close

Вход

Забыли?

вход по аккаунту

?

JPH05300598

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 JPH05300598
[0001]
TECHNICAL FIELD The present invention relates to a binaural processing method.
[0002]
BACKGROUND OF THE INVENTION When listening to audio reproduction sound with
headphones, usually, sound images gather in the head and are localized, and a sense of reality
can not be obtained. Therefore, the same level difference and time difference as when the dummy
head is placed in the original sound field and listened to by the human left and right ears.
Specifically, the transfer characteristic consisting of the amplitude characteristic (so-called
frequency characteristic) for each frequency, phase characteristic and group delay characteristic.
There is a method of performing so-called binaural recording by collecting sound with individual
microphones so as to cause 、, and listening a signal by binaural recording using a dummy head
with headphones. In this method, the sound image is localized forward, and almost the same
sense of realism as in the case of listening to music in the original sound field can be obtained.
However, in the case of this method, it is premised that the audio signal is binaurally recorded on
a recording medium such as a disc or a tape, and this method is not effective for a recording
medium on which a normal recording has already been made.
[0003]
10-05-2019
1
BACKGROUND A binaural processor is known as an apparatus for binaural processing of an
audio signal obtained by playing a recording medium that is not binaural recorded as if it was
binaurally recorded. This binaural processor is composed of digital filters such as FIR type filter
and IIR type filter, and direction information for localizing the sound image forward is provided
in advance as a coefficient of the digital filter. Direction information is given by an impulse
response (transfer characteristic), and the convolution of the audio signal to the impulse
response is performed in a digital filter. This impulse response places the sound source 2 at a
position where the sound image is to be localized relative to the dummy head (or human head) 1
as shown in FIG. The right and left microphones 3 and 4 are respectively measured by collecting
the sound.
[0004]
A monoural binaural processor is configured as shown in FIG. That is, the audio signal from the
sound source is supplied to the digital filters 12 and 13 of the binaural processor 11. The
impulse response of the right digital filter 12 is H1, and the impulse response of the left digital
filter 13 is H2. The output signals of the digital filters 12 and 13 are supplied to the headphones
16 via the amplifiers 14 and 15, respectively.
[0005]
When constructing a binaural processor for stereo, the relationship between the left and right
speakers from the listener is usually symmetrical in order to obtain the effect that listening with
headphones is listening with the left and right speakers. As shown in FIG. 3, the impulse response
between the left speaker 21 and the left ear of the listener 22 is H1, the impulse response
between the right speaker 23 and the left ear of the listener 22 is H2, right The impulse response
between the speaker 23 and the right ear of the listener 22 can be H1, and the impulse response
between the left speaker 21 and the right ear of the listener 22 can be H2. Therefore, as shown
in FIG. 4, the binaural processor 25 for stereo is composed of digital filters 26 to 29 and adders
30 and 31. The digital filters 26 and 27 are supplied with the audio signal of the left channel,
and the digital filters 28 and 29 are supplied with the audio signal of the right channel. The
impulse response of the digital filters 26 and 29 is H1, and the impulse response of the digital
filters 27 and 28 is H2. The adder 30 adds the output signals of the digital filters 26 and 28 and
supplies the result to the amplifier 32. The adder 31 adds the output signals of the digital filters
27 and 29 and supplies the result to the amplifier 33. The output signals of the amplifiers 32 and
33 become the left and right channel signals of the headphone 34.
10-05-2019
2
[0006]
In this configuration, the positional relationship between the listener and the left and right
speakers is considered to be symmetrical, and binaural processing is performed using only two
impulse responses H1 and H2. However, the positional relationship between the listener and the
left and right speakers is not necessarily symmetrical, and variations in the left and right
characteristics of the speakers and in particular the left and right asymmetry of human hearing
are not reflected. It is not enough to perform localization or high precision binaural processing.
[0007]
In addition, since the characteristics of headphones and the like used when performing binaural
processing are not taken into consideration, there is a problem that the characteristics of
headphones and the like overlap the acoustic characteristics obtained by binaural processing,
and the desired acoustic characteristics can not be obtained. .
