JP2003259499

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 JP2003259499
[0001]
The present invention relates to a method for reproducing virtual sound source from an acoustic
signal for transaural reproduction, in which virtual sound source reproduction is performed
using two speakers, through an ear speaker attached to the listener's ears. The present invention
relates to a conversion method and conversion apparatus for generating an acoustic signal for
binaural reproduction.
[0002]
2. Description of the Related Art As shown in FIG. 1, two-channel audio signals recorded in
content such as a compact disc are two speakers sp1 and sp2 arranged at regular intervals, and
the front of the listener M is shown. It assumes that it arranges in left and right object as an axis
and listens. However, even when viewing using the speakers sp1 and sp2 normally, in the case of
viewing at night etc., viewing using the speaker is a noise to the neighborhood, so reduce the
volume or use the ear speakers esp1, esp2 It is common to watch using.
[0003]
However, when the audio signals included in these contents are reproduced by the ear speakers
esp1 and esp2 attached to the listener's ears, the sound image is localized in the listener's head,
which is unnatural. It is known that the phenomenon of in-head localization accompanied by an
impression or a sense of discomfort may occur.
10-05-2019
1
[0004]
Also, in recent years, as shown in FIG. 2, in a content such as a DVD, it is assumed that a plurality
of speakers sp1 to sp5 are arranged around a listener M and reproduced, a medium including a
plurality of audio signals is also recorded. It is getting more and more.
Therefore, it is necessary to arrange and reproduce the number of speakers supposed to be
originally supposed around the listener. However, in practice, it may be difficult to arrange a
plurality of speakers due to the circumstances of the house. Therefore, a technique for simulating
a sound field reproduced through a plurality of speakers sp1 to sp5 using two speakers sp1 and
sp2 as shown in FIG. 1 has been proposed.
[0005]
Among these techniques, paying attention to the head acoustic transfer function, which is an
acoustic transfer function from the sound source position to the listener's ears, it is inputted to
the listener's ears by the sound reproduced from the two speakers sp1 and sp2. The one that
controls the sound is called a transaural system.
[0006]
Also when reproducing the sound by such transaural system using the ear speakers esp1 and
esp2 attached to the listener's both ears, a plurality of speakers sp1 to sp5 including
reproduction using two speakers sp1 and sp2 It is expected that the acoustic signal can be
generated so that the sound field reproduced using Tm can be perceived, and can be heard
without the aforementioned problem of in-head localization.
[0007]
As such a technique, as in the transaural system, one that controls the sound input to the
listener's both ears by the sound reproduced from the ear speaker focusing on the head acoustic
transfer function is called a binaural system.
This technique is known as a means to solve the problem of in-head localization.
10-05-2019
2
[0008]
From the foregoing, it is desirable to convert transaural speech into binaural speech in order to
be compatible with both transaural and binaural systems.
As a method of converting transaural speech into binaural speech, it is assumed to obtain
binaural speech by convolving it with transaural speech using a transfer function from a speaker
reproducing transaural speech to the listener's both ears.
[0009]
However, binaural speech is generated for transaural speech generated on the assumption of a
standard speaker arrangement as shown in FIG. 1 (an opening angle of about 90 to 60 degrees
connecting a listener and left and right speakers) as conventionally studied. In the case of
conversion to a complex, complicated filtering is involved, and it is difficult to cope with the
phenomenon that the above-described localization is increased or the influence of the ear
speaker to be reproduced can not be ignored.
[0010]
In addition, in the binaural system, the phenomenon that the sound generated for the purpose of
localization to the front of the listener moves upward than the intended direction when
reproduced from the ear speaker, or the phenomenon of elevation of the localization direction or
when reproducing There is a phenomenon that sound is deteriorated by acoustic distortion in a
route from the ear speaker of the speaker to the eardrum of the listener.
[0011]
The present invention has been proposed to solve the problems of the prior art as described
above, and its object is to facilitate the reproduction of an ear speaker from acoustic signals
generated for two speakers, ie, for transaural reproduction. That is, the present invention is to
provide a method and apparatus for converting an acoustic signal for binaural reproduction.
[0012]
SUMMARY OF THE INVENTION In order to achieve the above object, an apparatus for converting
an acoustic signal according to a first aspect of the present invention is a signal for transaural
reproduction of an acoustic signal output from a plurality of speakers from two speakers.
10-05-2019
3
Transaural reproduction signal converter for converting into, and transaural voice outputted
from the transaural reproduction signal converter, superimposed on a signal whose transfer
characteristic from two speakers to both ears of the listener is an element, And a transaural-tobinaural converter for converting into a binaural signal corresponding to an acoustic signal that
reaches the listener's ears when reproduced by the system.
