistical analysis model shown in
the following equation (3), which is learned in advance based on the explained variables and the
explanatory variables, is recorded as a gain adjustment model.
[0055]
[0056]
In Equation (3), a, b, c, d, and e are model parameters, respectively.
[0057]
Here, FIG. 4 shows an example of learning results of the gain adjustment model when X in the
equation (3) is a root mean square RMS_ab (n) and Y is a root mean square RMS_rb (n).
[0058]
In FIG. 4, the horizontal axis X axis and the vertical axis Y axis respectively indicate root mean
square RMS_ab (n) and root mean square RMS_rb (n).
Further, the concentration in each region indicates the gain adjustment amount F (X, Y), and as
the concentration is higher, the absolute value of the gain adjustment amount is larger, that is,
the suppression amount of the resonance band is larger.
11-04-2019
13
[0059]
In this example, as the ratio of the resonance band signal included in the input signal is
increased, the gain adjustment amount is increased, and the component of the resonance band is
more largely suppressed.
For example, when the value of root mean square RMS_ab (n) is a constant value, the larger the
power of the resonance band signal, that is, the larger the root mean square RMS_rb (n), the
larger the absolute value of the gain adjustment amount.
[0060]
As described above, by determining the appropriate gain adjustment amount based on the
relationship between the input signal and the resonance band signal, it is possible to prevent the
resonance band component from being suppressed more than necessary, and to perform more
appropriate gain adjustment. Will be able to
[0061]
When learning model parameters as a gain adjustment model, an audio signal in which
resonance occurs and an audio signal in which resonance does not occur are used as learning
data, and are optimal when the learning data is reproduced for each of a plurality of learning
data. The amount of gain adjustment is predetermined by human.
In addition, a resonance analysis feature value extracted from learning data is used as an
explanatory variable.
In this case, since it is possible to reflect the gain adjustment amount that a person actually feels
optimum, that is, the sense of an actual person in learning, it is possible to obtain an optimal gain
adjustment model.
[0062]
11-04-2019
14
The gain adjustment amount calculation unit 23 uses the gain adjustment model generated by
such learning to calculate root mean square RMS_ab (n) and root mean square RMS_rb (n as
resonance analysis feature amounts supplied from the analysis unit 22. The gain adjustment
amount is calculated based on.
Specifically, root mean square RMS_ab (n) and root mean square RMS_rb (n) are substituted into
the above-mentioned equation (3), and the value of F (X, Y) obtained as a result is taken as the
gain adjustment amount. Ru.
[0063]
Although the method using the statistical analysis model has been described as the method of
calculating the gain adjustment amount here, the gain adjustment amount may be calculated
using a simple linear or non-linear function or table.
[0064]
Returning to the explanation of the flowchart of FIG. 2, when the gain adjustment amount is
calculated in step S13, the process proceeds to step S14.
[0065]
In step S14, the gain adjustment unit 24 adjusts the gain of the resonance band signal by
converting the amplitude of the resonance band signal from the band division unit 21 based on
the gain adjustment amount from the gain adjustment amount calculation unit 23. The data is
supplied to the adding unit 25.
By this, gain adjustment of the resonance band in the input signal is performed.
[0066]
In step S15, the adding unit 25 adds the resonance band signal supplied from the gain adjusting
unit 24 and the resonance band out signal supplied from the band dividing unit 21 to generate
an output signal, and outputs the output signal.
11-04-2019
15
When the output signal is output, the resonance band component suppression process ends.
[0067]
As described above, the audio signal processing apparatus 11 extracts the resonance analysis
feature amount from the input signal and the resonance band signal, and performs the gain
adjustment of the resonance band signal by the gain adjustment amount obtained based on the
resonance analysis feature amount. Do.
[0068]
By thus determining the gain adjustment amount using the resonance analysis feature extracted
from the input signal and the resonance band signal, it is possible to more appropriately
suppress the resonance band component of the input signal and obtain higher quality speech.
become able to.
[0069]
Specifically, for example, for a signal whose signal level outside the resonance band is smaller
than that of the resonance band, that is, when the resonance is large, the gain is largely
suppressed.
Conversely, the gain is not suppressed for a signal not out of the resonance band with respect to
the signal level in the resonance band, that is, a signal that does not resonate or a signal that
does not care about resonance.
[0070]
As described above, according to the audio signal processing device 11, the optimum resonance
suppression can be performed according to the magnitude of the resonance.
As a result, it is possible to prevent the resonance band signal from being gain-suppressed more
than necessary, and to obtain a large volume and sound with good sound quality.
