close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JPH09185383

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 JPH09185383
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an
adaptive sound field control apparatus for actively controlling a sound field, and in particular, to
correct desired sound field characteristics such as flat, low boost, vocal band emphasis, stereo
sound image correction, etc. The present invention relates to an adaptive sound field control
apparatus that can be realized in
[0002]
2. Description of the Related Art When a speaker is driven by a music signal reproduced by a
reproducing apparatus such as a CD player, the response in the low range is insufficient unless
the aperture of the speaker is sufficiently large, and powerful bass can not be heard. There is a
thing. In addition, when the sound reproduction space is narrow as in a vehicle compartment, the
response of 500 Hz to 1 kHz may be dropped, and the vocal sound may be unclear. For this
reason, conventionally, a graphic equalizer is provided between the CD player and the power
amplifier, and the user raises the gain of the low frequency band and the vocal band of 500 Hz to
1 kHz so that the frequency-gain characteristic at the listening position is flat. I was trying to get
close. In addition, a correction filter with a fixed transfer characteristic (transfer function)
between the CD player and the power amplifier is provided so that the frequency characteristic at
the listening position becomes flat even if the user does not make a manual adjustment. Was.
[0003]
10-04-2019
1
However, it is very difficult for the user to manually adjust the frequency characteristic at the
listening position to be flat using the graphic equalizer. In addition, since the correction
characteristic (transfer function) has a fixed frequency characteristic that is realized by the
correction filter, the difference between the head shape of the individual user and the shape of
the pinna (difference in head transfer function) There is a problem that it is impossible to cope
with the problem and it is impossible to accurately realize flat frequency characteristics. On the
other hand, when the speakers are replaced with headphones, the influence of the space in the
vehicle interior will not be affected, but the frequency characteristics of the headphone speakers
will be undone, and the transmission frequency characteristics of the pinna will be corrected to
flatten the overall frequency characteristics. It is necessary to use a graphic equalizer and a
correction filter with a fixed transfer characteristic (transfer function). However, as described
above, it is very difficult for the user to manually adjust the frequency characteristic at the
listening position to be flat by manual operation using a graphic equalizer, and is realized by a
correction filter having a fixed transfer characteristic. Since the frequency characteristics are
fixed, it is not possible to cope with the difference in the shape of the pinna of individual users,
and it is not possible to accurately realize flat frequency characteristics. An object of the present
invention is to provide an adaptive sound field control device which can accurately realize
desired sound field characteristics in view of the problems of the prior art described above.
[0004]
In the adaptive sound field control apparatus according to the present invention, processing
means for processing an input signal with a variable transfer function, amplification means for
power-amplifying the output of the processing means, and One or more sound reproducing
means including electro-acoustic conversion means for converting the sound, at least one
microphone placed near one ear of the listener, and an input signal provided for each
microphone A reference processing means for processing with a desired transfer function to
obtain a reference signal, provided for each sound reproduction means, for at least one
microphone, processing so that the difference between the microphone output and the
corresponding reference signal is reduced And control means for adaptively controlling the
transfer function of the means.
[0005]
Further, the control means performs adaptive control of the transfer function of the processing
means temporarily at any predetermined time, and fixes the transfer function of the processing
means after completion of the adaptive control, and only when the control means performs
adaptive control, It is characterized in that each microphone is located near the listener's ear, and
the other time is provided with movable support means for retracting away from the vicinity of
10-04-2019
2
the listener's ear.
[0006]
Further, the electro-acoustic conversion means is a headphone speaker provided in a headphone,
and the microphone is fixed to the headphone so as to be positioned near the entrance of the ear
canal.
[0007]
Further, when there are a plurality of channels of input signals, the one or more sound
reproducing means are provided for each channel of the input signal, and the reference
processing means individually processes the input signal of each channel with a desired transfer
function Then, they are added to obtain a reference signal.
[0008]
Further, the reference processing means is characterized in that the listener can select one as a
desired transfer function from among a plurality of predetermined transfer functions prepared in
advance.
[0009]
Also, the electro-acoustic conversion means is characterized in that it is placed downward or
upward behind the listener's ear and below the height of the ear.
[0010]
According to the adaptive sound field control device of the present invention, the sound
reproduced by the one or more sound reproducing means is one of at least one microphone
placed near one of the listener's ears or each of both ears. Pick up with at least two microphones
placed near the.
Then, the control means is provided in each sound reproducing means so that the difference
between the reference signal obtained by processing the input signal with a desired transfer
function by the reference processing means provided for each microphone and the
corresponding microphone output is reduced. The transfer function of the processing means is
adaptively controlled.
10-04-2019
3
Thereby, regardless of what frequency-response characteristic the electro-acoustic conversion
means of the acoustic reproduction means or the acoustic reproduction space has, or what kind
of head transfer characteristic the listener has, it is possible to accurately obtain the desired
sound field of the listener Can be realized.
[0011]
Further, the control means performs adaptive control of the transfer function of the processing
means temporarily at any predetermined time, and fixes the transfer function of the processing
means after completion of the adaptive control.
And, only when the control means performs adaptive control, the movable support means locates
each microphone near the listener's ear and retracts the microphone away from the listener's ear
for another time.
As a result, for example, when the position of the listener's ear hardly moves as in the vehicle
compartment, while performing adaptive control to achieve desired sound field characteristics at
any predetermined time before the start of music listening or during music listening Other than
that, the microphone can be kept away from the listener's head and not get in the way as an
obstacle.
[0012]
In addition, electro-acoustic conversion is performed by a headphone speaker provided in the
headphone, and the microphone is fixed to the headphone and picks up a sound near the
entrance of the ear canal.
Thereby, even at the time of headphone reproduction, regardless of the individual difference of
the transfer function of the listener's pinna, it is possible to accurately realize the listener's
desired sound field.
10-04-2019
4
[0013]
Further, when there are a plurality of channels of input signals, the one or more sound
reproducing means are provided for each channel of the input signal, and the reference
processing means individually processes the input signal of each channel with a desired transfer
function After that, they are added to obtain a reference signal.
As a result, adaptive control of the transfer function of each processing means can be
simultaneously performed in parallel even for an input having a plurality of channels, such as a
stereo music signal, for example. It can be realized.
[0014]
In addition, the reference processing means allows the listener to select one desired transfer
function from among a plurality of predetermined transfer functions prepared in advance.
Thereby, the sound field characteristics can be variously changed according to the genre of
music, the type of music source, and the like.
[0015]
Also, the electro-acoustic conversion means is placed behind the listener's ear, below the height
of the ear, downward or upward. This makes the localization of the sound image unclear and
makes it possible not to be aware of the presence of the electro-acoustic conversion means.
[0016]
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a general circuit diagram
of an adaptive sound field control apparatus according to a first embodiment of the present
invention. An audio source device 10 such as a CD player and a tuner has two output terminals
LOUT and ROUT capable of outputting two-channel stereo audio signals of L and R. Reference
numerals 20 and 30 are sound reproducing means for reproducing sound based on input signals
from two terminals of LOUT and ROUT from the audio source device, of which 21 and 31 are A /
D converters, and 22 and 32 are It is an adaptive filter composed of FIR digital filters with
10-04-2019
5
variable filter coefficients and order I, and processes the outputs of the A / D converters 21 and
31 with variable transfer characteristics (transfer functions). 23 and 33 are D / A converters, 24
and 34 are power amplifiers, and 25 and 35 are L and R speakers (electrical-acoustic conversion
means), respectively.
[0017]
The two monitor microphones 40 and 50 are placed near the entrance of the ear canal of both
listeners' ears, and monitor the sound heard by the listener with flat frequency characteristics
(frequency-response characteristics). 41 and 51 are microphone amplifiers, 42 and 52 are A / D
converters, and 60 is a transfer characteristic (transfer function) under which the listener wants
to hear the audio signal output from the LOUT terminal. The reference processing means, which
sets a desired transfer characteristic (transfer function) as a reference, listens to what transfer
characteristic (transfer function) the audio signal output from the ROUT terminal is to be heard
with the right ear It is a reference processing means in which a transfer characteristic (transfer
function) desired by a person is set as a reference.
[0018]
The reference processing means 60 is composed of a delay unit 61 for delaying the output of the
A / D converter 21 with a delay time of ? L1 and an FIR digital filter 62 connected to the output
side of the delay unit. The reference processing means 70 is composed of a delay unit 71 for
delaying the output of the A / D converter 31 with a delay time ?Rr, and an FIR digital filter 72
connected to the output side of the delay unit.
[0019]
The sound based on the audio signal output from the LOUT terminal of the audio source device
10 should reach the left ear simultaneously with the sound based on the audio signal output
from the ROUT terminal reaching the right ear , ?Ll = ?Rr. The delay times .tau.L1 and .tau.Rr
are assumed to be more than the time the sound travels over the distance L between the speaker
farthest from the listener and the listener, assuming that the listener listens farthest from the
speakers 25 and 35. Value. For example, assuming that L = 3 m and the sound velocity at room
temperature = 340 m / s, it suffices to set ?Ll = ?Rr> 8.8 ms. The digital filters 62 and 72 have
transfer characteristics (transfer functions) having arbitrary frequency-gain characteristics
10-04-2019
6
desired by the listener, such as flat, low boost, high boost, vocal band emphasis, etc. As an
example, it is assumed that the flat frequency characteristic of 0 dB is set over the entire band.
[0020]
43 is a subtractor for calculating the deviation between the output of the A / D converter 42 and
the output of the reference processing means 60, and 53 is a subtraction for calculating the
deviation between the output of the A / D converter 52 and the output of the reference
processing means 70 An adaptive control unit 26 uses the output of the LOUT terminal as a
reference signal from the output of the LOUT terminal after A / D conversion and the output of
the subtractor 43 so that the output of the subtractor 43 becomes smaller. The filter coefficient
of the speaker is updated so that the sound from the speaker 25 can obtain a desired frequencyresponse characteristic (here, flat) at the entrance of the ear canal of the listener's left ear. An
adaptive control unit 36 is a filter of the adaptive filter 32 so that the output of the ROUT
terminal becomes a reference signal from the output of the ROUT terminal after A / D conversion
and the output of the subtractor 53 and the output of the subtractor 53 becomes small. The
coefficients are updated so that the sound from the speaker 35 provides the desired frequencyresponse characteristics (here flat) at the entrance to the ear canal of the listener's right ear.
[0021]
Based on the output from the LOUT terminal and the output from the subtractor 43, the adaptive
control unit 26 updates the filter coefficients of the adaptive filter 22 so that the output from the
subtractor 43 can be minimized using the output from the LOUT terminal as a reference signal.
An adaptive filter that minimizes the evaluation function of, for example, Je = {el (n)} 2 from the
discrete signal x L (n) output from the A / D converter 21 and the discrete signal el output from
the subtractor 43 The filter coefficients of 22 are determined, for example, by the LMS algorithm,
which is one of the steepest descent methods, and are updated and set in the adaptive filter 22.
[0022]
Assuming that the adaptive filter 22 is composed of an I-order FIR filter, the filter coefficient of
the ith filter of the adaptive filter 22 is wLi (where i = 0 to I), from the output point of the
adaptive filter 22 to the monitor microphone 40 Let cLl, j be the j-th filter coefficient when the
transfer function CLl of the transmission line of the above is embodied by a Jth-order FIR filter
10-04-2019
7
(where j = 0 to J), the adaptive control unit 26 performs adaptation at time tn Assuming that the
ith filter coefficient of the filter 22 is wLi (n), the filter coefficient wLi (n + 1) at time tn + 1 can be
expressed by the following equation: wLi (n + 1) = wLi (n) + ? и ?el (n) и qLl * ( n?i)... (1) ?: a
predetermined convergence coefficient, ?: a filter coefficient update setting is performed on the
adaptive filter 22 according to a predetermined weighting coefficient. The initial value wLi (0) of
the filter coefficient wLi is set to a predetermined value. Further, the filter coefficient cLl, j is to
be subjected to a predetermined calculation from the response signal of the output point A 'of the
A / D converter 42 when the M-sequence noise data is injected to the input point A of the D / A
converter 23. We identify with and ask for. The adaptive control unit 26 includes a built-in
memory (not shown) that temporarily stores filter coefficients at the time when adaptive control
is completed.
[0023]
Based on the output from the ROUT terminal and the output of the subtractor 53, the adaptive
control unit 36 updates the filter coefficients of the adaptive filter 32 so that the output of the
subtractor 53 can be minimized using the output from the ROUT terminal as a reference signal.
The adaptive filter 32 that minimizes the evaluation function of Je = {er (n)} 2 from the discrete
signal x (n) output from the A / D converter 31 and the discrete signal er output from the
subtractor 53 The filter coefficients are determined, for example, by the LMS algorithm, which is
one of the steepest descent methods, and are updated and set in the adaptive filter 32.
[0024]
Assuming that the adaptive filter 32 is an I-order FIR filter, the i-th filter coefficient of the
adaptive filter 32 is wRi (where i = 0 to I), from the output point of the adaptive filter 32 to the
monitor microphone 50 Let cRr, j be the j-th filter coefficient when the transfer function CRr of
the transmission line of the above is embodied by a J-th FIR filter (where j = 0 to J), the adaptive
control unit 36 performs adaptation at time tn Assuming that the ith filter coefficient of the filter
32 is wRi (n), the filter coefficient wRi (n + 1) at time tn + 1 is expressed by the following
equation: wRi (n + 1) = wRi (n) + ? и ?er (n) и qRr * n?i) иии (2) ?: a predetermined convergence
coefficient, ?: a filter coefficient update setting is performed on the adaptive filter 32 according
to a predetermined weighting coefficient. The initial value wRi (0) of the filter coefficient wRi is
set to a predetermined value. Further, the filter coefficient cRr, j is to be subjected to a
predetermined calculation from the response signal of the output point B 'of the A / D converter
52 when the M-sequence noise data is injected to the input point B of the D / A converter 33. We
identify with and ask for.