[0008]
SUMMARY OF THE INVENTION It is an object of the present invention to provide a binaural
processing method which performs binaural processing with high accuracy in line with the real
human left-right asymmetry.
Another object of the present invention is to provide a binaural processing method capable of
correctly obtaining desired acoustic characteristics without being affected by characteristics of
headphones or the like.
[0009]
According to the present invention, the binaural processing method according to the present
invention is characterized in that input audio signals of left and right channels localized at a
standard sound image position by sound output from left and right sound output points from the
left and right sound output points. Signal processing is applied to each of the left and right
channels based on the transfer characteristics according to the positional relationship between
the left ear and right ear listening points as the position where the audio output by the input
audio signal should be listened and the left and right audio output points. A binaural processing
method for generating an output audio signal, wherein the input audio signal on the left channel
10-05-2019
3
is subjected to signal processing based on a first transfer characteristic and a second transfer
characteristic, and the input audio signal on the right channel is processed. Signal processing is
performed on the basis of the third transfer characteristic and the fourth transfer characteristic,
respectively; A signal obtained based on the transfer characteristic and a signal obtained based
on the fourth transfer characteristic are added to obtain an output audio signal of the left
channel, and the signal obtained based on the second transfer characteristic A signal is obtained
by adding the signal and the signal obtained based on the third transfer characteristic to obtain
an output audio signal of the right channel.
[0010]
In the binaural processing method according to the present invention, the input audio signals
from the left and right sound output points with respect to the input audio signals of the left and
right channels localized at the standard sound image position by sound output from the left and
right sound output points. Based on the transfer characteristics according to the positional
relationship between the left and right ear listening points as the position where the sound
output by the signal should be heard and the left and right sound output points, the signal
processing is applied to output audio signals of the left and right channels. The audio output
means performs acoustic output according to the output audio signal, and the output audio
signal is subjected to signal processing based on the inverse characteristic of the transfer
characteristic of the audio output means. It is characterized by outputting sound.
[0011]
Furthermore, the binaural processing method according to the present invention localizes at a
predetermined sound image position obtained by intercepting the sound output from the sound
source by the recording means at the left ear and the right ear listening point or generating a
signal equivalent thereto. Audio signal of left and right channels is recorded on a recording
medium, the audio signal is reproduced from the recording medium by a reproducing means, and
an acoustic output means performs acoustic output according to the audio signal, The audio
signal reproduced by the means is subjected to signal processing on the basis of the inverse
characteristic of the transfer characteristic of the recording means, the reproducing means or the
sound output means to output the sound.
[0012]
The binaural processing method according to the present invention performs signal processing
on the input audio signal of the left channel based on the first transfer characteristic and the
second transfer characteristic, and performs the first on the input audio signal of the right
channel. The signal processing is performed based on the third transfer characteristic and the
fourth transfer characteristic, and the signal obtained based on the first transfer characteristic
10-05-2019
4
and the signal obtained based on the fourth transfer characteristic are added to the left. An
output audio signal of the channel is obtained, and a signal obtained based on the second
transfer characteristic and a signal obtained based on the third transfer characteristic are added
to obtain an output audio signal of the right channel.
[0013]
Further, according to the binaural processing method of the present invention, the binaural
processed output audio signal is subjected to signal processing based on the reverse
characteristic of the transfer characteristic of the sound output means, and sound output.
Furthermore, the binaural processing method according to the present invention performs signal
processing on an audio signal binaurally recorded by the recording unit and reproduced by the
reproduction unit based on the inverse characteristic of the transfer characteristic of the
recording unit, the reproduction unit or the sound output unit. Apply sound and output.
[0014]
Embodiments of the present invention will be described in detail below with reference to the
drawings.
FIG. 5 is a view showing the principle of the binaural processing method according to the present
invention, in which the parts equivalent to those in FIG. 3 are given the same reference numerals.
