close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2013102411

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 JP2013102411
Abstract: To automatically control the level of an audio signal under various environments by
optimum mapping control according to environmental sounds. SOLUTION: The characteristics of
an input signal are analyzed to generate an input sound feature, the characteristics of an
environmental sound are analyzed to generate an environmental sound feature, and the
generated input sound feature and the environmental sound feature are applied. Mapping control
information is generated as control information of amplitude conversion processing on an input
signal. Furthermore, the input signal is amplitude converted based on a linear or non-linear
mapping function determined by the mapping control information to generate an output signal.
The mapping control information is generated with reference to, for example, a model generated
in consideration of an input signal and an environmental sound. [Selected figure] Figure 2
Audio signal processing device, audio signal processing method, and program
[0001]
The present disclosure relates to an audio signal processing device, an audio signal processing
method, and a program. Specifically, the present invention relates to, for example, a method for
automatically controlling the reproduction level of an audio signal optimally for the user.
[0002]
For example, when playing movie content or music content with a large dynamic range of sound
11-04-2019
1
volume on a portable device incorporating a small speaker, not only the sound volume of the
sound as a whole is reduced, but also serifs with particularly small volume, etc. It becomes
difficult to hear. Specifically, for example, as shown in FIG. 1, (A) a PC equipped with a small
microphone and a small speaker (B) a portable terminal equipped with a small microphone and a
small speaker There is a limitation in the size of the speaker, and there is a problem that it
becomes difficult to hear small speech with small volume without obtaining sufficient volume
output.
[0003]
There are techniques to adjust the volume, such as normalization and automatic gain control, as
techniques to make it easier to hear the sound of these contents, but such volume control does
not look ahead for sufficiently long data, which may be audible. It becomes unstable control.
[0004]
In addition, there is also a technology for boosting the low volume part of the sound and
compressing the high volume part by compression processing of the dynamic range of the
volume.
However, in compression processing, it is difficult to obtain a high enhancement effect of sound
if the characteristics of volume boost and compression are made versatile, and in order to obtain
high effects, it is necessary to change the characteristics for each content. There is.
[0005]
For example, dynamic range compression in Dolby AC 3 (Audio Code number 3) boosts a signal
with a smaller sound pressure level based on the sound pressure level specified in dialog
normalization, and compresses a signal with a large sound pressure level. Technology. However,
in this technique, in order to obtain a sufficient effect, it is necessary to specify sound pressure
levels for dialog normalization, and characteristics of boost and compression at the time of
encoding an audio signal.
[0006]
Furthermore, in the case where the dynamic range of the sound volume is compressed, a
11-04-2019
2
technique has also been proposed that makes it easier to hear small sounds of the audio signal by
multiplying the audio signal by a coefficient determined by the average of the absolute values of
the audio signal. (See, for example, Patent Document 1).
[0007]
Unexamined-Japanese-Patent No. 05-275950 gazette
[0008]
Nowadays, users carry various portable devices with built-in small speakers in various
environments, such as various quiet environments and noisy environments, and listen to various
types of content such as movies, music, self-recorded content, etc. The
However, depending on the size of the surrounding environment sound, the same playback
volume may be too loud or too small.
Therefore, in these portable devices, a technique for automatically controlling the volume of
various contents in accordance with the size of the environmental sound is required.
[0009]
The present disclosure has been made in view of, for example, the above-mentioned
circumstances, and an audio signal processing device, an audio signal processing method, and a
program for automatically controlling the reproduction level of an audio signal optimally
according to the size of environmental sound. It is intended to be provided.
[0010]
According to a first aspect of the present disclosure, there is provided an input analysis unit that
analyzes characteristics of an input signal and generates an input sound feature, an environment
analysis unit that analyzes an environment sound characteristic and generates an environment
sound feature, A mapping control information generation unit that generates mapping control
information as control information of amplitude conversion processing on the input signal by
applying an input sound feature amount and the environmental sound feature amount; linear or
non-linear determined by the mapping control information And a mapping processing unit that
amplitude-converts the input signal based on a mapping function to generate an output signal.
11-04-2019
3
[0011]
Furthermore, in an embodiment of the audio signal processing device according to the present
disclosure, the mapping control information generation unit is configured to generate a
preliminary mapping control information by applying the input sound feature amount; The
mapping control information adjustment unit is configured to generate the mapping control
information to be output to the mapping processing unit by adjustment processing in which the
environmental sound feature amount is applied to the general mapping control information.
[0012]
Furthermore, in one embodiment of the audio signal processing device according to the present
disclosure, the input analysis unit calculates a root mean square calculated using a plurality of
continuous samples defined in advance as the input sound feature amount, and the environment
analysis The unit calculates a root mean square calculated using a plurality of continuous
samples of the environmental sound signal as the environmental sound feature, and the mapping
control information generator calculates the root mean square of the input signal which is the
input sound feature. The mapping control information is generated using the square root and the
root mean square of the environmental sound signal, which is the environmental sound feature.
[0013]
Furthermore, in one embodiment of the audio signal processing device according to the present
disclosure, the input sound feature amount and the environmental sound feature amount may be
a root mean square of the feature amount calculation target signal, a logarithm of root mean
square, or a root mean square. The square root, or the logarithm of the root mean square, or the
zero crossing rate of the signal, or the slope of the frequency envelope, or the result of their
weighted addition.
[0014]
Furthermore, in one embodiment of the audio signal processing device according to the present
disclosure, the environment analysis unit is characterized by a band signal having a high
occupancy of environmental sound divided by band division processing from a collected sound
signal acquired through a microphone. Analysis is performed to calculate the environmental
sound feature quantity.
[0015]
Furthermore, in an embodiment of the audio signal processing device according to the present
11-04-2019
4
disclosure, the audio signal processing device includes a band limiting unit that performs band
limiting processing of the signal subjected to the mapping processing in the mapping processing
unit, The signal after band limitation in the limiting unit is output through the speaker.
[0016]
Furthermore, in one embodiment of the audio signal processing device according to the present
disclosure, the mapping control information generation unit applies a mapping control model
generated by statistical analysis processing to which a learning signal including an input signal
and an environmental sound signal is applied. Generating the mapping control information;
[0017]
Furthermore, in an embodiment of the audio signal processing device of the present disclosure,
the mapping control model is data in which mapping control information is associated with
various input signals and an environmental sound signal.
[0018]
Furthermore, in an embodiment of the audio signal processing device according to the present
disclosure, the input signal is configured by a plurality of input signals of a plurality of channels,
and the mapping processing unit performs individual mapping processing for each input signal.
is there.
[0019]
Furthermore, in an embodiment of the audio signal processing device according to the present
disclosure, the audio signal processing device further includes, for the environmental sound
feature value generated by the environment analysis unit, the mapping processing signal
generated by the mapping processing unit. It has a gain adjustment unit that performs a
corresponding gain adjustment.
[0020]
Furthermore, a second aspect of the present disclosure is an audio signal processing method to
be executed in an audio signal processing device, comprising: an input analysis step of analyzing
characteristics of an input signal to generate an input sound feature; Analyzing and generating an
environmental sound feature amount; applying the input sound feature amount and the
environmental sound feature amount; and generating mapping control information as control
information of amplitude conversion processing for the input signal An audio signal processing
method comprising: a control information generating step; and a mapping processing step of
11-04-2019
5
amplitude converting the input signal based on a linear or non-linear mapping function
determined by the mapping control information to generate an output signal.
[0021]
Furthermore, a third aspect of the present disclosure is a program that causes an audio signal
processing apparatus to execute audio signal processing, analyzes an input signal characteristic,
and generates an input sound feature, and an environmental sound characteristic. Analyzing the
environment sound to generate an environmental sound feature, and applying the input sound
feature and the environmental sound feature to generate mapping control information as control
information of amplitude conversion processing for the input signal. The program may execute a
mapping control information generation step, and a mapping processing step of amplitude
converting the input signal based on a linear or non-linear mapping function determined by the
mapping control information to generate an output signal.
[0022]
Note that the program of the present disclosure is, for example, a program that can be provided
by a storage medium or communication medium that provides various types of program code in
a computer-readable format to a general-purpose system that can execute the program.
By providing such a program in a computer readable form, processing according to the program
is realized on the computer system.
[0023]
Other objects, features, and advantages of the present disclosure will become apparent from the
more detailed description based on the embodiments of the present disclosure described later
and the attached drawings.
In addition, in this specification, a system is a logical set composition of a plurality of devices, and
the device of each composition is not limited to what exists in the same case.
[0024]
11-04-2019
6
According to an embodiment of the present disclosure, the optimum mapping control can be
performed when the environmental sound is large or small, and the user's dissatisfaction that the
volume is insufficient or the distortion may be reduced, and the reproduction level of the audio
signal Can be optimally controlled automatically for the user under various circumstances.
[0025]
Specifically, for example, the characteristic of the input signal is analyzed to generate the input
sound feature, the characteristic of the environmental sound is analyzed, the environmental
sound feature is generated, and the generated input sound feature and the environmental sound
feature are applied. Then, mapping control information is generated as control information for
amplitude conversion processing on the input signal.
Furthermore, the input signal is amplitude converted based on a linear or non-linear mapping
function determined by the mapping control information to generate an output signal.
The mapping control information is generated with reference to, for example, a model generated
in consideration of an input signal and an environmental sound.
With these configurations, it becomes possible to automatically control the level of the audio
signal under various environments by optimum mapping control according to the environmental
sound.
[0026]
It is a figure explaining an example of a device provided with a small speaker.
It is a block diagram showing an example of an audio signal processing method in a 1st
embodiment of this indication.
It is a figure which shows the example of the frequency band breakdown at the time of band
division | segmentation of the sound collection signal in 1st-8th embodiment of this indication.