10-04-2019
8
[0025]
Reference numeral 80 denotes a system controller. When sound field correction is instructed by
the hand controller 81, first, a filter coefficient cLl, j for determining a transfer function CLl and a
filter coefficient cRr, j for determining a transfer function CRr are identified, and adaptive control
is performed. Set in sections 26 and 36. Next, the adaptive control units 26 and 36 are controlled
to sequentially perform adaptive control on the Lch side and the Rch side.
[0026]
Next, the operation of the first embodiment described above will be briefly described. Here, it is
assumed that the audio source device 10 is outputting a stereo audio signal, and any filter
coefficient that is not all zero is set in the adaptive filters 22 and 32. When a sound field
correction instruction is given by the hand controller 81, the system controller 80 controls the
adaptive control units 26 and 36 to make the filter coefficients of the adaptive filters 22 and 32
all zero, and the Lch side and Rch side speakers Stop the sound output of 25 and 35. In this state,
M-sequence noise data is injected for a fixed time at point A, pseudo-random noise is emitted
from the speaker 25, and the response output at point A 'is analyzed to determine the filter
coefficient cLl, j that determines the transfer function CLl. It is determined and set in the adaptive
control unit 26. Subsequently, M-sequence noise data is injected at a point B for a predetermined
time, the M-sequence noise sound is emitted from the speaker 35, the response output at the
point B 'is analyzed, and the filter coefficient cRr, j for determining the transfer function CRr It is
determined and set in the adaptive control unit 36.
[0027]
Next, the adaptive control unit 26 is controlled to start adaptive control on the Lch side. The
adaptive control unit 26 first sets each filter coefficient of the adaptive filter 22 to wLi (0). The
LOUT terminal output of the audio source device 10 is A / D converted by the A / D converter
21, and predetermined digital signal processing is performed by the adaptive filter 22. Then,
after D / A conversion is performed by the D / A converter 23, power is amplified by the power
amplifier 24 and sound is converted by the speaker 25. The sound output from the speaker 25 is
picked up by the monitor microphone 40 installed near the entrance of the ear canal of the left
ear of the listener, amplified by the microphone amplifier 41 and then A / D converted by the A /
10-04-2019
9
D converter 42 .
[0028]
On the other hand, the output of the A / D converter 21 is input to the reference processing
means 60, delayed by .tau.L1 by the delay unit 61, processed into a flat frequency characteristic
by the digital filter 62, and output as a reference signal. Then, the difference between the
reference signal and the output of the A / D converter 42 is obtained by the subtractor 43. After
initial setting of the filter coefficients for the adaptive filter 22, the adaptive control unit 26 uses
the output xL (n) of the A / D converter 21 and the deviation output el (n) from the subtractor 43
according to (1) The filter coefficients of the adaptive filter 22 are updated in real time such that
the square value of el (n) is minimized. A certain time has passed since the start of adaptive
control, and adaptive control is completed when the square value of el (n) becomes smaller than
a certain value. At this time, the sound radiated from the speaker 25 has a flat frequencyresponse characteristic at the entrance of the ear canal of the left ear of the listener.
[0029]
However, since adaptive control on the Rch side is not performed yet, adaptive control unit 26
can make adaptive control on the Rch side, so when the square value of el (n) becomes smaller
than a certain value, adaptive control unit 26 at that time After storing the filter coefficients wLi
(i = 0 to I) set in the adaptive filter 22 in the built-in memory (not shown), all the filter
coefficients of the adaptive filter 22 are changed to zero to stop the sound output of Lch And
outputs an adaptive control completion notification to the system controller 80. The system
controller 80 that has received the notification controls the adaptive control unit 36 to start the
adaptive control on the Rch side. The adaptive control unit 36 first sets each filter coefficient of
the adaptive filter 32 to wRi (0). The output of the ROUT terminal of the audio source device 10
is A / D converted by the A / D converter 31, and predetermined digital signal processing is
performed by the adaptive filter 32. Then, after being D / A converted by the D / A converter 33,
the power is amplified by the power amplifier 34 and sound is converted by the speaker 35. The
sound output from the speaker 35 is picked up by the monitor microphone 50 installed near the
entrance of the ear canal of the left ear of the listener, amplified by the microphone amplifier 51
and then A / D converted by the A / D converter 52 .
[0030]
10-04-2019
10
On the other hand, the output of the A / D converter 31 is input to the reference processing
means 70, delayed by .tau.Rr by the delay unit 71, processed into flat frequency characteristics
by the digital filter 72, and output as a reference signal. Then, the difference between the
reference signal and the output of the A / D converter 52 is obtained by the subtractor 53. The
adaptive control unit 36 initializes the filter coefficients of the adaptive filter 32 and then uses
the output xR (n) of the A / D converter 31 and the deviation output er (n) from the subtracter 53
according to (2) , Er (n) are updated in real time so that the square value of er (n) is minimized. A
certain time has passed since the start of adaptive control, and adaptive control is completed
when the square value of er (n) becomes smaller than a certain value. At this time, the sound
radiated from the speaker 35 has a flat frequency-response characteristic at the entrance of the
ear canal of the right ear of the listener.
[0031]
The adaptive control unit 36 outputs an adaptive control completion notification to the system
controller 80 when the square value of er (n) becomes smaller than a predetermined value. The
system controller 80 that has received the notification instructs the adaptive control unit 26 to
reset the filter coefficient. The adaptive control unit 26 that has received the instruction sets the
filter coefficient stored in the built-in memory in the adaptive filter 22 and returns the Lch side to
the state where the adaptive control is completed. As described above, the sound emitted from
the Lch speaker 25 has a flat frequency characteristic at the entrance of the ear canal of the left
ear of the listener, and the sound emitted from the Rch speaker 35 also has a flat frequency
characteristic at the entrance of the right ear external canal So, from here onwards, the listener
will listen regardless of what frequency-response characteristics the speakers 25 and 35 have,
what transfer function the sound reproduction space has, and what head transfer function the
listener has. A sound field having a desired flat frequency characteristic can be accurately
realized. Therefore, even if the diameter of the speakers 25 and 35 is small and the response in
the low band is bad, the bass can be heard on a grand scale. Further, even in the case where
there is a tendency that the vocal band of 500 Hz to 1 kHz falls due to the undulation of the
frequency characteristic of the acoustic space as in the passenger compartment, the vocal band
can be clearly heard. Also, even if the speakers 25 and 35 are provided on the left and right
doors of the vehicle compartment and the distance from the listener (driver) to each of the
speakers 25 and 35 is different, the listener's The arrival time and level of the sound to the left
and right ears can be matched, resulting in a natural stereo sound image. Also, if the listener
changes, or the listener changes the listening position even if the listener is the same, reset the
monitor microphones 40 and 50 to a new listener or a new listening position, and instruct sound
field correction with the hand controller 81 If so, it becomes possible to listen to a new listener
or a new listening position with desired sound field characteristics.
10-04-2019
11
[0032]
Further, as shown in FIG. 2, the speakers 25 and 35 of the sound reproducing means 20 and 30
are replaced with the headphone speakers 101 and 102 of the stereo headphone 100, and the
monitor microphones 40 and 50 are fixed to the stereo headphone 100. It may be located near
the ear canal entrances 105 and 106 of 104. Thereby, even at the time of headphone
reproduction, regardless of the individual difference in the transfer characteristic of the listener's
pinna, it is possible to accurately realize the listener's desired sound field. If the stereo
headphone 100 is a closed type, the sound emitted from the headphone speaker 101 (102) does
not reach the headphone speaker 102 (101) on the opposite side. Therefore, when performing
sound field control, the system controller 80 of FIG. 1 can simultaneously perform adaptive
control on the adaptive filters 22 and 32 while continuing sound output for both Lch and Rch.
[0033]
That is, when the listener puts a headphone 100 on the ear and a stereo audio signal is output
from the audio source device 10, the system controller 80 controls the adaptive control unit
when an instruction for sound field correction is given by the hand controller 81. 26 and 36 are
controlled to set each filter coefficient of the adaptive filters 22 and 32 to 0, and the M-sequence
noise data generated inside the system controller 80 is injected at points A and B for a certain
period of time. M-series noise is emitted. The sound emitted from the speaker 101 is picked up
only by the monitor microphone 40, and the sound emitted from the speaker 102 is picked up
only by the monitor microphone 50. Then, the transfer function CLl is analyzed from the
response output at point A ', the filter coefficient cLl, j is identified and set in the adaptive control
unit 26, the transfer function CRr is analyzed from the response output at point B', and the filter
coefficient cRr , J are identified and set in the adaptive control unit 36. Then, the injection of Msequence noise data is stopped. Next, the adaptive control unit 26 is controlled to start adaptive
control of the Lch-side adaptive filter 22, and at the same time, the adaptive control unit 36 is
controlled to start adaptive control of the Rch-side adaptive filter 32. Then, the adaptive control
units 26 and 36 set predetermined initial filter coefficients wLi (0) and wRi (0) in the adaptive
filters 22 and 32, respectively. At this time, the sound emitted from the headphone speaker 101
enters only the left ear and is picked up by the monitor microphone 40, and a deviation el from
the output of the reference processing means 60 is output from the subtractor 43. Further, the
sound emitted from the headphone speaker 102 enters only the right ear and is picked up by the
monitor microphone 50, and the difference er with the output of the reference processing means
70 is output from the subtractor 53. After initialization, the adaptive control units 26 and 36
10-04-2019
12
perform adaptive control on the adaptive filters 22 and 32 until el and er become smaller than or
equal to a predetermined size, and when el and er become smaller than or equal to a fixed size.
The filter coefficients of the adaptive filters 22 and 32 may be fixed, and control may be
completed (the adaptive control may be continued even after el and er have become smaller than
or equal to a certain size).
[0034]
FIG. 3 is a circuit diagram of a vehicle-mounted adaptive sound field control apparatus according
to a second embodiment of the present invention. The same reference numerals as in FIG. 1
denote the same parts in FIG. In the embodiment of FIG. 1, the sound emitted from the Lch-side
speaker can be listened to the desired sound field characteristic with the left ear of the listener,
and the sound emitted from the Rch-side speaker can be desired with the right ear of the listener
Although it was made possible to listen by characteristics, the embodiment of FIG. 3 can listen to
the sound emitted from the Lch side speaker with desired sound field characteristics by both the
left and right ears of the listener, and from the Rch side speaker The emitted sound can be
listened to a desired sound field characteristic in both the right ear and the left ear of the listener.
[0035]
In the reference processing means 60A, compared with FIG. 1, a delay unit 63 of delay time
.tau.R1 provided in series on the output side of the A / D converter 31 and an FIR type digital
connected to the output side of the delay unit. A filter 64, a gain adjusting multiplier 65
(multiplication coefficient gR1) connected to the output side of the digital filter, and a switch 66
for switching the output of the digital filter 62 and the multiplier 65 are added. The delay unit
63, the digital filter 64, and the multiplier 65 set the transfer characteristic desired by the
listener based on what transfer characteristic (transfer function) the audio signal output from the
ROUT terminal is desired to be heard by the left ear. It is to do. Further, in the reference
processing means 70A, a delay unit 73 of delay time ? Lr provided in series on the output side
of the A / D converter 21 as compared with FIG. 1, and an FIR type connected to the output side
of the delay unit. A digital filter 74, a multiplier for gain adjustment 75 (multiplication coefficient
gLr) connected to the output side of the digital filter, and a switch 76 for switching the output of
the digital filter 72 and the multiplier 75 are added. The delay unit 73, the digital filter 74, and
the multiplier 75 set the transfer characteristic desired by the listener based on what transfer
characteristic (transfer function) the audio signal output from the LOUT terminal is desired to be
heard by the right ear. It is to do.
10-04-2019
13
[0036]
The sound based on the audio signal output from the LOUT terminal of the audio source device
10 should reach the left ear simultaneously with the sound based on the audio signal output
from the ROUT terminal reaching the right ear The delay time ? L1 at the delay unit 61 is made
equal to the delay time ? Rr of the delay unit 71. The delay times .tau.L1 and .tau.Rr are
assumed to be more than the time the sound travels over the distance L between the speaker
farthest from the listener and the listener, assuming that the listener listens farthest from the
speakers 25 and 35. It is as good value. For example, assuming that L = 3 m and the sound
velocity at room temperature = 340 m / s, it suffices to set ?Ll = ?Rr> 8.8 ms. On the other
hand, assuming that the arrival time difference between the monitor microphones 50 and 40 for
the sound emitted from the speaker 35 is ??R1, and the arrival time difference between the
monitor microphones 40 and 50 for the sound emitted from the speaker 25 is ??Lr, then ?Rl
= ?Rr + ??Rl, ?Lr = It is set as ?Ll + ??Lr. The ratio of the sound reception level at the
monitor microphone 40 to the sound reception level at the monitor microphone 50 of the sound
emitted from the speaker 35 is mR1, and the monitor microphone 50 for the sound reception
level at the monitor microphone 40 of the sound emitted from the speaker 25. Assuming that the
ratio of the sound reception level in the case of (mLr) is gRL = mRl and gLr = mLr.
