close

Вход

Забыли?

вход по аккаунту

?

JP2009194769

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 JP2009194769
An ear canal resonance correction device capable of canceling resonance according to the
structure of each person's ear canal. An ear canal model comprising an earphone or headphone
and an attenuator (58, 60) according to the reflection coefficient of the eardrum and a delay
according to the distance between the earphone or headphone and the eardrum (62, 66) And an
operation unit (16) for performing a convolution operation on the impulse response of the
inverse filter and the sound source signal. [Selected figure] Figure 1
Ear canal resonance correction device and ear canal resonance correction method
[0001]
The present invention relates to an ear canal resonance correction device and an ear canal
resonance correction method that cancels out the resonance of the ear canal.
[0002]
When listening to music with earphones or headphones, a resonance phenomenon occurs
between the tympanic membrane and the earphones or headphones, which means that an
unnatural sound is heard.
Various systems for canceling this resonance have been realized (see, for example, Patent
Document 1, Patent Document 2, and Patent Document 3).
08-05-2019
1
[0003]
Patent Document 1 discloses a technique relating to the following out-of-head sound image
localization. FIGS. 2A and 2B of Patent Document 1 are principle explanatory diagrams for
realizing the out-of-head sound image localization. (A) represents listening by a speaker, and (b)
represents listening by a binaural earphone or stereo headphone. In (a), 101 is a sound source
signal, 103 is a speaker, and 102 is a microphone installed in the ear canal of a listener. In (b),
104 indicates an earphone or headphone, and 105 indicates a digital filter. The suffixes L and R
such as HRTFL and HRTFR indicate the left side and the right side.
[0004]
The principle of out-of-head sound localization is to electrically create the same transfer function
as the transfer function from the sound source in space to the tympanic membrane.
[0005]
However, it is difficult to easily capture the vibration signal on the tympanic membrane by the
sound wave from the living body by the electric signal, so the transfer function of the electric
signal from the sound source signal 101 to the tympanic membrane in FIG. Can not measure.
Therefore, the microminiature microphone 102 is attached to the ear canal of both ears, and the
transfer function from the sound source signal 101 input to the speaker 103 to the output of the
microphone 102, that is, the head acoustic transfer function in the left and right ears (HRTF:
Head Measure Related Transfer Function.
[0006]
Since the speaker 103 has frequency characteristics, the true transfer function of the electric
signal from the input of the speaker 103 to the output of the microphone 102 is HRTF / SPTF if
the transfer function of the speaker 103 is SPTF (Speaker Transfer Function). It is.
[0007]
On the other hand, in FIG. 2 (b) of Patent Document 1, in order to create a transfer function
equivalent to this using binaural earphones or stereo headphones 104, it is attached to the
08-05-2019
2
external ear canal from the input of binaural earphones or stereo headphones 104. The transfer
function up to the output of the microphone 102, ie, the ear canal transfer function (ECTF), is
measured, and the transfer function of the product of this ECTF and the transfer function of the
digital filter 105 matches the transfer function HRTF / SPTF. If so, the same listening signal as
the speaker listening can be reproduced at the location of the microphone 102 installed in the
ear canal.
[0008]
In Patent Document 1, an ear canal transfer function when an earphone or a headphone is
mounted is measured using an out-of-head sound image localization unit as shown in FIG. 5, and
the correction is performed using an adaptive equalization filter.
[0009]
As shown in FIG. 1 of Patent Document 1, the microphone 3 for picking up the sound in the ear
canal is integrally attached to a speaker of an earphone or a headphone.
The digital filter 11 is a digital filter that stores the impulse response of the HRTF / SPTF transfer
function, which has been measured in advance by the configuration as shown in FIG. 2A of
Patent Document 1.
[0010]
The reason why the band digital filter 13 is provided is as follows.
That is, the adaptive digital filter 12 and ECTF are connected in series, and if this output signal is
an impulse, the transfer function of the adaptive digital filter 12 is the inverse transfer function
(= 1 / ECTF) of ECTF.
However, the ECTF includes the speaker 1 and the microphone 3 and is attenuated outside the
band. Therefore, the transfer function of the adaptive digital filter 12, which is the inverse
transfer function of ECTF, has a large gain outside the band.