[0013]
The apparatus for converting an acoustic signal according to claim 2 is characterized in that, in
the conversion apparatus according to claim 1, the transaural reproduction signal converter has
a filter characteristic according to a reproduction system in which two speakers are arranged
close to each other. .
Here, the reproduction method in which two speakers are disposed close to each other means
that the opening angle connecting the listener and the left and right speakers is about 30 degrees
or less, and preferably, speaker arrangement with an opening angle of about 10 degrees is
adopted Stereo dipole reproduction method.
[0014]
The apparatus for converting an audio signal according to claim 3 is the apparatus for converting
an audio signal according to claim 1 or 2, wherein the transaural reproduction signal converter
converts the audio signal into the audio signal at the listener's ears. It is characterized in that it
comprises: a filter for giving information; and a reproduction equalizer for reproducing the
speaker, which cancels crosstalk occurring during reproduction of the speaker.
[0015]
The apparatus for converting an audio signal according to claim 4 is the apparatus for converting
an audio signal according to claim 1 or 2, wherein the transaural reproduction signal converter
converts the audio signal into the audio signal at the listener's ears. It is characterized in that it is
constituted by a filter having a characteristic in which a filter for giving information and a
reproduction equalizer for reproducing the speaker which cancels crosstalk generated at the time
of reproducing the speaker are integrated.
[0016]
The apparatus for converting an audio signal according to claim 5 is the apparatus for converting
10-05-2019
4
an audio signal according to claim 1 or 2, wherein the transaural-binaural converter comprises a
filter for transaural-binaural conversion and a path from an ear speaker to an eardrum of a
listener. And a reproduction equalizer for reproducing the ear speaker having the inverse
characteristic of
[0017]
The apparatus for converting an acoustic signal according to claim 6 is characterized in that, in
the conversion apparatus according to claim 5, the filter for transaural-binaural conversion leads
directly from the speaker to the listener's ear as a transfer function from the speaker to the
listener's ear. It is characterized in that it uses a combination of information on the route and the
route reflected on the floor to the listener's ears.
[0018]
In the method of converting an acoustic signal according to claim 7, the transaural reproduction
signal is superimposed on a signal whose transfer characteristic from the two speakers to the
listener's ears is an element, and is received by the transaural system. It is characterized in that it
is converted into a binaural signal corresponding to an acoustic signal reaching the listener's
ears.
[0019]
The method of converting an acoustic signal according to claim 8 is the conversion method
according to claim 7, in which the transaural reproduction signal is converted to a binaural
signal, and the direct transfer from the speaker to the listener is performed as a transfer function
from the speaker to the listener's ears. It is characterized in that it uses a combination of
information on the path leading to the ear and the path reflected on the floor to the both ears of
the listener.
[0020]
The method of converting an acoustic signal according to claim 9 is the conversion method
according to claim 7 or 8, wherein, in converting the transaural reproduction signal into a
binaural signal, an ear speaker as a transfer function from the speaker to the listener's ears. The
information processing apparatus is characterized by using information to which inverse
characteristics of the route from the listener to the tympanic membrane have been added.
[0021]
According to the present invention having the configuration as described above, an acoustic
10-05-2019
5
signal simulating a sound field reproduced by a plurality of speakers in reproduction using the
transaural system, that is, transaural voice, is also in-head in ear speaker reproduction. It is
possible to convert to an acoustic signal, ie binaural sound, which provides the effect of multiple
speaker reproduction without localization.
[0022]
BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be
specifically described below with reference to the drawings.
This embodiment uses, as a transaural system for generating an input signal, a reproduction
system in which two speakers formed by arranging two speakers close to each other in front of a
listener are arranged close to each other.
[0023]
In the reproduction method in which two speakers are arranged close to each other according to
the present embodiment, the speakers are arranged close to the front of the listener as shown in
FIG.
This is because in the transaural system with the standard speaker arrangement, the sound field
reproduction can be obtained only at the assumed control point, while in the reproduction
method in which two speakers are arranged closely, the speakers are received. It is characterized
in that it has good robustness to a listener's head rotation, movement, etc. from a geometric
arrangement in which it is brought close to the front of the listener.
For this reason, if the transaural voice input in this embodiment is generated for a reproduction
method in which two speakers are closely arranged, robustness is also expected for the
characteristics of the filter used when converting to binaural voice. it can.
[0024]
10-05-2019
6
As mentioned above, to convert transaural speech to binaural speech, the transfer function from
the two speakers from which the transaural speech is played back to the listener's ears is
convoluted to obtain binaural speech.