11-04-2019
7
It is an example of the mapping control information adjustment amount function graph in 1st
Embodiment of this indication.
It is an example of the mapping function graph in 1st Embodiment of this indication.
It is a block diagram which shows the example of the audio | voice signal processing method in
2nd Embodiment of this indication.
It is a block diagram which shows the example of the audio | voice signal processing method in
3rd Embodiment of this indication.
It is a block diagram which shows the example of the mapping control model learning method in
3rd Embodiment of this indication.
It is a flow chart which shows the example of the mapping control information grant method in a
3rd embodiment of this indication.
It is an example of the graph of the regression curve by the mapping control model in 3rd
Embodiment of this indication. It is a block diagram which shows the example of the audio |
voice signal processing method in 4th Embodiment of this indication. It is a block diagram which
shows the example of the mapping control model learning method in 4th Embodiment of this
indication. It is a flow chart which shows the example of the mapping control information grant
method in a 4th embodiment of this indication. It is a block diagram which shows the example of
the audio | voice signal processing method in 5th Embodiment of this indication. It is a block
diagram which shows the example of the audio | voice signal processing method in 6th
Embodiment of this indication. It is a block diagram which shows the example of the audio |
voice signal processing method in 7th Embodiment of this indication. It is a block diagram which
shows the example of the audio | voice signal processing method in 8th Embodiment of this
indication.
[0027]
11-04-2019
8
Hereinafter, the audio signal processing device, the audio signal processing method, and the
program of the present disclosure will be described in detail with reference to the drawings. Note
that the audio signal processing device of the present disclosure controls output sound from a
speaker such as a device including a small speaker as described above with reference to FIG. 1,
for example, and various ambient noises. Even in an environment where an environmental sound
such as that is generated, audio signal processing is performed to make it easier to hear the
output sound. Specifically, for example, processing for automatically controlling the reproduction
level of the audio signal optimally according to the environmental sound is performed.
[0028]
Several embodiments of the audio signal processing device according to the present disclosure
will be described sequentially according to the following items. 1. About the first embodiment
Second Embodiment 3. About 3rd Embodiment 4. About the fourth embodiment 5. Fifth
Embodiment About the sixth embodiment About the seventh embodiment
[0029]
[1. First Embodiment] A block diagram of an audio signal processing device according to a
first embodiment of the present disclosure is shown in FIG. The audio signal processing
apparatus 100 shown in FIG. 2 can be configured as an internal device of an information
processing apparatus such as (A) PC and (B) portable terminal described above with reference to
FIG. 1, for example. It can also be configured as an independent device connected to the audio
output device and processing the audio signal output from the audio output device.
[0030]
The audio signal processing apparatus 100 shown in FIG. 2 has the following configuration.
Input unit 101, input analysis / mapping control information determination unit 102,
microphone 111, band division unit 112, environment analysis unit 113, mapping control
information adjustment unit 114, mapping processing unit 121, band limitation unit 122,
speaker 123, configuration of these Have.
11-04-2019
9
[0031]
The input unit 101 is an input unit of an audio signal to be reproduced. For example, in an
information processing apparatus such as (A) a PC or (B) a portable terminal as shown in FIG. 1,
it is an input unit for an audio signal generated by a reproduction signal generation unit in the
information processing apparatus. Alternatively, it corresponds to an input unit or the like
connected to an audio output unit of an external audio reproduction device. The audio signal
processing apparatus shown in FIG. 2 includes a microphone 111 and a speaker 123 as in the PC
and the portable terminal shown in FIG.
[0032]
The reproduction target input signal input from the input unit 101 is input to the input signal
analysis / mapping control information determination unit 102. The input signal analysis /
mapping control information determination unit 102 analyzes the characteristics of the input
speech signal. Specifically, input signal analysis / mapping control information determination
section 102 performs root mean square RMS by N samples centering on the nth sample of the
input signal from input section 101 according to (Equation 1) shown below. n) Calculate and
output.
[0033]
... (Equation 1)
[0034]
In the above (formula 1), x is a reproduction target input signal input from the input unit 101,
and is, for example, data in which the audio level is normalized to a value of -1.0 to 1.0.
The input signal analysis / mapping control information determination unit 102 sets the
processing target signal as the nth sample signal, and uses N consecutive samples defined in
advance with the nth sample as a center, according to the above (Equation 1). The root mean
square RMS (n) is calculated as the feature quantity corresponding to the n-th sample.
11-04-2019
10
[0035]
The input signal analysis / mapping control information determination unit 102 supplies the root
mean square RMS (n) calculated according to (Equation 1) above to the mapping control
information adjustment unit 114 as the mapping control information α0 for the nth input
sample signal. .
[0036]
In the processing example described above, the mapping control information calculated by the
input signal analysis / mapping control information determination unit 102 is a processing
example using root mean square RMS (n).
However, as the mapping control information, in addition to the root mean square RMS (n),
various analysis feature quantities such as the t-th power of RMS (n) (t> 2), the zero crossing rate,
and the slope of the frequency envelope It is possible to use The mapping control information
adjustment unit 114 may be configured to generate the mapping control information α0 based
on data obtained by arbitrarily adding and combining the feature amounts related to these
various input signals, for example, the weighted addition result.
[0037]
The mapping control information adjustment unit 114 adjusts the mapping control information
according to the size of the environmental sound with respect to the mapping control
information α 0 input from the input signal analysis / mapping control information
determination unit 102.
[0038]
The environmental sound is a sound included in the sound collection signal from the microphone
111.
The signal (sound collection signal) collected by the microphone 111 includes an ambient pure
environmental sound and an output signal output from the speaker 123 of the audio signal
processing apparatus 100. That is, as shown in FIG. 3, the output signal from the speaker is
included together with the surrounding sound (environmental sound). In the following
11-04-2019
11
description, the environmental sound includes all sounds except the output signal from the
speaker 123 of the audio signal processing apparatus 100 from the collected sound signal of the
microphone 111. That is, the environmental sound includes various sounds and noises in the
surrounding area, and also includes, for example, a voice emitted by the user and noise generated
from the apparatus itself.
[0039]
FIG. 3 is an example of analysis data of a signal (sound collection signal) collected by the
microphone 111, in which the horizontal axis represents frequency and the vertical axis
represents power spectrum. For example, as an example, as shown in FIG. 3, a characteristic is
obtained that the ratio of the output signal from the speaker 123 is large in a band with a
frequency of 150 Hz or less for an environmental sound and in a band of 150 Hz or more. It
should be noted that the environmental sound and the speaker output signal are separated at a
boundary of frequency = 150 Hz shown in FIG. 3 because the output signal from the speaker 123
is band-limited by the band-limiting section 122 at the previous stage of the speaker 123. Do.
That is, the output signal from the speaker 123 is band-limited at a stage before being collected
by the microphone 111. Details of this band limiting process will be described later.
[0040]
A sound pickup signal by microphone 111 is divided into a low band signal of 150 Hz or less
which is a frequency band including only environmental sound in band division section 112, and
a high band signal including an output signal from speaker 123 in addition to the environmental
sound. Ru.
[0041]
In this processing example, the frequency is divided into two at 150 Hz according to the
characteristics described with reference to FIG. 3, but it may be divided into a band including
only environmental sound and a band other than that. It is good to divide by a suitable frequency.
If the band of the signal input from the input unit 101 is known in advance, division processing
may be performed according to the input signal. Specifically, for example, when the input signal
from the input unit 101 is a signal in which the low band and the high band are cut, the low
11-04-2019
12
band, the mid band, and the high band are divided into three, and each divided area unit The area
of only the environmental sound may be divided into the mixed area of the environmental sound
and the output signal from the speaker.
[0042]
The collected sound signal divided by the band division unit 112 is input to the environment
analysis unit 113. The environment analysis unit 113 calculates the feature amount of the
environmental sound. That is, in the present processing example, the feature amount of the low
band signal estimated that most of the collected sound signals divided by the band dividing unit
112 are composed of the environmental sound is calculated. Specifically, among the collected
sound signals divided in the same manner as (Equation 1) above, root mean square RMS (k ) Is
supplied to the mapping control information adjustment unit 114 as the analysis feature value.
[0043]
In addition to the root mean square RMS (k), the feature amount of the environmental sound in
the environment analysis unit 113 is the t-th power of RMS (n) (t> 2), the zero crossing rate, the
slope of the frequency envelope, etc. Alternatively, data obtained by arbitrarily adding and
combining various analysis feature amounts, for example, weighted addition results may be used.
[0044]
Also, when the band signal including only the environmental sound is only the high band or both
the low band and the high band, the analysis feature quantity of only the high band signal or the
analysis feature quantity obtained from the low band signal and the high band signal is applied
Do.
Depending on the mixture ratio of environmental sound, a weighted sum of the analysis
characteristic of the low band and the analysis characteristic of the high band may be calculated
as the final analysis characteristic of the environmental sound.
[0045]
In this embodiment, the analysis feature is obtained from the band division signal excluding the
11-04-2019
13
reproduction band of the speaker 123, but only the low band, only the high band, or both the low
band and the high band excluding the middle band From the analysis feature quantities of the
band split signal, it is also possible to obtain analysis feature quantities of signals not to be
analyzed in the mid band signal or the entire frequency band, using a function or a table or a
statistical model based on statistical analysis in advance.
[0046]
For example, when the high band is missing in two divisions, the low band signal is divided into a
plurality of sub-bands, and the mean and the slope of the root mean square of each sub-band
signal are used as explanatory variables. Regression estimation may be performed using the root
mean square of each subband signal when divided into two as the dependent variable, and the
result may be used as the final analysis feature value.
[0047]
Furthermore, although the microphone 111 is described here as a monaural microphone, the
microphone 111 may be configured as two or more microphones.
In that case, band division is performed for each microphone, and the respective signals are
supplied to the environment analysis unit 113.
