Note: Descriptions are shown in the official language in which they were submitted.
CA 02646278 2008-07-22
WO 2007/091870 PCT/KR2007/000730
Description
METHOD FOR ENCODING AND DECODING OBJECT-BASED
AUDIO SIGNAL AND APPARATUS THEREOF
Technical Field
[1] The present invention relates to methods and apparatuses for encoding and
decoding an audio signal, and more particularly, to methods and apparatuses
for
encoding and decoding an audio signal which can localize a sound image in a
desired
spatial location for each object audio signal.
[2]
Background Art
[3] In general, in a typical object-based audio encoding method, an object
encoder
generates a down-mix signal by down-mixing a plurality of object audio signals
and
generates parameter information including a plurality of pieces of information
extracted from the object audio signals. In a typical object-based audio
decoding
method, an object decoder restores a plurality of object audio signals by
decoding a
received down-mix signal using object-based parameter information, and a
renderer
synthesizes the object audio signals into a 2-channel signal or a multi-
channel signal
using control data, which is necessary for designating the positions of the
restored
object audio signals.
[4] However, the control data is simply inter-level information, and there is
a clear
limitation in creating 3D effects by performing sound image localization
simply using
level information.
[5]
Disclosure of Invention
Technical Problem
[6] The present invention provides methods and apparatuses for encoding and
decoding
an audio signal which can localize a sound image in a desired spatial location
for each
object audio signal.
[7]
Technical Solution
[8] According to an aspect of the present invention, there is provided a
method of
decoding an audio signal. The method includes extracting a down-mix signal and
object-based parameter information from an input audio signal, generating an
object-
audio signal using the down-mix signal and the object-based parameter
information,
and generating an object audio signal with three-dimensional (3D) effects by
applying
3D information to the object audio signal.
2
WO 2007/091870 PCT/KR2007/000730
[9] According to another aspect of the present invention, there is provided an
apparatus
for decoding an audio signal. The apparatus includes a demultiplexer which
extracts a
down-mix signal and object-based parameter information from an input audio
signal,
an object decoder which generates an object-audio signal using the down-mix
signal
and the object-based parameter information, and a renderer which generates a
three-
dimensional object audio signal with 3D effects by applying 3D information to
the
object audio signal.
[10] According to another aspect of the present invention, there is provided a
method of
decoding an audio signal. The method includes extracting a down-mix signal and
object-based parameter information from an input audio signal, generating
channel-
based parameter information by converting the object-based parameter
information,
generating an audio signal using the down-mix signal and the channel-based
parameter
information, and generating an audio signal with 3D effects by applying 3D in-
formation to the audio signal.
[11] According to another aspect of the present invention, there is provided
an apparatus
for decoding an audio signal. The apparatus includes a demultiplexer which
extracts a
down-mix signal and object-based parameter information from an input audio
signal, a
renderer which withdraws 3D information using index data and outputs the 3D in-
formation, a transcoder which generates channel-based parameter information
using
the object-based parameter information and the 3D information, and a multi-
channel
decoder which generates an audio signal using the down-mix signal and the
channel-
based parameter information and generates an audio signal with 3D effects by
applying
3D information to the audio signal.
[12] According to another aspect of the present invention, there is provided
an apparatus
for decoding an audio signal. The apparatus includes a demultiplexer which
extracts a
down-mix signal and object-based parameter information from an input audio
signal, a
renderer which withdraws 3D information using input index data and outputs the
3D
information, a transcoder which converts the object-based parameter
information into
channel-based parameter information, converts the 3D information into channel-
based
3D information and outputs the channel-based parameter information and the
channel-
based 3D information, and a multi-channel decoder which generates an audio
signal
using the down-mix signal and the channel-based parameter information and
generates
an audio signal with 3D effects by applying the channel-based 3D information
to the
audio signal.
[13] According to another aspect of the present invention, there is provided a
method of
encoding an audio signal. The method includes generating a down-mix signal by
down-mixing an object audio signal, extracting information regarding the
object audio
signal and generating object-based parameter information based on the
extracted in-
CA 02646278 2008-07-22
3
WO 2007/091870 PCT/KR2007/000730
formation, and inserting index data into the object-based parameter
information, the
index data being necessary for searching for 3D information which is used to
create 3D
effects for the object audio signal.
[14] According to another aspect of the present invention, there is provided a
computer-
readable recording medium having recorded thereon a program for executing one
of
the above-mentioned methods.