11-04-2019
16
[0071]
In addition, since it is not necessary to take hardware measures against resonance, there is no
limitation in design, cost is reduced accordingly, or cost is added to other parts to achieve high
functionality and high quality. And so on.
[0072]
Second Embodiment <Configuration Example of Audio Signal Processing Device> In the first
embodiment, although an example in which one resonance band signal is divided and extracted
from an input signal has been described, an integer multiple of the resonance frequency is
described. In general, the frequency of
[0073]
Therefore, resonance suppression may be performed more effectively by performing gain
adjustment on a plurality of resonance band signals.
In such a case, the audio signal processing device is configured, for example, as shown in FIG.
In FIG. 5, parts corresponding to those in FIG. 1 are given the same reference numerals, and the
description thereof will be omitted as appropriate.
[0074]
The audio signal processing device 51 shown in FIG. 5 includes a band division unit 61, an
analysis unit 22, a gain adjustment amount calculation unit 23, gain adjustment units 62-1 to 62K, and an addition unit 25. .
[0075]
The band division unit 61 divides the supplied input signal into K resonance band signals and
one out-of-resonance band signal.
11-04-2019
17
Here, the K resonance band signals are signals composed of different resonance band
components of the input signal, and both of the resonance bands are frequency bands in which
resonance occurs.
For example, each of the K resonance bands is a band that is an integral multiple of a
predetermined resonance band.
[0076]
The band division unit 61 supplies K resonance band signals to the analysis unit 22 and the gain
adjustment unit 62-1 to the gain adjustment unit 62-K, and supplies a resonance out-of-band
signal to the addition unit 25.
[0077]
The analysis unit 22 extracts a resonance analysis feature amount from each of the supplied
input signal and the K resonance band signals supplied from the band division unit 61, and
supplies the extracted feature amount to the gain adjustment amount calculation unit 23.
The gain adjustment amount calculation unit 23 calculates an optimum gain adjustment amount
for suppressing the resonance band component of the input signal for each of the K resonance
bands based on the resonance analysis feature amount supplied from the analysis unit 22, and
calculates the gain. The adjustment unit 62-1 to the gain adjustment unit 62-K are supplied.
[0078]
The gain adjusting unit 62-1 to the gain adjusting unit 62-K convert the amplitude of the
resonance band signal supplied from the band dividing unit 61 based on the gain adjustment
amount supplied from the gain adjustment amount calculating unit 23. The gain of the resonance
band signal is adjusted and supplied to the adding unit 25.
[0079]
Hereinafter, the gain adjusting unit 62-1 to the gain adjusting unit 62-K will be simply referred to
as the gain adjusting unit 62 unless it is necessary to distinguish them.
11-04-2019
18
[0080]
The addition unit 25 adds the resonance band signal supplied from the gain adjustment unit 621 to the gain adjustment unit 62-K and the resonance out-of-band signal supplied from the band
division unit 61 to generate an output signal, Output.
[0081]
<Description of Resonant Band Component Suppression Processing> Next, with reference to the
flowchart of FIG. 6, the resonant band component suppression processing performed by the
audio signal processing device 51 will be described.
[0082]
In step S41, the band division unit 61 divides the supplied input signal into K resonance band
signals of the resonance band and one out-of-resonance band signal, and the analysis section 22
and the gain adjustment unit The signal is supplied to 62 and the out-of-resonance band signal is
supplied to the adding unit 25.
[0083]
Specifically, K resonance band signals are supplied to the analysis unit 22, and the resonance
band signals of the first to K-th resonance bands respectively correspond to the gain adjustment
units 62-1 to 62-K. Supplied to
[0084]
In step S 42, the analysis unit 22 extracts the resonance analysis feature amount from the
supplied input signal and the K resonance band signals supplied from the band division unit 61,
and supplies the resonance adjustment feature amount to the gain adjustment amount
calculation unit 23.
[0085]
Specifically, for example, the analysis unit 22 performs the calculation of Equation (1) described
above to calculate the root mean square RMS_ab (n) of the input signal, and uses it as a
resonance analysis feature value.
[0086]
11-04-2019
19
Further, the analysis unit 22 performs the same calculation as the above-mentioned equation (2)
for each of the K resonance bands to calculate the root mean square RMS_rbi (n) of each
resonance band signal, and uses it as a resonance analysis feature value. .
Here, 1 ≦ i ≦ K, and the root mean square RMS_rbi (n) indicates the root mean square of the
resonance band signal which is a component of the ith resonance band.