What is important at this time is the characteristics of the transfer function from the speaker to
the listener's ears.
Assuming that a reproduction system in which two speakers are arranged close to each other is
assumed as the transaural system, since the speakers are arranged close to the front of the
listener, the transfer function from the speakers to both ears of the listener is shown in FIG. It
shows simple characteristics as shown in FIG.
As described above, assuming a reproduction method in which two speakers are closely
arranged, the filter characteristics are simple, and the values converge quickly as shown in FIG.
15 even in time series. It does not require many product-sum operations and storage areas like.
[0025]
On the other hand, assuming a conventional standard speaker arrangement, the transfer function
from the speaker to the listener's ears is as shown in FIG. 16 to FIG. 18, and a reproduction
method in which two speakers are arranged closely Peaks and dips are noticeable and show
complex characteristics as compared with the assumed case.
Therefore, when the conventional speaker arrangement is assumed, as shown in FIG. 18 also in
time series, the value does not converge quickly as in the speaker arrangement according to the
reproduction method in which two speakers in FIG. Implementation requires many multiplyaccumulate operations and storage areas.
[0026]
From this, by assuming a reproduction method in which two speakers are closely arranged as in
the present embodiment, transaural sound is converted into binaural sound only by applying
extremely simple filter processing as compared with the prior art. Can.
10-05-2019
7
In addition, since it is sufficient to perform filter processing with simple conversion processing, it
is possible to flexibly cope with optimization to a specific condition.
[0027]
However, the signal for transaural reproduction according to the present invention is not
necessarily limited to the reproduction method in which two speakers are closely arranged, and a
processing apparatus having high speed and a large capacity despite the characteristics as shown
in FIGS. By using it, it is also possible to use a signal for transaural reproduction generated by a
conventional reproduction method in which a pair of speakers are arranged at a distance.
[0028]
Hereinafter, the configuration of the present embodiment will be specifically described.
In addition, the symbol used for each formula of this Embodiment shall have the following
meaning.
[0029]
［１．
Description of Symbols] E: Transaural system and the sound that reaches the listener's tympanic
membrane when the desired number of speakers are arranged at a desired position and the
sound is reproduced.
S: Sound included in content or the like, or sound delivered and transmitted via a network or the
like.
V: Transfer characteristic (or a filter having such a transfer characteristic) from the position
where the sound is originally to be reproduced to the listener's ears. P: Reproduction equalizer
10-05-2019
8
for speaker reproduction. C: Acoustic transfer characteristics (or a filter having such transfer
characteristics) from the speaker to the listener's ears. These are used to convert transaural
sound into binaural sound. T: A transaural sound that has been processed and generated at V and
P or V ′ and is reproduced from a speaker of the transaural system. V ': A filter having
characteristics equivalent to the processing of V and P. W: A reproduction equalizer for ear
speaker reproduction. H: Transmission characteristics from the ear speaker to the eardrum of the
listener. C ′: A filter having characteristics equivalent to the processing of C and W. E ': sound
that reaches the eardrum of the listener when the converted binaural sound is reproduced using
an ear speaker. B: Sound reproduced from the ear speaker.
[0030]
［２． Overall Configuration] In FIG. 3, 1 is a converter for converting an acoustic signal S
output from a plurality of speakers into a signal T for transaural reproduction, and this converter
1 is used by both ears of a listener. A filter V for providing positional information for sound
image localization for conversion into an audio signal E, and a speaker reproduction reproduction
equalizer P for canceling crosstalk generated at the time of speaker reproduction. A transauralbinaural converter 3 is provided on the output side of the converter 1 via an analog-to-digital
converter (ADC) 2 that converts a transaural reproduction signal into a digital signal.
[0031]
This transaural-binaural converter 3 is a transaural system by superposing the input transaural
voice T on a filter whose transmission characteristics from the two respective speakers to the
listener's ears are an element. A signal (binaural signal) corresponding to an acoustic signal that
reaches the listener's ears when being reproduced is generated. The transfer function used by
this converter 3 has a simple characteristic as shown in FIGS.
[0032]
A digital-to-analog converter (DAC) 4 for converting the binaural signal generated by the
converter 3 into an analog signal is connected to the output side of the transaural-binaural
converter 3, and the ear speaker esp 1 , Esp 2 are reproduced to the listeners both ears.
[0033]
10-05-2019
9
［３．
Transaural Reproduction Signal Converter] Next, a specific configuration of the transaural
reproduction signal converter 1 will be described. That is, in the transaural system, a plurality of
acoustic signals S originally supposed to be reproduced from different speakers are
superimposed on the filter V having a characteristic based on the head acoustic transfer function.