Advantageous Effects
[15] As described above, according to the present invention, it is possible to
provide a
more vivid sense of reality than in typical object-based audio encoding and
decoding
methods during the reproduction of object audio signals by localizing a sound
image
for each of the object audio signals while making the utmost use of typical
object-
based audio encoding and decoding methods. In addition, it is possible to
create a high-
fidelity virtual reality by applying the present invention to interactive
games in which
position information of game characters manipulated via a network by game
players
varies frequently.
[16]
Brief Description of the Drawings
[17] FIG. 1 illustrates a block diagram of a typical object-based audio
encoding
apparatus;
[18] FIG. 2 is a block diagram of an apparatus for decoding an audio signal
according to
an embodiment of the present invention;
[19] FIG. 3 illustrates a flowchart illustrating an operation of the apparatus
illustrated in
FIG. 2;
[20] FIG. 4 illustrates a block diagram of an apparatus for decoding an audio
signal
according to another embodiment of the present invention;
[21] FIG. 5 illustrates a flowchart illustrating an operation of the apparatus
illustrated in
FIG. 4;
[22] FIG. 6 illustrates a block diagram of an apparatus for decoding an audio
signal
according to another embodiment of the present invention;
[23] FIG. 7 illustrates the application of three-dimensional (3D) information
to frames
by the apparatus illustrated in FIG. 6;
[24] FIG. 8 illustrates a block diagram of an apparatus for decoding an audio
signal
according to another embodiment of the present invention; and
[25] FIG. 9 illustrates a block diagram of an apparatus for decoding an audio
signal
according to another embodiment of the present invention
Best Mode for Carrying Out the Invention
[26] The present invention will hereinafter be described more fully with
reference to the
CA 02646278 2008-07-22
4
WO 2007/091870 PCT/KR2007/000730
accompanying drawings, in which exemplary embodiments of the invention are
shown.
[27] Methods and apparatuses for encoding and decoding an audio signal
according to
the present invention can be applied to, but not restricted to, object-based
audio
encoding and decoding processes. In other words, methods and apparatuses for
encoding and decoding an audio signal according to the present invention
according to
the present invention can also be applied to various signal processing
operations, other
than those set forth herein, as long as the signal processing operations meet
a few
conditions. Methods and apparatuses for encoding and decoding an audio signal
according to the present invention according to the present invention can
localize
sound images of object audio signals in desired spatial locations by applying
three-
dimensional (3D) information such as a head related transfer function (HRTF)
to the
object audio signals.
[28] FIG. 1 illustrates a typical object-based audio encoding apparatus.
Referring to
FIG. 1, the object-based audio encoding apparatus includes an object encoder
110 and
a bitstream generator 120.
[29] The object encoder 110 receives N object audio signals, and generates an
object-
based down-mix signal and object-based parameter information including a
plurality of
pieces of information extracted from the N object audio signals. The plurality
of pieces
of information may be energy difference and correlation values.
[30] The bitstream generator 120 generates a bitstream by combining the object-
based
down-mix signal and the object-based parameter information generated by the
object
encoder 110. The bitstream generated by the bitstream generator 120 may
include
default mixing parameters necessary for default settings for a decoding
apparatus. The
default mixing parameters may include index data necessary for searching for
3D in-
formation such as an HRTF, which can be used to create 3D effects.
[31] FIG. 2 illustrates an apparatus for decoding an audio signal according to
an
embodiment of the present invention. The apparatus illustrated in FIG. 2 may
be
designed by combining the concept of HRTF-based 3D binaural localization to a
typical object-based encoding method. A HRTF is a transfer function which
describes
the transmission of sound waves between a sound source at an arbitrary
location and
the eardrum, and returns a value that varies according to the direction and
altitude of
the sound source. If a signal with no directivity is filtered using the HRTF,
the signal
may be heard as if it were reproduced from a certain direction.
[32] Referring to FIG. 2, the apparatus includes a demultiplexer 130, an
object decoder
140, a renderer 150, and a 3D information database 160.
[33] The demultiplexer 130 extracts a down-mix signal and object-based
parameter in-
formation from an input bitstream. The object decoder 140 generates an object
audio
signal based on the down-mix signal and the object-based parameter
information. The
CA 02646278 2008-07-22
5
WO 2007/091870 PCT/KR2007/000730
3D information database 160 is a database which stores 3D information such as
an
HRTF, and searches for and outputs 3D information corresponding to input index
data.
The renderer 150 generates a 3D signal using the object audio signal generated
by the
object decoder 140 and the 3D information output by the 3D information
database 160.
[34] FIG. 3 illustrates an operation of the apparatus illustrated in FIG. 2.