08-05-2019
3
[0011]
Then, if the convolution operation result of each impulse response of the adaptive digital filter 12
and the ECTF is used as the impulse response of the band digital filter 13, the tap coefficient
value of the adaptive digital filter 12 or the impulse response value can be determined stably.
That is, if the band of the band digital filter 13 is made to pass a band narrower than the band of
the adaptive digital filter 12, the subtractor 14 cancels out-of-band part of the transfer function
from the adaptive digital filter 12 and stabilizes. You can find the solution you
[0012]
As described above, in Patent Document 1, the characteristics of the ear canal are corrected
using an adaptive equalization filter. In order to correct correctly, it is desirable that the
microphone 3 have a flat frequency characteristic in a band. This is because if the inverse
transfer function is created by the adaptive digital filter 12 using the ECTF including the
characteristics of the microphone, there is a possibility that the eardrum becomes an unnatural
sound. Also, I have to examine the position to attach the microphone. There is no problem if the
microphone mounting position is the tympanic membrane position, but if, for example, acquiring
the characteristics at the tip of the earphone or headphone (a position not at the end of the ear
canal) as shown in FIG. In order to pick up the sound, the characteristic in which the valley (dip)
is generated is acquired, which is different from the characteristic to be picked up by the
tympanic membrane. Therefore, listening to the sound corrected by the filter created using the
adaptive equalization filter with this characteristic results in a strange sound.
[0013]
Patent Document 2 discloses a technology for canceling the influence of standing waves
generated in earphones or headphones and the tympanic membrane. In order to cancel the
standing wave, it is desirable to measure the vibration signal on the tympanic membrane to
determine the ear canal transmission characteristic. However, it is difficult to measure a vibration
signal directly near the tympanic membrane by installing a microphone at the tympanic
membrane position of a person. Therefore, in Patent Document 2, a microphone is installed at the
tympanic membrane position of the pseudo head to measure the ear canal transfer function. And
based on the measured characteristic, the filter which cancels the standing wave which arises
with earphones or headphones and an eardrum is created.
08-05-2019
4
[0014]
However, since the acoustic impedance of the ear canal of a person and the eardrum are different
among individuals, and hence the characteristics of the ear canal transfer function are different
for each person, the position where the resonance frequency is generated is for each person. In
addition, since it differs depending on the right and left, it is necessary to make correction
according to the individual, and there is little possibility that a correction filter satisfying
everyone can be created using characteristics acquired by the pseudo head. There is also a
system in which some general characteristics are prepared and the characteristics that are
suitable to you are selected, but it is difficult for the user to select the correction filter that
matches his characteristics. , It is unlikely that the choice is perfect.
[0015]
Patent Document 3 discloses a resonance frequency component reducing means for reducing the
sound level in the vicinity of the resonance frequency of the human ear in order to prevent a
decrease in hearing when listening to music or the like at a large volume using earphones or
headphones. It is provided in the front of the electro-acoustic conversion means. Therefore, the
sound level of the resonance frequency component of the ear is prevented from becoming
excessive. In the register of the resonant frequency component reduction circuit, parameters are
set such that the component of the measured resonant frequency is reduced. Details of the
determination of this parameter are not described. As a general method of this determination,
there is known a method of using an inverse filter of resonance data actually measured as
described in Patent Document 1, a method of creating a filter close to data measured by a
parametric equalizer or the like, and the like. ing. However, these methods have the following
problems.
[0016]
1) Since it is impossible to place a microphone at the position of the eardrum, the characteristics
can not be measured accurately, and the sound quality is deteriorated by convolving the inverse
filter generated from the measured data.
[0017]
08-05-2019
5
2) Because of the large number of parameters and very difficult tuning, it may not be possible to
produce the desired characteristics.
Also, even if a desired amplitude characteristic is obtained, it is extremely difficult to accurately
represent the phase. JP-A-2000-92589 (paragraph 0047, FIG. 1, FIG. 2) JP-A-2002-209300
(paragraph 0040, FIG. 1) JP-A-9-187093 (paragraph 0024, FIG. 2)
[0018]
As described above, the conventional ear canal resonance correction device can not easily
perform correction according to the structure of the ear canal of each person.