Further, in addition to the above-mentioned analysis feature quantities, differences and
correlations of signals from the microphones, estimated sound source directions and the like may
be used as analysis feature quantities.
[0048]
The environmental sound feature quantity which is the feature quantity of the environmental
sound calculated by the environmental analysis unit 113 is input to the mapping control
information adjustment unit 114.
[0049]
The mapping control information adjustment unit 114 receives, from the input signal analysis /
11-04-2019
14
mapping control information determination unit 102, the mapping control information α0 that
is the feature amount for the nth input sample signal, and the feature of the environmental sound
calculated by the environment analysis unit 113. Enter the amount.
These are all, for example, values of root mean square RMS calculated according to (Equation 1)
described above.
[0050]
The mapping control information adjustment unit 114 adjusts the mapping control information
α0 which is the feature amount for the nth input sample signal based on the environmental
sound feature amount obtained from the environment analysis unit 113, and supplies the
mapping control information α0 to the mapping processing unit 121. Do.
[0051]
The mapping control information adjustment unit 114 obtains the mapping control information
adjustment amount y using, for example, a non-linear function such as the following (Expression
2).
x is the environmental sound feature value RMS (k).
[0052]
... (Equation 2) In addition, p, q and r are parameters prescribed beforehand.
[0053]
FIG. 4 shows a graph corresponding to the above (formula 2).
The graph of FIG. 4 is a graph in which the horizontal axis (x) and the vertical axis (y) are set as
follows. x: environmental sound feature amount RMS (k) y: mapping control information
adjustment amount It is a graph which shows these correspondences.
11-04-2019
15
[0054]
The horizontal axis (x) corresponds to the power (db) of the environmental sound. It means that
the power of the environmental sound increases as you move to the right. As the environmental
sound becomes larger, the mapping control information adjustment amount y becomes smaller,
and as the environmental sound becomes smaller, the mapping control information adjustment
amount y becomes larger.
[0055]
In this embodiment, the non-linear function shown in the above (Equation 2) is used in the
process of calculating the mapping control information adjustment amount y, but a linear or nonlinear function representing the relationship between the environmental sound feature and the
mapping control information adjustment amount Functions or tables or linear regression models
or non-linear regression models may be used.
[0056]
The mapping control information adjustment unit 114 further uses the mapping control
information adjustment amount y calculated by (Equation 2) and further uses a function such as
(Equation 3) shown below to perform input analysis / mapping control information
determination unit The mapping control information α0, which is a feature for the input sample
signal input from 102, is adjusted.
[0057]
... (Equation 3)
[0058]
In the above (Equation 3), α 0 is mapping control information RMS (n) which is a feature amount
for the input sample signal input from the input analysis / mapping control information
determination unit 102, and α is mapping control information after adjustment, is there.
[0059]
As described above with reference to FIG. 4, as the environmental sound increases, the mapping
11-04-2019
16
control information adjustment amount y decreases, and as the environmental sound decreases,
the mapping control information adjustment amount y increases.
Therefore, the value of the adjusted mapping control information α is adjusted as follows.
As the environmental sound becomes larger, the value of mapping control information α after
adjustment becomes smaller, and as the environmental sound becomes smaller, the value of
mapping control information α after adjustment becomes larger.
[0060]
In this embodiment, as the process of calculating the adjusted mapping control information α,
the mapping control information adjustment amount y calculated by (Expression 2) is added to
the mapping control information α0 which is the feature amount for the input sample signal and
However, these values may be multiplied to calculate the adjusted mapping control information
α according to the above equation, for example, α = α0 × y.
In addition, linear or non-linear functions or tables or linear regression models or non-linear
regression models may be used.
[0061]
As described above, the mapping control information adjustment unit 114 applies the
environmental sound feature quantity x (= RMS (k)) and uses the non-linear function (FIG. 4)
shown in (Expression 2) to adjust the mapping control information adjustment quantity. Further,
y is obtained, and the adjustment value of mapping control information α0 which is a feature for
the input sample signal input from input analysis / mapping control information determination
unit 102 using mapping control information adjustment amount y, that is, adjustment mapping
control information Calculate α.
[0062]
The adjustment mapping control information α calculated by the mapping control information
adjusting unit 114 is input to the mapping processing unit 121.
11-04-2019
17
The mapping processing unit 121 converts the amplitude of the reproduction target input signal
input from the input unit 101 and outputs the converted signal to the band limiting unit 122
using a non-linear function such as the following (Expression 4) as the mapping function.
[0063]
... (Equation 4)
[0064]
In the above (Equation 4), x is, for example, an input sample signal obtained by normalizing the
power in the range of −1.0 to 1.0, α is the adjusted mapping supplied from the mapping control
information adjustment unit 114 Control information,
[0065]
The graph of (Expression 4) is shown in FIG.
The horizontal axis is x, that is, a normalized signal x of −1.0 to 1.0, and the vertical axis is f (x),
that is, the output f (x) calculated according to the above (Equation 4). (X),
[0066]
FIG. 5 exemplifies the values of the adjusted mapping control information α supplied from the
mapping control information adjusting unit 114 with respect to these three types, α = 50, α = 5,
α = 3.
The smaller the adjusted mapping control information α, the larger the amplification amount is
set.
[0067]
11-04-2019
18
As described above with reference to (Expression 3), the value of the adjusted mapping control
information α is adjusted as follows.
As the environmental sound becomes larger, the value of mapping control information α after
adjustment becomes smaller, and as the environmental sound becomes smaller, the value of
mapping control information α after adjustment becomes larger. Therefore, the amplification
amount is set to be larger as the environmental sound is larger, and the amplification amount is
set to be smaller as the environmental sound is smaller.
[0068]
As described above, the audio signal processing device 100 according to the present disclosure
executes a process of changing the amplification amount of the input signal by changing the
adjusted mapping control information α according to the environmental sound. The influence on
the input signal by the change processing of the amplification amount changes according to the
size of the mapping control information α0 (= RMS (n)) which is the feature amount for the n-th
input sample signal, for example. That is, when the RMS (n) is small, amplitude conversion to
which a steep mapping function is applied is performed on the n-th input sample signal, and
when the RMS (n) is large, a loose characteristic is obtained. Amplitude conversion to which the
mapping function is applied will be performed.
[0069]
Also, the amount of amplification changes depending on the magnitude of the environmental
sound. That is, as understood from FIG. 4, FIG. 5 and (Equation 3) and (Equation 4) described
above, the feature amount RMS (k) (x in FIG. 4) of the environmental sound is large, that is, the
environmental sound is large. The value of mapping control information α after adjustment
becomes smaller as it becomes, the amplification amount as the adjustment amount increases as
shown in FIG. 5, and the adjustment process of mapping control information according to the size
of the environmental sound is executed.
[0070]
11-04-2019
19
Although a non-linear function is used as the mapping function in this embodiment, a linear
function or an exponential function may be used, and for the input of −1.0 ≦ x ≦ 1.0, −1.0 ≦ f
(x It is possible to apply any function that satisfies the condition of ≦ 1.0. It is better to use a
mapping function that is suitable for processing effects and hearing.
[0071]
Here, although the mapping control information α is derived for each sample of the input signal
and the amplitude conversion is controlled in the mapping control unit, for example, the control
information α is derived for every two or more consecutive samples. The amplitude conversion
in the mapping control unit may be controlled.
[0072]
As described above, the mapping processing unit 121 converts the amplitude of the reproduction
target input signal input from the input unit 101 by using the above-described (Expression 4),
that is, the non-linear function as shown in FIG. Output to the part 122.
[0073]
Finally, the band limiting unit 122 applies a band limiting filter to the amplitude-converted input
signal output from the mapping processing unit 121 to generate a band-limited output signal.
For example, low pass cut processing is performed.
Specifically, for example, when reproducing with the small speaker 123 which is the output unit,
processing is performed to cut the low band to such an extent that the difference in auditory
sense is small even compared to before band limitation.
[0074]
The band limiting unit 122 may perform band limiting on the reproduction target input signal
instead of performing band limiting on the amplitude converted input signal output from the
mapping processing unit 121. Furthermore, when the reproducible band is limited by the
performance of the speaker 123, that is, when the band is naturally limited at the time of speaker
11-04-2019
20
reproduction, the band restriction process may not be performed again. Further, although the
frequency cut by the band limiting unit is only the low band here, only the high band or both the
low band and the high band may be cut. It is preferable to limit the frequency band to a
frequency band suitable for hearing and analysis in the above-mentioned environment analysis
unit 113.
[0075]
As described above, the sound collection signal acquired by the microphone 111 is divided into
bands, and the appropriate mapping control information adjustment amount is determined from
the analysis result of the environmental sound, so that the optimal mapping control information
according to the size of the environmental sound Thus, it is possible to realize the optimum
reproduction level control according to the environment for the user.
[0076]
[2.
Second Embodiment] A block diagram of an audio signal processing device according to a second
embodiment of the present invention is shown in FIG. The audio signal processing apparatus 200
shown in FIG. 6 includes an input unit 201, an input analysis / mapping control information
determination unit 202, a microphone 211, a band division unit 212, an environment analysis
unit 213, a mapping processing unit 221, a band restriction unit 222, and a speaker 223. , Have.
[0077]
The difference from the audio signal processing apparatus 100 of the first embodiment
described with reference to FIG. 2 is that the mapping control information adjustment unit 114
shown in FIG. 2 is omitted. In the audio signal processing apparatus 200 of the second
embodiment shown in FIG. 6, the input analysis / mapping control information determination
unit 202 generates final mapping control information α to be output to the mapping processing
unit 221.
[0078]
11-04-2019
21
The processing of other configurations is the same as that of the first embodiment. That is, the
sound collection signal acquired by the microphone 211 is divided into bands, and the
environment analysis unit analyzes the sound collection signal to obtain the environmental sound
feature value RMS (k).
