JPH0319513

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 JPH0319513
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an
acoustic transfer characteristic control apparatus for controlling the acoustic transfer
characteristic of an audio reproduction output signal, and more particularly to an apparatus for
controlling the acoustic transfer characteristic of sound field space such as graphic equalizer and
surround processor. The present invention relates to a suitable acoustic transfer characteristic
control device. 2. Description of the Related Art Conventionally, various control devices have
been developed to make the reproduction signal from an audio reproduction device more
realistic. A typical example is a graphic equalizer that divides the frequency characteristics into
multiple frequency bands and can be corrected independently for each band, and also to increase
the phase difference of the signal transfer time in order to increase the sense of reality The
surround processor etc. which were used are known. [The problem to be solved by the invention]
As shown by the Fletcher-Manson curve in FIG. 14, the human ear has an auditory sensitivity
characteristic that the low range and high range become difficult to hear when the sound
pressure level is low. . However, the above-mentioned conventional control device only performs
correction with the correction curve set at the time of adjustment regardless of the level of the
sound pressure level. On the other hand, a loudness control circuit is known which boosts the
low-high range at the time of small volume to correct the auditory sensitivity characteristic
accompanying the high and low of the sound pressure level, but the correction curve obtained by
this is uniquely determined by the circuit constant. It was fixed and I could not make fine
adjustments. In addition, when the above-mentioned graphic equalizer, surround processor, etc.
are provided with the same function as the loudness control circuit and automation of optimum
adjustment is considered to correct human auditory sensitivity characteristics, any situation is set
in any device, It is necessary to store characteristic patterns corresponding thereto, and there is a
problem that a memory with a large storage capacity is required. Therefore, an object of the
08-05-2019
1
present invention is to provide an acoustic transfer characteristic control device capable of being
adapted by a simple operation to the listener's auditory sensitivity characteristic that changes
with the volume change of reproduced sound. [Means for Solving the Problems] In order to solve
the above problems, the present invention is characterized by an input unit for inputting data of
a determinant which determines the acoustic transfer characteristic of an audio reproduction
output signal, and a fuzzy based on the determinant data. The arithmetic unit for calculating the
adjustment amount of the acoustic transfer characteristic of the audio reproduction output signal
by inference, and the operation unit for operating the acoustic transfer characteristic of the audio
reproduction output signal based on the calculated adjustment amount The acoustic transfer
characteristic control device is characterized in that the determining factor is a factor that
corrects auditory characteristics that differ depending on the volume level (hereinafter referred
to as a volume factor.
Was configured to include. [Operation] According to the present invention, by inputting the
determination factor data for determining the acoustic transfer characteristic from the input
section and further inputting the volume factor data, the operation section is fuzzy based on the
input determination factor data. Perform an operation. As determinants other than the said
volume factor data, a frequency characteristic factor, a reverberation characteristic factor, etc.
are mentioned, for example. By performing fuzzy inference in the operation unit, an adjustment
amount is calculated to be a characteristic more suitable for human sense based on a rule
including ambiguity, and the operation unit outputs an audio reproduction output signal based
on the calculated value. Do the operation of As a result, it is possible to obtain a sound
transmission characteristic that conforms to the listener's auditory sensitivity characteristic that
depends on the volume change with a simple operation, and to obtain a sound transmission
characteristic that more conforms to the listener's sense. In addition, it is not necessary to
prepare in memory or the like all characteristic patterns assumed to perform fuzzy inference
based on given determinant data, and since the calculation is simple, adjustment is possible in
real time, and the obtained characteristics It is possible to make a very simple configuration with
respect to the diversity of Embodiments of the present invention will be described in detail with
reference to FIGS. 1 to 13. First Embodiment FIG. 1 shows a schematic block diagram when the
present invention is applied to a graphic equalizer. First, an overview will be described. Fuzzy
graphic equalizer (hereinafter referred to as fuzzy equalizer. 1) a display / operation unit 2 for
displaying data and inputting data, a fuzzy operation unit 3 for operating, by fuzzy inference, a
change amount of a determination factor which determines an acoustic transfer characteristic
based on the input data; And a determining factor changing unit 6 for changing the determining
factor of the audio input signal 5 according to the calculation result of the fuzzy calculating unit
3. The operator inputs the relative change amount of the decision factor other than the volume
factor to the display operation unit 2 and the fuzzy operation unit automatically takes in the
volume change from the position of the mover of the volume or the like. The fuzzy operation unit
3 performs fuzzy inference based on the determination factor change amount, the volume
08-05-2019
2
change amount, and the membership function stored in the storage unit 4 to obtain an absolute
change amount of each determination factor. Thereafter, the decision factor changing unit 6
changes the decision factor of the audio input signal 5 on the basis of the calculation result of the
fuzzy calculation unit 3 and outputs it as the audio output signal 7. Next, I will describe the
details.