Referring to
FIGS. 2 and 3, when a bitstream transmitted by an apparatus for encoding an
audio
signal is received (S 170), the demultiplexer 130 extracts a down-mix signal
and object-
based parameter information from the bitstream (S172). The object decoder 140
generates an object audio signal using the down-mix signal and the object-
based
parameter information (S 174).
[35] The renderer 150 withdraws 3D information from the 3D information
database 160
using index data included in control data, which is necessary for designating
the
positions of object audio signals (S 176). The renderer 150 generates a 3D
signal with
3D effects by performing a 3D rendering operation using the object audio
signal
provided by the object decoder 110 and the 3D information provided by the 3D
in-
formation database 160 (S178).
[36] The 3D signal generated by the renderer 150 may be a 2-channel signal
with three
or more directivities and can thus be reproduced as a 3D stereo sound by 2-
channel
speakers such as headphones. In other words, the 3D signal generated by the
renderer
150 may be reproduced by 2-channel speakers so that a user can feel as if the
3D
down-mix signal were reproduced from a sound source with three or more
channels.
The direction of a sound source may be determined based on at least one of the
difference between the intensities of two sounds respectively input to both
ears, the
time interval between the two sounds, and the difference between the phases of
the two
sounds. Therefore, the 3D renderer 150 can generate a 3D signal based on how
the
humans can determine the 3D position of a sound source with their sense of
hearing.
[37] An apparatus for encoding an audio signal may include index data
necessary for
withdrawing 3D information in default mixing parameter information for default
settings. In this case, the renderer 150 may withdraw 3D information from the
3D in-
formation database 160 using the index data included in the default mixing
parameter
information.
[38] An apparatus for encoding an audio signal may include, in control data,
index data,
which is necessary for searching for 3D information such as an HRTF that can
be used
to create 3D effects for an object signal. In other words, mixing parameter
information
included in control data used by an apparatus for encoding an audio signal may
include
not only level information but also index data necessary for searching for 3D
in-
formation. The mixing parameter information may be time information such as
inter-
channel time difference information, position information, or a combination of
the
CA 02646278 2008-07-22
6
WO 2007/091870 PCT/KR2007/000730
level information and the time information.
[39] If there are a plurality of object audio signals and 3D effects need to
be added to
one or more of the plurality of object audio signals, 3D information
corresponding to
given index data is searched for and withdrawn from the 3D information
database 160,
which stores 3D information specifying the target positions of the object
audio signals
to which the 3D effects are to be added. Then, the 3D renderer 150 performs a
3D
rendering operation using the withdrawn 3D information so that the 3D effects
can be
created. 3D information regarding all object signals may be used as mixing
parameter
information. If 3D information is applied only to a few object signals, level
in-
formation and time information regarding object signals, other than the few
object
signals, may also be used as mixing parameter information.
[40] FIG. 4 illustrates an apparatus for decoding an audio signal according to
another
embodiment of the present invention. Referring to FIG. 4, the apparatus
includes a
multi-channel decoder 270, instead of an object decoder.
[41] More specifically, the apparatus includes a demultiplexer 230, a
transcoder 240, a
renderer 250, a 3D information database 260, and the multi-channel decoder
270.
[42] The demultiplexer 230 extracts a down-mix signal and object-based
parameter in-
formation from an input bitstream. The renderer 250 designates the 3D position
of each
object signal using 3D information corresponding to index data included in
control
data. The transcoder 230 generates channel-based parameter information by syn-
thesizing object-based parameter information and 3D position information of
each
object audio signal provided by the renderer 250. The multi-channel decoder
270
generates a 3D signal using the down-mix signal provided by the demultiplexer
230
and the channel-based parameter information provided by the transcoder 230.
[43] FIG. 5 illustrates an operation of the apparatus illustrated in FIG. 4.
Referring to
FIGS. 4 and 5, the apparatus receives a bitstream (S280). The demultiplexer
230
extracts an object-based down-mix signal and object-based parameter
information
from the received bitstream (S282). The renderer 250 extracts index data
included in
control data, which is used to designate the positions of object audio
signals, and
withdraws 3D information corresponding to the index data from the 3D
information
database 260 (S284). The positions of the object audio signals primarily
designated by
default mixing parameter information may be altered by designating 3D
information
corresponding to desired positions of the object audio signals using mixing
control
data.
[44] The transcoder 230 generates channel-based parameter information
regarding M
channels by synthesizing object-based parameter information regarding N object
signals, which is transmitted by an apparatus for encoding an audio signal,
and 3D
position information of each of the object signals, which is obtained using 3D
in-
CA 02646278 2008-07-22
7
WO 2007/091870 PCT/KR2007/000730
formation such as an HRTF by the renderer 250 (S286).