[0037]
When the positional relationship between the speakers 25 and 35 and the listener is fixed, the
monitor microphones 40 and 50 are attached to the entrance to the ear canal of the dummy
head in advance in ??R1 and mR1 and a pulse sound of a certain level is emitted from the
speaker 35 The value obtained by measuring the arrival time difference between the monitor
microphones 50 and 40 and the sound reception level difference at the time of setting is set as
an approximation value, and ?? Lr and mLr are pulse sound of a certain level from the speaker
25. The value obtained by measuring the arrival time difference between the monitor
microphones 40 and 50 and the sound reception level difference when the light is emitted is set
as an approximate value. The digital filters 62, 64, 72, 74 have transfer characteristics (transfer
functions) with any frequency-gain characteristics desired by the listener, such as flat, low boost,
high boost, vocal band emphasis, etc. Here, as an example, it is assumed that the flat frequency
characteristic of 0 dB is set over the entire band.
[0038]
10-04-2019
14
The adaptive control unit 26A uses the output of the LOUT terminal as a reference signal from
the output of the LOUT terminal after A / D conversion and the outputs of the subtractors 43 and
53 so that both the outputs of the subtractors 43 and 53 decrease. The filter coefficients of 22
are updated so that the sound from the speaker 25 provides the desired frequency-response
characteristics (here flat) at the entrance to the ear canal of the listener's left and right ears. From
the discrete signal x L (n) output from the A / D converter 21 and the discrete signal el output
from the subtractor 43 and the discrete signal er output from the subtracter 53, for example, Je =
{el (n) The filter coefficient of the adaptive filter 22 that minimizes the evaluation function of 2 +
{er (n)} 2 is determined by, for example, the LMS algorithm, which is one of the steepest descent
methods, and updated in the adaptive filter 22.
[0039]
Assuming that the adaptive filter 22 is composed of an I-order FIR filter, the filter coefficient of
the ith filter of the adaptive filter 22 is wLi (where i = 0 to I), from the output point of the
adaptive filter 22 to the monitor microphone 40 Filter coefficient of the jth order when the
transfer function CLl of the transmission line is embodied by the Jth-order FIR filter, cLl, j, the
transfer function CLr of the transmission line from the output point of the adaptive filter 22 to
the monitor microphone 50 Jth Assuming that the j-th filter coefficient when embodied by the
FIR filter of this embodiment is cLr, j (where j = 0 to J), the adaptive control unit 26A determines
the i-th filter coefficient of the adaptive filter 22 at time tn. As wLi (n), the filter coefficient wLi (n
+ 1) at time tn + 1 is expressed by the following equation: wLi (n + 1) = wLi (n) + ? и ?el (n) и qLl
* (n?i) + ? и ?er (n) QLr * (n?i) (3) ?: predetermined convergence coefficient, ?: update
setting of the filter coefficient is performed on the adaptive filter 22 in accordance with the
predetermined weighting coefficient. The initial value wLi (0) of the filter coefficient wLi is set to
a predetermined value. Further, the filter coefficient cLl, j is to be subjected to a predetermined
calculation from the response signal of the output point A 'of the A / D converter 42 when the Msequence noise data is injected to the input point A of the D / A converter 23. We identify with
and ask for. The adaptive control unit 26A includes a built-in memory (not shown) that
temporarily stores filter coefficients at the time when adaptive control is completed.
[0040]
The adaptive control unit 36A uses the output of the ROUT terminal as a reference signal from
the output of the ROUT terminal after A / D conversion and the outputs of the subtractors 53 and
10-04-2019
15
43 so that both the outputs of the subtractors 53 and 43 decrease. The filter coefficients of 32
are updated so that the sound from the speaker 35 provides the desired frequency-response
characteristics (here flat) at the entrance to the ear canal of the listener's left and right ears. From
the discrete signal xR (n) output from the A / D converter 31 and the discrete signal el output
from the subtractor 43 and the discrete signal er output from the subtractor 53, for example, Je
= {el (n) The filter coefficients of the adaptive filter 32 that minimize the evaluation function of 2
+ {er (n)} 2 are determined by, for example, the LMS algorithm, which is one of the steepest
descent methods, and updated in the adaptive filter 32.
[0041]
Assuming that the adaptive filter 32 is an I-order FIR filter, the i-th filter coefficient of the
adaptive filter 32 is wRi (where i = 0 to I), from the output point of the adaptive filter 32 to the
monitor microphone 50 Filter coefficient of the jth order when the transfer function CRr of the
transmission line is embodied by the J-th FIR filter, cRr, j, the transfer function CRl of the
transmission line from the output point of the adaptive filter 22 to the monitor microphone 40
Jth Assuming that the j-th filter coefficient when embodied by the FIR filter of c is cRl, j (where j =
0 to J), the adaptive control unit 36A determines the i-th filter coefficient of the adaptive filter 32
at time tn. As wRi (n), the filter coefficient wRi (n + 1) at time tn + 1 is expressed by the following
equation: wRi (n + 1) = wRi (n) + ? и ?er (n) и qRr * (n?i) + ? и ?el (n) QRl * (n-i) (4) ?: a
predetermined convergence coefficient, ?: a filter coefficient update setting is performed on the
adaptive filter 32 according to a predetermined weighting coefficient. The initial value wRi (0) of
the filter coefficient wRi is set to a predetermined value. Further, the filter coefficient cRr, j is to
be subjected to a predetermined calculation from the response signal of the output point B 'of the
A / D converter 52 when the M-sequence noise data is injected to the input point A of the D / A
converter 33. The filter coefficient cRl, j is determined from the response signal of the output
point A 'of the A / D converter 42 when the M-sequence noise data is injected to the input point
A of the D / A converter 33. Identification is obtained by performing an operation.
[0042]
When sound field correction is instructed by the hand controller 81, the system controller 80A
first determines the filter coefficient cLl, j that determines the transfer function CLl, the filter
coefficient cLr, j that determines the transfer function CLr, and the filter coefficient that
determines the transfer function CRr The filter coefficients cRl, j for determining cRr, j and the
transfer function CRl are identified and set in the adaptive control units 26A and 36A. Next, the
adaptive control units 26A and 36A and the reference processing means 60A and 70A are
10-04-2019
16
controlled to sequentially perform adaptive control on the Lch side and the Rch side. The other
components are configured the same as in FIG.
[0043]
Next, the operation of the second embodiment described above will be briefly described. Here, it
is assumed that the audio source device 10 is outputting a stereo audio signal, and any filter
coefficient that is not all zero is set in the adaptive filters 22 and 32. Further, adaptive control of
the adaptive filters 22 and 32 is performed in order. When a sound field correction instruction is
given by the hand controller 81, the system controller 80A controls the adaptive control units
26A and 36A to make all the filter coefficients of the adaptive filters 22 and 32 zero, and the Lch
side and Rch side speakers Stop the sound output of 25 and 35. In this state, M-sequence noise
data is injected at a point A for a fixed time, an M-sequence noise sound is emitted from the
speaker 25, and a response output at the point A 'is analyzed to determine a filter coefficient cLl,
j that determines a transfer function CLl. Then, the response output at point B 'is analyzed to
obtain a filter coefficient cLr, j for determining the transfer function CLr, which is set in the
adaptive control unit 26A. Subsequently, M-sequence noise data is injected at a point B for a
fixed time, M-sequence noise sound is emitted from the speaker 35, the response output at the
point B 'is analyzed, and a filter coefficient cRr, j for determining a transfer function CRr is
obtained. Also, the response output at the point A 'is analyzed to obtain a filter coefficient cRl, j
for determining a transfer function CRl, which is set in the adaptive control unit 36A.
[0044]
Next, the switches 66 and 76 of the reference processing means 60A and 70A are on the L side,
and a reference transfer characteristic is set for listening to the sound based on the audio signal
on the Lch side with desired sound field characteristics with the left and right ears. Do. Then, the
control unit 26A controls the adaptive control unit 26A to start adaptive control on the Lch side.
The adaptive control unit 26A first sets each filter coefficient of the adaptive filter 22 to wLi (0).
The LOUT terminal output of the audio source device 10 is A / D converted by the A / D
converter 21, and predetermined digital signal processing is performed by the adaptive filter 22.
Then, after D / A conversion is performed by the D / A converter 23, power is amplified by the
power amplifier 24 and sound is converted by the speaker 25. The sound output from the
speaker 25 is picked up by monitor microphones 40 and 50 installed near the entrance of the
ear canal of the left and right ears of the listener, amplified by the microphone amplifiers 41 and
51 respectively, and then A / D converted. A / D conversion is performed by the units 42 and 52.
10-04-2019
17
[0045]
On the other hand, the output of the A / D converter 21 is input to the reference processing
means 60A, delayed by ?L1 by the delay unit 61, processed into flat frequency characteristics
by the digital filter 62, and processed as a reference signal for the left ear It is output. Then, the
difference between the reference signal and the output of the A / D converter 42 is obtained by
the subtractor 43. The output of the A / D converter 21 is input to the reference processing
means 70A, delayed by ?Lr (= ?Ll + ??Lr) by the delay unit 71, processed into a flat
frequency characteristic by the digital filter 74, and further multiplied by the multiplier 75 Is
multiplied by the multiplier gLr and output as a reference signal for the right ear. Then, the
difference between the reference signal and the output of the A / D converter 52 is obtained by
the subtractor 53. After initial setting of the filter coefficients to the adaptive filter 22, the
adaptive control unit 26A uses xL (n), el (n) and er (n) to set the square value of el (n) according
to (3). The filter coefficients of the adaptive filter 22 are updated in real time so that the sum
with the square value of er (n) is minimized. A certain time has passed since the start of adaptive
control, and the adaptive control is completed when the sum of the square value of el (n) and the
square value of er (n) becomes smaller than a certain value. At this time, the sound radiated from
the speaker 25 has a flat frequency-response characteristic at the entrance of the ear canal of the
left and right ears of the listener. Also, since the sound reaches the right ear delayed by ?? Lr
from the left ear and the level is reduced to gLr times, a good sense of direction can be obtained.
[0046]
However, since adaptive control on the Rch side has not been performed yet, the adaptive control
unit 26A adds the square value of el (n) and the square value of er (n) to enable adaptive control
on the Rch side. When it becomes smaller than a certain value, the filter coefficient wLi (i = 0 to I)
set in the adaptive filter 22 at that time is stored in the built-in memory (not shown), and then
the filter coefficient of the adaptive filter 22 is stored. All are changed to zero to stop the sound
output of Lch, and an adaptive control completion notification is output to the system controller
80A. The system controller 80A that has received the notification sets the switches 76 and 66 of
the reference processing means 70A and 60A to the R side, and listens to the sound based on the
audio signal on the Rch side with the desired sound field characteristics with the right and left
ears. Set the reference transfer characteristics for Then, it controls the adaptive control unit 36A
to start adaptive control on the Rch side. The adaptive control unit 36A first sets each filter
coefficient of the adaptive filter 32 to wRi (0). The output of the ROUT terminal of the audio
source device 10 is A / D converted by the A / D converter 31, and predetermined digital signal
10-04-2019
18
processing is performed by the adaptive filter 32. Then, after being D / A converted by the D / A
converter 33, the power is amplified by the power amplifier 34 and sound is converted by the
speaker 35. The sound output from the speaker 35 is picked up by monitor microphones 50 and
40 installed near the entrance of the ear canal of the listener's right and left ears, amplified by
the microphone amplifiers 51 and 41 respectively, and then A / D converted. A / D conversion is
performed by the units 52 and 42.
[0047]
On the other hand, the output of the A / D converter 31 is input to the reference processing
means 70A, delayed by ?Rr by the delay unit 71, processed to a flat frequency characteristic by
the digital filter 72, and processed as a reference signal for the right ear It is output. Then, the
difference between the reference signal and the output of the A / D converter 52 is obtained by
the subtractor 53. The output of the A / D converter 21 is input to the reference processing
means 60A, delayed by ?R1 (= ?Rr + ??R1) by the delay unit 61, processed into a flat
frequency characteristic by the digital filter 64, and further multiplied by the multiplier 65 The
multiplier gR1 is multiplied by the signal and output as a reference signal for the left ear. Then,
the difference between the reference signal and the output of the A / D converter 42 is obtained
by the subtractor 43. After the initial setting of the filter coefficients to the adaptive filter 32, the
adaptive control unit 36A uses xR (n), el (n) and er (n) to set the square value of eR (n) according
to (4). The filter coefficients of the adaptive filter 32 are updated in real time so that the sum
with the square value of e L (n) is minimized. After a certain time has passed since the start of the
adaptive control, the adaptive control is completed when the sum of the square value of eR (n)
and the square value of eL (n) becomes smaller than a certain value. At this time, the sound
radiated from the speaker 35 has a flat frequency-response characteristic at the entrance of the
ear canal of the right and left ears of the listener. Also, since the sound reaches the left ear
delayed by ??R1 from the right ear and the level is reduced by a factor of gR1, a good sense of
direction of the sound image can be obtained. When the square value of er (n) + el (n) becomes
smaller than a predetermined value, the adaptive control unit 36A outputs an adaptive control
completion notification to the system controller 80A. The system controller 80A that has
received the notification instructs the adaptive control unit 26A to reset the filter coefficient. The
adaptive control unit 26A that has received the instruction sets the filter coefficient stored in the
built-in memory in the adaptive filter 22, and returns the Lch side to the state where the adaptive
control is completed.
[0048]
10-04-2019
19
As described above, the sound from Lch speaker 25 has a flat frequency characteristic at the
entrance of the ear canal of the left and right ears of the listener, and the sound from Rch
speaker 35 is also at the entrance of the ear canal of right and left ears. Since the frequency
characteristics become flat, the listener subsequently determines what frequency-response
characteristics the speakers 25 and 35 have, what transfer characteristics the sound
reproduction space has, and what head transfer characteristics the listener has Regardless, it is
possible to more accurately realize a sound field having a flat frequency characteristic desired by
the listener. Also, even if the speakers 25 and 35 are provided on the left and right doors of the
vehicle compartment and the distance from the listener (driver) to each of the speakers 25 and
35 is different, the listener's The arrival time and level of the sound to the left and right ears can
be made equal, and in addition, the sound from the speaker 25 arrives at the left ear delayed
from the right ear and at a smaller level and comes out from the speaker 35 The sound arrives at
the right ear delayed from the left ear and at a reduced level, so a very natural stereo image can
be obtained.