Desired position information to be localized when reproduced. Then, the sound E at the listener's
ears is controlled to be equivalent to the following equation. [Equation 1] E = SV equation 1
[0034]
However, as shown by C2 and C3 in FIG. 5, in the system from the two speakers to the listener's
ears during the speaker reproduction, the crosstalk from the left speaker to the right ear causes
the sound to reach the left ear from the left speaker Problems occur. [Equation 2] PC = I equation
2
[0035]
Therefore, processing for eliminating problems such as crosstalk occurring at the time of speaker
reproduction is performed in the speaker reproduction reproduction equalizer P satisfying the
above equation. [Equation 3] T = SVP equation 3
[0036]
That is, FIG. 6 shows an example of the filter V and the reproduction equalizer P, which converts
the outputs of the plurality of speakers sp1 to sp5 into the outputs to the left and right speakers
sp1 and sp2 and After the outputs from these filters V1 to V10 are added to the left and right
speakers sp1 and sp2, respectively, reproduction equalizers P1 to P4 for canceling crosstalk are
provided.
[0037]
As means for producing the above-mentioned transaural reproduction signal, as shown in FIG. 6
and FIG. 7 below, in addition to providing the filter V and the reproduction equalizer P in two
stages as shown in FIG. It is also possible to perform processing equivalent to these two
10-05-2019
10
processings by one filter V'1 to V'10.
[0038]
Further, in the reproduction method in which two speakers are arranged close to each other,
when designing a filter for performing these two processes at one time, the following method is
used as an integrated method of optimization by least squares and regularization in the
frequency domain. In some cases, the value calculated from equation 5 may be used.
[0039]
Here, β is referred to as a regularization parameter, and is a variable that determines how much
weight is assigned to the term of effort penalty.
A represents the transfer function from the desired position where the sound image is to be
localized to the listener's ears.
This design method is effective in reducing transient peaks and dips that are often problematic in
audio reproduction systems, and for calculating an audibly superior filter.
[0040]
［４．
Trans-Aural-Binaural Converter] Next, a specific configuration of the trans-aural-binaural
converter 3 will be described according to FIG. That is, the transaural-binaural converter 3 is
composed of transaural-binaural conversion filters C1, C2, C3 and C4, and ear speaker
reproduction reproduction equalizers W1 and W2.
[0041]
Here, as the transaural-binaural conversion filters C1, C2, C3 and C4, as shown in FIGS. 13 to 15,
speakers sp1 and sp2 assuming a reproduction system in which two speakers are arranged in
proximity to each other are used as listeners Although it is desirable to use one having the
10-05-2019
11
characteristics up to the ear, it is also possible to use one having the characteristics from the
speaker assuming the standard speaker arrangement to the listener's both ears as shown in FIGS.
Although possible, for the reasons described below, in the present embodiment, a filter having
the characteristics as described in FIG. 19 and later is used.
[0042]
That is, as described above, it is generally known that in ear speaker reproduction, when it is
intended to localize the sound in front of the listener, the perceived sound image has a strong
tendency to rise above the listener.
In addition, there may be cases where distortion generated in the path from the ear speakers
esp1 and esp2 to the listener's both ears to the eardrum of the listener can not be ignored.
[0043]
Therefore, in the present embodiment, when the sound image rises above the listener, the
transfer function from the speaker to the listeners ears directly from the speakers sp1 and sp2 as
shown in FIG. Use a combination of information on the route to the ground and the route
reflecting on the floor to the listener's ears. This can suppress an increase in the sound image
perceived by the listener. The characteristic of the filter used for conversion in this embodiment
is shown in FIGS.
[0044]
The filters C1, C2, C3 and C4 having the characteristics shown in FIGS. 19 to 21 have filter
characteristics in which the information from the speaker directly to the listener's ears and the
route reflected from the floor to the listener's ears are combined. Have. Therefore, when this
filter is used, it is reflected from the speaker to the floor, which is considered as a factor of rising
of the sound image, by using a filter that reflects information from the speaker to the floor and
brings information of the route to the listeners ears Since the sound reaching to both ears can be
reproduced, it is expected to suppress the phenomenon that the sound image rises.
10-05-2019
12
[0045]
With regard to the influence of distortion and the like in the path from the ear speakers esp1 and
esp2 to the eardrum of the listener, reception is performed from the ear speakers esp1 and esp2
by the ear speaker reproduction reproduction equalizers W1 and W2 satisfying Eq. 6 and Eq. It
can be coped with by removing the influence of the route leading to the eardrum of the listener.