In this embodiment, the positional relationship between the listener 22 and the left and right
speakers 21 and 23 as sound output points is not assumed to be symmetrical as in the prior art
in order to perform high-accuracy binaural processing. The impulse responses between the left
and right ears of the listener 22 and the left and right speakers 21 and 23 are individually set.
[0015]
That is, as shown in the figure, the impulse response between the left speaker 21 and the left ear
of the listener 22 is H1 L (first transfer characteristic), between the left speaker 21 and the right
10-05-2019
5
ear of the listener 22. Impulse response H2 L (second transfer characteristic), impulse response
between the right speaker 23 and the right ear of the listener 22 H1 R (third transfer
characteristic), the right speaker 23 and the left of the listener 22 The impulse response between
the user and the ear is taken as H2R (fourth transfer characteristic).
[0016]
FIG. 6 is a schematic block diagram showing a binaural processor configured on the basis of
these set impulse responses, and parts equivalent to those in FIG. 4 are assigned the same
reference numerals.
In the figure, a binaural processor 41 for stereo comprises digital filters 42 to 45 and adders 46
and 47.
The digital filters 42 and 43 are supplied with the audio signal of the left channel, and the digital
filters 44 and 45 are supplied with the audio signal of the right channel. The impulse response of
the digital filter 42 is H1L, the impulse response of the digital filter 43 is H2L, the impulse
response of the digital filter 44 is H2R, and the impulse response of the digital filter 45 is H1R.
The adder 46 adds the output signals of the digital filters 42 and 44 and supplies the result to
the amplifier 32. The adder 47 adds the output signals of the digital filters 43 and 45 and
supplies the result to the amplifier 33. Then, the output signals of the amplifiers 32 and 33
become the left and right channel signals of the headphones 34.
[0017]
In FIGS. 7 and 8, audio signals are output as sounds by the left and right speakers 21 and 23 as
shown in FIG. 5, and the sound outputs are intercepted by the left and right microphones worn
on both ears of a real person (listener 22). The intercept sound pressure level SPL characteristic,
that is, the frequency characteristic with respect to the frequency f in each impulse response is
shown. At this time, the distance between the speakers 21 and 23 and the listener 22 is 2 m, and
the speakers 22 and 23 are disposed at an angle of 30 ° to the left and right from the front
center line of the listener 22. Therefore, the positional relationship between the left and right
speakers and the listening point is symmetrical. Further, H1L, H1R, H2L, and H2R attached to the
curves correspond to the respective impulse responses in FIG. 5 and FIG.
10-05-2019
6
[0018]
As can be seen from FIGS. 7 and 8, even if the sound output by the left and right speakers 21 and
23 arranged symmetrically with respect to the listener 22, impulse responses H1L and H1R
which should be theoretically equivalent to each other, And H2L and H2R have different
characteristics due to the left-right asymmetry of the listener's 22 hearing. In the present
embodiment, since each impulse response in the binaural processor 41 is individually set as
shown in FIG. 6 based on the above different characteristics, a fine binaural of the texture that
also takes into consideration the left-right asymmetry of the listener's 22 hearing. It can be
processed. Therefore, it is possible to perform extremely accurate binaural processing.
[0019]
In the above embodiment, the impulse responses H1L, H2L, H1R and H2R are obtained by each
digital filter of the binaural processing apparatus, but various impulse responses are prepared,
and these are appropriately switched to produce various acoustic effects. Also good. That is, the
digital filter is configured by a DSP (Digital Signal Processor), a memory in which plural types of
impulse response (coefficient for FIR) data are stored in advance, and a CPU (Central Processing
Unit) are provided around it. When the desired coefficient data is selected for the memory by
this, the data is taken in the data area in the DSP, and the DSP can carry out the signal processing
in accordance therewith. If such coefficient data is generated for each room type or person
condition at the time of listening, the user can achieve better sound effects by designating these
to be optimum. .