[0079]
As in the first embodiment, the input signal analysis / mapping control information
determination unit 202 analyzes the characteristics of the reproduction target input signal input
from the input unit 201, and obtains the input sound feature value RMS (n). Then, from the input
sound feature amount RMS (n) and the environmental sound feature amount RMS (k), the
mapping control information α is obtained using a function of (Expression 5) shown below and
supplied to the mapping processing unit 221.
[0080]
... (Equation 5) a and b are parameters prescribed beforehand.
[0081]
In this embodiment, mapping control information is obtained using the function of the above
(Equation 5) from the input sound feature quantity RMS (n) and the environmental sound feature
quantity RMS (k) only in the input signal analysis / mapping control information determination
unit 202. α is determined and supplied to the mapping processing unit 221.
[0082]
Although RMS (n) and RMS (k) are shown as analysis feature quantities of the input signal and
the environmental sound also in this second embodiment, other analysis features similar to those
described in the first embodiment You may use quantity.
[0083]
Similar to the first embodiment described above, the mapping processing unit 221 uses a nonlinear function such as (Equation 4) described above as the mapping function.
11-04-2019
22
In Equation (4), x is a normalized input sample signal in the range of −1.0 to 1.0, and α is
mapping control information.
[0084]
Thereafter, mapping processing is performed as in the first embodiment of the present invention,
band limitation is performed in the band limiting unit 222, and an output signal is output
through the speaker 223.
[0085]
As described above, the sound collection signal is divided into bands, the environmental sound is
analyzed, and the mapping control information is obtained based on the analysis feature amount
to obtain the optimal mapping control information according to the size of the environmental
sound. It is possible to realize the optimum reproduction level control according to the
environment for the user.
[0086]
[3.
About Third Embodiment] A block diagram of an audio signal processing device 300 according to
a third embodiment of the present disclosure is shown in FIG.
The audio signal processing device 300 shown in FIG. 7 has the following configuration.
Input unit 301, input analysis unit 302, mapping control information determination unit 303,
mapping control model 304 (storage unit), microphone 311, band division unit 312,
environment analysis unit 313, mapping control information adjustment unit 321, mapping
processing unit 322, The band limiting unit 323 and the speaker 324 have these configurations.
[0087]
In FIG. 7, the reproduction target input signal input from the input unit 301 is supplied to the
input analysis unit 302, and its characteristics are analyzed.
11-04-2019
23
[0088]
The input analysis unit 302 calculates the root mean square RMS (n) of N samples centered on
the n-th sample of the input signal from the input unit 301 according to (Equation 1) described
above in the first embodiment. It is calculated as an input sound feature amount for the third
reproduction target input signal and supplied to the mapping control information determination
unit 303.
The analysis feature quantity is not limited to RMS (n), and the other analysis feature quantities
described above may be used or optionally added / combined.
[0089]
Next, the mapping control information determination unit 303 obtains mapping control
information corresponding to the input analysis feature amount using the mapping control model
304 generated by learning processing to be executed in advance, and supplies the mapping
control information adjustment unit 321 with it. Do.
[0090]
The mapping control model 304 is generated in advance based on learning processing, that is,
statistical analysis to which learning data is applied.
A method of generating the mapping control model 304 will be described with reference to FIG.
FIG. 8 is a diagram showing the configuration of a learning process for generating the mapping
control model 304, that is, a learning device 350 for executing statistical analysis.
[0091]
The learning device 350 shown in FIG. 8 includes an input unit 351, a mapping control
11-04-2019
24
information addition unit 352, a mapping processing unit 353, a band limiting unit 354, a
speaker 355, an input analysis unit 356, a mapping control model learning unit 357, and a
recording unit 358. Configured In the learning device 350, a learning sound source signal used
for learning the mapping control model is supplied to the mapping control information adding
unit 352, the input analysis unit 356, and the mapping processing unit 353.
[0092]
The input unit 351 includes, for example, a button or the like operated by the user, and supplies
a signal corresponding to the user's operation to the mapping control information attachment
unit 352. The mapping control information adding unit 352 adds mapping control information to
each sample of the supplied learning sound source signal according to the signal from the input
unit 351, and supplies the mapping control information to the mapping processing unit 353 or
the mapping control model learning unit 357.
[0093]
The mapping processing unit 353 performs mapping processing on the supplied learning sound
source signal using the mapping control information from the mapping control information
adding unit 352, and supplies the learning output signal obtained as a result to the band limiting
unit 354. Do. The band limiting unit 354 performs band limiting processing such as low band
cutting, for example, and supplies a processing signal to the speaker 355. The speaker 355
reproduces the sound based on the learning output signal generated by the mapping processing
unit 353.
[0094]
The input analysis unit 356 analyzes the characteristics of the supplied learning sound source
signal, and supplies an analysis feature indicating the analysis result to the mapping control
model learning unit 357. The mapping control model learning unit 357 obtains a mapping
control model by statistical learning using the analysis feature amount from the input analysis
unit 356 and the mapping control information from the mapping control information adding unit
352, and supplies the mapping control model to the recording unit 358.
11-04-2019
25
[0095]
The recording unit 358 records the mapping control model supplied from the mapping control
model learning unit 357. The mapping control model recorded in the recording unit 358 in this
manner is recorded as the mapping control model 304 in the storage unit of the audio signal
processing apparatus 300 shown in FIG. 7.
[0096]
Note that the learning device 350 shown in FIG. 8 can be configured inside the audio signal
processing device 300 shown in FIG. 7 or can be configured as an external device. When the
learning device 350 shown in FIG. 8 is configured inside the audio signal processing device 300
shown in FIG. 7, the components of the learning device shown in FIG. 8 are common to the
components of the audio signal processing device 300 shown in FIG. Regarding the components,
the components of the audio signal processing device 300 can be applied as the components of
the learning device.
[0097]
Next, the learning process by the learning device 350 shown in FIG. 8 will be described with
reference to the flowchart shown in FIG. In this learning process, one or more learning sound
source signals are supplied to the learning device 350. Further, in this case, the input analysis
unit 356, the mapping processing unit 353, the speaker 355, and the like correspond to the
input analysis unit 321, the mapping processing unit 322, and the like of the audio signal
processing apparatus 300 to which the mapping control model obtained by learning is supplied.
The same as each block. That is, block characteristics and processing algorithms are the same.
[0098]
In step S11, the input unit 351 receives input or adjustment of mapping control information
from the user. For example, when a learning sound source signal is input, the mapping
processing unit 353 supplies the supplied learning sound source signal to the speaker 355, and
outputs sound based on the learning sound source signal. Then, while listening to the output
voice, the user operates the input unit 351 with a predetermined sample of the learning sound
11-04-2019
26
source signal as a processing target sample, and instructs to assign mapping control information
to the processing target sample.
[0099]
The instruction to assign mapping control information is performed, for example, by the user
directly inputting the mapping control information or designating a desired one from several
pieces of mapping control information. Alternatively, the user may instruct to assign the mapping
control information by instructing adjustment of the mapping control information once
designated.
[0100]
When the user operates the input unit 351 in this manner, the mapping control information
adding unit 352 adds mapping control information to the processing target sample according to
the user's operation. Then, the mapping control information adding unit 352 supplies the
mapping control information added to the processing target sample to the mapping processing
unit 353. In step S12, the mapping processing unit 353 performs mapping processing on the
processing target sample of the supplied learning sound source signal using the mapping control
information supplied from the mapping control information adding unit 352, and the obtained
result is obtained. The learning output signal is supplied to the speaker 355.
[0101]
For example, the mapping processing unit 353 substitutes the sample value x of the processing
target sample of the learning sound source signal into the non-linear mapping function f (x)
shown in (Expression 4) described above to perform amplitude 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 processing target sample of the learning output signal.
[0102]
Note that, in (Expression 4), it is assumed that the sample value x of the learning sound source
11-04-2019
27
signal is normalized so as to be a value from −1 to 1. Further, in (Expression 4), α indicates
mapping control information.
[0103]
Such a mapping function f (x) is a function that changes more sharply as the mapping control
information α is smaller as shown in FIG. In FIG. 5, the horizontal axis indicates the sample value
x of the learning sound source signal, and the vertical axis indicates the value of the mapping
function f (x). The mapping function f (x) when the mapping control information α is “3”,
“5”, and “50” is represented.
[0104]
As can be seen from FIG. 5, as the mapping control information α is smaller, a mapping function
f (x) with a larger amount of change in f (x) with respect to the change in sample value x is
generally used to convert the amplitude of the learning sound source signal. Is done. When the
mapping control information α is changed as described above, the amplification amount of the
learning sound source signal is changed.
[0105]
Returning to the explanation of the flowchart in FIG. 9, in step S13, the speaker 355 reproduces
the learning output signal supplied from the mapping processing unit 353. More specifically, the
learning output signal obtained by performing the mapping process on a predetermined section
including the processing target sample is reproduced. Here, the section to be reproduced is, for
example, a section including samples for which mapping control information has already been
designated. In this case, each sample of the section to be processed is subjected to mapping
processing using mapping control information defined for the samples, and the learning output
signal obtained as a result is reproduced.
[0106]
Thus, when the learning output signal is reproduced, the user evaluates the effect of the mapping
11-04-2019
28
process while listening to the sound output from the speaker 355. That is, it is evaluated whether
the sound volume of the learning output signal is appropriate. Then, the user operates the input
unit 351 and instructs adjustment of the mapping control information from the result of the
evaluation, or determines that the specified mapping control information is optimum and decides
the specified mapping control information. To direct.