[0087]
In step S43, the gain adjustment amount calculation unit 23 calculates an optimal gain
adjustment amount for suppressing the resonance band component of the input signal for each
of the K resonance bands based on the resonance analysis feature amount supplied from the
analysis unit 22. , To the gain adjustment unit 62.
[0088]
For example, the gain adjustment amount calculation unit 23 records the gain adjustment model
shown in the equation (3) described above for each of the K resonance bands, more specifically,
model parameters.
At this time, the gain adjustment model of the i-th (where 1 ≦ i ≦ K) resonance band is the gain
adjustment of the root mean square RMS_rbi (n) and the i-th resonance band for the explanatory
variable Y and the explained variable, respectively. The quantity is used and learned.
[0089]
The gain adjustment amount calculation unit 23 expresses the root mean square RMS_ab (n) and
the root mean square RMS_rbi (n) of the ith resonance band using the model parameters of the
ith resonance band (3 And substitute the value of F (X, Y) obtained as a result thereof into the
gain adjustment amount of the ith resonance band.
Then, the gain adjustment amount calculation unit 23 supplies the obtained gain adjustment
amount of the i-th resonance band to the gain adjustment unit 62-i (where 1 ≦ i ≦ K).
11-04-2019
20
[0090]
Note that the root mean square of another resonance band may be used to calculate the gain
adjustment amount of the ith resonance band.
In such a case, for example, the gain adjustment model of the i-th resonance band is a statistical
analysis model represented by the following equation (4).
[0091]
[0092]
In Equation (4), X is root mean square RMS_ab (n), and Y 1 to Y K are root mean square RMS_rb
1 (n) to root mean square RMS_rbK (n).
Therefore, in this example, with the root mean square RMS_ab (n) and the root mean square
RMS_rb1 (n) to the root mean square RMS_rbK (n) of the K resonance bands as explanatory
variables, the gain adjustment amount of the ith resonance band The learning is performed by
using as an explained variable.
Further, a, b, c 1 to c K, d 1 to d K, e are model parameters, respectively.
[0093]
In this case, the gain adjustment amount calculation unit 23 substitutes the root mean square
RMS_ab (n) and the K root mean square RMS_rbi (n) into the equation (4), and obtains the
obtained F (X, Y 1). ,..., YK) is set as the gain adjustment amount of the ith resonance band.
[0094]
Furthermore, for example, one gain adjustment model is learned from the root mean square
RMS_ab (n) and the K root mean square RMS_rbi (n), and the gain adjustment model is
11-04-2019
21
commonly used in each of the K resonance bands. May be
[0095]
In step S44, the gain adjustment unit 62 performs gain adjustment of the resonance band signal
by converting the amplitude of the resonance band signal from the band division unit 61 based
on the gain adjustment amount from the gain adjustment amount calculation unit 23. The data is
supplied to the adding unit 25.
Specifically, the gain adjustment unit 62-i (where 1 ≦ i ≦ K) performs gain adjustment of the
resonance band signal of the i-th resonance band using the gain adjustment amount of the i-th
resonance band.
[0096]
In step S45, the adding unit 25 adds the K resonance band signals supplied from the K gain
adjusting units 62 and the resonance out-of-band signal supplied from the band dividing unit 61
to generate an output signal, Output.
When the output signal is output, the resonance band component suppression process ends.
[0097]
As described above, the audio signal processing apparatus 51 extracts the resonance analysis
feature amount from the input signal and the plurality of resonance band signals, and resonates
for each resonance band with the gain adjustment amount obtained based on the resonance
analysis feature amount. Adjust the gain of the band signal.
As a result, it is possible to appropriately suppress the resonance band component of the input
signal for each resonance band, and to obtain voice of higher quality.
11-04-2019
22
[0098]
Third Embodiment <Configuration Example of Audio Signal Processing Device> In the above, in
order to suppress the gain of the resonance band, the gain adjustment amount is calculated
directly from the resonance analysis feature amount, and the amplitude of the resonance band
signal is obtained. For example, a notch filter may be used for the input signal to suppress the
gain of the resonance band.
[0099]
In such a case, the audio signal processing device is configured, for example, as shown in FIG.
In FIG. 7, parts corresponding to the case in FIG. 1 are given the same reference numerals, and
the description thereof will be omitted as appropriate.
[0100]
An audio signal processing apparatus 91 shown in FIG. 7 includes a band division unit 21, an
analysis unit 22, a resonance control information calculation unit 101, and a resonance control
unit 102.