The auditory sensitivity characteristic corresponding to the sound pressure level of human
beings is known to make it difficult to hear the high bass region as the sound pressure level
decreases (see FIG. 14). In the present embodiment, attention is focused on the frequency
characteristics and the volume, and the determination factors related to these are controlled.
When fuzzy inference is performed for the volume factor, the center frequency of the equalizer
and the current volume are used as the input condition, and the fuzzy operation unit outputs the
relative operation amount of the boost amount of the center frequency as the output operation
amount. FIG. 2 shows a membership function for performing fuzzy inference based on the
volume factor. The membership function which shows a center frequency in FIG. 2 (a) is shown.
The number of divisions of the audio frequency band of the equalizer is 7, the number of
operation steps for each frequency band is ± 5, and the input operation amount of each factor is
± 10. The vertical axis is the grade (unknown number) representing the degree, and the
horizontal axis is the frequency (H2). L (Low) M (Middle) H (fligh) is a membership function
representing low frequency, medium frequency, and high frequency, respectively, and the degree
becomes higher as the grade is closer to 1. For example, the frequency of 400 Hz represents a
slightly lower middle frequency because the grade of the membership function L is about 0.3 and
the grade of the membership function M is about 0.7. FIG. 2 (b) shows a membership function
representing the current volume. The vertical axis is the grade (unknown number) representing
the degree, and the horizontal axis is the volume. S (Small), N (Nora + al), and B (Big) are
membership functions indicating a small volume, a medium volume, and a large volume,
respectively. For example, the volume at point p indicates that the current volume is slightly low
and medium because the grade of membership function S is approximately O, 3 and the grade of
membership function N is approximately 0.7. There is. FIG. 2 (C) shows a membership function
representing an output operation amount. The vertical axis represents grade (unknown number)
representing the degree, and the horizontal axis represents the amount of output operation. FIG.
3 is an explanatory view of an example of the production rule of the volume factor. The
respective symbols are the same as the symbols in FIG. 2, so the detailed description will be
omitted. For example, to explain the case of rule 3, the center frequency is high (H), the current
volume is small volume (S), and the output operation amount in this case is about 3 (fuzzy
amount) Indicates that.
On the other hand, it has been reported that human auditory characteristics can be expressed by
at least three determinants independent of each other. In this embodiment, three factors are
08-05-2019
3
controlled: an aesthetic factor that affects the aesthetic factor of the reproduced sound as a
determinant other than the volume factor, a powerful factor that affects the power of the
reproduced sound, and a depth factor that affects the depth of the reproduced sound. Do. Each
determinant depends on frequency characteristics, and its rough definition is as follows.