[0049]
In the second embodiment described above, the adaptive control on the Lch side and the Rch side
is sequentially performed, but the time required for sound field correction may be shortened by
simultaneously performing the adaptive control. In this case, the adaptive sound field control
device of FIG. 3 is modified as shown in FIG. 4 and in the reference processing means 60B and
70B, the output of the digital filter 62 and the output of the multiplier 65 are added by the adder
67 and a subtractor 43 By outputting the sound data to the Lch and Rch from the speakers 25
and 35, the reference transfer characteristic is set to simultaneously listen to the sound of the
Lch and Rch radiated from the speakers 25 and 35 with a desired sound field characteristic. Also,
by adding the output of the digital filter 72 and the output of the multiplier 75 by the adder 77
and outputting the result to the subtractor 53, the Rch and Lch sounds radiated from the
speakers 35 and 25 are simultaneously desired by the right ear. Set the reference transfer
characteristic for listening in the sound field characteristic.
[0050]
In the configuration of FIG. 4, when an instruction for sound field correction is given by the hand
controller 81, the system controller 80A controls the adaptive control units 26A and 36A to
make all of the filter coefficients of the adaptive filters 22 and 32 zero. The sound output of the
speakers 25 and 35 on the side and the Rch side is stopped. Then, as in the case of FIG. 3, Msequence noise data is injected at points A and B for a predetermined time, and the response
10-04-2019
20
outputs of points A ? and B ? are analyzed to determine transfer functions CLl and CLr. cLl, j,
find the filter coefficient cLr, j that determines the transfer function CLr, set it in the adaptive
control unit 26A, find the filter coefficient cRr, j that determines the transfer function CRr, and
find the filter coefficient cRl, j that determines the transfer function CRl, apply Set in the control
unit 36A.
[0051]
Next, the adaptive control units 26A and 36A are controlled to simultaneously start adaptive
control on the Lch side and the Rch side. The adaptive control unit 26A first sets each filter
coefficient of the adaptive filter 22 to wLi (0). The LOUT terminal output of the audio source
device 10 is A / D converted by the A / D converter 21, and predetermined digital signal
processing is performed by the adaptive filter 22. Then, after D / A conversion is performed by
the D / A converter 23, power is amplified by the power amplifier 24 and sound is converted by
the speaker 25. Furthermore, the adaptive control unit 36A first sets each filter coefficient of the
adaptive filter 32 to wRi (0). The output of the ROUT terminal of the audio source device 10 is A
/ D converted by the A / D converter 31, and predetermined digital signal processing is
performed by the adaptive filter 32. Then, after being D / A converted by the D / A converter 33,
the power is amplified by the power amplifier 34 and sound is converted by the speaker 35. The
sounds output from the speakers 25 and 35 are picked up by the monitor microphones 40 and
50, respectively amplified by the microphone amplifiers 41 and 51, and then A / D converted by
the A / D converters 42 and 52.
[0052]
The sound components of Lch and Rch are added to the output of the A / D converter 42, and the
output of the processing means 60B for reference is subtracted by the subtractor 43 to obtain
the desired left ear in stereo sound reproduction. The deviation between the sound field
characteristic of and the actual sound field characteristic is determined. Further, the acoustic
component of Rch and the acoustic component of Lch are added to the output of A / D converter
52, and the output of processing means 70B for reference is subtracted by subtractor 53,
thereby reproducing stereo sound. The deviation between the desired and actual sound field
characteristics for the right ear is determined. After initial setting of the filter coefficients to the
adaptive filter 22, the adaptive control unit 26A uses xL (n), el (n) and er (n) to set the square
value of el (n) according to (3). The filter coefficients of the adaptive filter 22 are updated in real
time so that the sum with the square value of er (n) is minimized. After a certain time has passed
since the start of adaptive control, adaptive control is completed when the sum of the square
10-04-2019
21
value of el (n) and the square value of er (n) becomes smaller than a certain value (at this time,
adaptation) The control unit 26A may fix the filter coefficient of the adaptive filter 22 to stop the
adaptive control, but may continue. In parallel to this, after the adaptive control unit 36A
initializes the filter coefficients of the adaptive filter 32, it uses xL (n), er (n) and el (n) and uses er
(n) according to (3). The filter coefficients of the adaptive filter 32 are updated in real time so as
to minimize the sum of the squared value of y) and the squared value of el (n). After a certain
time has elapsed since the start of adaptive control, adaptive control is completed when the sum
of the square value of er (n) and the square value of el (n) becomes smaller than a certain value
(at this time, adaptive) The control unit 26A may fix the filter coefficient of the adaptive filter 22
to stop the adaptive control, but may continue. When the adaptive control of Lch and Rch is
completed, the sound radiated from the speaker 25 has a flat frequency-response characteristic
at the entrance of the ear canal of the listener's left ear and right ear. Also, since the sound
reaches the right ear delayed by ?? Lr from the left ear and the level is reduced to gLr times, a
good sense of direction can be obtained. Similarly, the sound radiated from the speaker 35 has a
flat frequency-response characteristic at the entrance to the ear canal of the right and left ears of
the listener. In addition, since the sound reaches the left ear delayed by ??R1 from the right ear
and the level is reduced to gR1 times, a very natural stereo sound image can be obtained.
[0053]
Here, in the example of FIG. 3 and FIG. 4, the delay time .tau.L1 of the delay unit 61 of the
reference processing unit 60A, 60B, the delay time .tau.R1 of the delay unit 63, the multiplication
coefficient gR1 of the multiplier 65, the reference processing unit 70A, The delay time ?Rr of
the delay unit 71 of 70 B, the delay time ?Lr of the delay unit 73, and the multiplication
coefficient gLr of the multiplier 75 are set to appropriate fixed values, but these can be made
variable and the listener instructs the sound field correction It is also possible to re-measure and
set each time it is done. Specifically, when the system controller 80A injects M-sequence noise
data at a point A for a certain period of time to obtain filter coefficients cL1 and cLr that
determine CL1 and CLr, the response delay time at the point A 'is measured The ?L1 of the
delay unit 61 is set, the response delay time at the point B ? is measured, and the ?R1 of the
delay unit 66 is set. Further, the ratio of the response output level at point B 'to the response
output level at point A' is measured to set the multiplication coefficient gR1 of the multiplier 65.
Similarly, when M-sequence noise data is injected for a fixed time at point B in order to obtain
filter coefficients cR1 and cR1 that determine CRr and CR1, the response delay time at point B 'is
measured and ?Rr of the delay unit 71 is calculated. Then, the response delay time at the point
A ? is measured to set ?R1 of the delay unit 76. Also, the ratio of the response output level at
point C ? to the response output level at point B ? is measured to set the multiplication
coefficient gLr of the multiplier 75. In this way, since the reference signal corresponding to the
shape of the head of each listener is obtained, the sound image localization becomes better, and
10-04-2019
22
the listener changes, or the listener changes the listening position even if the listener is the same.
In this case, a new listener or a new listening can be performed by setting the monitor
microphones 40 and 50 again with a new listener or setting them at a new listening position and
instructing the sound field correction with the hand controller 81. It becomes possible to listen
with desired sound field characteristics under good sound image localization even in position.
[0054]
In the first embodiment of FIG. 1 and the modification of FIG. 2 and the second embodiment of
FIG. 3 and the modification of FIG. 4 described above, the audio source device 10 receives stereo
audio signals (audio signals of 2 channels of L and R) Although the output case has been
described as an example, the present invention can be applied to the case where the same 1-ch
monaural audio signal is output from the LOUT terminal and the ROUT terminal. In this case, for
example, the sound reproduction means 30 on the ROUT terminal side, the monitor microphone
50, the microphone amplifier 51, the A / D converter 52, the subtracter 53, and the adaptive
control units 36 and 36A are omitted, and the reference processing means 60A, 60B. The
subtractor 63, the digital filter 64, the multiplier 65, the subtractor 73 in the reference
processing means 70A, 70B, the digital filter 74, the multiplier 75 are omitted, and only the
sound based on the audio signal from the LOUT terminal is Adaptive sound field control may be
performed.
[0055]
Also, in the example of FIGS. 1 to 4, the digital filters 62, 64, 72, 74 are fixed to flat frequencygain characteristics, but flat, low boost, high boost, vocal band emphasis (500 Hz to 1 kHz boost
It may be possible to selectively set one characteristic selected by the listener from among a
plurality of sound field characteristics such as various surround sound fields such as reflection
sound addition and / or reverberation addition. That is, the digital filters 62, 64, 72, 74 of the
reference processing means 60, 60A, 60B, 70, 70A, 70B can switch and set the filter coefficients,
and transmit multiple types to the digital filters 62, 64, 72, 74. A plurality of types of filter
coefficients for setting the characteristics (transfer function) are stored, for example, in the builtin memory (not shown) of the system controller 80, 80A. Then, when the listener performs an
operation to select one of the desired transfer characteristics with the on-hand operation device
81 before instructing the sound field correction, the system controller 80, 80A converts the
corresponding filter coefficient to the digital filter 62. , 64, 72, 74 are set. In this way, by
switching the transfer characteristics of the reference processing means 60, 60A, 60B, 70, 70A,
70B, not only the flat, but also the low boost and high according to the genre of music, the type
10-04-2019
23
of music source, etc. The sound field characteristics can be variously changed, such as boost,
vocal band emphasis (500 Hz to 1 kHz boost), reflection addition and / or reverberation addition.
[0056]
In each of the examples shown in FIGS. 1, 3 and 4, as shown in FIG. 5, two pivoting support
members 92 are provided rotatably on both sides of a headrest 91 provided at the upper end of
the seat 90, Monitor microphones 40 and 50 are provided at the tip of each rotation support
member 92, and a rotation transmission mechanism is connected to the rotation shaft 93 of the
rotation support member 92 to rotate the rotation support member 92 by receiving the
rotational force of the motor. . Then, the rotation support member 92 is made to stand by
upwards beforehand (state of the broken line E in FIG. 5), and when the listener instructs the
sound field correction, the system controller 80, 80A controls the motor to rotate The support
member 92 is turned 90 ░ forward to position the monitor microphones 40 and 50 respectively
near the entrance of the ear canal of the user's left and right ears. After that, the system
controller 80, 80A performs the identification of the required ones of the transfer functions CLl,
CLr, CRl, CRr and the adaptive sound field control in the same manner as described above, and
controls the motor to rotate after completion. The support member 92 is turned 90 ░ until it is
upward, and kept in the original position. In this way, the head can be moved undisturbed by the
monitor microphones 40 and 50 except during execution of adaptive control.
[0057]
In any of the examples shown in FIGS. 1, 3 and 4, when the speakers 25 and 35 are installed on
both sides behind the listener's ear, as shown in FIG. Of the enclosure 98 fixed to the lower end
of the hanger 97 fixed to the ceiling 96 as shown in FIG. Inside, the acoustic emission ports of
the speakers 25 and 35 are placed downward so that they are below the height of the ears, and
the diffuser 99 is placed under the cone paper to be omnidirectional (preferably in the horizontal
plane). It may be made omnidirectional in all directions). When the speakers 25 and 35 enter the
view of the listener, the position where the sound image is localized is drawn to the installation
location of the speakers 25 and 35. However, according to the example of FIGS. 6 and 7, even
though the speakers 25 and 35 are present just near the ears, they can not be made aware of
their existence, and the sound image can be localized only by the sound heard in both ears.
Therefore, desired sound field characteristics can be easily obtained. In FIG. 6, the diffuser 95 is
installed on the cone paper of the speakers 25 and 35 so as to be omnidirectional at least in the
horizontal plane, or the diffuser 99 is installed on the cone paper of the speakers 25 and 35 in
10-04-2019
24
FIG. If it becomes omnidirectional (preferably omnidirectional in all directions) at least in the
horizontal plane, the positions of the speakers 25, 25 become more difficult to understand, so the
presence of the speakers 25, 35 is made thinner. This makes it easier to obtain desired sound
field characteristics.
[0058]
Also, in the examples of FIGS. 1 and 3, instead of performing the sound field control using the
audio signal output from the audio source device 10, the system controller 80, 80A uses the Msequence noise data generated internally. The accuracy of the sound field control may be
improved. That is, before the start of audio output from the audio source device 10, when sound
field correction is instructed by the hand operation unit 81, the system controllers 80 and 80A
control the adaptive control units 26 and 26A to make the adaptive filter 22 The 0-order
coefficient is 1, the other coefficients are 0, and M-sequence noise data generated inside the
system controller 80, 80A is injected to the input side of the adaptive filter 22 (at this time, the 0order of the adaptive filter 22). Because the coefficient of is 1, the other coefficients are 0, which
is equivalent to the injection at point A), the speaker 25 emits an M-sequence noise sound. Then,
first, the transfer function CLl is analyzed from the response output at the point A ', and the filter
coefficient cLl, j is identified (in the case of FIG. 3, the transfer function CLr is further analyzed
from the response output at the point B', and the filter coefficient cLr , J) are set in the adaptive
control unit 26, 26A. Next, while the injection of M-sequence noise data continues (in the case of
FIG. 3, the switch 66 of the reference processing means 60A is switched to the L side), the
adaptive control units 26 and 26A are controlled to make the Lch side similar to the above.
