close

Вход

Забыли?

вход по аккаунту

?

JPH08294195

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 JPH08294195
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a
signal processing method for reproducing an input signal from an electroacoustic transducer
without loss.
[0002]
2. Description of the Related Art The sound reproduction characteristic of a speaker is generally
an amplitude characteristic having a peak dip. In order to reduce this peak dip, it has been
conventional to correct the reproduction characteristics using a digital filter. FIG. 16 is a block
diagram showing an example of a conventional signal processing apparatus of this type, which is
disclosed, for example, in Japanese Examined Patent Publication No. 2-52886.
[0003]
In FIG. 16, 1 is a speaker, 2 is a non-recursive digital filter, 3 is an input of an analog signal, 4 is
an A / D converter for converting an analog signal to a digital signal, and 5 is for converting a
digital signal to an analog signal D / A converter, 6 is an amplifier for converting an analog signal
whose characteristic has been changed by the non-recursive digital filter 2 into a drive signal for
the speaker 1, 7 is a microphone for characteristic measurement, 8 is a frequency amplitude
characteristic of the speaker 1, 9 Is the correction characteristic of the speaker 1 determined by
measurement, and is detected by the time measurement data of the filter coefficient of the nonrecursive digital filter 2, and 10 by the microphone 7 for characteristic measurement of the
08-05-2019
1
speaker 1 using the correction characteristic 9 Integrated regeneration characteristics.
[0004]
FIG. 17 shows the amplitude frequency characteristic 8 of the speaker 1 in which the horizontal
axis represents a linear frequency display and the vertical axis represents a logarithmic
amplitude, and FIG. 18 a correction characteristic for correcting the amplitude frequency
characteristic of the speaker 1 of the non-recursive digital filter 9 and FIG. 19 show the overall
reproduction characteristic 10 detected by the characteristic measurement microphone when the
speaker 1 having the characteristic of FIG. 17 is reproduced using the non-recursive digital filter
2 having the characteristic of FIG. It is.
[0005]
FIG. 20 shows the amplitude frequency characteristics of the speaker 101 which is the same type
as the speaker 1 and another individual, for example, the same type and different production lot.
FIG. 21 shows a reproduction overall characteristic 10 detected by the characteristic
measurement microphone 7 when the speaker 101 having the amplitude frequency
characteristic 8 of FIG. 20 is reproduced using the correction characteristic 9 of FIG. 18 with the
configuration of FIG. It is.
[0006]
Next, the operation of the above conventional signal processing method will be described.
In FIG. 16, an analog signal input from the input terminal 3 is converted into a digital signal by
the A / D converter 4 and then input to the non-recursive digital filter 2. In the non-recursive
digital filter 2, the characteristics of FIG. 18 are convoluted in real time with respect to the input
digital signal. The digital signal in which the characteristics of FIG. 18 are convoluted is
converted back to an analog signal by the D / A converter 5 and input to the amplifier 6. When
the speaker 1 having the characteristic of FIG. 17 is driven by the amplifier 6 using a signal in
which the characteristic of FIG. 18 is convoluted, the reproduction amplitude characteristic of the
characteristic measurement microphone 7 has no peak dip as shown in FIG. Overall regeneration
characteristics are obtained.
08-05-2019
2
[0007]
However, even if speakers of the same type are reproduced using different individual speakers
101 as shown in FIG. 20 using the correction characteristic 9 shown in FIG. 18, the
comprehensive reproduction characteristic 10 detected by the characteristic measurement
microphone 7 is flat. It is not a disordered characteristic.
[0008]
The non-recursive digital filter 2 convolves a preset characteristic into an input signal.
On the other hand, the overall characteristics of the reproduction of the speakers are different
even if they are the same type of speaker, and the characteristics are different from each other,
and the configuration shown in FIG. Even when it is performed, there is a problem that the peak
dip of the overall regeneration characteristic is larger than that without the acyclic digital filter 2.
[0009]
Also, measuring and inputting the characteristics of the non-recursive digital filter 2 each time
using different speakers 101 of the same type, it is necessary to measure the filter characteristics
only a few times of the speakers to be used. There has also been the problem of increasing time
and cost.
[0010]
The present invention has been made for the purpose of solving the above-mentioned problems,
and it is an object of the present invention to provide a signal processing method capable of
correcting the overall reproduction characteristics of a speaker regardless of variations in
amplitude frequency characteristics of individual speakers. I assume.