[45] The multi-channel decoder 270 generates an audio signal using the object-
based
down-mix signal provided by the demultiplexer 230 and the channel-based
parameter
information provided by the transcoder 230, and generates a multi-channel
signal by
perfonning a 3D rendering operation on the audio signal using 3D information
included in the channel-based parameter information (S290).
[46] FIG. 6 illustrates an apparatus for decoding an audio signal according to
another
embodiment of the present invention. The apparatus illustrated in FIG. 6 is
different
from the apparatus illustrated in FIG. 4 in that a transcoder 440 transmits
channel-
based parameter information and 3D information separately to a multi-channel
decoder
470. In other words, the transcoder 440, unlike the transcoder 240 illustrated
in FIG. 4,
transmits channel-based parameter information regarding M channels, which is
obtained using object-based parameter information regarding N object signals,
and 3D
information, which is applied to each of the N object signals, to the multi-
channel
decoder 470, instead of transmitting channel-based parameter information
including
3D information.
[47] Referring to FIG. 7, channel-based parameter information and 3D
information have
their own frame index data. Thus, the multi-channel decoder 470 can apply 3D
in-
formation to a predetermined frame of a bitstream by synchronizing the channel-
based
parameter information and the 3D information using the frame indexes of the
channel-
based parameter information and the 3D information. For example, referring to
FIG. 7,
3D information corresponding to index 2 can be applied to the beginning of
frame 2
having index 2.
[48] Even if 3D information is updated over time, it is possible to determine
where in
channel-based parameter information the 3D information needs to be applied to
by
referencing a frame index of the 3D information. In other words, the
transcoder 440
may insert frame index information into channel-based parameter information
and 3D
information, respectively, in order for the multi-channel decoder 470 to
temporally
synchronize the channel-based parameter information and the 3D information.
[49] FIG. 8 illustrates an apparatus for decoding an audio signal according to
another
embodiment of the present invention. The apparatus illustrated in FIG. 8 is
different
from the apparatus illustrated in FIG. 6 in that the apparatus illustrated in
FIG. 8
further includes a preprocessor 543 and an effect processor 580 in addition to
a de-
multiplexer 530, a transcoder 547, a renderer 550, and a 3D information
database 560,
and that the 3D information database 560 is included in the renderer 550.
[50] More specifically, the structures and operations of the demultiplexer
530, the
transcoder 547, the renderer 560, the 3D information database 560, and the
multi-
channel decoder 570 are the same as the structures and operations of their
respective
CA 02646278 2008-07-22
8
WO 2007/091870 PCT/KR2007/000730
counterparts illustrated in FIG. 6. Referring to FIG. 8, the effect processor
580 may
add a predetermined effect to a down-mix signal. The preprocessor 543 may
perform a
preprocessing operation on, for example, a stereo down-mix signal, so that the
position
of the stereo down-mix signal can be adjusted. The 3D information database 560
may
be included in the renderer 550.
[51] FIG. 9 illustrates an apparatus for decoding an audio signal according to
another
embodiment of the present invention. The apparatus illustrated in FIG. 9 is
different
from the apparatus illustrated in FIG. 8 in that a unit 680 for generating a
3D signal is
divided into a multi-channel decoder 670 and a memory 675. Referring to FIG.
9, the
multi-channel decoder 670 copies 3D information, which is stored in an
inactive
memory of the multi-channel decoder 670, to the memory 675, and the memory 675
performs a 3D rendering operation using the 3D information. The 3D information
copied to the memory 675 may be updated with 3D information output by a
transcoder
647. Therefore, it is possible to generate a 3D signal using desired 3D
information
without any modifications to the structure of multi-channel decoder 670.
[52] The present invention can be realized as computer-readable code written
on a
computer-readable recording medium. The computer-readable recording medium may
be any type of recording device in which data is stored in a computer-readable
manner.
Examples of the computer-readable recording medium include a ROM, a RAM, a CD-
ROM, a magnetic tape, a floppy disc, an optical data storage, and a carrier
wave (e.g.,
data transmission through the Internet). The computer-readable recording
medium can
be distributed over a plurality of computer systems connected to a network so
that
computer-readable code is written thereto and executed therefrom in a
decentralized
manner. Functional programs, code, and code segments needed for realizing the
present invention can be easily construed by one of ordinary skill in the art.
[53] Other implementations are within the scope of the following claims.
Industrial Applicability
[54] The present invention can be applied to various object-based audio
decoding
processes and can provide a vivid sense of reality during the reproduction of
object
audio signals by localizing a sound image for each of the object-audio
signals.
CA 02646278 2008-07-22