[0020]
The mounting positions of the left and right microphones at the time of measurement of the
impulse response are not limited to mounting on the dummy head or the human ear, for example,
for use in simulation with a reduced model at the time of acoustic field design such as holes. The
standard dummy head may be worn on a reduced size, or worn on the ear of the animal's dummy
head for studying animal hearing.
[0021]
On the other hand, FIG. 9 is a block diagram showing a binaural processing apparatus configured
10-05-2019
7
to avoid the influence of unnecessary characteristics such as headphones on the sound output.
In the figure, the reproduction device 51 is an audio signal obtained by intercepting the sound
output at a predetermined standard position with respect to the sound source position, or an
equivalent signal generation (audio signal forming a standard sound image localization)
Reproduces the audio signal from the recording medium on which the recording has been
recorded, and supplies this to the binaural processing circuit 52. The binaural processing circuit
52 performs binaural processing as described above on the supplied audio signal and supplies
the audio signal to the characteristic cancellation circuit 53. The characteristic cancellation
circuit 53 supplies an input signal to the output circuit 54 based on an impulse response set as
described later. The output circuit 54 drives the headphones 55 in accordance with the output of
the characteristic cancel circuit 53.
[0022]
Now, it is assumed that the impulse response of each circuit and device is that the reproducing
device 51 is H51, the binaural processing circuit 52 is H52, the characteristic canceling circuit
53 is H53, the output circuit 54 is H54, and the headphone 55 is H55. In the case of canceling,
the inverse characteristic of the impulse response H55 of the headphone 55 is set as the impulse
response characteristic H53 of the characteristic cancel circuit 53. すなわち、H53=H55-1
…(1)となる。 Since the characteristic cancellation circuit 53 cancels the amplitude, phase
and group delay characteristics of the headphone 55 in the input signal based on this inverse
characteristic, the acoustic output of the headphone 55 is not influenced by the characteristics of
the headphone itself, and can not be binaural. Sound effects as intended in the processing circuit
52 can be obtained.
[0023]
Now, focusing on the amplitude characteristic (so-called frequency characteristic) among the
transmission characteristics of the headphone 55, for example, assuming that the middle region
is a rising characteristic, the frequency characteristic of the transmission characteristic of the
characteristic cancellation circuit 53 is the middle region It becomes the characteristic which fell.
The other amplitude and group delay characteristics are also canceled. Furthermore, if the
impulse response H51 of the reproduction device 51 is also taken into consideration here, the
impulse response characteristic H53 of the characteristic cancellation circuit 53 is H53 = H51-1
* H55-1 (2). By doing this, it is possible to obtain an acoustic effect that is not influenced by the
10-05-2019
8
characteristics of the reproduction device 51 in addition to the headphones 55. Of course, the
impulse response characteristic of the output circuit 54 can also be handled in exactly the same
way, and in this case H53 = H51-1 * H54-1 * H55-1 (3). In addition, * represents the convolution
of an impulse response. Also, the impulse response of each device and circuit is individually
measured in advance prior to system configuration.
[0024]
Similarly, in view of various impulse response characteristics in the recording system, the
impulse response of the characteristic cancellation circuit 53 may be set in the configuration as
shown in FIG. In FIG. 10, reference numeral 61 denotes, for example, a tuning fork as a sound
source, and an audio output from the tuning fork 61 is intercepted by a microphone mounted on
the dummy head 62, and an audio signal obtained by intercepting is recorded via the amplifier
63. It is supplied to the device 64. The recording device 64 records the supplied audio signal on
the recording medium in a predetermined format. Then, the reproduction circuit 51 reads the
recording medium and reproduces the signal. Since the audio signal recorded using the dummy
head 62 is recorded on the recording medium, the binaural processing circuit is not necessary in
the reproduction system. The other configuration is the same as that of FIG.