Aesthetic factor: In the increasing direction of factor amount, it shows a gentle tone that cuts the
high and low range, and in the decreasing direction, it becomes a bright tone with Pousto of the
high and low range. Force factor: In the increasing direction of factor setting, the whole area is
boosted by emphasizing Nakashiro in particular, and in the decreasing direction, Nakashiro is cut
the other way. Depth factor: Takagi is cut and boost is boosted in the increasing direction of
factor amount, and low section is cut and boost is boosted in the decreasing direction. In this
case, as the input condition used for the fuzzy inference, three of the center frequency of the
equalizer, the Phost amount of the equalizer, and the input operation amount of the factor whose
operation is desired among the three factors are used. The relative operation amount is output as
an output operation amount. FIG. 4 shows a membership function for performing fuzzy inference
based on the above three factors. FIG. 4 (a) shows a membership function representing the center
frequency. The vertical axis is grade (unknown number) representing the degree, and the
horizontal axis is frequency (Hz). L (Low), M (Middle) and H (High) are membership functions
representing low frequency, medium frequency, and high frequency, respectively, and the grade
becomes higher as the grade is closer to 1. For example, the frequency of 400 Hz represents a
slightly lower middle frequency because the grade of the membership function L is about 0.3 and
the grade of the membership function M is about 0.7. FIG. 4 (b) shows a membership function
representing the current boost amount. The vertical axis is a grade (unknown number) indicating
the degree, and the horizontal axis is a boost amount. M (MinIls), Z (ze + o), and P (Plug) are
membership functions representing a negative boost amount, a zero, and a positive boost
amount, respectively. For example, if the current boost amount is −1, the current boost amount
is approximately negative O because the grade of membership function M is approximately 0.3
and the grade of membership function 2 is approximately 0.8. It shows that it is.
FIG. 4 (c) shows a membership function representing the amount of input operation. The vertical
axis represents the grade (unknown number) representing the degree, and the horizontal axis
represents the input operation amount. The other points are the same as in FIG. 4 (b). FIG. 4 (d)
shows a membership function representing an output operation amount. The vertical axis
represents grade (unknown number) representing the degree, and the horizontal axis represents
the amount of output operation. FIG. 5 shows an explanatory diagram of an example of
production rules of an aesthetic factor. The respective symbols are the same as the symbols in
FIG. 2 and thus the detailed description will be omitted. For example, in the case of rule N + 14,
when the center frequency is a low frequency (L), the current boost amount of that frequency is
0 (Z), and the input operation amount is negative (M), ie, the aesthetic factor is to be reduced.
This indicates that the output manipulated variable in this case is about 2 (fuzzy quantity). 6 and
7 show an example of the production rules of the force factor and the depth factor, respectively.
08-05-2019
4
The contents are the same as in the case of the production factor of the aesthetic factor in FIG. In
order to actually perform the fuzzy inference using the production rules of FIGS. 3 and 5 to 7, the
following procedure is performed. In the following, it is an inference method in the case of using
the win-mix barycentric method, and inferring a volume factor, and subsequently inferring a
determinant other than the volume factor will be described. a) Inference method of volume factor
The input condition of the center frequency and the current volume is applied to each
membership function, and the grade value of the intersection point is the degree of matching the
membership function has with the input condition. 1) The input condition of the center
frequency and the current volume is applied to each membership function, and the grade value
of the intersection point is the degree of matching the membership function has with the input
condition. 2) In the production rule to which the above two membership functions apply, the one
showing the smaller value of the two matching degrees is called the matching degree (the rule
matching degree) that the production rule has with respect to the input condition. )とする。 3)
The output quantities of all production rules are weighted by the rule conformity of the
production rules, and in the case of the same output operation quantity, the maximum value
among them is taken as the fuzzy inference result, and the center of gravity is calculated. As the
result of the inference (win-m breath I center of gravity method). b) Inference method for
determinants other than loudness factor The deduction of determinants other than loudness
factor is also performed as follows, similarly to the loudness factor.
1) Apply three input conditions of center frequency, current boost amount, and input operation
amount of factor desired to operate to each membership function, and the membership function
has the grade value of the intersection for the human power condition The degree of conformity.