[0101]
The resonance control information calculation unit 101 calculates resonance control information
as information for suppressing a resonance band component based on the resonance analysis
feature amount supplied from the analysis unit 22, and supplies the resonance control
information to the resonance control unit 102.
The resonance control unit 102 suppresses the resonance band component of the supplied input
signal based on the resonance control information supplied from the resonance control
information calculation unit 101, and outputs the resulting signal as an output signal.
[0102]
11-04-2019
23
For example, when filter processing using notch filter coefficients, that is, notch filter processing
is performed in the resonance control unit 102 as processing for suppressing the resonance band
component of the input signal, notch filter coefficients are calculated as resonance control
information.
In the following, description will be continued on the assumption that notch filter processing is
performed in the resonance control unit 102.
[0103]
<Description of Resonant Band Component Suppression Processing> Next, the operation of the
audio signal processing device 91 will be described.
Hereinafter, the resonance band component suppression processing by the audio signal
processing device 91 will be described with reference to the flowchart in FIG.
[0104]
In step S71, the band division unit 21 divides the supplied input signal into the resonance band
signal and the out-of-resonance band signal, and supplies the resonance band signal to the
analysis unit 22.
[0105]
In step S 72, the analysis unit 22 extracts a resonance analysis feature value from the supplied
input signal and the resonance band signal supplied from the band division unit 21, and supplies
the resonance analysis feature value to the resonance control information calculation unit 101.
For example, the root mean square RMS_ab (n) of the input signal and the root mean square
RMS_rb (n) of the resonance band signal are calculated as resonance analysis feature quantities.
[0106]
11-04-2019
24
In step S 73, the resonance control information calculation unit 101 calculates resonance control
information based on the resonance analysis feature amount supplied from the analysis unit 22,
and supplies the resonance control information to the resonance control unit 102.
Here, a notch filter coefficient is calculated as resonance control information.
[0107]
In step S74, the resonance control unit 102 suppresses the resonance band component of the
supplied input signal based on the resonance control information supplied from the resonance
control information calculation unit 101, and outputs the resultant output signal. .
Specifically, a notch filter coefficient is used as resonance control information, notch filtering is
performed on the input signal, and a resonance band component of the input signal is
suppressed.
That is, gain adjustment of the resonance band is performed directly on the input signal.
[0108]
When the output signal is output, the resonance band component suppression process ends.
[0109]
As described above, the audio signal processing apparatus 91 extracts the resonance analysis
feature amount from the input signal and the resonance band signal, and adaptively resonates
the input signal with the resonance control information obtained based on the resonance analysis
feature amount. Perform gain adjustment of band components.
Thus, the resonance band component of the input signal can be suppressed more appropriately
in consideration of not only the resonance band but also the entire band of the input signal.
11-04-2019
25
As a result, higher quality speech can be obtained.
[0110]
Fourth Embodiment <Configuration Example of Audio Signal Processing Device> Further,
although the example in which the gain adjustment of the resonance band component of the
input signal is performed has been described above, the input signal in addition to the gain
adjustment of the resonance band component It may be possible to emphasize the reproduction
level of
In such a case, the audio signal processing device is configured, for example, as shown in FIG.
In FIG. 9, parts corresponding to the case in FIG. 1 are given the same reference numerals, and
the description thereof will be omitted as appropriate.
[0111]
The audio signal processing device 131 shown in FIG. 9 includes a band division unit 21, an
analysis unit 22, a gain adjustment unit 24, an addition unit 25, a reproduction level emphasis
analysis unit 141, a mapping control information determination unit 142, a gain adjustment
amount calculation unit 143, And a mapping processing unit 144.
[0112]
The reproduction level emphasis analysis unit 141 extracts the reproduction level emphasis
analysis feature amount from the supplied input signal, and supplies the extracted feature value
to the mapping control information determination unit 142.
[0113]
The mapping control information determination unit 142 obtains mapping control information
from the reproduction level emphasis analysis feature amount supplied from the reproduction
level emphasis analysis unit 141 using the mapping control model learned in advance based on
statistical analysis, and the gain adjustment amount The information is supplied to the
11-04-2019
26
calculation unit 143 and the mapping processing unit 144.
Here, the mapping control information is information indicating the degree of emphasis when
emphasizing the reproduction level of the input signal.