[0025]
In this configuration, assuming that the impulse response is Hd1, the space between the tuning
fork 61 and the dummy head 62 is H62, the microphone mounted on the dummy head 62 is
H62, the amplifier 63 is H63, and the recording device 64 is H64. The impulse response
characteristic H53 of H53 = H62-1 * H63-1 * H64-1 * H51-1 * H54-1 * H55-1 (4)
[0026]
FIG. 11 shows an audio output of a binaurally recorded audio signal by a speaker, and the parts
equivalent to those in FIG. 10 are denoted by the same reference numerals.
In FIG. 11, the audio signal from the output circuit 54 is supplied to the left and right speakers
56R and 56L, and produces an acoustic output toward the left and right ears of the listener 57.
Further, as in FIG. 10, since the audio signal recorded by using the dummy head is recorded on
the recording medium, the binaural processing circuit is not necessary in the reproduction
system.
10-05-2019
9
[0027]
In this configuration, assuming that the impulse response is H56 for the left and right speakers
56R and 56L, and Hd3 for the space between the speakers 56R and 56L and the left and right
ears of the listener 57, the impulse response H53 of the characteristic cancellation circuit 53 is
H53 = H62-1 * H63-1 * H64-1 * H51-1 * H54-1 * H56-1 * Hd3-1 (5). Here, sound output not
influenced by unnecessary characteristics can be obtained, and cancellation of Hd 3 can release
crosstalk between left and right sound outputs generated when outputting a binaurally recorded
audio signal to a speaker. Such a configuration is a basic configuration that performs processing
in the case of listening to a speaker without listening to headphones for binaural recorded audio
signals, that is, trans-aural processing, and the characteristic cancellation circuit 53 includes the
speakers 56R and 56L. , And the listener 57, depending on the inverse characteristic of the
impulse response Hd3 of the space, but in actuality it is as follows.
[0028]
FIG. 12 is a block diagram showing such a transoral processing circuit. In the transaction
processing circuit 101, the left channel audio signal is supplied to one input end of the adder
102 and also supplied to the other input end of the adder 104 through the inverting buffer 103.
Further, the right channel audio signal is supplied to one input end of the adder 104 and is also
supplied to the other input end of the adder 102 through the non-inversion buffer 105.
[0029]
The output of the adder 102 is supplied to one input end of the adder 107 after passing through
a digital filter 106 having an impulse response 1 / (h1 + h2), and the other of the adder 109 via
the non-inversion buffer 108. Supplied to the input end of the The output of the adder 104 is
supplied to one input end of the adder 109 after passing through a digital filter 106 having an
impulse response 1 / (h1-h2), and the other of the adder 107 via the inverting buffer 111.
Supplied to the input end of the
[0030]
10-05-2019
10
The output audio signal of the adder 107 is supplied to a speaker 56L via an amplifier 54L. The
output audio signal of the adder 109 is supplied to a speaker 56R via an amplifier 54R. The
speakers 56L and 56R provide an acoustic output to the listener 57. In this configuration, the
sound levels L ′ ′ and R ′ ′ reaching the left ear and the right ear of the listener 57 have L
′ and R ′ as the sound output levels of the left and right speakers, respectively. Assuming that
the positional relationship with the mosquitoes 56L and 56R is symmetrical, L ′ ′ = L′ h1 +
R′h2 (6) R ′ ′ = R′h1 + L′ h2 (7). Here, assuming that the impulse response of the digital
filter 106 in the transaction processing circuit 101 is 1 / (h1 + h2) and the impulse response of
the digital filter 110 is 1 / (h1-h2), the left and right speaker outputs for the input audio signals
L and R are The level is L '= (L + R) / (h1 + h2) + (L-R) / (h1-h2) (8) R' = (L + R) / (h1 + h2) + (R-L)
/ (h1-h2) ) ... (9). However, the impulse responses for the amplifiers 54L and 54R and the
speakers 56L and 56R are neglected. Substituting these equations (8) and (9) into equations (6)
and (7), L ′ ′ = 2L (10) R ′ ′ = 2R (11), and only the original signal component is the listener
57 Will reach your ears. Thus, the listener 57 can release the binaural recorded audio signal
from the crosstalk that occurs when the speaker output. Although the internal configuration of
the transaction processing circuit 101 is a so-called cross-over and looks complicated, it is
advantageous because it requires less calculation when processing with the DSP.