Adaptive control of the adaptive filter 22 is performed (when the adaptive control is started, the
filter coefficient of the adaptive filter 22 is set to wLi (0)). After the adaptive control is completed,
the filter coefficients of the adaptive filter 22 are left at the time of completion of the adaptive
control.
[0059]
Next, the injection of M-sequence noise data to the adaptive filter 22 is stopped, and the adaptive
control units 36 and 36A are controlled to set the zeroth coefficient of the adaptive filter 32 to 1
and the other coefficients to 0. The M-sequence noise data generated inside the system controller
80, 80A is injected to the input side of (at this time, since the 0th-order coefficient of the
adaptive filter 22 is 1 and the other coefficients are 0, Equivalent), the speaker 35 emits an Msequence noise sound. Then, first, the transfer function CRr is analyzed from the response output
at the B 'point, and the filter coefficient cRr, j is identified (in the case of FIG. 3, the transfer
10-04-2019
25
function CRl is further analyzed from the response output at the A' point, the filter coefficient cRl
, J) are set in the adaptive control unit 36, 36A. Subsequently, while the injection of M-sequence
noise data continues (in the case of FIG. 3, the switch 76 of the reference processing means 70A
is switched to the R side), the adaptive control units 36 and 36A are controlled to similarly
operate the Rch side. The adaptive control of the adaptive filter 32 is performed (when the
adaptive control is started, the filter coefficient of the adaptive filter 32 is set to wRi (0)). After
the adaptive control is completed, the filter coefficients of the adaptive filter 32 remain at the
completion of the adaptive control, and the injection of M-sequence noise data is stopped. After
that, if a desired audio signal is output from the audio source device 10, it is possible to listen to
the desired sound field characteristics without interruption of the music on the way.
[0060]
Also in the example of FIG. 4, instead of performing the sound field control using the audio signal
output from the audio source device 10, using the M-sequence noise data generated internally by
the system controller 80 A, the sound field control The accuracy may be improved. That is,
before the start of audio output from the audio source device 10, when sound field correction is
instructed by the hand operation device 81, the system controller 80A controls the adaptive
control unit 26A to set the zeroth coefficient of the adaptive filter 22. 1, and the other
coefficients to 0, M-sequence noise data generated inside the system controller 80A is injected to
the input side of the adaptive filter 22, and an M-sequence noise sound is emitted from the
speaker 25. Then, first, transfer function CLl is analyzed from the response output at point A ',
filter coefficient cLl, j is identified, and transfer function CLr is analyzed from the response
output at point B', and filter coefficient cLr, j is identified. And set in the adaptive control unit
26A. In addition, another type of M generated in the system controller 80A at the input side of
the adaptive filter 32 by controlling the adaptive control unit 36A to set the 0th coefficient of the
adaptive filter 32 to 1 and the other coefficients to 0. The series noise data is injected, and the M
series noise sound is emitted from the speaker 35. Then, first, the transfer function CRr is
analyzed from the response output at point B ', the filter coefficient cRr, j is identified, and the
transfer function CRl is analyzed from the response output at point A', and the filter coefficient
cRl, j is identified. Is set in the adaptive control unit 36A.
[0061]
Then, while the injection of the two types of M-sequence noise data continues, the adaptive
control units 26A and 36A are controlled to perform the adaptive control of the Lch side
adaptive filter 22 and the adaptive control of the Rch side adaptive filter 32 as described above.
10-04-2019
26
At the same time, the filter coefficient of the adaptive filter 22 is set to wLi (0) and the filter
coefficient of the adaptive filter 32 is set to wRi (0). After the adaptive control is completed, the
filter coefficients of each of the adaptive filters 22 and 32 remain at the completion of the
adaptive control, and the injection of the two types of M-sequence noise data is stopped. After
that, if a desired audio signal is output from the audio source device 10, it is possible to listen to
the desired sound field characteristics without interruption of the music on the way.
[0062]
Further, in the example of FIGS. 1 to 4, when an instruction operation of sound field correction is
performed by the hand controller 81, the system controller 80, 80 A identifies the necessary
ones of the transfer functions CLl, CLr, CRr, Crl. The adaptive control unit 26, 26A, 36, 36A
performs the adaptive control, but when the audio source device 10 is put into the PLAY state or
turned on to start the output of the audio signal, In conjunction with this, among the transfer
functions CLl, CLr, CRr, and Crl, necessary ones may be identified, and the adaptive control units
26, 26A, 36, 36A may be made to perform adaptive control adaptive control.
[0063]
Further, in the example of FIGS. 3 and 4, the adaptive control units 26A and 36A may alternately
scan and capture the outputs el (n) and er (n) of the subtractors 43 and 53 at high speed.
Further, in the examples of FIG. 1, FIG. 3 and FIG. 4, the configuration of the reference processing
means 60, 60A, 60B is only an example, and for example, the digital filters 62, 64, 72, 74 are FIR
type. Instead of the digital filter, an IIR digital filter may be used, and the delay processing of the
delay unit 61 (71) may be performed by the digital filter 62 (72), or the delay unit 63 (digital
filter 64 (74)). The delay process of 73) and the multiplication process of the multiplier 64 (74)
may be performed together. Further, in the example of FIGS. 1 to 4, the speakers 25 and 35 may
have flat frequency-response characteristics.
[0064]
According to the adaptive sound field control device of the present invention, the sound
reproduced by the one or more sound reproducing means is at least one microphone or both ears
placed near one of the listener's ears. Pick up with at least two microphones placed near each of
the. Then, the control means is provided in each sound reproducing means so that the difference
10-04-2019
27
between the reference signal obtained by processing the input signal with a desired transfer
function by the reference processing means provided for each microphone and the
corresponding microphone output is reduced. The transfer function of the processing means is
adaptively controlled. Thereby, regardless of what frequency-response characteristic the electroacoustic conversion means of the acoustic reproduction means or the acoustic reproduction
space has, and which head-transmission characteristic the listener has, the sound field desired by
the listener can be accurately determined. It can be realized.
[0065]
Further, the control means performs adaptive control of the transfer function of the processing
means temporarily at any predetermined time, and fixes the transfer function of the processing
means after completion of the adaptive control. And, only when the control means performs
adaptive control, the movable support means locates each microphone near the listener's ear and
retracts the microphone away from the listener's ear for another time. As a result, for example,
when the position of the listener's ear hardly moves as in the vehicle compartment, while
performing adaptive control to achieve desired sound field characteristics at any predetermined
time before the start of music listening or during music listening Other than that, the microphone
can be kept away from the listener's head and not get in the way as an obstacle.
[0066]
In addition, electro-acoustic conversion is performed by a headphone speaker provided in the
headphone, and the microphone is fixed to the headphone and picks up a sound near the
entrance of the ear canal. Thereby, even at the time of headphone reproduction, regardless of the
individual difference in the transfer characteristic of the listener's pinna, it is possible to
accurately realize the listener's desired sound field.
[0067]
Further, when there are a plurality of channels of input signals, the one or more sound
reproducing means are provided for each channel of the input signal, and the reference
processing means individually processes the input signal of each channel with a desired transfer
function After that, they are added to obtain a reference signal. As a result, adaptive control of
the transfer function of each processing means can be simultaneously performed in parallel even
10-04-2019
28
for an input having a plurality of channels, such as a stereo music signal, for example. It can be
realized.
[0068]
In addition, the reference processing means allows the listener to select one desired transfer
function from among a plurality of predetermined transfer functions prepared in advance.
Thereby, the sound field characteristics can be variously changed according to the genre of
music, the type of music source, and the like.
[0069]
Also, the electro-acoustic conversion means is placed behind the listener's ear, below the height
of the ear, downward or upward. This makes the localization of the sound image unclear and
makes it possible not to be aware of the presence of the electro-acoustic conversion means.
[0070]
Brief description of the drawings
[0071]
1 is a block diagram of an adaptive sound field control apparatus according to a first
embodiment of the present invention.
[0072]
2 is a configuration diagram of a headphone according to a modification of the first embodiment.
[0073]
3 is a block diagram of an adaptive sound field control apparatus according to a second
embodiment of the present invention.
[0074]
10-04-2019
29
4 is a block diagram of an adaptive sound field control device according to a modification of the
second embodiment.
[0075]
5 is an explanatory view showing a specific example of the installation method of the monitor
microphone.
[0076]
6 is an explanatory view showing a specific example of the installation method of the speaker.
[0077]
7 is an explanatory view showing another specific example of the installation method of the
speaker.
[0078]
Explanation of sign
[0079]
DESCRIPTION OF SYMBOLS 10 Audio source apparatus 20, 30 Sound reproduction means 22,
32 Adaptive filter 25, 35 Speaker 26, 26A, 36, 36A40, 50 Monitor microphone 43, 53 Subtractor
60, 60A, 60B, 70, 70A, 70B Reference processing means 80, 80A System controller 81 Hand
controller 81
10-04-2019
30
ay unit.
[0019]
The sound based on the audio signal output from the LOUT terminal of the audio source device
10 should reach the left ear simultaneously with the sound based on the audio signal output
from the ROUT terminal reaching the right ear , ?Ll = ?Rr. The delay times .tau.L1 and .tau.Rr
are assumed to be more than the time the sound travels over the distance L between the speaker
farthest from the listener and the listener, assuming that the listener listens farthest from the
speakers 25 and 35. Value. For example, assuming that L = 3 m and the sound velocity at room
temperature = 340 m / s, it suffices to set ?Ll = ?Rr> 8.8 ms. The digital filters 62 and 72 have
transfer characteristics (transfer functions) having arbitrary frequency-gain characteristics
10-04-2019
6
desired by the listener, such as flat, low boost, high boost, vocal band emphasis, etc. As an
example, it is assumed that the flat frequency characteristic of 0 dB is set over the entire band.
[0020]
43 is a subtractor for calculating the deviation between the output of the A / D converter 42 and
the output of the reference processing means 60, and 53 is a subtraction for calculating the
deviation between the output of the A / D converter 52 and the output of the reference
processing means 70 An adaptive control unit 26 uses the output of the LOUT terminal as a
reference signal from the output of the LOUT terminal after A / D conversion and the output of
the subtractor 43 so that the output of the subtractor 43 becomes smaller. The filter coefficient
of the speaker is updated so that the sound from the speaker 25 can obtain a desired frequencyresponse characteristic (here, flat) at the entrance of the ear canal of the listener's left ear. An
adaptive control unit 36 is a filter of the adaptive filter 32 so that the output of the ROUT
terminal becomes a reference signal from the output of the ROUT terminal after A / D conversion
and the output of the subtractor 53 and the output of the subtractor 53 becomes small. The
coefficients are updated so that the sound from the speaker 35 provides the desired frequencyresponse characteristics (here flat) at the entrance to the ear canal of the listener's right ear.
[0021]
Based on the output from the LOUT terminal and the output from the subtractor 43, the adaptive
control unit 26 updates the filter coefficients of the adaptive filter 22 so that the output from the
subtractor 43 can be minimized using the output from the LOUT terminal as a reference signal.
An adaptive filter that minimizes the evaluation function of, for example, Je = {el (n)} 2 from the
discrete signal x L (n) output from the A / D converter 21 and the discrete signal el output from
the subtractor 43 The filter coefficients of 22 are determined, for example, by the LMS algorithm,
which is one of the steepest descent methods, and are updated and set in the adaptive filter 22.
[0022]
Assuming that the adaptive filter 22 is composed of an I-order FIR filter, the filter coefficient of
the ith filter of the adaptive filter 22 is wLi (where i = 0 to I), from the output point of the
adaptive filter 22 to the monitor microphone 40 Let cLl, j be the j-th filter coefficient when the
transfer function CLl of the transmission line of the above is embodied by a Jth-order FIR filter
10-04-2019
7
(where j = 0 to J), the adaptive control unit 26 performs adaptation at time tn Assuming that the
ith filter coefficient of the filter 22 is wLi (n), the filter coefficient wLi (n + 1) at time tn + 1 can be
expressed by the following equation: wLi (n + 1) = wLi (n) + ? и ?el (n) и qLl * ( n?i)... (1) ?: a
predetermined convergence coefficient, ?: a filter coefficient update setting is performed on the
adaptive filter 22 according to a predetermined weighting coefficient. The initial value wLi (0) of
the filter coefficient wLi is set to a predetermined value. Further, the filter coefficient cLl, j is to
be subjected to a predetermined calculation from the response signal of the output point A 'of the
A / D converter 42 when the M-sequence noise data is injected to the input point A of the D / A
converter 23. We identify with and ask for. The adaptive control unit 26 includes a built-in
memory (not shown) that temporarily stores filter coefficients at the time when adaptive control
is completed.
[0023]
Based on the output from the ROUT terminal and the output of the subtractor 53, the adaptive
control unit 36 updates the filter coefficients of the adaptive filter 32 so that the output of the
subtractor 53 can be minimized using the output from the ROUT terminal as a reference signal.
The adaptive filter 32 that minimizes the evaluation function of Je = {er (n)} 2 from the discrete
signal x (n) output from the A / D converter 31 and the discrete signal er output from the
subtractor 53 The filter coefficients are determined, for example, by the LMS algorithm, which is
one of the steepest descent methods, and are updated and set in the adaptive filter 32.