2) In the production rule to which the above three membership functions apply, the one showing
the minimum value among the three fitness levels is called the fitness level (rule fitness degree)
that the production rule has with respect to the input condition. )とする。 3) The output
quantities of all production rules are weighted by the rule conformity of the production rules,
and in the case of the same output operation quantity, the maximum value among them is taken
as the fuzzy inference result, and the center of gravity is calculated. Inference results (win 1 old I
centroid method). FIG. 8 shows an example of the actual fuzzy inference for the volume factor. In
this example, the center frequency is 400 Hz and the current volume is a. In FIG. 8 (a), the
production rule shame indicates that the degree of rule conformity is not 0, and the case of O
does not. For example, to explain the case of production rule shame 1, the intersection point of
the membership function L of the center frequency with 400 Hz, ie, the goodness of fit is
approximately 0.3, and similarly the goodness of fit with the membership function M of the
current volume. Is approximately 0.4, and the rule conformity of the production rule NIll is
approximately 0.3, which is the minimum value among them. Similarly, for each production rule
whose conformity is not zero, the conformity of each rule is determined. Next, the output amount
corresponding to the production rule shame is obtained. For example, the output amount of
production rule shame 1 is +5 (see FIG. 3). The membership function of the output operation
08-05-2019
5
amount corresponding to the obtained output amount is obtained from FIG. 2 (c), and weighting
is performed with the degree of rule conformity (in practice, triangles with hatched portions in
FIG. 8 (a) are obtained. ). FIG. 8 (b) shows the fuzzy inference result of the total output operation
amount. In this case, the output operation amount is weighted for all production rules and
superimposed on each other (that is, in the same output operation amount, the maximum value
of the weighted output operation amount is the output operation amount in the output operation
amount) Is the fuzzy inference result of the total output manipulated variable. Next, the center of
gravity of the total output manipulated variable fuzzy inference result obtained in FIG. 8 (b) is
determined, and the solution is taken as the inference result. In this case, the output control
amount is approximately +0.6.
FIG. 9 shows an example of the actual fuzzy inference for other determinants. In this example,
the center frequency is 400 Hz, the current equalizer boost amount is 2, and the input factor of
the force factor is +2. In FIG. 9 (a), as in FIG. 8 (a), the production rule shame indicates that the
degree of rule conformity is not zero and not zero. For example, in the case of production rule
shame 3, the intersection point of the membership function L of the center frequency with 400
Hz, ie, the fitness is approximately 0.3, and the fitness with the membership function M of the
current Poost amount in the same manner. The degree is about 0.8, and the matching degree
with the input operation amount P is about 0.2, and the rule matching degree of the production
rule k3 is about 0.2 which is the minimum value among them. Similarly, each rule matching
degree is determined for a production rule whose matching degree is not O. Next, the output
amount corresponding to the production rule magic is obtained. For example, the output amount
of the production rule magnet 3 is 8 (see FIG. 6). The membership function of the output
operation amount corresponding to the obtained output amount is obtained from FIG. 4 (d), and
weighting is performed with the degree of rule conformity (in practice, triangles with hatched
portions in FIG. 9 (a) are obtained. ). FIG. 9 (b) shows the fuzzy inference result of the total output
manipulated variable. In this case, the output operation amount is weighted for all production
rules and superimposed on each other (that is, in the same output operation amount, the
maximum value of the weighted output operation amount is the output operation amount in the
output operation amount) Is the fuzzy inference result of the total output manipulated variable.
Next, the center of gravity of the total output manipulated variable fuzzy inference result
obtained in FIG. 9 (b) is determined, and the solution is taken as the inference result. In this case,
the output control amount is approximately +1. The operation will be described with reference to
the flowchart of FIG. A variable Band indicating the center frequency is set to 1 (step SL). This
Band takes values from 1 to 7, and each indicates the center frequency of the frequency band
divided into seven. A memory R 1 (B a n d) storing the fuzzy inference result for the volume
factor is set to O (step 82). The value obtained by adding 1 to the value of Band is taken as the
value of Band (step S 3), and if the value of B in d is greater than 7, the process proceeds to step
S 5. The process of step S4 is repeated.