[0107]
In step S14, based on the signal according to the user's operation supplied from the input unit
351, the mapping control information adding unit 352 determines whether or not optimum
mapping control information has been obtained. For example, when determination of mapping
control information is instructed by the user, it is determined that optimal mapping control
information has been obtained.
[0108]
If it is determined in step S14 that optimal mapping control information has not been obtained
yet, that is, if adjustment of mapping control information is instructed, the process returns to
step S11, and the above-described process is repeated.
[0109]
In this case, new mapping control information is added to the sample to be processed, and the
mapping control information is evaluated.
As described above, by evaluating the effect of the mapping process while actually listening to
the speech of the learning output signal, it is possible to provide the optimum mapping control
information in terms of hearing.
[0110]
On the other hand, if it is determined in step S14 that optimal mapping control information has
been obtained, the process proceeds to step S15. In step S15, the mapping control information
adding unit 352 supplies the mapping control model learning unit 357 with the mapping control
11-04-2019
29
information added to the processing target sample.
[0111]
In step S16, the input analysis unit 356 analyzes the characteristics of the supplied learning
sound source signal, and supplies the analysis feature obtained as a result to the mapping control
model learning unit 357. For example, assuming that the n-th sample of the learning sound
source signal is a sample to be processed, the input analysis unit 356 performs the operation of
(Equation 1) described above to calculate the root mean square for the n-th sample of the
learning sound source signal. RMS (n) is calculated as an analysis feature of the n-th sample.
[0112]
In the present example, in (Expression 1), x (m) indicates the sample value (the value of the
learning sound source signal) of the m-th sample of the learning sound source signal. Further, in
(Expression 1), it is assumed that the value of the learning sound source signal, that is, the
sample value of each sample of the learning sound source signal is normalized such that −1 ≦ x
(m) ≦ 1.
[0113]
Therefore, the root mean square RMS (n) takes the logarithm of the root of the root mean square
of the sample values of the samples contained in the section, for the section consisting of N
consecutive samples centered on the nth sample, Is obtained by multiplying the value obtained
by the constant "20".
[0114]
The value of the root mean square RMS (n) obtained in this manner is smaller as the absolute
value of the sample value of each sample in a specific section centered on the nth sample of the
learning source signal to be processed is smaller. , Become smaller.
That is, the lower the volume of the sound of the entire specific section including the sample to
be processed of the learning sound source signal, the smaller the root mean square RMS (n).
11-04-2019
30
[0115]
Although the root mean square RMS (n) has been described as an example of the analysis feature,
the analysis feature is a t-th power value of RMS (n) (where t t 2) or a zero crossing rate of the
learning sound source signal The inclination of the frequency envelope of the learning sound
source signal may be used, or a combination of them, for example, a weighted addition result
may be used.
[0116]
As described above, when the analysis feature quantity is supplied from the input analysis unit
356 to the mapping control model learning unit 357, the mapping control model learning unit
357 determines the analysis feature quantity obtained for the sample to be processed and the
sample The mapping control information is associated with and temporarily recorded.
[0117]
In step S17, the learning device 51 determines whether a sufficient number of mapping control
information has been obtained.
For example, when a set of analysis feature amounts and mapping control information
temporarily recorded is obtained by a sufficient number to learn the mapping control model, it is
determined that a sufficient number of mapping control information has been obtained. Ru.
[0118]
If it is determined in step S17 that a sufficient number of mapping control information has not
been obtained, the process returns to step S11, and the above-described process is repeated.
That is, the next sample of the sample to be processed at the current time of the learning sound
source signal is set as a new processing target sample and mapping control information is added,
or mapping control information for the new learning sound source signal sample Is given. Also,
mapping control information may be added to the samples of the learning sound source signal by
different users.
11-04-2019
31
[0119]
If it is determined in step S17 that a sufficient number of mapping control information has been
obtained, in step S18, the mapping control model learning unit 357 uses a set of temporarily
stored analysis feature amounts and mapping control information. And learn the mapping control
model.
[0120]
For example, the mapping control model learning unit 357 determines that the mapping control
information α can be obtained from the analysis feature amount by performing the calculation
of (Expression 6) shown below, and sets the function shown in (Expression 6) as a mapping
control model, This is determined by learning.
[0121]
... (Equation 6)
[0122]
In Equation (6), x represents an analysis feature value, and a, b and c are constants.
In particular, the constant c is an offset term uncorrelated with the analysis feature x.
[0123]
In this case, the mapping control model learning unit 66 sets the square value of the root mean
square RMS (n) and the root mean square RMS (n) corresponding to x and x <2> in (Equation 6)
as explanatory variables, and performs mapping control. The linear regression model is learned
by the least squares method using the information α as an explained variable, and model
parameters a, b, c are obtained.
[0124]
Thereby, for example, the result shown in FIG. 10 is obtained.
11-04-2019
32
In FIG. 10, the vertical axis represents mapping control information α, and the horizontal axis
represents root mean square RMS (n) as an analysis feature.
In FIG. 10, the curve shows the value of the mapping control information α determined for each
value of each analysis feature, that is, the graph of the function shown in the above (Equation 6).
[0125]
In this example, the volume of the sound of the speech signal such as the learning sound source
signal and the input signal is smaller, and the value of the mapping control information α is
smaller as the analysis feature amount is smaller.
If the constants a, b, and c in the function ax <2> + bx + c for obtaining mapping control
information from the analysis feature amount are determined by learning as described above, the
mapping control model learning unit 357 maps these constants. The recording unit 358 is
supplied as a model parameter of the control model and recorded.
[0126]
When the mapping control model obtained by learning is recorded in the recording unit 358, the
learning process ends. The mapping control model recorded in the recording unit 358 is then
recorded as the mapping control model 304 in the recording unit of the audio signal processing
apparatus 300 shown in FIG. 7 and used for the mapping process.
[0127]
As described above, the learning device 350 shown in FIG. 8 performs mapping by learning using
a plurality of learning sound source signals for each audio signal processing device 300 shown in
FIG. 7 and mapping control information designated by a plurality of users. Find a control model.
[0128]
11-04-2019
33
Therefore, using the obtained mapping control model, it is possible to obtain statistically
optimum mapping control information with respect to the audio signal processing apparatus 300
regardless of the input signal to be reproduced and the user who listens to the reproduced
sound. become able to.
In particular, if learning is performed using only mapping control information assigned by one
user, it is possible to generate a mapping control model that can obtain mapping control
information optimal for the user.
[0129]
In the above, the case of performing input and adjustment of mapping control information for
each sample with respect to the learning sound source signal has been described as an example,
but mapping control information for each two or more consecutive samples of the learning sound
source signal Input or adjustment may be performed.
[0130]
In addition, although a quadratic expression related to RMS (n) is used as the mapping control
model here, a third or higher order function may be used.
Also, although it has been described that root mean square RMS (n) and its square value are used
as explanatory variables of the mapping control model, it is possible to optionally add or combine
other analysis feature quantities as explanatory variables. Good. For example, as other analysis
feature quantities, the t-power value (where t 3 3) of the root mean square RMS (n), the zero
crossing rate of the learning sound source signal, the inclination of the frequency envelope of the
learning sound source signal, etc. can be considered .
[0131]
Thus, the mapping control information determination unit 303 shown in FIG. 7 is the mapping
control model 304 obtained by the learning processing described with reference to FIGS. 8 to 9,
for example, the root mean square as the analysis feature shown in FIG. Using the
correspondence data between the square root RMS (n) and the mapping control information α,
the mapping control information α that is optimal for the analysis feature value input from the
input analysis unit 302 is calculated and output to the mapping control information adjustment
11-04-2019
34
unit 321 .
[0132]
Next, the mapping control information adjustment unit 321 adjusts the mapping control
information according to the size of the environmental sound with respect to the mapping
control information α obtained from the mapping control information determination unit 303.
This process is similar to that of the first embodiment.
[0133]
Thereafter, as in the first embodiment described above, the mapping processing unit 322
performs mapping processing, the band limitation unit 323 performs band limitation, and an
output signal is output via the speaker 324.
[0134]
As described above, the audio signal processing apparatus 300 according to the third
embodiment adjusts the mapping control information based on the analysis result of the
environmental sound in addition to using the mapping control model based on the statistical
analysis in advance. The optimal mapping control information can be obtained according to the
size of the environmental sound, and the user can realize the optimal reproduction level control
according to the environmental sound.
[0135]
[4.
Fourth Embodiment] A block diagram of an audio signal processing device 400 according to a
fourth embodiment of the present invention is shown in FIG.
The audio signal processing device 400 shown in FIG. 11 has the following configuration. Input
unit 401, input analysis unit 402, mapping control information determination unit 403, mapping
11-04-2019
35
control model 404 (storage unit), microphone 411, band division unit 412, environment analysis
unit 413, mapping processing unit 421, band restriction unit 422, speaker 423 , Have these
configurations.
[0136]
The difference from the configuration described with reference to FIG. 7 is that the mapping
control information adjustment unit 321 shown in FIG. 7 is omitted. Furthermore, the mapping
control model 404 (storage unit) is different from the data shown in FIG. 7 in that the data is
generated in consideration of the environmental sound.
[0137]
In this embodiment, the mapping control information determination unit 403 is configured to
generate mapping control information to be applied in the mapping processing unit 221. In the
audio signal processing apparatus 400 shown in FIG. 11, the input signal input from the input
unit 401 is supplied to the input analysis unit 402, and its characteristics are analyzed. Next, as
in the first embodiment of the present invention, the sound collection signal input via the
microphone 411 is divided into bands by the band division unit 412 and analyzed by the
environment analysis unit 413. The input sound feature amount from the input analysis unit 402
and the environmental sound feature amount from the environment analysis unit 413 are
supplied to the mapping control information determination unit 403. These processes are similar
to the processes described in the first to third embodiments.