[0011]
A signal processing method according to claim 1 selects amplitude values at a plurality of
adjacent frequency points of the reproduction inverse characteristic of the electroacoustic
transducer determined by measurement, and selects the selected amplitude value. The
reproduction general characteristics of the electroacoustic transducer are changed using the
08-05-2019
3
amplitude characteristics obtained by averaging.
[0012]
The signal processing apparatus according to claim 2 changes the amplitude value of the
frequency point having an amplitude value larger than the predetermined amplitude value of the
reproduction inverse characteristic of the electroacoustic transducer determined by
measurement to an arbitrary value, The reproduction characteristic of the electroacoustic
transducer is changed using the amplitude characteristic.
[0013]
A signal processing method according to a third aspect of the present invention is the signal
processing method for changing the characteristics, comprising: selecting amplitude values at
adjacent plural frequency points of the reproduction inverse characteristic of the electroacoustic
transducer determined by measurement; After averaging the values, the amplitude characteristic
at an arbitrary frequency point is changed to an arbitrary value to change the overall
reproduction characteristic of the electroacoustic transducer.
[0014]
A signal processing method according to claim 4 is the signal processing method according to
claim 1 or 3, wherein the number of amplitude values to be used for averaging selected
according to the frequency can be changed.
[0015]
A signal processing method according to claim 5 measures the reproduction inverse
characteristic of the plurality of electro-acoustic transducers, and changes the reproduction
characteristic of the electro-acoustic transducer using the amplitude characteristic on which the
averaging is performed. .
[0016]
In the first aspect of the present invention, a moving average is obtained in the direction of the
frequency axis of the correction characteristic of the speaker obtained by measurement, and it is
set as a new correction characteristic and input to the non-recursive digital filter to correct the
overall reproduction characteristic of the speaker Is done.
[0017]
08-05-2019
4
According to the second aspect of the present invention, the peak characteristic of the correction
characteristic of the speaker obtained by the measurement is suppressed to obtain a new
correction characteristic, which is input to the non-recursive digital filter to correct the general
reproduction characteristic of the speaker.
[0018]
According to the invention of claim 3, the correction characteristic of the speaker obtained by the
measurement is obtained as a moving average in the direction of the frequency axis, and the
peak characteristic is further suppressed to obtain a new correction characteristic, which is input
to the non-recursive digital filter The correction of the overall characteristic of reproduction is
performed.
[0019]
According to the fourth aspect of the present invention, in obtaining the moving average in the
frequency axis direction, the number of samples to be averaged is changed according to the
frequency, and the frequency resolution for averaging is corrected in accordance with the
auditory characteristics. It will be.
[0020]
According to the invention of claim 5, the averaging is performed from the reproduction
characteristics of the plurality of speakers, the correction characteristic is determined, and a new
correction characteristic is obtained, which is input to the non-recursive digital filter to correct
the reproduction general characteristic of the speaker.
[0021]
Embodiments of the present invention will be described below with reference to the
accompanying drawings.
Example 1
FIG. 1 is a block diagram showing the configuration of a signal processing apparatus according to
a first embodiment of the present invention. The same reference numerals as in FIG. 16 denote
the same parts, and therefore the description thereof will be omitted here.
08-05-2019
5
11 is a fast Fourier transform computing unit (hereinafter referred to as "FFT") for converting
time axis information of the measured correction characteristic into frequency axis information,
and 12 is discrete on the frequency axis obtained by FFT 11. A selector for selecting the values of
a plurality of adjacent successive amplitude characteristics obtained, 13 is an averaging operator
for calculating an average value of the values selected by the selector 12, and 15 is a row where
averaging is performed on the frequency axis It is an inverse fast Fourier transform operator
(hereinafter referred to as “inverse FFT”) that performs an operation to convert it into time
axis information after it is received.
[0022]
FIG. 2 is a diagram for explaining the operation of the selector 12 and the averaging calculator
13.
In FIG. 2A, part of the amplitude characteristic calculated by the FFT 11 is enlarged.
Reference numeral 22 denotes values discretely present on the frequency axis obtained by the
FFT operation, and shows the case where amplitude values of three discrete frequencies are
selected by the selector 12 and averaged by the averaging operator 13.
FIG. 2 (b) 23 is an averaged value.
[0023]
FIG. 3 is a graph showing the correction characteristic obtained by calculating the amplitude
frequency characteristic of the speaker 1 shown in FIG. 17 by the calculation result of the
selector 12 and the averaging calculator 13. FIG. 4 uses the averaged correction characteristic
shown in FIG. 20 shows the overall reproduction characteristics when the characteristic
correction of the speaker 1 is performed, and FIG. 5 shows the overall reproduction
characteristics when the characteristic correction of another speaker 101 of the same type
whose amplitude frequency characteristic is shown in FIG. FIG.