[0024]
Assuming that the adaptive filter 32 is an I-order FIR filter, the i-th filter coefficient of the
adaptive filter 32 is wRi (where i = 0 to I), from the output point of the adaptive filter 32 to the
monitor microphone 50 Let cRr, j be the j-th filter coefficient when the transfer function CRr of
the transmission line of the above is embodied by a J-th FIR filter (where j = 0 to J), the adaptive
control unit 36 performs adaptation at time tn Assuming that the ith filter coefficient of the filter
32 is wRi (n), the filter coefficient wRi (n + 1) at time tn + 1 is expressed by the following
equation: wRi (n + 1) = wRi (n) + ? и ?er (n) и qRr * n?i) иии (2) ?: a predetermined convergence
coefficient, ?: a filter coefficient update setting is performed on the adaptive filter 32 according
to a predetermined weighting coefficient. The initial value wRi (0) of the filter coefficient wRi is
set to a predetermined value. Further, the filter coefficient cRr, j is to be subjected to a
predetermined calculation from the response signal of the output point B 'of the A / D converter
52 when the M-sequence noise data is injected to the input point B of the D / A converter 33. We
identify with and ask for.
10-04-2019
8
[0025]
Reference numeral 80 denotes a system controller. When sound field correction is instructed by
the hand controller 81, first, a filter coefficient cLl, j for determining a transfer function CLl and a
filter coefficient cRr, j for determining a transfer function CRr are identified, and adaptive control
is performed. Set in sections 26 and 36. Next, the adaptive control units 26 and 36 are controlled
to sequentially perform adaptive control on the Lch side and the Rch side.
[0026]
Next, the operation of the first embodiment described above will be briefly described. Here, it is
assumed that the audio source device 10 is outputting a stereo audio signal, and any filter
coefficient that is not all zero is set in the adaptive filters 22 and 32. When a sound field
correction instruction is given by the hand controller 81, the system controller 80 controls the
adaptive control units 26 and 36 to make the filter coefficients of the adaptive filters 22 and 32
all zero, and the Lch side and Rch side speakers Stop the sound output of 25 and 35. In this state,
M-sequence noise data is injected for a fixed time at point A, pseudo-random noise is emitted
from the speaker 25, and the response output at point A 'is analyzed to determine the filter
coefficient cLl, j that determines the transfer function CLl. It is determined and set in the adaptive
control unit 26. Subsequently, M-sequence noise data is injected at a point B for a predetermined
time, the M-sequence noise sound is emitted from the speaker 35, the response output at the
point B 'is analyzed, and the filter coefficient cRr, j for determining the transfer function CRr It is
determined and set in the adaptive control unit 36.
[0027]
Next, the adaptive control unit 26 is controlled to start adaptive control on the Lch side. The
adaptive control unit 26 first sets each filter coefficient of the adaptive filter 22 to wLi (0). The
LOUT terminal output of the audio source device 10 is A / D converted by the A / D converter
21, and predetermined digital signal processing is performed by the adaptive filter 22. Then,
after D / A conversion is performed by the D / A converter 23, power is amplified by the power
amplifier 24 and sound is converted by the speaker 25. The sound output from the speaker 25 is
picked up by the monitor microphone 40 installed near the entrance of the ear canal of the left
ear of the listener, amplified by the microphone amplifier 41 and then A / D converted by the A /
10-04-2019
9
D converter 42 .
[0028]
On the other hand, the output of the A / D converter 21 is input to the reference processing
means 60, delayed by .tau.L1 by the delay unit 61, processed into a flat frequency characteristic
by the digital filter 62, and output as a reference signal. Then, the difference between the
reference signal and the output of the A / D converter 42 is obtained by the subtractor 43. After
initial setting of the filter coefficients for the adaptive filter 22, the adaptive control unit 26 uses
the output xL (n) of the A / D converter 21 and the deviation output el (n) from the subtractor 43
according to (1) The filter coefficients of the adaptive filter 22 are updated in real time such that
the square value of el (n) is minimized. A certain time has passed since the start of adaptive
control, and adaptive control is completed when the square value of el (n) becomes smaller than
a certain value. At this time, the sound radiated from the speaker 25 has a flat frequencyresponse characteristic at the entrance of the ear canal of the left ear of the listener.
[0029]
However, since adaptive control on the Rch side is not performed yet, adaptive control unit 26
can make adaptive control on the Rch side, so when the square value of el (n) becomes smaller
than a certain value, adaptive control unit 26 at that time After storing the filter coefficients wLi
(i = 0 to I) set in the adaptive filter 22 in the built-in memory (not shown), all the filter
coefficients of the adaptive filter 22 are changed to zero to stop the sound output of Lch And
outputs an adaptive control completion notification to the system controller 80. The system
controller 80 that has received the notification controls the adaptive control unit 36 to start the
adaptive control on the Rch side. The adaptive control unit 36 first sets each filter coefficient of
the adaptive filter 32 to wRi (0). The output of the ROUT terminal of the audio source device 10
is A / D converted by the A / D converter 31, and predetermined digital signal processing is
performed by the adaptive filter 32. Then, after being D / A converted by the D / A converter 33,
the power is amplified by the power amplifier 34 and sound is converted by the speaker 35. The
sound output from the speaker 35 is picked up by the monitor microphone 50 installed near the
entrance of the ear canal of the left ear of the listener, amplified by the microphone amplifier 51
and then A / D converted by the A / D converter 52 .
[0030]
10-04-2019
10
On the other hand, the output of the A / D converter 31 is input to the reference processing
means 70, delayed by .tau.Rr by the delay unit 71, processed into flat frequency characteristics
by the digital filter 72, and output as a reference signal. Then, the difference between the
reference signal and the output of the A / D converter 52 is obtained by the subtractor 53. The
adaptive control unit 36 initializes the filter coefficients of the adaptive filter 32 and then uses
the output xR (n) of the A / D converter 31 and the deviation output er (n) from the subtracter 53
according to (2) , Er (n) are updated in real time so that the square value of er (n) is minimized. A
certain time has passed since the start of adaptive control, and adaptive control is completed
when the square value of er (n) becomes smaller than a certain value. At this time, the sound
radiated from the speaker 35 has a flat frequency-response characteristic at the entrance of the
ear canal of the right ear of the listener.
[0031]
The adaptive control unit 36 outputs an adaptive control completion notification to the system
controller 80 when the square value of er (n) becomes smaller than a predetermined value. The
system controller 80 that has received the notification instructs the adaptive control unit 26 to
reset the filter coefficient. The adaptive control unit 26 that has received the instruction sets the
filter coefficient stored in the built-in memory in the adaptive filter 22 and returns the Lch side to
the state where the adaptive control is completed. As described above, the sound emitted from
the Lch speaker 25 has a flat frequency characteristic at the entrance of the ear canal of the left
ear of the listener, and the sound emitted from the Rch speaker 35 also has a flat frequency
characteristic at the entrance of the right ear external canal So, from here onwards, the listener
will listen regardless of what frequency-response characteristics the speakers 25 and 35 have,
what transfer function the sound reproduction space has, and what head transfer function the
listener has. A sound field having a desired flat frequency characteristic can be accurately
realized. Therefore, even if the diameter of the speakers 25 and 35 is small and the response in
the low band is bad, the bass can be heard on a grand scale. Further, even in the case where
there is a tendency that the vocal band of 500 Hz to 1 kHz falls due to the undulation of the
frequency characteristic of the acoustic space as in the passenger compartment, the vocal band
can be clearly heard. Also, even if the speakers 25 and 35 are provided on the left and right
doors of the vehicle compartment and the distance from the listener (driver) to each of the
speakers 25 and 35 is different, the listener's The arrival time and level of the sound to the left
and right ears can be matched, resulting in a natural stereo sound image. Also, if the listener
changes, or the listener changes the listening position even if the listener is the same, reset the
monitor microphones 40 and 50 to a new listener or a new listening position, and instruct sound
field correction with the hand controller 81 If so, it becomes possible to listen to a new listener
or a new listening position with desired sound field characteristics.
10-04-2019
11
[0032]
Further, as shown in FIG. 2, the speakers 25 and 35 of the sound reproducing means 20 and 30
are replaced with the headphone speakers 101 and 102 of the stereo headphone 100, and the
monitor microphones 40 and 50 are fixed to the stereo headphone 100. It may be located near
the ear canal entrances 105 and 106 of 104. Thereby, even at the time of headphone
reproduction, regardless of the individual difference in the transfer characteristic of the listener's
pinna, it is possible to accurately realize the listener's desired sound field. If the stereo
headphone 100 is a closed type, the sound emitted from the headphone speaker 101 (102) does
not reach the headphone speaker 102 (101) on the opposite side. Therefore, when performing
sound field control, the system controller 80 of FIG. 1 can simultaneously perform adaptive
control on the adaptive filters 22 and 32 while continuing sound output for both Lch and Rch.
[0033]
That is, when the listener puts a headphone 100 on the ear and a stereo audio signal is output
from the audio source device 10, the system controller 80 controls the adaptive control unit
when an instruction for sound field correction is given by the hand controller 81. 26 and 36 are
controlled to set each filter coefficient of the adaptive filters 22 and 32 to 0, and the M-sequence
noise data generated inside the system controller 80 is injected at points A and B for a certain
period of time. M-series noise is emitted. The sound emitted from the speaker 101 is picked up
only by the monitor microphone 40, and the sound emitted from the speaker 102 is picked up
only by the monitor microphone 50. Then, the transfer function CLl is analyzed from the
response output at point A ', the filter coefficient cLl, j is identified and set in the adaptive control
unit 26, the transfer function CRr is analyzed from the response output at point B', and the filter
coefficient cRr , J are identified and set in the adaptive control unit 36. Then, the injection of Msequence noise data is stopped. Next, the adaptive control unit 26 is controlled to start adaptive
control of the Lch-side adaptive filter 22, and at the same time, the adaptive control unit 36 is
controlled to start adaptive control of the Rch-side adaptive filter 32. Then, the adaptive control
units 26 and 36 set predetermined initial filter coefficients wLi (0) and wRi (0) in the adaptive
filters 22 and 32, respectively. At this time, the sound emitted from the headphone speaker 101
enters only the left ear and is picked up by the monitor microphone 40, and a deviation el from
the output of the reference processing means 60 is output from the subtractor 43. Further, the
sound emitted from the headphone speaker 102 enters only the right ear and is picked up by the
monitor microphone 50, and the difference er with the output of the reference processing means
70 is output from the subtractor 53. After initialization, the adaptive control units 26 and 36
10-04-2019
12
perform adaptive control on the adaptive filters 22 and 32 until el and er become smaller than or
equal to a predetermined size, and when el and er become smaller than or equal to a fixed size.
The filter coefficients of the adaptive filters 22 and 32 may be fixed, and control may be
completed (the adaptive control may be continued even after el and er have become smaller than
or equal to a certain size).
[0034]
FIG. 3 is a circuit diagram of a vehicle-mounted adaptive sound field control apparatus according
to a second embodiment of the present invention. The same reference numerals as in FIG. 1
denote the same parts in FIG. In the embodiment of FIG. 1, the sound emitted from the Lch-side
speaker can be listened to the desired sound field characteristic with the left ear of the listener,
and the sound emitted from the Rch-side speaker can be desired with the right ear of the listener
Although it was made possible to listen by characteristics, the embodiment of FIG. 3 can listen to
the sound emitted from the Lch side speaker with desired sound field characteristics by both the
left and right ears of the listener, and from the Rch side speaker The emitted sound can be
listened to a desired sound field characteristic in both the right ear and the left ear of the listener.
[0035]
In the reference processing means 60A, compared with FIG. 1, a delay unit 63 of delay time
.tau.R1 provided in series on the output side of the A / D converter 31 and an FIR type digital
connected to the output side of the delay unit. A filter 64, a gain adjusting multiplier 65
(multiplication coefficient gR1) connected to the output side of the digital filter, and a switch 66
for switching the output of the digital filter 62 and the multiplier 65 are added. The delay unit
63, the digital filter 64, and the multiplier 65 set the transfer characteristic desired by the
listener based on what transfer characteristic (transfer function) the audio signal output from the
ROUT terminal is desired to be heard by the left ear. It is to do. Further, in the reference
processing means 70A, a delay unit 73 of delay time ? Lr provided in series on the output side
of the A / D converter 21 as compared with FIG. 1, and an FIR type connected to the output side
of the delay unit. A digital filter 74, a multiplier for gain adjustment 75 (multiplication coefficient
gLr) connected to the output side of the digital filter, and a switch 76 for switching the output of
the digital filter 72 and the multiplier 75 are added. The delay unit 73, the digital filter 74, and
the multiplier 75 set the transfer characteristic desired by the listener based on what transfer
characteristic (transfer function) the audio signal output from the LOUT terminal is desired to be
heard by the right ear. It is to do.
10-04-2019
13
[0036]
The sound based on the audio signal output from the LOUT terminal of the audio source device
10 should reach the left ear simultaneously with the sound based on the audio signal output
from the ROUT terminal reaching the right ear The delay time ? L1 at the delay unit 61 is made
equal to the delay time ? Rr of the delay unit 71. The delay times .tau.L1 and .tau.Rr are
assumed to be more than the time the sound travels over the distance L between the speaker
farthest from the listener and the listener, assuming that the listener listens farthest from the
speakers 25 and 35. It is as good value. For example, assuming that L = 3 m and the sound
velocity at room temperature = 340 m / s, it suffices to set ?Ll = ?Rr> 8.8 ms. On the other
hand, assuming that the arrival time difference between the monitor microphones 50 and 40 for
the sound emitted from the speaker 35 is ??R1, and the arrival time difference between the
monitor microphones 40 and 50 for the sound emitted from the speaker 25 is ??Lr, then ?Rl
= ?Rr + ??Rl, ?Lr = It is set as ?Ll + ??Lr. The ratio of the sound reception level at the
monitor microphone 40 to the sound reception level at the monitor microphone 50 of the sound
emitted from the speaker 35 is mR1, and the monitor microphone 50 for the sound reception
level at the monitor microphone 40 of the sound emitted from the speaker 25. Assuming that the
ratio of the sound reception level in the case of (mLr) is gRL = mRl and gLr = mLr.