[0138]
Next, in the mapping control information determination unit 403, mapping control information is
obtained from the analysis feature amount using the mapping control model 404 generated by
the learning form in consideration of the environmental sound, and supplied to the mapping
processing unit 421.
[0139]
The mapping control model 404 is generated, for example, in the learning device 500 shown in
FIG.
11-04-2019
36
The learning device 500 illustrated in FIG. 12 includes an input unit 501, a mapping control
information adding unit 502, a mapping processing unit 503, a band limiting unit 504, a speaker
505, an input analysis unit 506, a mapping control model learning unit 507, a recording unit
508, and a microphone. A band separation unit 512, an environment analysis unit 513, and an
environmental sound speaker 531 are provided. The environmental sound speaker 531 may be a
speaker of an external device. In the learning device 500, a learning sound source signal used for
learning the mapping control model is supplied to the mapping control information adding unit
502, the input analysis unit 506, and the mapping processing unit 503. In addition, a learning
environmental sound signal is input to the microphone 511 via the environmental sound speaker
531.
[0140]
The input unit 501 includes, for example, a button operated by the user, and supplies a signal
corresponding to the user's operation to the mapping control information attachment unit 502.
The mapping control information adding unit 502 adds mapping control information to each
sample of the supplied learning sound source signal according to the signal from the input unit
501, and supplies the mapping control information to the mapping processing unit 503 or the
mapping control model learning unit 507.
[0141]
Mapping processing section 503 performs mapping processing on the supplied learning sound
source signal using the mapping control information from mapping control information adding
section 502, and supplies the obtained learning output signal to band limiting section 504. Do.
The band limiting unit 504 performs band limiting processing such as low band cutting, for
example, and supplies a processing signal to the speaker 505. The speaker 505 reproduces the
sound based on the learning output signal generated by the mapping processing unit 503.
[0142]
The input analysis unit 506 analyzes the characteristics of the supplied learning sound source
signal, and supplies an analysis feature indicating the analysis result to the mapping control
11-04-2019
37
model learning unit 507. In addition, the sound collection signal including the environmental
sound input through the microphone 511 and the output signal of the speaker 505 is separated
into a low band signal and a high band signal constituted by the environmental sound in the
band division unit 512, and the environment analysis unit 513 Generates a feature of the
environmental sound, for example, RMS (k). The processes of the microphone 511 to the
environment analysis unit 513 are the same as the processes performed by the microphone to
the environment analysis unit of the first embodiment.
[0143]
The mapping control model learning unit 357 receives the analysis feature corresponding to the
reproduction target learning sound signal from the input analysis unit 356, the environment
sound feature corresponding to the learning environment sound from the environment analysis
unit 513, and the mapping control information giving unit 502. The mapping control model is
obtained by statistical learning using the mapping control information of (1), and is supplied to
the recording unit 508.
[0144]
The recording unit 508 records the mapping control model supplied from the mapping control
model learning unit 507.
The mapping control model recorded in the recording unit 508 in this manner is recorded as the
mapping control model 404 in the storage unit of the audio signal processing apparatus 400
shown in FIG.
[0145]
Note that the learning device 500 shown in FIG. 12 can be configured inside the audio signal
processing device 400 shown in FIG. 11, and can also be configured as an external device. When
the learning device 500 shown in FIG. 12 is configured inside the audio signal processing device
400 shown in FIG. 11, the components of the learning device shown in FIG. 12 are common to
the components of the audio signal processing device 400 shown in FIG. Regarding the
components, the components of the audio signal processing device 400 can be applied as the
components of the learning device.
11-04-2019
38
[0146]
Next, the learning process by the learning device 500 shown in FIG. 12 will be described with
reference to the flowchart shown in FIG. As shown in step S01 of the flowchart shown in FIG. 13,
at the start of the learning process, for example, the environmental sound is reproduced from the
environmental sound speaker 531 shown in FIG. 12 in the viewing room, and the mapping
control information is input or adjusted under that environment. Accept
[0147]
The process of steps S11 to S17 is the same process as the process of steps S11 to S17 shown in
FIG. 9 described above with reference to the flowchart of FIG. By these processes, the input
sound feature quantity is obtained by the analysis process of the characteristics of the learning
sound source signal under one environmental sound reproduced in step S01. Also, the sound
pickup signal under the environment being reproduced is divided into bands, and the
characteristics of the divided signals are analyzed to obtain an environmental sound feature
value. This is repeated in the same environment until a sufficient number of mapping control
information is obtained.
[0148]
Then, in step S21, after a sufficient number of mapping control information is obtained, the next
environmental sound is reproduced, and a sufficient number of mapping control information is
similarly collected under the environment. Do this with a sufficient number of environmental
sounds. For example, m different learning environment sounds SRS1 to SRSm are prepared in
advance, and a sufficient number of mapping control information is collected under the
environment of these m different learning environment sounds SRS1 to SRSm. After reproducing
a sufficient number of environmental sounds, the mapping control model is learned in step S22.
[0149]
In the learning apparatus 350 in the third embodiment described above with reference to FIG. 8,
only the input sound feature amount of the learning sound source corresponding to the
reproduction target sound input from the input analysis unit 356 is used as an explanatory
variable. The learning device 500 shown in FIG. 12 explains both the input sound feature amount
11-04-2019
39
of the learning sound source corresponding to the reproduction target sound and the
environment sound feature amount from the environment analysis unit 513 analyzed
corresponding to the learning environment sound. Find the mapping control model as
[0150]
The mapping control model calculated in the present embodiment is the correspondence data
between the root mean square RMS (n) as the analysis feature of the reproduction target signal
described above with reference to FIG. 10, and the mapping control information α. The
correspondence data is further configured by a plurality of data set for each of the environmental
sounds (the above-mentioned learning environmental sounds SRS1 to SRSm).
Alternatively, the root mean square RMS (n) as the analysis feature of the reproduction target
signal, the root mean square RMS (k) as the analysis feature of the environmental sound, and the
mapping control information α It may be set as dimensional data. In the present embodiment, a
mapping control model is generated which makes it possible to obtain the optimum mapping
control information α from the analysis feature of the reproduction target signal and the
analysis feature of the environmental sound.
[0151]
In addition, although the example which set the speaker which outputs an environmental sound
as a monaural speaker was demonstrated in the learning apparatus shown in FIG. 12, an
environmental sound may be reproduced with a speaker of 2 or more channels. Alternatively, the
mapping control information may be input or adjusted under the actual environment.
[0152]
As described above, the mapping control information determination unit 403 illustrated in FIG.
11 uses the mapping control model 404 obtained by the learning process described with
reference to FIGS. The optimal mapping control information α corresponding to the amount and
the environmental sound feature value input from the environment-friendly seat unit 513 is
calculated and output to the mapping processing unit 421.
[0153]
11-04-2019
40
Next, the mapping processing unit 421 performs the same mapping processing as in the second
embodiment described above, and outputs the mapping processing result to the band limiting
unit 422.
The band limiting unit 422 performs band limiting similar to that of the first embodiment
described above, and outputs an output signal through the speaker 423.
[0154]
As described above, the audio signal processing apparatus 400 according to the present
embodiment shown in FIG. 11 has a configuration in which a mapping control model based on
statistical processing in which learning processing in advance, that is, learning data is applied, is
applied. The mapping control model in this embodiment uses both the analysis result of the input
signal which is the reproduction target signal and the analysis result of the environmental sound
as an explanatory variable, and the optimal mapping control information according to the size of
the environmental sound Thus, it is possible to realize the optimum reproduction level control
according to the environment for the user.
[0155]
[5. 5. Regarding the Fifth Embodiment Next, the fifth embodiment of the audio signal
processing device of the present disclosure will be described with reference to FIG. In the audio
signal processing device 600 shown in FIG. 14, the input signal to be reproduced is constituted
by a plurality of signals of the left channel and the right channel. As described above, when the
number of channels of the audio signal is two or more, the volume balance changes when the
independent amplitude conversion is performed for each channel. Therefore, it is desirable to
perform the same amplitude conversion for all the channels.
[0156]
The audio signal processing apparatus 600 shown in FIG. 14 has an input section 601 for the left
channel input signal, an input section 602 for the right channel input signal, and an input
11-04-2019
41
analysis section 603 for analyzing the left and right channel input signals. Furthermore, a
mapping control information determination unit 604 that applies mapping control model 605
based on the input sound feature amount from input analysis unit 603 to determine mapping
control information, and a storage unit that stores mapping control model 605 are included. The
mapping control model is data similar to the mapping control model 404 shown in FIG. 11 used
in the fourth embodiment described above.
[0157]
Further, audio signal processing apparatus 600 shown in FIG. 14 has the following configuration.
A microphone 611 for acquiring an environmental sound, a band division unit 612 which
performs band division by inputting a collected sound signal from the microphone 611, An
environment for acquiring a feature amount of a low band signal including an environmental
sound generated by the band division unit 612 The analysis unit 613 has these configurations.
These configurations are the same as those of the first embodiment described above.
[0158]
Further, audio signal processing apparatus 600 shown in FIG. 14 has the following configuration.
Mapping processing unit 621 that performs mapping processing of left channel input signal,
band limiting unit 622 that performs band limiting processing on mapping result of left channel
input signal, speaker 623 that outputs band limitation result of left channel input signal, right
channel input A mapping processing unit 631 that performs signal mapping processing, a band
limiting unit 632 that performs band limiting processing on the mapping processing result of the
right channel input signal, and a speaker 633 that outputs the band limiting result of the right
channel input signal have these configurations.
[0159]
The characteristics of the reproduction target input signals of the left and right channels input
from the input units 601 and 602 are analyzed by the input analysis unit 603, and input sound
feature amounts common to the left and right channels are obtained. The band division unit 612
performs band division on the signal input from the microphone 611, and the environment
analysis unit 613 analyzes the characteristics to obtain an environmental sound feature value.