[0031]
Further, FIG. 13 shows a case where special acoustic effects are aimed by applying the abovedescribed binaural processing, and the same reference numerals are given to parts equivalent to
those in FIG. In the figure, it is a desktop microphone 65 installed at a predetermined standard
position for the sound source that intercepts the sound output of the tuning fork 61. The desktop
microphone 65 supplies an audio signal obtained by interception to the recording device 64 via
the amplifier 63. In addition, it is a monaural signal until reproduction.
[0032]
In this configuration, since binaural recording is not performed, and a standard recording format
using a desktop microphone 65, it is necessary to perform binaural processing in the
reproduction system in order to listen to the audio signal recorded by this in binaural. While, in
order to obtain an auditory sensation by driving the actual speakers 56R and 56L as described
above, the impulse response H52 of the binaural processing circuit 52 is set to H52 = H56 * Hd3
(12), and a tuning fork The impulse response H53 of the characteristic cancellation circuit 53 is
H53 = Hd2-1 * H65-1 * H63-1 * H64-1 * H51-, when the impulse response of the space between
10-05-2019
11
61 and the table microphone 65 is Hd2. 1 * H54-1 * H55-1 (13). Note that canceling Hd2
corresponds to listening to the tuning fork 61 in an anechoic room. By doing this, it is possible to
obtain a feeling of hearing as being heard by the real speakers 56R, 56L at the predetermined
position while listening with headphones. Moreover, it can be said that such treatment is a
modification of binaural treatment.
[0033]
In the above embodiment, although the impulse response of the headphone is obtained by each
digital filter of the binaural processing circuit, a plurality of impulse responses corresponding to
the type of headphone are prepared, and they may be appropriately switched according to the
connected headphones. good. This configuration can also be implemented using the abovedescribed DSP and a memory and a CPU in which plural types of impulse response (coefficient
for FIR) data are stored in advance.
[0034]
On the other hand, as microphones for video cameras, small microphones were attached to the
left and right ears of the photographer to perform video recording and binaural recording, and
when playing back while listening to the sound with headphones during playback, it became a
photographer You can enjoy the image and sound effects with your senses. In addition, during
such recording or reproduction, transversal processing as described above and various
characteristic cancellation can be performed as needed.
[0035]
As described above, according to the binaural processing method of the present invention, the
input audio signal of the left channel is subjected to signal processing based on the first transfer
characteristic and the second transfer characteristic, and the input audio of the right channel is
processed. The signal is subjected to signal processing based on the third transfer characteristic
and the fourth transfer characteristic, and the signal obtained based on the first transfer
characteristic and the signal obtained based on the fourth transfer characteristic are Addition is
performed to obtain an output audio signal of the left channel, and a signal obtained based on
the second transfer characteristic and a signal obtained based on the third transfer characteristic
are added to obtain an output audio signal of the right channel Therefore, it is possible to
perform accurate binaural processing in line with the asymmetry of a real human being.
10-05-2019
12
[0036]
Further, according to the binaural processing method of the present invention, since the signal
processing is performed on the output audio signal based on the inverse characteristic of the
transfer characteristic of the acoustic output means and the acoustic output is performed, the
desired acoustic characteristic is not affected by the characteristics of the headphone etc. You
can get it right.
Furthermore, according to the binaural processing method of the present invention, the audio
signal reproduced from the binaural recorded by the recording means and reproduced by the
reproduction means is subjected to signal processing based on the inverse characteristic of the
transfer characteristic of the recording means, the reproduction means or the sound output
means. Since the sound output is performed, desired sound characteristics can be correctly
obtained without being influenced by the characteristics of the sound recording system and the
reproduction system in addition to the sound output means.
10-05-2019
13
Документ
Категория
Без категории
Просмотров
0
Размер файла
25 Кб
Теги
jph05300598
1/--страниц
Пожаловаться на содержимое документа