[0024]
08-05-2019
6
Next, the operation of the signal processing device of the first embodiment will be described.
The amplitude frequency characteristic 8 of the speaker 1 shown in FIG. 17 can be obtained by
reproducing an impulse signal from the speaker 1 and performing an FFT operation on an
impulse response detected and obtained by the characteristic measurement microphone 7.
Next, the correction characteristic 9 of the speaker 1 is extracted from the difference between
the flat amplitude characteristic with a phase straight line and the amplitude frequency
characteristic 8 and is obtained as frequency axis information or time axis information.
The extracted correction characteristic 9 is nothing but the correction characteristic measured by
the characteristic measurement microphone 7.
The time information of the correction characteristic 9 of the speaker 1 measured by the
characteristic measurement microphone 7 is once converted into frequency axis information
using the FFT 11.
As discrete values obtained on the frequency axis, as shown in FIG. 2 (a), a plurality of adjacent
values 22 are selected by the selector 12, and these selected value groups are averaged
separately for each of them. Perform averaging.
The averaged value 23 becomes the amplitude value of the new correction characteristic 9, and
the new correction characteristic becomes as shown in FIG.
The new correction characteristic is converted from frequency axis information to time axis
information by the inverse FFT 15 and is input to the non-recursive digital filter 2 as the
correction characteristic.
[0025]
08-05-2019
7
The analog signal input from the input terminal 3 is converted to a digital signal by the A / D
converter 4 and then input to the non-recursive digital filter 2. In the non-recursive digital filter
2, the correction characteristic input from the inverse FFT 15 is convoluted in real time with
respect to the input digital signal, and the digital signal convoluted with the correction
characteristic is a D / A converter The signal is converted back to an analog signal by 5 and input
to the amplifier 6. When the speaker 1 having the amplitude frequency characteristic shown in
FIG. 17 is driven by the amplifier 6 using the signal in which the correction characteristic shown
in FIG. 3 is convoluted, the overall reproduction characteristic detected by the characteristic
measuring microphone 7 is shown in FIG. The result is as shown in FIG. 4, and a peak-dip
suppressed overall regeneration characteristic is obtained as compared with FIG. 17.
[0026]
Even when the speaker 101 having the amplitude frequency characteristic shown in FIG. 20 is
reproduced using the same correction characteristic, as shown in FIG. 5, the comprehensive
reproduction characteristic in which the peak dip is suppressed as compared with that before the
correction is performed. Is obtained. Therefore, in the case of a speaker system having the same
structure and using speakers of the same type, it is possible to suppress the peak dip of the
overall reproduction characteristic without using a filter having different characteristics for each
speaker.
[0027]
Example 2 6 is a block diagram showing the configuration of a signal processing apparatus
according to a second embodiment of the present invention, in which the same reference
numerals as in FIG. 1 denote the same parts, and 16 is a level greater than the average amplitude
level of the correction characteristic. And a level changer 17 for changing the amplitude level of
the frequency domain detected by the peak detector 16.
[0028]
FIG. 7 shows a correction characteristic obtained by changing the peak level of the correction
characteristic measured for a loudspeaker having an average amplitude frequency characteristic
of the same type by the level changer 17. FIG. 8 shows the correction characteristic shown in FIG.
FIG. 9 is a view showing a reproduction overall characteristic in which the speaker 1 is corrected,
and FIG. 9 is a view showing the reproduction overall characteristic in the case where the
speaker 101 of the same type is reproduced using the correction characteristic shown in FIG.
08-05-2019
8
[0029]
Next, the operation of the second embodiment will be described.
The correction characteristic measured for a speaker having an average amplitude frequency
characteristic is converted from time axis information to frequency axis information using the
FFT 11. After the amplitude characteristic is determined as the frequency axis information, the
peak detector 16 detects an amplitude value larger than the average amplitude value. From the
value and frequency detected by the peak detector 16, the amplitude value is changed by the
level changer 17 to, for example, an average amplitude value. The characteristic whose amplitude
value has been changed is converted to time axis information by the inverse FFT 15 and input to
the non-recursive digital filter 2.
[0030]
When the speaker 1 is reproduced using the correction characteristic in which the amplitude
characteristic has been changed by the level changer 17, the result is as shown in FIG. Further,
even if reproduction is performed using the same correction characteristic using different
speakers 101 of the same type, a large peak does not occur as shown in FIG.
[0031]
Here, although it demonstrated using an average amplitude value as a predetermined amplitude
value, any value may be sufficient.