[0114]
The gain adjustment amount calculation unit 143 determines a gain adjustment amount based on
the resonance analysis feature amount supplied from the analysis unit 22 and the mapping
control information supplied from the mapping control information determination unit 142, and
the gain adjustment unit 24. Supply to
[0115]
The mapping processing unit 144 performs mapping processing on the signal supplied from the
adding unit 25 based on the mapping control information supplied from the mapping control
information determining unit 142, and emphasizes the reproduction level of the input signal.
The mapping processing unit 144 outputs the input signal whose reproduction level is
emphasized as a final output signal.
In the mapping process, a linear or nonlinear mapping function determined by the mapping
control information is used to perform linear or nonlinear amplitude conversion on the input
signal.
[0116]
<Description of Resonant Band Component Suppression Processing> Next, the operation of the
audio signal processing device 131 will be described. Hereinafter, the resonance band
component suppression processing by the audio signal processing device 131 will be described
with reference to the flowchart in FIG. In addition, since the process of step S101 and step S102
is the same as the process of step S11 of FIG. 2, and step S12, the description is abbreviate |
omitted.
11-04-2019
27
[0117]
In step S103, the reproduction level emphasis analysis unit 141 extracts the reproduction level
emphasis analysis feature amount from the supplied input signal, and supplies the extracted
feature value to the mapping control information determination unit 142.
[0118]
Specifically, for example, the reproduction level emphasis analysis unit 141 performs the same
calculation as Expression (1) to calculate the root mean square RMS (n) of the input signal as the
reproduction level emphasis analysis feature value.
[0119]
Here, the input signal is analyzed without pre-processing.
That is, extraction of the feature amount is performed from the input signal without preprocessing.
Therefore, although the root mean square RMS (n) determined as the reproduction level
emphasis analysis feature amount is equivalent to the root mean square RMS_ab (n) determined
as the resonance analysis feature amount, pre-processing such as DC cutting is performed as
necessary. After the process is performed, calculation of the reproduction level emphasis analysis
feature value may be performed.
[0120]
In addition, t-power value (where t 2 2) of root mean square RMS (n), zero crossing rate, slope of
frequency envelope, etc. are added as a reproduction level emphasis analysis feature value, or
any combination thereof May be used as a reproduction level emphasis analysis feature amount.
[0121]
In step S104, the mapping control information determination unit 142 obtains mapping control
information from the reproduction level emphasis analysis feature amount supplied from the
reproduction level emphasis analysis unit 141 using the mapping control model, and the gain
adjustment amount calculation unit 143 and the mapping process It supplies to the part 144.
11-04-2019
28
[0122]
For example, the mapping control information determination unit 142 uses a plurality of
mapping control information prepared for learning as an explained variable, and uses the
reproduction level emphasis analysis feature prepared for learning as an explanatory variable for
linear regression by the least squares method. A mapping control model generated by
performing model learning is recorded in advance.
The mapping control model is, for example, a coefficient of a quadratic function represented by a
curve C21 of FIG.
[0123]
In FIG. 11, the vertical axis represents mapping control information, and the horizontal axis
represents root mean square RMS (n) as a reproduction level emphasis analysis feature value.
[0124]
In FIG. 11, a curve C21 indicates the value of mapping control information determined for the
value of each reproduction level emphasis analysis feature value.
In this example, as the volume of the sound of the input signal is smaller and the reproduction
level emphasis analysis feature amount is smaller, the value of the mapping control information
is also smaller.
[0125]
The mapping control information determination unit 142 obtains mapping control information
by substituting the reproduction level emphasis analysis feature amount into a function
determined by a model parameter recorded as a mapping control model.
[0126]
Returning to the description of the flowchart of FIG. 10, when the mapping control information is
11-04-2019
29
obtained in step S104, the process of step S105 is performed thereafter.
[0127]
In step S105, the gain adjustment amount calculation unit 143 calculates the gain adjustment
amount based on the resonance analysis feature amount from the analysis unit 22 and the
mapping control information from the mapping control information determination unit 142, and
the gain adjustment unit 24 calculates the gain adjustment amount. Supply.
[0128]
The gain adjustment amount calculation unit 143 uses the gain adjustment amount F (X, Y) as an
explained variable, and adds to the root mean square RMS_ab (n) and the root mean square
RMS_rb (n) as resonance analysis feature amounts, and further mapping. The control information
is also used as an explanatory variable, and a gain adjustment model learned is recorded.
[0129]
For example, in the gain adjustment model obtained by learning, in equation (3), X is a sum of
root mean square RMS_ab (n) and a predetermined value β obtained from mapping control
information, and Y is root mean square RMS_rb (n) The sum of the predetermined values β is
used.