[0019]
An object of the present invention is to provide an ear canal resonance correction device that can
cancel resonance depending on the structure of each person's ear canal.
[0020]
According to one aspect of the present invention, there is provided an ear canal resonance
correction device comprising: an ear canal model comprising an earphone or a headphone and an
attenuator according to a reflection coefficient of an eardrum and a delay according to a distance
between the earphone or the headphone and an eardrum. Inverse filter creation means for
creating an inverse filter of an ear canal model, and an operation unit for performing a
convolution operation on an impulse response of the inverse filter and a sound source signal.
[0021]
According to one aspect of the present invention, there is provided an ear canal resonance
correction method comprising: outputting a sound source signal from an earphone or
headphone; determining a frequency characteristic of an audio signal collected by a microphone
disposed in the ear canal; The delay time of the delay unit of the ear canal model provided with
the corresponding attenuator and the delay according to the distance between the earphone or
the headphone and the tympanic membrane is set to the time according to the resonance
frequency obtained from the frequency characteristic. An input signal is input to a series
connection circuit of an adaptive equalization filter and the ear canal model, and the adaptive
equalization filter is adjusted so as to minimize an error between an ideal signal of the input
signal and an output of the series circuit; A convolution operation is performed on the impulse
response of the adaptive equalization filter and the sound source signal.
08-05-2019
6
[0022]
As described above, according to the present invention, the resonance frequency is detected from
the frequency characteristic of the ear canal acquired by the microphone installed at a position
other than the resonance node, and the attenuator corresponding to the reflection coefficient of
the earphone or headphone and the tympanic membrane, The delay time of the delay unit of the
sound wave propagation model in the ear canal provided with a delay device according to the
distance between the earphone or headphone and the tympanic membrane The earphone or
headphone and tympanic membrane determined from the resonance wavelength obtained from
the above resonance frequency Adaptively equalize the inverse filter using a model in which the
time is set according to the distance between the two and cancel the resonance phenomenon of
the ear canal acoustic characteristic of each person by convoluting the impulse response and the
sound source signal of this inverse filter Can.
[0023]
Hereinafter, an embodiment of an ear canal resonance correction apparatus according to the
present invention will be described with reference to the drawings.
[0024]
FIGS. 1A and 1B are diagrams showing an example of the configuration of an ear canal resonance
correction apparatus according to a first embodiment of the present invention.
An audio signal collected by the microphone 12 is input to the correction filter generation unit
14.
On the other hand, the right ear sound source signal and the left ear sound source signal are
input to the convolution unit 16.
The correction filter generation unit 14 analyzes the input audio signal to create a correction
filter.
The correction filter has a frequency characteristic such that a dip can occur near the resonance
08-05-2019
7
frequency to cancel the resonance.
The tap coefficients of the correction filter are set in the convolution unit 16 in the example
shown in FIG. 1A, and once set in the memory 18 in the example shown in FIG.
However, even the configuration shown in (b) can be folded without writing in the memory 18.
The convolution operation unit 16 performs convolution operation processing on the sound
source signals of the left and right ears using the set tap coefficients. This gives a signal whose
resonance has been canceled.
[0025]
The microphone 12 is attached to the earphone or headphone 20 as shown in FIG. As described
above, when the microphone 12 is disposed at a position other than the end of the ear canal and
characteristics are acquired, a sound is picked up at a node of a standing wave, resulting in a dip
as shown in FIGS. 3 and 4. The characteristic is acquired, and it differs from the characteristic
which picks up sound with a tympanic membrane. FIG. 3 shows the ear canal characteristics of a
person's left ear and right ear. FIG. 4 shows the ear canal characteristics of the left ear of a
plurality of persons.
[0026]
If the microphone 12 is placed at a position other than the end of the ear canal and
characteristics are acquired, the characteristics become different from those in FIGS. 3 and 4.