11-04-2019
42
[0160]
The input sound feature amount generated by the input analysis unit 603 and the environmental
sound feature amount generated by the environment analysis unit 613 are supplied to the
mapping control information determination unit 604. The mapping control information
determination unit 604 determines mapping control information by applying the same mapping
control model 605 as that of the fourth embodiment described above with reference to FIG. This
mapping control information is the same mapping control information for the left and right
channels.
[0161]
The mapping control information is output to a mapping processing unit 621 that performs
mapping processing for the left channel input signal, a mapping processing unit 631 that
performs mapping processing for the right channel input signal, and these two mapping
processing units. Do the processing. Thereafter, the band limiting sections 622 and 632 perform
band limitation on the signals of the respective channels subjected to the mapping process, and
output an output signal through the speakers 623 and 633. Although the configuration shown in
FIG. 14 is an example in which the input signal has two channels, in the case of three or more
input signals, an input unit for each channel, a mapping processing unit, a band limiting unit, and
a speaker are provided. Just do it.
[0162]
As described above, in the case where there are a plurality of input signals, common mapping
control information is generated, and the common mapping control information is applied to
perform the same amplitude conversion in all channels. With such processing, it is possible to
realize an audio signal processing method and apparatus capable of enhancing the reproduction
level of the audio signal without changing the volume balance between the channels.
[0163]
[6. Sixth Embodiment] Next, the configuration and processing of an audio signal processing
11-04-2019
43
device 700 according to a sixth embodiment of the present disclosure will be described with
reference to FIG. The audio signal processing apparatus 700 shown in FIG. 15 inputs a
reproduction target input signal input through the input unit 701 to the band division filter 702,
separates the input signal into a high band signal and a low band signal, and performs
processing. It has composition. The other configuration is the same as that of the fourth
embodiment described above with reference to FIG.
[0164]
Voice and music have different characteristics depending on the frequency band. Therefore, by
performing analysis suitable for each frequency band, analysis feature quantities more suitable
for processing and hearing can be obtained.
[0165]
In the audio signal processing device 700 shown in FIG. 15, the reproduction target input signal
input from the input unit 701 is divided into a low-pass signal and a high-pass signal bandlimited at around 300 Hz by the band division filter 702, and the input analysis unit 703
Supplied to Then, the input analysis unit 703 performs different analysis on each of the low band
signal and the high band signal, and obtains a common analysis feature amount from the results.
[0166]
The input analysis unit 703 performs different analysis on the low band signal and the high band
signal according to, for example, (Equation 7) to (Equation 9) shown below, and obtains a
common analysis feature value from the results. (Expression 7) is a calculation expression of root
mean square RMS_l (n) as a feature amount corresponding to the n-th sample of the low-pass
signal. (Expression 8) is a calculation expression of the root mean square RMS_h (n) as the
feature amount corresponding to the n-th sample of the high frequency band signal. The root
mean square RMS_l (n) and RMS_h (n) of N and M samples centered on the nth sample of each
band split signal are calculated, respectively.
[0167]
... (Equation 7)
11-04-2019
44
[0168]
... (Equation 8)
[0169]
In the above (Equation 7) and (Equation 8), x_l and x_h are a low-pass signal and a high-pass
signal obtained by the band division filter from the input signal x to be reproduced, It is assumed
that the signal is normalized with 0.
[0170]
The input analysis unit 703 calculates the feature amount RMS_l (n) of the low band signal
calculated according to the above (Equation 7), the feature amount RMS_h (n) of the high band
signal calculated according to the above (Equation 8) Weighted addition is performed using
weights a and b defined in advance according to (Expression 9) shown below, to obtain an
analysis feature RMS '(n) common to the low band signal and the high band signal.
The weights a and b are, for example, = a = b = 0.5.
[0171]
... (equation 9)
[0172]
Let RMS ′ (n) determined according to the above (Equation 9) be an analysis feature of the
input signal to be reproduced.
The RMS ′ (n) obtained here is supplied to the mapping control information determination unit
704 as an input sound feature amount for the n-th reproduction target input signal.
[0173]
11-04-2019
45
In the above (Equation 9), although the weights a and b are equalized, it may be set so as to apply
a large weight to a signal of a specific band.
Further, in the above processing example, the processing is performed by dividing the frequency
band of the input signal into two at 300 Hz, but within the band limitation by the band limiting
unit 722, another frequency such as 200 Hz, 400 Hz, 1 kHz, or 3.4 kHz The analysis feature
value may be obtained from the signal divided in the above, or the signal divided into three or
more band signals.
Furthermore, separate analysis may be performed on the input signal and the band division
signal, and the results may be combined to form an analysis feature. It is good to use what is
suitable for the effect of processing and mapping control as an analysis feature-value. Further,
although filters are used for band division here, signals of each band may be generated on the
frequency axis.
[0174]
The input analysis unit 703 supplies the analysis feature value obtained in this manner to the
mapping control information determination unit 704.
[0175]
Hereinafter, the mapping control information is obtained by applying the mapping control model
705 similar to that of the fourth embodiment described above with reference to FIG.
The mapping control information is output to the mapping processing unit 721 and mapping
processing is performed. Thereafter, the band limiting unit 722 performs band limitation on the
signal subjected to the mapping process, and outputs an output signal through the speaker 723.
[0176]
In this embodiment, the feature quantities corresponding to the respective bands of the input
11-04-2019
46
signal are individually acquired, and the weighted addition result of the feature quantities is
calculated as the feature quantity for the input signal. As described above, analysis suitable for
processing and hearing can be obtained by performing analysis suitable for each frequency band.
[0177]
[7. Seventh Embodiment] Next, a configuration and processing of an audio signal processing
device 800 according to a seventh embodiment of the present disclosure will be described with
reference to FIG. The audio signal processing device 800 shown in FIG. 16 is configured to
perform linear gain adjustment according to the size of the environmental sound after
performing mapping processing according to the characteristics of the input signal.
[0178]
A block diagram of an audio signal processing device 800 according to a seventh embodiment of
the present disclosure is shown in FIG. The audio signal processing device 800 shown in FIG. 16
has the following configuration. Input unit 801, input analysis mapping control information
determination unit 802, microphone 811, band division unit 812, environment analysis unit 813,
gain adjustment amount determination unit 814, mapping processing unit 821, gain adjustment
unit 822, band limitation unit 823, speaker 824 , Have these configurations.
[0179]
The difference from the second embodiment described with reference to FIG. 6 is that a gain
adjustment amount determination unit 814 and a gain adjustment unit 822 are added. The other
configurations and processes are similar to those of the second embodiment.
[0180]
An input analysis / mapping control information determination unit 802 calculates mapping
control information from the reproduction target input signal input through the input unit 801.
The mapping processing unit 821 performs mapping processing based on the mapping control
11-04-2019
47
information, and supplies this to the gain adjusting unit 822.
[0181]
The processing of the microphone 811 to the band division unit 812 to the environment analysis
unit 813 is the same processing as that of the first embodiment described above. The analysis
feature amount of the environmental sound is determined in the environment analysis unit 813
and supplied to the gain adjustment amount determination unit 814.
[0182]
The gain adjustment amount determination unit 814 determines the gain adjustment amount
from the analysis feature of the environmental sound obtained from the environment analysis
unit 813 using a table or a function or a statistical model based on statistical analysis in advance.
[0183]
The gain adjustment amount determination unit 814 obtains the gain adjustment amount by the
following process, for example.
An environmental sound feature that is an analysis feature of an environmental sound obtained
from the environmental analysis unit 813, that is, a root mean square RMS (k sample) with K
samples centered on the k-th sample of the low band signal containing only the environmental
sound Assuming that x), the gain adjustment amount y is obtained using a linear function of
(Expression 10) shown below.
[0184]
... (Equation 10)
[0185]
Although the root mean square RMS (k) is used as the environmental sound feature amount here,
11-04-2019
48
other feature amounts and combinations thereof may be used as in the above-described
embodiments.
Also, although the linear function shown in (Equation 10) was used to calculate the gain
adjustment amount y, using a non-linear function or table representing the relationship between
the environmental sound feature amount and the gain adjustment amount or a linear regression
model or nonlinear regression model Also good.
[0186]
The gain adjustment amount determination unit 814 thus calculates the gain adjustment amount
y according to the feature amount of the environmental sound, and outputs the gain adjustment
amount y to the gain adjustment unit 822. The gain adjustment unit 822 linearly performs gain
adjustment on the mapping processing signal input from the mapping processing unit 821 based
on the gain adjustment amount input from the gain adjustment amount determination unit 814.
[0187]
Finally, the band limiting unit 823 applies a band limiting filter to the gain-adjusted mapping
processing signal to generate a band limited output signal, and outputs the output signal through
the speaker 824. In the configuration of the present embodiment, it is possible to obtain an
output signal whose gain has been adjusted according to the size of the environmental sound.
[0188]
[8. Eighth Embodiment] Next, an eighth embodiment of the present disclosure will be
described with reference to FIG. The audio signal processing device 900 shown in FIG. 17 is the
same as the audio signal processing device 400 according to the fourth embodiment described
above with reference to FIG. 11 as in the seventh embodiment described with reference to FIG. It
has a configuration in which a gain adjustment amount determination unit 914 and a gain
adjustment unit 922 are added.
[0189]
11-04-2019
49
Audio signal processing apparatus 900 shown in FIG. 17 has the following configuration. Input
unit 901, input analysis unit 902, mapping control information determination unit 903, mapping
control model 904 (storage unit), microphone 911, band division unit 912, environment analysis
unit 913, gain adjustment amount determination unit 914, mapping processing unit 921, A gain
adjusting unit 922, a band limiting unit 923, and a speaker 924 have these configurations.