08-05-2019
6
Set Band again to 1 (Step 35), and set the value of the variable Va Iue (Band), which indicates the
output manipulated variable at each center frequency, to the value obtained by subtracting R 1 (B
a n d) from the current equalizer boost amount. A value obtained by adding the king to the value
of Bind is set as the value of Band (Step S7), and it is determined whether the value of Band is
greater than 7 (Step S8). The process shifts to the process of S9, and in the case of 7 or less, the
process of steps S6 to step S8 is repeated. The Band is again set to 1 (step S9), the output
operation amount is fuzzy inferred from the B a n d s volume, and stored in the memory R 1 (B a
n d) (step S10). The value obtained by adding 1 to the value of Band is taken as the value of Band
(step S11), and it is determined whether the value of Band is greater than 7 (step S12). If it is
greater than 7, processing proceeds to step S13. If the number is 7 or less, the processing of step
SIO to step 312 is repeated. The Band is again set to 1 (Step 313), and the Band, Value (Band),
the output operation amount is fuzzy inferred from the input operation amount, and stored in the
memory R2 (B a n d) (Step S14). The value obtained by adding 1 to the value of Band is taken as
the value of Band (step S15), and it is determined whether the value of Band is greater than 7
(step S16). If it is greater than 7, processing proceeds to step S17. , 7 or less, the processing of
step 814 to step S16 is repeated. The Band is again set to 1 (step S17), and the equalizer is set to
the sum value of the memory R 1 (B a n d), the memory R 2 CB and] and the value CB and l (step
S18). The value obtained by adding 1 to the value of Band is taken as the value of Band (step
S19), and it is determined whether the value of Band is greater than 7 (step S20). If it is greater
than 7, processing proceeds to step S21. , 7 or less, the processing of step SL8 to step S20 is
repeated. It is determined whether the input amount other than the current input factor has
changed or the input factor selection switch has been pressed (step S21), and if the input amount
has changed or the input factor selection switch has been pressed, the step The processing of 85
to step 821 is repeated, and in the other cases, the processing of step 89 to step S21 is repeated.
As described above, it is possible to obtain a desired tone by sequentially making changes by
other or the same determinant based on the obtained acoustic transfer characteristics. It is also
possible to configure so as to change the acoustic transfer characteristic based on the fuzzy
operation result of all the input determinants after inputting a plurality of determinants in
advance and performing the fuzzy operation for each input determinant. . Second Embodiment
FIG. 11 shows a block diagram of a second embodiment of the present invention. The same parts
as those of the embodiment of FIG. 1 are designated by the same reference numerals and their
detailed description will be omitted. What differs from the embodiment of FIG. 1 is that the
feedback signal 8 of the audio reproduction output signal is used as a determining factor. Even if
the audio input signal 5 is changed based on the determining factor to which the graphic
equalizer is input, depending on the reproduction capability of the ambe 9 and the skier 10
connected to the subsequent stage of the graphic equalizer, desired acoustic transfer
characteristics Can not get. Therefore, by taking the audio reproduction output signal 1l as the
feedback signal 8 and performing a fuzzy operation with this as a determination factor, it is
08-05-2019
7
possible to substantially make the acoustic transfer characteristic as desired. The other points are
the same as in the embodiment of FIG. In the above, the audio reproduction output signal 11 is
taken in as the feedback signal 8. However, the reproduction frequency characteristics of the
amb 9 and the scaker 10 etc. are stored in the storage unit 4 in advance, and they are used as a
determination factor. It is also possible to configure to change the acoustic transfer
characteristics. Third Embodiment FIG. 12 shows a third embodiment of the present invention.
The same parts as those in the embodiment of FIG. 9 are designated by the same reference
numerals and their detailed description will be omitted. The difference from the embodiment of
FIG. 11 is that the sound pressure level storage unit 12 for storing the sound pressure level of
the audio reproduction output signal 11 in the fuzzy equalizer 1 and the output when the
acoustic transfer characteristic is changed by performing the fuzzy operation. This is a point
provided with the output sound pressure level adjustment unit 13 that measures the sound
pressure level and matches it with the stored output sound pressure level. Thereby, even if the
sound transfer characteristic by the fuzzy arithmetic processing changes, a constant sound
pressure level can be maintained. Fourth Embodiment FIG. 13 shows a block diagram of a fourth
embodiment in which the present invention is applied to a simple surround processor. The
sounds heard in theaters are not only direct sounds, but also include many indirect sounds that
are reflected intricately by surrounding walls and ceilings, so the reproduced sound in normal
playback systems reproduces the sense of the sound field It is difficult.