However, the peak frequency (resonance frequency) substantially matches whether the sound is
collected at the eardrum position or the earphone or headphone position. It will be described
using FIG. 5 and FIG. 6 that the resonance frequency of the frequency characteristic collected
near the tympanic membrane and the resonance frequency of the frequency characteristic
collected at a position other than the tympanic membrane coincide. FIG. 5 is a diagram showing
an outline of an experiment using a simulated external ear canal. The artificial ear canal 22 is a
cylindrical tube that simulates the human ear canal. The experiment was performed by inserting
a very small inner microphone 24 inside the simulated external ear canal 22, and attaching an
eardrum microphone 26 and an earphone or headphone 28 to both ends of the tube. White noise
having a uniform frequency spectrum was output from the earphone or headphone 28, and was
collected by the inner microphone 24 and the tympanic membrane microphone 26, and the
08-05-2019
8
frequency spectra were compared. FIG. 6 is a diagram showing frequency characteristics of the
tympanic membrane microphone and the inner microphone acquired in this experiment. Thus,
although the characteristic acquired by the inner microphone causes a valley (dip) at the node of
the standing wave, the frequency at which the resonance peak occurs substantially matches the
characteristic acquired by the eardrum earphone or the headphone. Therefore, the frequency
characteristic acquired by the microphone 12 changes according to the installation position of
the microphone, so even if the inverse filter of the acquired frequency characteristic is made, the
correct inverse filter can not be made, and the resonance phenomenon is canceled accurately. Is
difficult. However, since the resonance frequency is correct, it is possible to cancel the resonance
phenomenon by correcting using only this.
[0027]
The microphone 24 may be installed inside the earphone or headphone 28 or at a position away
from the earphone or headphone 28. However, it is necessary to install the microphone 24 so
that the valley (dip) does not occur at the peak frequency (resonance frequency).
[0028]
FIG. 7 is a flowchart showing the flow of processing of the correction filter creation unit 14. For
example, as shown in FIG. 2, an earphone or headphone 20 equipped with a microphone 12 is
inserted into the external ear canal, and a sound source signal is output from the earphone or
headphone 20 and collected by the microphone 12 (block 32). Here, it is desirable that the sound
source signal output from the earphone or headphone 20 be a signal having a uniform frequency
spectrum, such as white noise. However, it may be a pink noise-like signal that is attenuated in a
certain band. Alternatively, TSP (Time-Stretched Pulse) may be used.
[0029]
At block 34, the collected audio signal is converted from the time domain to the frequency
domain. At block 36, a resonance peak is detected on the frequency axis. From the frequency
characteristics as shown in FIG. 3, for example, a first peak between 5 kHz and 10 kHz and a
second peak between 10 kHz and 15 kHz are detected for each of the left and right ears.
08-05-2019
9
[0030]
A correction filter is created for each of the left and right ears so as to create a dip at the
frequency at which the peaks occur in order to cancel out the two peaks for each of the left and
right ears detected (block 38). Although creation of a correction filter may be made with a
parametric equalizer or a graphic equalizer, here, a model is used to create a correction filter. The
details will be described later.
[0031]
In block 40, the correction filter creation unit 14 stores the tap coefficients of the created
correction filters for each of the left and right ears directly in the convolution operation unit 16
or temporarily stored in the memory 18 and then sets the tap coefficients in the convolution
operation unit 16.
[0032]
The convolution operation unit 16 performs a convolution operation of the left and right data
(tap coefficients indicating an impulse response) transferred from the correction filter generation
unit 14 or the memory 18 with the left and right sound source signal, and the right ear signal
and the left with resonance canceled. Create an ear signal.
[0033]
In this way, a filter that cancels the peak of resonance actually measured in each person's ear
canal is created, tap coefficients indicating the impulse response are set in the convolution unit
16, and the left and right sound source signals are convolutionally calculated. , The peaks in FIG.
3 are smoothed.
[0034]
In the above description, the microphones are attached to both the left and right earphones or
headphones to acquire the characteristics of the left ear and the right ear and create the
respective correction filters. However, the characteristics of only one ear are acquired and the
characteristics thereof A configuration may be adopted in which the correction filter created
using is combined with the sound source of both ears.
[0035]
Such correction filter creation processing may be performed, for example, each time the audio
08-05-2019
10
player is activated, or may be performed by the user arbitrarily operating, or the period set by
the user is exceeded. It may be performed when it is started after that.