[0037]
When the positional relationship between the speakers 25 and 35 and the listener is fixed, the
monitor microphones 40 and 50 are attached to the entrance to the ear canal of the dummy
head in advance in ??R1 and mR1 and a pulse sound of a certain level is emitted from the
speaker 35 The value obtained by measuring the arrival time difference between the monitor
microphones 50 and 40 and the sound reception level difference at the time of setting is set as
an approximation value, and ?? Lr and mLr are pulse sound of a certain level from the speaker
25. The value obtained by measuring the arrival time difference between the monitor
microphones 40 and 50 and the sound reception level difference when the light is emitted is set
as an approximate value. The digital filters 62, 64, 72, 74 have transfer characteristics (transfer
functions) with any frequency-gain characteristics desired by the listener, such as flat, low boost,
high boost, vocal band emphasis, etc. Here, as an example, it is assumed that the flat frequency
characteristic of 0 dB is set over the entire band.
[0038]
10-04-2019
14
The adaptive control unit 26A uses the output of the LOUT terminal as a reference signal from
the output of the LOUT terminal after A / D conversion and the outputs of the subtractors 43 and
53 so that both the outputs of the subtractors 43 and 53 decrease. The filter coefficients of 22
are updated so that the sound from the speaker 25 provides the desired frequency-response
characteristics (here flat) at the entrance to the ear canal of the listener's left and right ears. From
the discrete signal x L (n) output from the A / D converter 21 and the discrete signal el output
from the subtractor 43 and the discrete signal er output from the subtracter 53, for example, Je =
{el (n) The filter coefficient of the adaptive filter 22 that minimizes the evaluation function of 2 +
{er (n)} 2 is determined by, for example, the LMS algorithm, which is one of the steepest descent
methods, and updated in the adaptive filter 22.
[0039]
Assuming that the adaptive filter 22 is composed of an I-order FIR filter, the filter coefficient of
the ith filter of the adaptive filter 22 is wLi (where i = 0 to I), from the output point of the
adaptive filter 22 to the monitor microphone 40 Filter coefficient of the jth order when the
transfer function CLl of the transmission line is embodied by the Jth-order FIR filter, cLl, j, the
transfer function CLr of the transmission line from the output point of the adaptive filter 22 to
the monitor microphone 50 Jth Assuming that the j-th filter coefficient when embodied by the
FIR filter of this embodiment is cLr, j (where j = 0 to J), the adaptive control unit 26A determines
the i-th filter coefficient of the adaptive filter 22 at time tn. As wLi (n), the filter coefficient wLi (n
+ 1) at time tn + 1 is expressed by the following equation: wLi (n + 1) = wLi (n) + ? и ?el (n) и qLl
* (n?i) + ? и ?er (n) QLr * (n?i) (3) ?: predetermined convergence coefficient, ?: update
setting of the filter coefficient is performed on the adaptive filter 22 in accordance with the
predetermined weighting coefficient. The initial value wLi (0) of the filter coefficient wLi is set to
a predetermined value. Further, the filter coefficient cLl, j is to be subjected to a predetermined
calculation from the response signal of the output point A 'of the A / D converter 42 when the Msequence noise data is injected to the input point A of the D / A converter 23. We identify with
and ask for. The adaptive control unit 26A includes a built-in memory (not shown) that
temporarily stores filter coefficients at the time when adaptive control is completed.
[0040]
The adaptive control unit 36A uses the output of the ROUT terminal as a reference signal from
the output of the ROUT terminal after A / D conversion and the outputs of the subtractors 53 and
10-04-2019
15
43 so that both the outputs of the subtractors 53 and 43 decrease. The filter coefficients of 32
are updated so that the sound from the speaker 35 provides the desired frequency-response
characteristics (here flat) at the entrance to the ear canal of the listener's left and right ears. From
the discrete signal xR (n) output from the A / D converter 31 and the discrete signal el output
from the subtractor 43 and the discrete signal er output from the subtractor 53, for example, Je
= {el (n) The filter coefficients of the adaptive filter 32 that minimize the evaluation function of 2
+ {er (n)} 2 are determined by, for example, the LMS algorithm, which is one of the steepest
descent methods, and updated in the adaptive filter 32.
[0041]
Assuming that the adaptive filter 32 is an I-order FIR filter, the i-th filter coefficient of the
adaptive filter 32 is wRi (where i = 0 to I), from the output point of the adaptive filter 32 to the
monitor microphone 50 Filter coefficient of the jth order when the transfer function CRr of the
transmission line is embodied by the J-th FIR filter, cRr, j, the transfer function CRl of the
transmission line from the output point of the adaptive filter 22 to the monitor microphone 40
Jth Assuming that the j-th filter coefficient when embodied by the FIR filter of c is cRl, j (where j =
0 to J), the adaptive control unit 36A determines the i-th filter coefficient of the adaptive filter 32
at time tn. As wRi (n), the filter coefficient wRi (n + 1) at time tn + 1 is expressed by the following
equation: wRi (n + 1) = wRi (n) + ? и ?er (n) и qRr * (n?i) + ? и ?el (n) QRl * (n-i) (4) ?: a
predetermined convergence coefficient, ?: a filter coefficient update setting is performed on the
adaptive filter 32 according to a predetermined weighting coefficient. The initial value wRi (0) of
the filter coefficient wRi is set to a predetermined value. Further, the filter coefficient cRr, j is to
be subjected to a predetermined calculation from the response signal of the output point B 'of the
A / D converter 52 when the M-sequence noise data is injected to the input point A of the D / A
converter 33. The filter coefficient cRl, j is determined from the response signal of the output
point A 'of the A / D converter 42 when the M-sequence noise data is injected to the input point
A of the D / A converter 33. Identification is obtained by performing an operation.
[0042]
When sound field correction is instructed by the hand controller 81, the system controller 80A
first determines the filter coefficient cLl, j that determines the transfer function CLl, the filter
coefficient cLr, j that determines the transfer function CLr, and the filter coefficient that
determines the transfer function CRr The filter coefficients cRl, j for determining cRr, j and the
transfer function CRl are identified and set in the adaptive control units 26A and 36A. Next, the
adaptive control units 26A and 36A and the reference processing means 60A and 70A are
10-04-2019
16
controlled to sequentially perform adaptive control on the Lch side and the Rch side. The other
components are configured the same as in FIG.
[0043]
Next, the operation of the second embodiment described above will be briefly described. Here, it
is assumed that the audio source device 10 is outputting a stereo audio signal, and any filter
coefficient that is not all zero is set in the adaptive filters 22 and 32. Further, adaptive control of
the adaptive filters 22 and 32 is performed in order. When a sound field correction instruction is
given by the hand controller 81, the system controller 80A controls the adaptive control units
26A and 36A to make all the filter coefficients of the adaptive filters 22 and 32 zero, and the Lch
side and Rch side speakers Stop the sound output of 25 and 35. In this state, M-sequence noise
data is injected at a point A for a fixed time, an M-sequence noise sound is emitted from the
speaker 25, and a response output at the point A 'is analyzed to determine a filter coefficient cLl,
j that determines a transfer function CLl. Then, the response output at point B 'is analyzed to
obtain a filter coefficient cLr, j for determining the transfer function CLr, which is set in the
adaptive control unit 26A. Subsequently, M-sequence noise data is injected at a point B for a
fixed time, M-sequence noise sound is emitted from the speaker 35, the response output at the
point B 'is analyzed, and a filter coefficient cRr, j for determining a transfer function CRr is
obtained. Also, the response output at the point A 'is analyzed to obtain a filter coefficient cRl, j
for determining a transfer function CRl, which is set in the adaptive control unit 36A.
[0044]
Next, the switches 66 and 76 of the reference processing means 60A and 70A are on the L side,
and a reference transfer characteristic is set for listening to the sound based on the audio signal
on the Lch side with desired sound field characteristics with the left and right ears. Do. Then, the
control unit 26A controls the adaptive control unit 26A to start adaptive control on the Lch side.
The adaptive control unit 26A first sets each filter coefficient of the adaptive filter 22 to wLi (0).
The LOUT terminal output of the audio source device 10 is A / D converted by the A / D
converter 21, and predetermined digital signal processing is performed by the adaptive filter 22.
Then, after D / A conversion is performed by the D / A converter 23, power is amplified by the
power amplifier 24 and sound is converted by the speaker 25. The sound output from the
speaker 25 is picked up by monitor microphones 40 and 50 installed near the entrance of the
ear canal of the left and right ears of the listener, amplified by the microphone amplifiers 41 and
51 respectively, and then A / D converted. A / D conversion is performed by the units 42 and 52.
10-04-2019
17
[0045]
On the other hand, the output of the A / D converter 21 is input to the reference processing
means 60A, delayed by ?L1 by the delay unit 61, processed into flat frequency characteristics
by the digital filter 62, and processed as a reference signal for the left ear It is output. Then, the
difference between the reference signal and the output of the A / D converter 42 is obtained by
the subtractor 43. The output of the A / D converter 21 is input to the reference processing
means 70A, delayed by ?Lr (= ?Ll + ??Lr) by the delay unit 71, processed into a flat
frequency characteristic by the digital filter 74, and further multiplied by the multiplier 75 Is
multiplied by the multiplier gLr and output as a reference signal for the right ear. Then, the
difference between the reference signal and the output of the A / D converter 52 is obtained by
the subtractor 53. After initial setting of the filter coefficients to the adaptive filter 22, the
adaptive control unit 26A uses xL (n), el (n) and er (n) to set the square value of el (n) according
to (3). The filter coefficients of the adaptive filter 22 are updated in real time so that the sum
with the square value of er (n) is minimized. A certain time has passed since the start of adaptive
control, and the adaptive control is completed when the sum of the square value of el (n) and the
square value of er (n) becomes smaller than a certain value. At this time, the sound radiated from
the speaker 25 has a flat frequency-response characteristic at the entrance of the ear canal of the
left and right ears of the listener. Also, since the sound reaches the right ear delayed by ?? Lr
from the left ear and the level is reduced to gLr times, a good sense of direction can be obtained.
[0046]
However, since adaptive control on the Rch side has not been performed yet, the adaptive control
unit 26A adds the square value of el (n) and the square value of er (n) to enable adaptive control
on the Rch side. When it becomes smaller than a certain value, the filter coefficient wLi (i = 0 to I)
set in the adaptive filter 22 at that time is stored in the built-in memory (not shown), and then
the filter coefficient of the adaptive filter 22 is stored. All are changed to zero to stop the sound
output of Lch, and an adaptive control completion notification is output to the system controller
80A. The system controller 80A that has received the notification sets the switches 76 and 66 of
the reference processing means 70A and 60A to the R side, and listens to the sound based on the
audio signal on the Rch side with the desired sound field characteristics with the right and left
ears. Set the reference transfer characteristics for Then, it controls the adaptive control unit 36A
to start adaptive control on the Rch side. The adaptive control unit 36A first sets each filter
coefficient of the adaptive filter 32 to wRi (0). The output of the ROUT terminal of the audio
source device 10 is A / D converted by the A / D converter 31, and predetermined digital signal
10-04-2019
18
processing is performed by the adaptive filter 32. Then, after being D / A converted by the D / A
converter 33, the power is amplified by the power amplifier 34 and sound is converted by the
speaker 35. The sound output from the speaker 35 is picked up by monitor microphones 50 and
40 installed near the entrance of the ear canal of the listener's right and left ears, amplified by
the microphone amplifiers 51 and 41 respectively, and then A / D converted. A / D conversion is
performed by the units 52 and 42.
[0047]
On the other hand, the output of the A / D converter 31 is input to the reference processing
means 70A, delayed by ?Rr by the delay unit 71, processed to a flat frequency characteristic by
the digital filter 72, and processed as a reference signal for the right ear It is output. Then, the
difference between the reference signal and the output of the A / D converter 52 is obtained by
the subtractor 53. The output of the A / D converter 21 is input to the reference processing
means 60A, delayed by ?R1 (= ?Rr + ??R1) by the delay unit 61, processed into a flat
frequency characteristic by the digital filter 64, and further multiplied by the multiplier 65 The
multiplier gR1 is multiplied by the signal and output as a reference signal for the left ear. Then,
the difference between the reference signal and the output of the A / D converter 42 is obtained
by the subtractor 43. After the initial setting of the filter coefficients to the adaptive filter 32, the
adaptive control unit 36A uses xR (n), el (n) and er (n) to set the square value of eR (n) according
to (4). The filter coefficients of the adaptive filter 32 are updated in real time so that the sum
with the square value of e L (n) is minimized. After a certain time has passed since the start of the
adaptive control, the adaptive control is completed when the sum of the square value of eR (n)
and the square value of eL (n) becomes smaller than a certain value. At this time, the sound
radiated from the speaker 35 has a flat frequency-response characteristic at the entrance of the
ear canal of the right and left ears of the listener. Also, since the sound reaches the left ear
delayed by ??R1 from the right ear and the level is reduced by a factor of gR1, a good sense of
direction of the sound image can be obtained. When the square value of er (n) + el (n) becomes
smaller than a predetermined value, the adaptive control unit 36A outputs an adaptive control
completion notification to the system controller 80A. The system controller 80A that has
received the notification instructs the adaptive control unit 26A to reset the filter coefficient. The
adaptive control unit 26A that has received the instruction sets the filter coefficient stored in the
built-in memory in the adaptive filter 22, and returns the Lch side to the state where the adaptive
control is completed.