Equation 6 WH = I equation 6 Equation 7 B = TCW equation 7
[0046]
［５． Another Embodiment of Trans-Aural-Binaural Converter] The embodiment of FIG. 9 is
provided with the two filters C and W separately, but in the present invention, one of the two
filters has one of these characteristics. It is also possible to perform processing using a filter.
That is, the embodiment of FIG. 10 uses the filters C'1 to C'4 having characteristics equivalent to
the two filters C and W, and these filters C'1 to C'4 and C The characteristics of '2' and C'3 are
shown in FIGS.
[0047]
The filters having the characteristics shown in FIGS. 22 to 24 are the filters C1 to C4 shown in
FIGS. 13 to 15 (characteristics in the case of the speaker arrangement according to the
reproduction method in which two speakers are arranged close to each other) It is a filter having
the characteristics of the colorizers W1 and W2. This has an inverse characteristic of the path
from the ear speakers esp1 and esp2 to the eardrum of the listener, and therefore suppresses the
phenomenon that the sound is distorted while reaching the eardrum of the listener from the ear
speaker.
[0048]
Also, the transfer function from the speaker to the listener's ears is combined with information
on the route from the speaker to the listener's ears directly and the route reflected on the floor to
the ears, and further to the regeneration equalizer for ear speaker reproduction. A filter having
characteristics to be processed is shown in FIGS. This filter is a filter having a characteristic in
consideration of the reflection component from the floor as shown in FIG. 8 (a filter having the
10-05-2019
13
characteristics as shown in FIGS. 19 to 21) and a reproduction equalizer W1 for ear speaker
reproduction as shown in FIG. , And W2, it is possible to suppress the rise of the sound image
and to suppress the phenomenon that the sound is distorted while reaching the eardrum of the
listener from the ear speaker.
[0049]
［６． Operation of the Embodiment] In the present embodiment having the above
configuration, the signal processed by the transaural-binaural converter 3 is converted from the
digital signal to the analog signal by the DAC 4 and passes through the amplifier 5 to the ear. It is
reproduced from the speakers esp1 and esp2.
[0050]
The reproduced signal is equivalent to transaural reproduction at the listener's ears as shown in
FIG. 12, and is perceived as sound coming from a desired position. That is, as is apparent from
the equations described below, the desired sound E at the listener's ears and the sound E
'reproduced after being converted to binaural sound are equivalent.
[0051]
[Equation 8] E '= BH equation 8 E' = TCWH equation 9 [Equation 10] E '= TC equation 10
[Equation 11] E' = SVPC equation 11 [Equation 12] E ' = SV Eq. 12 E '= E Eq. 13
[0052]
［５．
Effects of the Embodiment] According to the present embodiment having the configuration as
described above, the following effects can be expected. (1) Transoral voice is transmitted or
recorded when used for 2 channel voice media widely used widely in the past, for example, CDs
and TVs. At the time of listening, the listener may play back from the speaker as it is when
listening to the speaker, and in the case of using the ear speaker, the listener can listen only by
performing the conversion process. As a result, multi-channel content can be provided even in
10-05-2019
14
2ch audio media.
[0053]
(2) Distribution of multi-channel content in a network is a large obstacle in bandwidth of a
transmission path, and distribution of multi-channel content requiring a large bandwidth is not
easy. However, since transaural voice has two channels no matter how many channels of the
original content are, it is possible to clear the bandwidth problem of the transmission path which
has been a major obstacle. Also in this case, it is possible to cope with either speaker
reproduction or ear speaker reproduction according to the listener's request.
[0054]
(3) Audio equipment handling multi-channel content such as a DVD player is expected to
simultaneously support reproduction with two speakers and reproduction with an ear speaker.
However, when processing for speaker reproduction and processing for ear speaker are
performed in completely different processes, many hardware and software resources are
required to simultaneously perform processing for speaker reproduction and ear speaker
reproduction. It will be necessary. However, according to the present invention, since it is
possible to generate the sound for ear speaker reproduction with less resources from the sound
for reproduction using two speakers, the sound for speaker reproduction and the sound for ear
speaker reproduction simultaneously It becomes possible to offer.
[0055]
As described above, according to the present invention, the audio signal generated for the two
speaker reproduction, ie, the transaural reproduction, is simply subjected to the simple
conversion processing, and thus the listening through the ear speaker, ie, the binaural. At the
time of reproduction, it becomes possible to obtain the same effect as at the time of transaural
reproduction.
[0056]
Also, only by applying the conversion processing of the present invention to the content in which
the audio signal that has been processed for transaural reproduction in advance is recorded, it is
possible to hear through the ear speaker without causing problems such as localization in the
head. It becomes.
10-05-2019
15
10-05-2019
16