Here, the predetermined value β is a value obtained from the equation β = h (α) which
converts the mapping control information α so as to have the same dimension as the root mean
square, that is, the mapping control information to the function h (α). It is the value obtained by
substitution.
[0130]
Also in such a gain adjustment model, as in the case of the gain adjustment model shown in
equation (3), the gain adjustment amount increases as the ratio of the resonance band signal
included in the input signal increases, and the component of the resonance band Will be more
greatly suppressed.
[0131]
11-04-2019
30
Further, the mapping processing unit 144 in the latter stage performs mapping processing such
that the reproduction level of the input signal is emphasized as the mapping control information
is smaller.
Therefore, in the gain adjustment model, the smaller the mapping control information, the larger
the amount of gain adjustment.
That is, the component of the resonance band is suppressed in advance according to the
reproduction level emphasis amount in the mapping process.
[0132]
In addition, for example, a gain adjustment model represented by the following equation (5) may
be used.
In equation (5), X is root mean square RMS_ab (n), Y is root mean square RMS_rb (n), and Z is
mapping control information.
Moreover, in Formula (5), a, b, c, d, e, f, and g are model parameters, respectively.
[0133]
[0134]
The gain adjustment amount calculation unit 143 sets the root mean square RMS_ab (n) and the
root mean square RMS_rb (n) as the resonance analysis feature amount to the function (formula)
indicated by the gain adjustment model recorded in advance, and the mapping control
information Alternatively, the gain adjustment amount is calculated by substituting the value β
determined by the mapping control information.
[0135]
In step S106, the gain adjustment unit 24 adjusts the gain of the resonance band signal by
11-04-2019
31
converting the amplitude of the resonance band signal from the band division unit 21 based on
the gain adjustment amount from the gain adjustment amount calculation unit 143. The data is
supplied to the adding unit 25.
[0136]
In step S107, the adding unit 25 adds the resonance band signal supplied from the gain adjusting
unit 24 and the resonance band outside signal supplied from the band dividing unit 21 and
supplies the added signal to the mapping processing unit 144.
Hereinafter, the signal output from the adding unit 25, that is, the signal obtained by adding the
resonance band signal and the signal outside the resonance band is also referred to as an
addition signal.
[0137]
In step S108, the mapping processing unit 144 performs mapping processing on the addition
signal supplied from the addition unit 25 using the mapping control information supplied from
the mapping control information determination unit 142.
[0138]
For example, the mapping processing unit 144 substitutes the sample value x of the n-th sample
to be processed in the addition signal into the non-linear mapping function f (x) shown in the
following equation (6) to obtain the amplitude of the addition signal. Do the conversion.
That is, the value obtained by substituting the sample value x into the mapping function f (x) is
taken as the sample value of the n-th sample of the final output signal.
In this case, it is assumed that the sample value x of the addition signal is normalized to be a
value from -1.0 to 1.0.
Further, in the equation (6), α indicates the value of the mapping control information.
11-04-2019
32
[0139]
[0140]
Such a mapping function f (x) is a function that changes more sharply as the mapping control
information is smaller as shown in FIG.
In FIG. 12, the horizontal axis indicates the sample value x of the addition signal, and the vertical
axis indicates the value of the mapping function f (x).
Curves f11 to f13 represent mapping functions f (x) when the mapping control information α is
“3”, “5”, and “50”, respectively.
[0141]
As can be seen from FIG. 12, as the mapping control information is smaller, a mapping function f
(x) in which the amount of change in f (x) relative to the change in sample value x is larger is
used to perform amplitude conversion of the added signal. Be When the mapping control
information is changed as described above, the amplification amount for the addition signal
(input signal) is changed.
[0142]
For example, when the root mean square RMS (n) is determined as the reproduction level
emphasis analysis feature quantity, the reproduction level emphasis analysis feature quantity
decreases as the volume of the audio of the input signal decreases, so the value of the mapping
control information also decreases Become. Also, the smaller the mapping control information,
the larger the slope of the mapping function f (x).
[0143]
11-04-2019
33
Therefore, as the volume of the speech of the input signal is smaller as a whole, the mapping
function f (x) of the characteristic that changes more steeply in most sections of the sample value
x including the sample value x = 0 is used, Amplitude conversion of the addition signal (input
signal) is performed.
[0144]
As a result, in a section where the volume of the input signal as a whole is small, the amplitude of
the input signal is converted so that the sound having a small volume is converted to a sound
having a larger volume, and the reproduction level of the input signal is emphasized.