[0190]
An input analysis unit 902 analyzes the characteristics of the reproduction target input signal
input from the input unit 901, and obtains an input sound feature amount. The band division unit
912 divides the signal input from the microphone 911 into bands, and the environment analysis
unit 913 analyzes the characteristics of the signal to obtain an environmental sound feature
value.
[0191]
The input sound feature amount generated by the input analysis unit 902 and the environmental
sound feature amount generated by the environment analysis unit 913 are supplied to the
mapping control information determination unit 903. The mapping control information
determination unit 903 obtains mapping control information by applying the same mapping
control model 904 as that of the fourth embodiment described above with reference to FIG. The
mapping control information is output to the mapping processing unit 921 and mapping
processing is performed.
[0192]
The gain adjustment amount determination unit 914 calculates the gain adjustment amount y
according to the feature amount of the environmental sound and outputs it to the gain
adjustment unit 922, as in the seventh embodiment described above with reference to FIG. . The
gain adjustment unit 922 linearly performs gain adjustment on the mapping processing signal
input from the mapping processing unit 921 based on the gain adjustment amount input from
the gain adjustment amount determination unit 914.
11-04-2019
50
[0193]
Finally, the band limiting unit 923 applies a band limiting filter to the gain-adjusted mapping
processing signal to generate a band-limited output signal, and outputs the output signal through
the speaker 924. In the configuration of the present embodiment, it is possible to obtain an
output signal whose gain has been adjusted according to the size of the environmental sound.
[0194]
[9. Summary of Configuration of Present Disclosure] The embodiments of the present
disclosure have been described in detail with reference to the specific embodiments. However, it
is obvious that those skilled in the art can make modifications and substitutions of the
embodiment without departing from the scope of the present disclosure. That is, the present
invention has been disclosed in the form of exemplification, and should not be construed as
limiting. In order to determine the scope of the present disclosure, the claims should be referred
to.
[0195]
The technology disclosed in the present specification can have the following configurations. (1)
Analysis of the characteristics of the input signal, an input analysis unit that generates an input
sound feature amount, an environment analysis unit that analyzes the characteristics of an
environmental sound, and generates an environmental sound feature amount; A mapping control
information generation unit that generates mapping control information as control information
of amplitude conversion processing on the input signal by applying an environmental sound
feature amount; and the linear or non-linear mapping function determined by the mapping
control information. An audio signal processing apparatus comprising: a mapping processing unit
that amplitude-converts an input signal and generates an output signal.
[0196]
(2) The mapping control information generation unit generates a mapping control information
determination unit that generates preliminary mapping control information by applying the input
11-04-2019
51
sound feature amount, and the environmental sound with respect to the preliminary mapping
control information. The audio signal processing device according to (1), further including: a
mapping control information adjustment unit that generates the mapping control information to
be output to the mapping processing unit by adjustment processing to which a feature amount is
applied. (3) The input analysis unit calculates a root mean square calculated using a plurality of
continuous samples defined in advance as the input sound feature amount, and the environment
analysis unit calculates an environmental sound signal as the environmental sound feature
amount Calculating the root mean square calculated using the plurality of continuous samples,
and the mapping control information generation unit is configured to calculate the root mean
square of the input signal as the input sound feature amount and the environmental sound as the
environmental sound feature amount. The audio signal processing device according to (1) or (2),
wherein the mapping control information is generated using a root mean square of a signal.
[0197]
(4) The input sound feature quantity and the environmental sound feature quantity are root
mean square of the feature quantity calculation target signal, or a logarithm of root mean square,
or a root mean square, or a root mean square of logarithm The audio signal processing device
according to any one of (1) to (3), which is a zero crossing rate of a signal, a slope of a frequency
envelope, or a weighted addition result thereof. (5) The environment analysis unit performs
feature analysis of a band signal having a high occupancy ratio of environmental sound divided
by band division processing from a collected sound signal acquired via a microphone to calculate
the environmental sound feature value The audio signal processing device according to any one
of (1) to (4). (6) The audio signal processing apparatus includes a band limiting unit that
performs band limiting processing of the signal subjected to the mapping processing in the
mapping processing unit, and the speaker is used as the signal after band limiting in the band
limiting unit. The audio signal processing device according to any one of the above (1) to (5),
which outputs the signal via
[0198]
(7) The mapping control information generation unit generates the mapping control information
by applying a mapping control model generated by statistical analysis processing to which a
learning signal including an input signal and an environmental sound signal is applied. (6) The
audio | voice signal processing apparatus in any one. (8) The audio signal processing device
according to (7), wherein the mapping control model is data in which mapping control
information is associated with various input signals and an environmental sound signal. (9) The
11-04-2019
52
input signal is composed of a plurality of input signals of a plurality of channels, and the
mapping processing unit is configured to execute individual mapping processing for each input
signal. The audio signal processing device as described. (10) The audio signal processing
apparatus further includes a gain adjustment unit that performs gain adjustment on the mapping
processing signal generated by the mapping processing unit according to the environmental
sound feature value generated by the environment analysis unit. The audio signal processing
device according to any one of (1) to (9).
[0199]
Furthermore, the method of the process performed in the above-mentioned apparatus etc. and
the program which performs a process are also contained in the structure of this indication.
[0200]
In addition, the series of processes described in the specification can be performed by hardware,
software, or a combined configuration of both.
When software processing is to be performed, the program recording the processing sequence is
installed in memory in a computer built into dedicated hardware and executed, or the program is
executed on a general-purpose computer capable of executing various processing. It is possible
to install and run. For example, the program can be recorded in advance on a recording medium.
The program can be installed from a recording medium to a computer, or can be installed in a
recording medium such as a built-in hard disk by receiving a program via a network such as a
LAN (Local Area Network) or the Internet.
[0201]
The various processes described in the specification may not only be performed in chronological
order according to the description, but also may be performed in parallel or individually
depending on the processing capability of the apparatus executing the process or the necessity.
Further, in the present specification, a system is a logical set configuration of a plurality of
devices, and the devices of each configuration are not limited to those in the same housing.
[0202]
11-04-2019
53
As described above, according to the configuration of an embodiment of the present disclosure,
the optimum mapping control can be performed when the environmental sound is large or small,
and the user's dissatisfaction that the volume is insufficient or distortion may be reduced. It is
possible to automatically control the reproduction level of the audio signal optimally for the user
even under various circumstances.
[0203]
Specifically, for example, the characteristic of the input signal is analyzed to generate the input
sound feature, the characteristic of the environmental sound is analyzed, the environmental
sound feature is generated, and the generated input sound feature and the environmental sound
feature are applied. Then, mapping control information is generated as control information for
amplitude conversion processing on the input signal.
Furthermore, the input signal is amplitude converted based on a linear or non-linear mapping
function determined by the mapping control information to generate an output signal. The
mapping control information is generated with reference to, for example, a model generated in
consideration of an input signal and an environmental sound. With these configurations, it
becomes possible to automatically control the level of the audio signal under various
environments by optimum mapping control according to the environmental sound.
[0204]
DESCRIPTION OF SYMBOLS 100 audio | voice signal processing apparatus 101 input part 102
input analysis / mapping control-information determination part 111 microphone 112 band
division part 113 environment analysis part 114 mapping control information adjustment part
121 mapping processing part 122 band-limit part 123 speaker 200 audio | voice signal
processing apparatus 201 input Part 202 Input analysis / mapping control information
determination part 211 Microphone 212 Band division part 213 Environment analysis part 221
Mapping processing part 222 Band limitation part 223 Speaker 300 Audio signal processing
device 301 Input part 302 Input analysis part 303 Mapping control information determination
part 311 Microphone 312 band division unit 313 environment analysis unit 321 mapping
control information adjustment unit 322 mapping processing unit 323 band limitation unit 324
speaker 350 learning device 351 input unit 352 mapping control Information giving unit 353
Mapping processing unit 354 Band-limiting unit 355 Speaker 356 Input analysis unit 357
Mapping control model learning unit 358 Recording unit 400 Audio signal processing device
401 Input unit 402 Input analysis unit 403 Mapping control information determination unit 404
Mapping control model 411 Microphone 412 band division unit 413 environment analysis unit
11-04-2019
54
421 mapping processing unit 422 band limitation unit 423 speaker 500 learning device 501
input unit 502 mapping control information addition unit 503 mapping processing unit 504
band limitation unit 505 speaker 506 input analysis unit 507 mapping control model learning
unit 508 recording unit 511 microphone 512 band division unit 513 environment analysis unit
531 environmental sound speaker 600 audio signal processing device 601 input unit 602 input
unit 603 input analysis unit 60 4 mapping control information determination unit 605 mapping
control model 611 microphone 612 band division unit 613 environment analysis unit 621, 631
mapping processing unit 622, 632 band limitation unit 623, 633 speaker 700 audio signal
processing device 701 input unit 702 band division filter 703 input Analysis unit 704 Mapping
control information determination unit 711 Microphone 712 Band division unit 713
Environment analysis unit 721 Mapping processing unit 722 Band limitation unit 723 Speaker
800 Audio signal processing device 801 Input unit 802 Input analysis / mapping control
information determination unit 811 Microphone 812 Band division Part 813 Environment
analysis part 814 Gain adjustment amount determination part 821 Mapping processing part 822
Gain adjustment part 823 Band limitation part 824 Speaker 900 Audio signal processing device
901 Input part 902 Power analysis unit 903 mapping control information determination unit
911 microphone 912 band division unit 913 environment analysis unit 914 gain adjustment
amount determination unit 921 mapping unit 922 gain adjuster 923 band restriction unit 924
speaker
11-04-2019
55
Документ
Категория
Без категории
Просмотров
0
Размер файла
76 Кб
Теги
description, jp2013102411
1/--страниц
Пожаловаться на содержимое документа