[0036]
In the above description, the microphone 12 for acquiring the characteristics of the ear canal, the
correction filter generation unit 14, and the convolution operation unit 16 for performing
convolution operation on the sound source signal have been described as an integral unit. There
is no need.
For example, the sound source signal acquired by the microphone 12 may be taken into another
device, such as a personal computer (PC), and the correction filter may be created by software
processing on the PC.
[0037]
The same applies to the case of reproducing music, and in addition to performing the correction
processing in real time by mounting the convolution operation unit 16 in the player and
reproducing, for example, the player performs the resonance correction processing on the
original sound source signal by software processing on the PC. It may be transferred to
[0038]
According to the ear canal resonance correction device shown in FIG. 1, a valley (dip) can be
generated at the frequency at which the picked-up characteristic peaks occur without using an
adaptive equalization filter to correct the measured ear canal transfer function By creating the
correction filter as described above, it is possible to cancel the resonance that occurs in the
eardrum or the earphone or headphones without using an expensive microphone installed in the
eardrum.
Since the correction filter can be created without examining the position where the microphone
is installed, the design period can be shortened.
By attaching a microphone to earphones or headphones, acquiring resonance characteristics
08-05-2019
11
generated by individual earphones or headphones and the tympanic membrane, and creating a
correction filter according to the characteristics, ear canal resonance different depending on
individual ear canal characteristics and insertion state You can cancel the characteristic.
By acquiring both left and right characteristics and creating a correction filter for each
characteristic, different ear canal resonance characteristics can be canceled in the left and right
ears.
[0039]
Next, the correction filter creation processing (generation of the correction filter of block 38 in
FIG. 7) of the correction filter creation unit 14 in FIG. 1 will be described.
As described above, although the frequency characteristic changes according to the installation
position of the microphone but the resonance frequency does not change, a correction filter is
created using only the resonance frequency from the measured frequency characteristic.
Therefore, in this embodiment, the measured data (frequency characteristics) is not used as it is,
but the sound wave propagation in the ear canal with the reflection coefficient of the earphone
or headphone and tympanic membrane and the propagation time of sound wave between the
earphone or headphone and tympanic membrane as a parameter By creating a model and
generating an inverse filter of this model, an individualized correction of the resonance
characteristics of the ear canal is realized.
[0040]
A one-dimensional model of the ear canal model is shown in FIG. The sound wave propagation
model in the ear canal shown in FIG. 8 is an attenuator 60 that represents the reflection
coefficient of the tympanic membrane, 58 that represents the reflection coefficient of the
earphone or headphone, the distance between the earphone or headphone and tympanic
membrane (the sound wave between the earphone or headphone and tympanic membrane And
an adder 64 for adding the input audio signal output from the earphone or headphone and the
signal reflected by the earphone or headphone (output of the attenuator 58). . The reflection
coefficient of earphones or headphones and tympanic membrane differs depending on the
08-05-2019
12
person, but here general values are used. The distance between the earphone or the headphone
and the tympanic membrane can be determined from the speed of sound and the wavelength by
determining the wavelength of the sound wave from the measured resonant frequency.
[0041]
The ear canal acoustic characteristic as shown in FIG. 9 can be obtained from the sound wave
propagation model in the ear canal as described above. (A) is an amplitude characteristic and (b)
is a phase characteristic.
[0042]
Next, an inverse filter is generated according to the model shown in FIG. 10 from the obtained
ear canal acoustic characteristics. As shown in FIG. 10, the input signal is input to the adaptive
equalization filter 72 and the delay unit 78. The output of the adaptive equalization filter 72 is
input to a filter 74 representing the ear canal acoustic characteristic (model of FIG. 8). The delay
time of the delay unit 78 is a delay time when the input signal passes through the adaptive
equalization filter 72 and the ear canal acoustic characteristic filter 74. Therefore, the input
signal through the delay unit 78 becomes the expected value of the input signal through the
adaptive equalization filter 72 and the ear canal acoustic characteristic filter 74. The outputs of
the delay unit 78 and the ear canal acoustic characteristic filter 74 are input to the subtractor 76.