The surround processor includes a display / operation unit 2 for displaying data and inputting
data, a fuzzy operation unit 3 for calculating, by fuzzy inference, a change amount of a
determination factor for determining an acoustic transfer characteristic based on the input data,
and a fuzzy operation From the storage unit 4 for storing membership functions etc. required for
the above, the decision factor changing unit 6 for changing the decision factor, the audio input
signal (L) 19 and the audio input signal (R) 20, L-R component or t, A surround component
extractor SR for extracting + R6 minutes, a low-pass filter LPFI, LPF2 and LPF3 for preventing
unnecessary high frequency components from entering a delay circuit described later, and a
delay circuit Di for generating a delay time of about 15 to 3 Qms, D2 and a compression circuit
S1 and an expansion circuits E1 and E2 that work as noise reduction. There. The surround
processor processes the surround components extracted by the surround extractor with a lowpass filter, a compression circuit, an expansion circuit and a delay circuit, and outputs as
surround output (L) signal 17 and surround output (R) signal 18. By playing with Sukika, it
makes you feel as if there is an indirect sound, creating a sense of sound from the theater.
However, since the reverberation characteristics are very different depending on the conditions
such as the room, the desired surround feeling can not be obtained only by the preset surround
information, and depending on the reproduction volume, the listener's hearing characteristics
will make the room as expected. I can not get a feeling. Therefore, the fuzzy surround processor
14 takes in the audio reproduction output signal 11, measures the reverberation characteristic,
and performs the fuzzy operation using the actual reverberation characteristic and the desired
08-05-2019
8
reverberation characteristic as the deciding factor, to obtain each low-pass filter LPF and each By
changing the characteristics of the delay circuit, desired reverberation characteristics can be
obtained, and the presence can be more effectively obtained based on the reproduction volume.
Although a simple example is shown here, in recent years, in order to obtain higher quality
reverberation, D SP (DigitxlSi (ntl P + oce gate or)) is often used. In such a case, the parameters
given to the DSP will be determined by fuzzy inference. [Effects of the Invention] The present
invention is configured to include the factor that corrects the auditory characteristic different
depending on the volume level in the deciding factor that determines the acoustic transfer
characteristic of the audio reproduction output signal. This has the effect of being able to obtain
sound transmission characteristics that match the auditory sensitivity characteristics of
In addition, it is not necessary to prepare in memory or the like every characteristic pattern
assumed to perform fuzzy inference based on given determinant data, and since the calculation is
simple, adjustment is possible in real time, and the characteristics obtained can be obtained The
effect is achieved that the configuration can be made very simple with respect to diversity.
Furthermore, since the correction based on the volume change is performed independently of the
other determinants, the number of rules in the fuzzy operation can be reduced.
[0002]
Brief description of the drawings
[0003]
FIG. 1 is a schematic block diagram of the first embodiment of the present invention, FIG. 2 is an
explanatory diagram of a membership function for a volume factor of the present invention, and
FIG. 3 is an explanatory diagram of an example of production rules of the volume factor. Fig. 4 is
an explanatory view of membership functions of determinants other than the loudness factor of
the present invention, Fig. 5 is an explanatory view of an example of production rules of an
aesthetic factor, and Fig. 6 is an explanatory view of an example of production rules of a
powerful factor. Fig. 7 is an explanatory view of an example of a production factor of a depth
factor, Fig. 8 is an explanatory view of a fuzzy inference for a loudness factor, Fig. 9 is an
explanatory view of a fuzzy reasoning for determinants other than the loudness factor, and Fig.
10 FIG. 11 is a block diagram of the second embodiment of the present invention, FIG. 12 is a
block diagram of the third embodiment of the present invention, and FIG. 13 is a fourth
embodiment of the present invention. Example block diagram, 14 is an explanatory diagram of
the hearing sensitivity.
08-05-2019
9
1 ... fuzzy equalizer 2 ... display operation unit 3 ... fuzzy operation unit 4 ... storage unit 5 ... audio
input signal 6 ... decision factor change unit 7 ... audio output signal 8 · · · · · · Feedback signal 9 · ·
· Amplifier 10 · · · Sukiaka 11 · · · audio reproduction output signal 12 · · · sound pressure level
storage unit 13 · · · · output sound pressure level adjustment unit
08-05-2019
10