[0048]
10-04-2019
19
As described above, the sound from Lch speaker 25 has a flat frequency characteristic at the
entrance of the ear canal of the left and right ears of the listener, and the sound from Rch
speaker 35 is also at the entrance of the ear canal of right and left ears. Since the frequency
characteristics become flat, the listener subsequently determines what frequency-response
characteristics the speakers 25 and 35 have, what transfer characteristics the sound
reproduction space has, and what head transfer characteristics the listener has Regardless, it is
possible to more accurately realize a sound field having a flat frequency characteristic desired by
the listener. Also, even if the speakers 25 and 35 are provided on the left and right doors of the
vehicle compartment and the distance from the listener (driver) to each of the speakers 25 and
35 is different, the listener's The arrival time and level of the sound to the left and right ears can
be made equal, and in addition, the sound from the speaker 25 arrives at the left ear delayed
from the right ear and at a smaller level and comes out from the speaker 35 The sound arrives at
the right ear delayed from the left ear and at a reduced level, so a very natural stereo image can
be obtained.
[0049]
In the second embodiment described above, the adaptive control on the Lch side and the Rch side
is sequentially performed, but the time required for sound field correction may be shortened by
simultaneously performing the adaptive control. In this case, the adaptive sound field control
device of FIG. 3 is modified as shown in FIG. 4 and in the reference processing means 60B and
70B, the output of the digital filter 62 and the output of the multiplier 65 are added by the adder
67 and a subtractor 43 By outputting the sound data to the Lch and Rch from the speakers 25
and 35, the reference transfer characteristic is set to simultaneously listen to the sound of the
Lch and Rch radiated from the speakers 25 and 35 with a desired sound field characteristic. Also,
by adding the output of the digital filter 72 and the output of the multiplier 75 by the adder 77
and outputting the result to the subtractor 53, the Rch and Lch sounds radiated from the
speakers 35 and 25 are simultaneously desired by the right ear. Set the reference transfer
characteristic for listening in the sound field characteristic.
[0050]
In the configuration of FIG. 4, when an instruction for sound field correction is given by the hand
controller 81, the system controller 80A controls the adaptive control units 26A and 36A to
make all of the filter coefficients of the adaptive filters 22 and 32 zero. The sound output of the
speakers 25 and 35 on the side and the Rch side is stopped. Then, as in the case of FIG. 3, Msequence noise data is injected at points A and B for a predetermined time, and the response
10-04-2019
20
outputs of points A ? and B ? are analyzed to determine transfer functions CLl and CLr. cLl, j,
find the filter coefficient cLr, j that determines the transfer function CLr, set it in the adaptive
control unit 26A, find the filter coefficient cRr, j that determines the transfer function CRr, and
find the filter coefficient cRl, j that determines the transfer function CRl, apply Set in the control
unit 36A.
[0051]
Next, the adaptive control units 26A and 36A are controlled to simultaneously start adaptive
control on the Lch side and the Rch side. The adaptive control unit 26A first sets each filter
coefficient of the adaptive filter 22 to wLi (0). The LOUT terminal output of the audio source
device 10 is A / D converted by the A / D converter 21, and predetermined digital signal
processing is performed by the adaptive filter 22. Then, after D / A conversion is performed by
the D / A converter 23, power is amplified by the power amplifier 24 and sound is converted by
the speaker 25. Furthermore, the adaptive control unit 36A first sets each filter coefficient of the
adaptive filter 32 to wRi (0). The output of the ROUT terminal of the audio source device 10 is A
/ D converted by the A / D converter 31, and predetermined digital signal processing is
performed by the adaptive filter 32. Then, after being D / A converted by the D / A converter 33,
the power is amplified by the power amplifier 34 and sound is converted by the speaker 35. The
sounds output from the speakers 25 and 35 are picked up by the monitor microphones 40 and
50, respectively amplified by the microphone amplifiers 41 and 51, and then A / D converted by
the A / D converters 42 and 52.
[0052]
The sound components of Lch and Rch are added to the output of the A / D converter 42, and the
output of the processing means 60B for reference is subtracted by the subtractor 43 to obtain
the desired left ear in stereo sound reproduction. The deviation between the sound field
characteristic of and the actual sound field characteristic is determined. Further, the acoustic
component of Rch and the acoustic component of Lch are added to the output of A / D converter
52, and the output of processing means 70B for reference is subtracted by subtractor 53,
thereby reproducing stereo sound. The deviation between the desired and actual sound field
characteristics for the right ear is determined. After initial setting of the filter coefficients to the
adaptive filter 22, the adaptive control unit 26A uses xL (n), el (n) and er (n) to set the square
value of el (n) according to (3). The filter coefficients of the adaptive filter 22 are updated in real
time so that the sum with the square value of er (n) is minimized. After a certain time has passed
since the start of adaptive control, adaptive control is completed when the sum of the square
10-04-2019
21
value of el (n) and the square value of er (n) becomes smaller than a certain value (at this time,
adaptation) The control unit 26A may fix the filter coefficient of the adaptive filter 22 to stop the
adaptive control, but may continue. In parallel to this, after the adaptive control unit 36A
initializes the filter coefficients of the adaptive filter 32, it uses xL (n), er (n) and el (n) and uses er
(n) according to (3). The filter coefficients of the adaptive filter 32 are updated in real time so as
to minimize the sum of the squared value of y) and the squared value of el (n). After a certain
time has elapsed since the start of adaptive control, adaptive control is completed when the sum
of the square value of er (n) and the square value of el (n) becomes smaller than a certain value
(at this time, adaptive) The control unit 26A may fix the filter coefficient of the adaptive filter 22
to stop the adaptive control, but may continue. When the adaptive control of Lch and Rch is
completed, the sound radiated from the speaker 25 has a flat frequency-response characteristic
at the entrance of the ear canal of the listener's left ear and right ear. Also, since the sound
reaches the right ear delayed by ?? Lr from the left ear and the level is reduced to gLr times, a
good sense of direction can be obtained. Similarly, the sound radiated from the speaker 35 has a
flat frequency-response characteristic at the entrance to the ear canal of the right and left ears of
the listener. In addition, since the sound reaches the left ear delayed by ??R1 from the right ear
and the level is reduced to gR1 times, a very natural stereo sound image can be obtained.
[0053]
Here, in the example of FIG. 3 and FIG. 4, the delay time .tau.L1 of the delay unit 61 of the
reference processing unit 60A, 60B, the delay time .tau.R1 of the delay unit 63, the multiplication
coefficient gR1 of the multiplier 65, the reference processing unit 70A, The delay time ?Rr of
the delay unit 71 of 70 B, the delay time ?Lr of the delay unit 73, and the multiplication
coefficient gLr of the multiplier 75 are set to appropriate fixed values, but these can be made
variable and the listener instructs the sound field correction It is also possible to re-measure and
set each time it is done. Specifically, when the system controller 80A injects M-sequence noise
data at a point A for a certain period of time to obtain filter coefficients cL1 and cLr that
determine CL1 and CLr, the response delay time at the point A 'is measured The ?L1 of the
delay unit 61 is set, the response delay time at the point B ? is measured, and the ?R1 of the
delay unit 66 is set. Further, the ratio of the response output level at point B 'to the response
output level at point A' is measured to set the multiplication coefficient gR1 of the multiplier 65.
Similarly, when M-sequence noise data is injected for a fixed time at point B in order to obtain
filter coefficients cR1 and cR1 that determine CRr and CR1, the response delay time at point B 'is
measured and ?Rr of the delay unit 71 is calculated. Then, the response delay time at the point
A ? is measured to set ?R1 of the delay unit 76. Also, the ratio of the response output level at
point C ? to the response output level at point B ? is measured to set the multiplication
coefficient gLr of the multiplier 75. In this way, since the reference signal corresponding to the
shape of the head of each listener is obtained, the sound image localization becomes better, and
10-04-2019
22
the listener changes, or the listener changes the listening position even if the listener is the same.
In this case, a new listener or a new listening can be performed by setting the monitor
microphones 40 and 50 again with a new listener or setting them at a new listening position and
instructing the sound field correction with the hand controller 81. It becomes possible to listen
with desired sound field characteristics under good sound image localization even in position.
[0054]
In the first embodiment of FIG. 1 and the modification of FIG. 2 and the second embodiment of
FIG. 3 and the modification of FIG. 4 described above, the audio source device 10 receives stereo
audio signals (audio signals of 2 channels of L and R) Although the output case has been
described as an example, the present invention can be applied to the case where the same 1-ch
monaural audio signal is output from the LOUT terminal and the ROUT terminal. In this case, for
example, the sound reproduction means 30 on the ROUT terminal side, the monitor microphone
50, the microphone amplifier 51, the A / D converter 52, the subtracter 53, and the adaptive
control units 36 and 36A are omitted, and the reference processing means 60A, 60B. The
subtractor 63, the digital filter 64, the multiplier 65, the subtractor 73 in the reference
processing means 70A, 70B, the digital filter 74, the multiplier 75 are omitted, and only the
sound based on the audio signal from the LOUT terminal is Adaptive sound field control may be
performed.
[0055]
Also, in the example of FIGS. 1 to 4, the digital filters 62, 64, 72, 74 are fixed to flat frequencygain characteristics, but flat, low boost, high boost, vocal band emphasis (500 Hz to 1 kHz boost
It may be possible to selectively set one characteristic selected by the listener from among a
plurality of sound field characteristics such as various surround sound fields such as reflection
sound addition and / or reverberation addition. That is, the digital filters 62, 64, 72, 74 of the
reference processing means 60, 60A, 60B, 70, 70A, 70B can switch and set the filter coefficients,
and transmit multiple types to the digital filters 62, 64, 72, 74. A plurality of types of filter
coefficients for setting the characteristics (transfer function) are stored, for example, in the builtin memory (not shown) of the system controller 80, 80A. Then, when the listener performs an
operation to select one of the desired transfer characteristics with the on-hand operation device
81 before instructing the sound field correction, the system controller 80, 80A converts the
corresponding filter coefficient to the digital filter 62. , 64, 72, 74 are set. In this way, by
switching the transfer characteristics of the reference processing means 60, 60A, 60B, 70, 70A,
70B, not only the flat, but also the low boost and high according to the genre of music, the type
10-04-2019
23
of music source, etc. The sound field characteristics can be variously changed, such as boost,
vocal band emphasis (500 Hz to 1 kHz boost), reflection addition and / or reverberation addition.
[0056]
In each of the examples shown in FIGS. 1, 3 and 4, as shown in FIG. 5, two pivoting support
members 92 are provided rotatably on both sides of a headrest 91 provided at the upper end of
the seat 90, Monitor microphones 40 and 50 are provided at the tip of each rotation support
member 92, and a rotation transmission mechanism is connected to the rotation shaft 93 of the
rotation support member 92 to rotate the rotation support member 92 by receiving the
rotational force of the motor. . Then, the rotation support member 92 is made to stand by
upwards beforehand (state of the broken line E in FIG. 5), and when the listener instructs the
sound field correction, the system controller 80, 80A controls the motor to rotate The support
member 92 is turned 90 ░ forward to position the monitor microphones 40 and 50 respectively
near the entrance of the ear canal of the user's left and right ears. After that, the system
controller 80, 80A performs the identification of the required ones of the transfer functions CLl,
CLr, CRl, CRr and the adaptive sound field control in the same manner as described above, and
controls the motor to rotate after completion. The support member 92 is turned 90 ░ until it is
upward, and kept in the original position. In this way, the head can be moved undisturbed by the
monitor microphones 40 and 50 except during execution of adaptive control.
[0057]
In any of the examples shown in FIGS. 1, 3 and 4, when the speakers 25 and 35 are installed on
both sides behind the listener's ear, as shown in FIG. Of the enclosure 98 fixed to the lower end
of the hanger 97 fixed to the ceiling 96 as shown in FIG. Inside, the acoustic emission ports of
the speakers 25 and 35 are placed downward so that they are below the height of the ears, and
the diffuser 99 is placed under the cone paper to be omnidirectional (preferably in the horizontal
plane). It may be made omnidirectional in all directions). When the speakers 25 and 35 enter the
view of the listener, the position where the sound image is localized is drawn to the installation
location of the speakers 25 and 35. However, according to the example of FIGS. 6 and 7, even
though the speakers 25 and 35 are present just near the ears, they can not be made aware of
their existence, and the sound image can be localized only by the sound heard in both ears.
Therefore, desired sound field characteristics can be easily obtained. In FIG. 6, the diffuser 95 is
installed on the cone paper of the speakers 25 and 35 so as to be omnidirectional at least in the
horizontal plane, or the diffuser 99 is installed on the cone paper of the speakers 25 and 35 in
10-04-2019
24
FIG. If it becomes omnidirectional (preferably omnidirectional in all directions) at least in the
horizontal plane, the positions of the speakers 25, 25 become more difficult to understand, so the
presence of the speakers 25, 35 is made thinner. This makes it easier to obtain desired sound
field characteristics.
[0058]
Also, in the examples of FIGS. 1 and 3, instead of performing the sound field control using the
audio signal output from the audio source device 10, the system controller 80, 80A uses the Msequence noise data generated internally. The accuracy of the sound field control may be
improved. That is, before the start of audio output from the audio source device 10, when sound
field correction is instructed by the hand operation unit 81, the system controllers 80 and 80A
control the adaptive control units 26 and 26A to make the adaptive filter 22 The 0-order
coefficient is 1, the other coefficients are 0, and M-sequence noise data generated inside the
system controller 80, 80A is injected to the input side of th
Документ
Категория
Без категории
Просмотров
0
Размер файла
54 Кб
Теги
description, jph09185383
1/--страниц
Пожаловаться на содержимое документа