The adaptive equalization filter 72 self-learns so that the error output from the subtractor 76 is
minimized. The characteristic of the adaptive equalization filter 72 when the error output from
the subtractor 76 is minimum is the inverse filter of the ear canal acoustic characteristic filter 74.
Although various specific examples can be considered for the adaptive equalization filter 72,
here, as an example, white noise is used for the input signal, and LMS is used for the adaptive
algorithm.
[0043]
Assuming that the ear canal acoustic characteristic shown in FIG. 9 is the characteristic of the ear
canal acoustic characteristic filter 74, the characteristic of the adaptive equalization filter 72 is as
shown in FIG. Therefore, if the correction filter generation unit 14 generates a correction filter
having the characteristics shown in FIG. 11, the convolution operation unit 16 can exactly cancel
the resonance phenomenon of the ear canal acoustic characteristic of each person.
08-05-2019
13
[0044]
The above operation is performed for each of the left and right ears to create a correction filter
for each of the left and right ears.
[0045]
The following is a method of further improving the characteristics.
In the model of FIG. 8, as shown in the frequency characteristic of FIG. 9A, resonance (peak)
occurs in the low band near 0 Hz where no resonance actually occurs. For this reason, the
frequency characteristic of the inverse filter created from this model is also attenuated in the low
band as shown in FIG. 11A, and the sound quality is degraded. As the cause of this, it is
conceivable that the model of FIG. 8 does not consider that the tension (elastic modulus) of the
tympanic membrane changes with frequency, that is, the frequency dependence of the acoustic
impedance is present. Therefore, in order to add the frequency dependence of the acoustic
impedance of the tympanic membrane, the sound wave propagation model in the ear canal (FIG.
12) in which the filter 80 is added to the output of the attenuator 60 showing the reflection
coefficient of the tympanic membrane in the model of FIG. .
[0046]
It is known that the elastic modulus of the polymer constituting the tympanic membrane is small
mainly at low frequencies and increases as the frequency increases. A high-pass filter 80 as
shown in FIG. 13 is added in consideration of the elastic modulus of the tympanic membrane
(frequency dependency of acoustic impedance) with reference to this.
[0047]
As a result, as shown in FIG. 14, the resonance in the low region is suppressed as shown in FIG.
14, and the inverse filter having no drop in the low region can be realized as shown in FIG. This
can improve the deterioration of sound quality that may occur in the model shown in FIG.
08-05-2019
14
[0048]
By using the models shown in FIGS. 8 and 12, desired characteristics can be easily obtained
simply by tuning the reflection coefficient and the length. By generating an inverse filter from a
sound wave propagation model in the ear canal in line with the physical phenomenon, an inverse
filter with appropriate phase characteristics can be obtained. Even if it is not possible to
accurately acquire the ear canal characteristics, it is possible to create an inverse filter without
sound quality deterioration. By using the resonance data measured for each individual, individual
differences in the ear canal or tympanic membrane can be reflected in the correction filter. The
resonance data measured by the left and right ears can reflect the difference in acoustic
characteristics of the left and right ears. It is possible to reflect the difference in the resonance
characteristics depending on the type of earphones or headphones and the wearing condition of
each individual.
[0049]
The mounting locations of the correction filter creation unit 14 and the convolution operation
unit 16 in FIG. 1 will be described with reference to FIG.
[0050]
When built in the player 90, the tap coefficients of the correction filter generated by the
correction filter generation unit 14 are stored in the memory 18, and the sound source signal
read from a flash memory, a hard disk or the like (not shown) After being corrected, it is output
to the earphone or headphone 94.
Alternatively, when incorporated in the player 90, it is possible to correct the sound source signal
when downloading it and store the corrected sound source signal in a memory or the like. Also, it
may be built in the remote control 92, the earphone or the headphone 94. In any case, the
microphone 12 is attached to the earphone or headphone 20 as shown in FIG.