[0032]
Example 3
FIG. 10 is a block diagram showing the configuration of a signal processing apparatus according
to a third embodiment of the present invention. The same reference numerals as in FIGS. 1 and 6
denote the same or corresponding parts. 11 shows the correction characteristics of the speaker 1
08-05-2019
9
obtained by the signal processing method according to the third embodiment, and FIG. 12 shows
the overall reproduction characteristics of the speaker 1 corrected using the correction
characteristics of FIG. These are figures which show the reproduction | regeneration general
characteristic at the time of reproducing | regenerating the speaker 101 with the same
correction | amendment characteristic.
[0033]
Next, the operation of the third embodiment will be described. The time information of the
correction characteristic of the speaker 1 measured by the characteristic measurement
microphone 7 is once converted into frequency axis information using the FFT 11. As discrete
values obtained on the frequency axis, as shown in FIG. 2, a plurality of adjacent values are
selected by the selector 12 and the selected value group is averaged by the averaging operator
13. The averaged value is further detected by the peak detector 16 as an amplitude value larger
than the average amplitude value, and the level changer 17 changes the amplitude value to, for
example, an average amplitude value. The characteristic whose amplitude value has been
changed is converted to time axis information by the inverse FFT 15 and input to the nonrecursive digital filter 2.
[0034]
After averaging, by changing the peak amplitude value, the obtained correction characteristic is
as shown in FIG. 11, and the reproduction characteristic of the speaker 1 using this characteristic
is as shown in FIG. In addition, when another speaker 101 of the same type is reproduced using
the same correction characteristic, a comprehensive reproduction characteristic shown in FIG. 13
is obtained.
[0035]
Example 4 FIG. 14 is a diagram showing the signal processing operation of the selection means
12. In the FFT 11, discrete frequency amplitude values are obtained at equal intervals on the
frequency axis at linear scale. The selector 12 selects amplitude values at a plurality of discrete
frequencies, but in the first embodiment, the number to be selected is constant regardless of the
frequency. On the other hand, human auditory senses have different resolutions in the low band
and the high band. Therefore, in the present embodiment, the number of amplitude values used
08-05-2019
10
for averaging is changed depending on the frequency when performing the averaging operation.
[0036]
As shown in FIG. 14, according to the increase of the frequency, for example, the selected
number is increased logarithmically. The average of the end points is calculated using a value
obtained by folding the frequency. As an example, when the frequency axis is divided at regular
intervals by 1024 points, the area to be averaged is divided into 10 by the logarithmic axis, and
the same number of averages are taken in the same division.
[0037]
In the fourth embodiment, division of the frequency is set to the logarithmic axis, but division
based on another function axis may be performed.
[0038]
Example 5
FIG. 15 is a diagram for explaining the operation of means for obtaining a correction
characteristic by averaging a plurality of correction characteristics obtained by measurement of
the signal processing method according to the fifth embodiment of the present invention. In the
figure, 9a to 9c are time axis displays of a plurality of correction characteristics obtained by
measurement, 18 is an addition averaging means, and 9h is a time axis display of correction
characteristics obtained for the correction characteristics after addition averaging.
[0039]
The reproduction characteristics of the speaker of the same type are determined from a plurality
of individuals. Even if speakers are of the same type, variations occur in the characteristics.
However, the general characteristics are almost the same. The correction characteristics 9a to 9c
of the plurality of speakers are obtained by measurement, and the time axis information thereof
is added and averaged by the number of the measured speakers. At this time, the addition
average is determined such that the time of the maximum value of the time axis information is
08-05-2019
11
the same time among a plurality of information to be added. By inputting the correction
characteristic 9h obtained by the averaging to the non-recursive digital filter 2 shown in FIG. 1,
the same filter coefficient is used even if the individual speaker changes as in the first
embodiment. The overall regeneration characteristics can be corrected.
[0040]
Since the present invention is configured as described above, the following effects can be
obtained.
[0041]
According to the signal processing method of claim 1, since the amplitude characteristic of the
correction characteristic of the speaker obtained by measurement is configured to be averaged
with the amplitude value at the adjacent frequency, the speaker is changed to an individual with
different types. Even in this case, it is possible to make the fluctuation range of the amplitude
characteristic smaller than that before the correction, and it is possible to correct the total
reproduction characteristic of the speaker of the same type having the dispersion with only one
filter characteristic.
[0042]
According to the signal processing method of claim 2, since the peak correction level of the
correction characteristic is suppressed and a new correction characteristic is supplied, even if the
speaker is changed to a different individual with the same type, it is possible to perform the
correction before the correction. Also, the maximum value of the amplitude generated in the
amplitude characteristic can be lowered, and even if the dip characteristic varies, the overall
reproduction characteristic can be improved with only one filter characteristic.