As a result, even in the case where, for example, content having a large dynamic range of volume
is reproduced by a device having a small speaker, it is possible to make it easy to hear small
sounds that have not been conventionally audible.
[0145]
In addition, even in a section where the volume is large as a whole in the input signal, the
mapping function f (x) having a suitably steep characteristic is used for a signal having a small
sample value x therein, and the input signal (addition signal) Amplitude conversion is performed.
[0146]
As a result, even in a section where the volume of the input signal is large as a whole, the input
signal is amplitude-converted such that the small volume sound in the section is converted to a
large volume, and the reproduction level of the input signal is emphasized.
This makes it possible to hear even louder sounds that have been reproduced relatively large
conventionally.
[0147]
Furthermore, by changing the non-linear mapping function f (x) according to the mapping
11-04-2019
34
control information, it is possible to realize amplitude conversion with a higher degree of
freedom. That is, according to the characteristics of the entire specific section of the input signal,
by using the non-linear function of the most effective characteristic as the mapping function, not
only the characteristic of the section including the sample to be processed but Amplitude
conversion can also be performed in consideration of the size.
[0148]
Note that the mapping function f (x) used in the mapping process is not limited to the nonlinear
function, but may be a function that satisfies −1 ≦ f (x) ≦ 1 for sample values x that satisfy −1
≦ x ≦ 1. For example, any function such as a linear function or an exponential function may be
used. For example, one having a high evaluation of the effect of the mapping process or one that
is suitable for hearing may be used as the mapping function.
[0149]
When the mapping processing unit 144 converts the amplitude of the addition signal by the
mapping processing into an output signal, the mapping processing unit 144 outputs the obtained
output signal, and the resonance band component suppression processing ends.
[0150]
As described above, the audio signal processing apparatus 131 extracts the resonance analysis
feature quantity from the input signal and the resonance band signal, and extracts the
reproduction level emphasis analysis feature quantity from the input signal to obtain mapping
control information.
Then, the audio signal processing device 131 adjusts the gain of the resonance band signal with
the gain adjustment amount obtained based on the resonance analysis feature amount and the
mapping control information, and further, based on the mapping control information, the
resonance band signal and the resonance band The amplitude conversion of the addition signal
obtained by adding the external signal is performed.
[0151]
11-04-2019
35
As described above, by obtaining the gain adjustment amount using the resonance analysis
feature amount and the mapping control information, it is possible to more appropriately
suppress the resonance band component of the input signal and obtain higher quality voice. In
particular, when combined with the reproduction level emphasizing technology, since the
reproduction level can be emphasized while the resonance is effectively suppressed, it is possible
to realize louder and powerful sound reproduction.
[0152]
By the way, the series of processes described above can be executed by hardware or software.
When the series of processes are performed by software, a program that configures the software
is installed on a computer. Here, the computer includes, for example, a general-purpose computer
capable of executing various functions by installing a computer incorporated in dedicated
hardware and various programs.
[0153]
FIG. 13 is a block diagram showing an example of a hardware configuration of a computer that
executes the series of processes described above according to a program.
[0154]
In the computer, a CPU 501, a ROM 502, and a RAM 503 are mutually connected by a bus 504.
[0155]
Further, an input / output interface 505 is connected to the bus 504.
An input unit 506, an output unit 507, a recording unit 508, a communication unit 509, and a
drive 510 are connected to the input / output interface 505.
[0156]
11-04-2019
36
The input unit 506 includes a keyboard, a mouse, a microphone, an imaging device, and the like.
The output unit 507 includes a display, a speaker, and the like. The recording unit 508 includes a
hard disk, a non-volatile memory, and the like. The communication unit 509 is formed of a
network interface or the like. The drive 510 drives removable media 511 such as a magnetic
disk, an optical disk, a magneto-optical disk, or a semiconductor memory.
[0157]
In the computer configured as described above, the CPU 501 loads, for example, the program
recorded in the recording unit 508 into the RAM 503 via the input / output interface 505 and
the bus 504, and executes the above-described series. Processing is performed.
[0158]
The program executed by the computer (CPU 501) can be provided by being recorded on, for
example, a removable medium 511 as a package medium or the like.
Also, the program can be provided via a wired or wireless transmission medium such as a local
area network, the Internet, or digital satellite broadcasting.
[0159]
In the computer, the program can be installed in the recording unit 508 via the input / output
interface 505 by attaching the removable media 511 to the drive 510. Also, the program can be
received by the communication unit 509 via a wired or wireless transmission medium and
installed in the recording unit 508. In addition, the program can be installed in advance in the
ROM 502 or the recording unit 508.