[0051]
As described above, according to this embodiment, the resonance frequency is detected from the
frequency characteristic of the ear canal acquired by the microphone installed at an arbitrary
position, and the attenuator according to the earphone or the headphone and the eardrum's
08-05-2019
15
reflection coefficient, and the earphone Or, as the delay time of the delay unit of the sound wave
propagation model in the external auditory canal provided with the delay unit according to the
distance between the headphone and the tympanic membrane, the earphone or the headphone
and the tympanic membrane obtained from the resonant wavelength obtained from the resonant
frequency Adaptive equalization (identification) of the inverse filter using a model in which the
time is set according to the distance of L, and correcting the frequency characteristic of the
sound source signal using this inverse filter, the resonance phenomenon of the ear canal acoustic
characteristic of each person Can be canceled exactly.
[0052]
If correction is performed using an inverse filter generated from actual measurement data
without using a model, it is impossible to place a microphone at the position of the tympanic
membrane, so the characteristics can not be measured accurately, and the sound quality is
degraded by the correction. .
[0053]
Furthermore, by adding a high pass filter to this model in consideration of the frequency
dependency of the acoustic impedance of the tympanic membrane, it is possible to realize an
inverse filter with no drop in the low frequency range, and to realize resonance correction with
less deterioration in sound quality.
[0054]
If an inverse filter is created using a parametric equalizer, it may not be possible to produce the
desired characteristics because of the large number of parameters and very difficult tuning.
Even if a desired amplitude characteristic is obtained, it is extremely difficult to create an inverse
filter that accurately reflects the phase, so the correction makes the phase information unnatural
(aberrant phase rotation). .
However, according to the model of the present embodiment, phase information can also be
correctly obtained.
[0055]
08-05-2019
16
The present invention is not limited directly to the above-described embodiment. In practice, the
structural elements can be modified and embodied without departing from the spirit of the
invention.
In addition, various inventions can be formed by appropriate combinations of a plurality of
components disclosed in the above embodiments. For example, some components may be deleted
from all the components shown in the embodiment. Furthermore, components in different
embodiments may be combined as appropriate.
[0056]
The present invention is also embodied as a computer-readable recording medium storing a
program for causing a computer to function as a predetermined means or causing a computer to
realize a predetermined function, in order to cause the computer to execute the predetermined
means. You can also
[0057]
FIG. 5 is a schematic view of the ear canal resonance correction according to an embodiment of
the present invention.
FIG. 2 is a view showing an example of the arrangement position of the microphone of FIG. 1;
The figure which shows the frequency characteristic of the right and left of a certain person
obtained from the sound collected with the microphone of FIG. The figure which shows the
frequency characteristic of the left ear of several people obtained from the sound collected with
the microphone of FIG. The figure which shows the outline of experiment using a simulated ear
canal for comparing the frequency characteristic of a tympanic membrane microphone and an
inner microphone. The figure which shows the frequency characteristic of the tympanic
microphone and inner microphone which were obtained by experiment. 7 is a flowchart showing
the operation of the correction filter creation unit of FIG. The figure which shows an example of
the sound wave propagation model of an ear canal. The figure which shows the acoustic
frequency characteristic of the ear canal obtained from the model of FIG. FIG. 9 is a schematic
diagram of creating an inverse filter using the model of FIG. 8; The figure which shows the
frequency characteristic of the inverse filter of FIG. The figure which shows the other example of
the sound wave propagation model of an ear canal. The figure which shows the frequency
08-05-2019
17
characteristic of the high-pass filter showing the frequency dependence of the acoustic
impedance of the tympanic membrane used for the model of FIG. The figure which shows the
acoustic frequency characteristic of the ear canal obtained from the model of FIG. The figure
which shows the frequency characteristic of the inverse filter obtained from the model of FIG.
The figure which shows the example of implementation of this embodiment.
Explanation of sign
[0058]
12 microphone 14 correction filter creation unit 16 convolution unit 52 earpiece or headphone
54 ear canal 56 tympanic membrane 58, 60 attenuator 62, 66 delayer 64 adder 72 ... adaptive
equalization filter, 74 ... ear canal acoustic characteristic filter.
08-05-2019
18
Документ
Категория
Без категории
Просмотров
0
Размер файла
31 Кб
Теги
jp2009194769
1/--страниц
Пожаловаться на содержимое документа