[0043]
According to the signal processing method of claim 3, after the correction characteristics
obtained by measurement are averaged with the amplitude values at adjacent frequencies, the
peak correction level is suppressed and supplied as new correction characteristics. Even if the
speaker is changed to a different individual in the same kind, the maximum value of the
fluctuation range and the amplitude generated in the amplitude characteristic can be made lower
than before correction, and the fluctuation of the approximate amplitude characteristic and the
dispersion of the dip characteristic occur. Also, with only one filter characteristic, the overall
regeneration characteristic can be improved.
08-05-2019
12
[0044]
According to the signal processing method of claim 4, since the number used for averaging the
correction characteristics is changed to the frequency resolution according to the frequency,
averaging of all the bands is uniformly performed in terms of hearing. .
[0045]
According to the signal processing method of claim 5, the frequency amplitude characteristics of
the plurality of speakers are measured and obtained, and the characteristic obtained by
averaging the respective correction characteristics is used as the correction characteristic of the
speaker, so that the FFT operation is performed. Thus, it is possible to obtain a filter
characteristic that can make the fluctuation range of the amplitude characteristic smaller than
before correction.
[0046]
Brief description of the drawings
[0047]
FIG. 1 is a block diagram showing the configuration of a signal processing apparatus according to
a first embodiment of the present invention.
[0048]
FIG. 2 is a diagram for explaining the operation of the averaging operator of the first
embodiment.
[0049]
FIG. 3 is a diagram showing correction characteristics after a change is made by the signal
processing method of the first embodiment.
[0050]
FIG. 5 is a view showing a total reproduction characteristic of the speaker 1 using the correction
characteristic of FIG. 3;
[0051]
FIG. 5 is a view showing a total reproduction characteristic of the speaker 101 using the
08-05-2019
13
correction characteristic of FIG. 3;
[0052]
FIG. 6 is a block diagram showing the configuration of a signal processing apparatus according to
a second embodiment of the present invention.
[0053]
FIG. 7 is a diagram showing an average correction characteristic of the speaker after a change is
made by the signal processing method of the second embodiment.
[0054]
FIG. 8 is a diagram showing a total reproduction characteristic of the speaker 1 using the
correction characteristic of FIG. 7;
[0055]
FIG. 10 is a diagram showing a total reproduction characteristic of the speaker 101 using the
correction characteristic of FIG. 7;
[0056]
FIG. 10 is a block diagram showing the configuration of a signal processing apparatus according
to a third embodiment of the present invention.
[0057]
FIG. 11 is a diagram showing the correction characteristic of the speaker 1 after the change of
the third embodiment is added.
[0058]
FIG. 12 is a diagram showing a total reproduction characteristic of the speaker 1 using the
correction characteristic of FIG.
[0059]
FIG. 13 is a diagram showing a total reproduction characteristic of the speaker 101 using the
correction characteristic of FIG. 11;
08-05-2019
14
[0060]
FIG. 14 is a diagram for explaining the operation of the selector according to the fourth
embodiment of the present invention.
[0061]
FIG. 15 is a view for explaining means for averaging the correction characteristics of the fifth
embodiment of the present invention;
[0062]
FIG. 16 is a block diagram showing a configuration of a conventional signal processing
apparatus.
[0063]
FIG. 17 is a diagram showing an amplitude frequency characteristic of the speaker 1.
[0064]
FIG. 18 is a diagram showing correction characteristics of the speaker 1.
[0065]
19 is a diagram showing a total reproduction characteristic of the speaker 1 using the correction
characteristic of FIG.
[0066]
FIG. 20 is a diagram showing an amplitude frequency characteristic of the speaker 101.
[0067]
21 is a diagram showing a total reproduction characteristic of the speaker 101 using the
correction characteristic of FIG.
[0068]
Explanation of sign
08-05-2019
15
[0069]
2 non-recursive digital filter, amplitude frequency characteristic of 8 speakers, 9 correction
characteristics, 10 reproduction general characteristics, 11 FFT (fast Fourier transform
arithmetic unit), 12 selectors, 13 averaging arithmetic unit, 15 inverse FFT (inverse fast Fourier
analysis Conversion calculator), 16 peak detectors, 17 amplitude value modifiers, 18 averaging
means.
08-05-2019
16
Документ
Категория
Без категории
Просмотров
0
Размер файла
25 Кб
Теги
jph08294195
1/--страниц
Пожаловаться на содержимое документа