[0160]
Note that the program executed by the computer may be a program that performs processing in
chronological order according to the order described in this specification, in parallel, or when
necessary, such as when a call is made. It may be a program to be processed.
11-04-2019
37
[0161]
Further, the embodiments of the present technology are not limited to the above-described
embodiments, and various modifications can be made without departing from the scope of the
present technology.
[0162]
For example, the present technology can have a cloud computing configuration in which one
function is shared and processed by a plurality of devices via a network.
[0163]
Further, each step described in the above-described flowchart can be executed by one device or
in a shared manner by a plurality of devices.
[0164]
Furthermore, in the case where a plurality of processes are included in one step, the plurality of
processes included in one step can be executed by being shared by a plurality of devices in
addition to being executed by one device.
[0165]
Further, the effects described in the present specification are merely examples and are not
limited, and other effects may be present.
[0166]
Furthermore, the present technology can also be configured as follows.
[0167]
(1) A band division unit that divides a voice input signal into bands to generate a resonance band
signal, an analysis unit that extracts each feature amount from the resonance band signal and the
input signal, and the features of the resonance band signal A gain adjustment amount calculation
unit that calculates a gain adjustment amount of the resonance band of the input signal based on
the amount and the feature amount of the input signal; and the resonance band of the input
signal based on the gain adjustment amount An audio signal processing device comprising: a gain
11-04-2019
38
adjustment unit that performs gain adjustment.
(2) The audio signal processing device according to (1), wherein the gain adjustment amount
calculation unit calculates the gain adjustment amount based on the feature amount using a
statistical analysis model generated by learning.
(3) The gain adjustment amount calculation unit calculates the gain adjustment amount for each
of the plurality of resonance bands, and the gain adjustment unit performs gain adjustment for
each of the plurality of resonance bands (1) or (2) The audio signal processing apparatus as
described in 2.).
(4) Mapping which determines the mapping control information based on the reproduction level
emphasis feature quantity extracted from the input signal and information for obtaining mapping
control information from the reproduction level emphasis feature quantity generated by learning
The apparatus further includes a control information determination unit, and a mapping
processing unit that performs amplitude conversion on the input signal whose gain has been
adjusted based on a linear or non-linear mapping function determined by the mapping control
information, and the gain adjustment amount calculation unit The audio signal processing device
according to any one of (1) to (3), wherein the gain adjustment amount is calculated based on
mapping control information and the feature amount extracted by the analysis unit.
(5) The audio signal processing device according to any one of (1) to (4), wherein the gain
adjustment amount increases as the ratio of the resonance band signal included in the input
signal increases.
(6) The band division unit divides the input signal into the resonance band signal and the out-ofresonance band signal, and the gain adjustment unit adjusts the gain of the resonance band
signal, and the gain band is adjusted. The audio signal processing device according to any one of
(1) to (3), further including an addition unit that adds a signal and the signal outside the
resonance band.
(7) A voice input signal is divided into bands to generate a resonance band signal, feature
quantities of each of the resonance band signal and the input signal are extracted, and the
feature quantities of the resonance band signal and the input signal An audio signal processing
method comprising the steps of: calculating a gain adjustment amount of a resonance band of the
11-04-2019
39
input signal based on the feature amount; and performing gain adjustment of the resonance band
in the input signal based on the gain adjustment amount.
(8) A speech input signal is divided into bands to generate a resonance band signal, each feature
quantity is extracted from the resonance band signal and the input signal, and the feature
quantity of the resonance band signal and the input signal Based on the feature amount, calculate
the gain adjustment amount of the resonance band of the input signal, and make the computer
execute processing including the step of performing the gain adjustment of the resonance band
in the input signal based on the gain adjustment amount program.
[0168]
DESCRIPTION OF SYMBOLS 11 audio signal processing apparatus, 21 band division part, 22
analysis part, 23 gain adjustment amount calculation part, 24 gain adjustment part, 25 addition
part, 61 band division part, 62-1 to 62-K, 62 gain adjustment part, 101 Resonance control
information calculation unit, 102 resonance control unit, 141 reproduction level emphasis
analysis unit, 142 mapping control information determination unit, 143 gain adjustment amount
calculation unit, 144 mapping processing unit
11-04-2019
40
Документ
Категория
Без категории
Просмотров
0
Размер файла
52 Кб
Теги
description, jp2015050685
1/--страниц
Пожаловаться на содержимое документа