DESCRIPTION
DATA PROCESSING APPARATUS, THE METHOD AND CODING
APPARATUS
TECHNICAL FIELD
[0001] The present invention relates to a data
processing apparatus, the method and a coding apparatus
for performing quantization on image data.
BACKGROUND ART
[0002] In recent years, apparatuses based on methods,
such as the Moving Picture Experts Group (MPEG) for using
image data as digital and compressing by discrete cosine
transformation and other orthogonal transformations and
motion compensation by using redundancy peculiar to image
information for the purpose of efficiently transferring
and accumulating information, have been widespread both
in information distribution by broadcast stations and in
information receiving by general households. In the MPEG
method, transformation coefficients are generated by
performing orthogonal transformation on image data to be
coded and quantization is performed on the transformation
coefficients by a predetermined quantization scale, and
then the quantized image data are coded.
In the MPEG method, the quantization scale is
determined based on a degree of complexity of an image to
be coded, so that the more complex the image becomes, the
smaller the value becomes.
[0003] Following to the MPEG method, coding methods
called the H.264 and JVT (Joined Video Team) for
realizing a still higher compression rate have been
proposed.
In the JVT method coding apparatus, coding in the
JVT method is performed after decoding image data coded
by the MPEG in some cases.
DISCLOSURE OF THE INVENTION
PROBLEMS TO BE SOLVED BY THE INVENTION
[0004] When performing quantization without
considering a quantization scale used in the MPEG method
coding apparatus in the JVT method coding apparatus of
the related art explained above, there is a disadvantage
that, for example, an extremely larger quantization scale
than the quantization scale used in the MPEG method
coding apparatus is selected and information held by the
MPEG method is lost by rough quantization, so that the
image quality is deteriorated in some cases.
Inversely, there is a disadvantage in the JVT
method coding apparatus of the related art explained
above that an extremely smaller quantization scale than
the quantization scale used in the MPEG method coding
apparatus is selected and a large number of bits are
assigned to less information, so that the coding
efficiency declines without improving the image quality.
The same disadvantages may arise also in coding
methods other than the MPEG method and the JVT method.
[0005] It is desired to provide a data processing
apparatus, the method and a coding apparatus for, when
performing second quantization on data to be processed
and obtained by performing inverse quantization after
performing first quantization, suitably performing the
above second quantization in terms of image quality and a
coding efficiency.
MEANS OF SOLVING THE PROBLEMS
[0006] To solve the above disadvantages of the
related art explained above, according to a first
invention, there is provided a data processing apparatus
for performing a second quantization on data to be
processed and obtained by performing inverse quantization
after performing a first quantization by a first
quantization scale, comprising a quantization scale
generation means for generating a second quantization
scale based on the first quantization scale; and a
quantization means for performing the second quantization
on the data to be processed based on the second
quantization scale generated by the quantization scale
generation means.
[0007] An operation of the data processing apparatus
of the first invention is as below.
First, the quantization scale generation means
generates the second quantization scale based on the
first quantization scale,
Next, the quantization means performs the second
quantization on the data to be processed based on the
second quantization scale generated by the quantization
scale generation means.
[0008] According to a second invention, there is
provided a data processing method for performing the
second quantization on data to be processed and obtained
by performing inverse quantization after performing the
first quantization by the first quantization scale,
including: a first step of generating a second
quantization scale based on the first quantization scale;
and a second step of performing the second quantization
on the data to be processed based on the second
quantization scale generated in the first step.
[0009] According to a third invention, there is
provided a coding apparatus, comprising a decoding means
for generating decoding data by decoding coding data
generated by performing coding on motion image data by
the first coding method and obtained by performing the
first quantization based on the first quantization scale
in the coding step; the quantization scale generation
means for generating the second quantization scale based
on the first quantization scale; and a quantization means
for performing second quantization on the decoding data
based on the second quantization scale generated by the
quantization scale generation means in a step of
performing coding in a second coding method which is
different from the first coding method on the decoding
data generated by the decoding means.
[0010] An operation of the coding apparatus of the
third invention is as below.
First, a decoding means generates decoding data by
decoding coding data generated by performing coding on
moving image data by the first coding method and obtained
by performing the first quantization based on the first
quantization scale in the above coding step.
Next, the quantization scale generation means
generates the second quantization scale based on the
first quantization scale.
Next, the quantization means performs the second
quantization on the decoding data based on the second
quantization scale generated by the quantization scale
generation means in a step of coding the decoding data
generated by the decoding means by a second coding method
which is different from the first coding method.
[0011] According to a fourth invention, there is
provided a data processing apparatus for performing the
second quantization on data to be processed and obtained
by performing the inverse quantization after performing
the first quantization by the first quantization scale,
comprising the quantization scale generation circuit for
generating a second quantization scale based on the first
quantization scale; and the quantization circuit for
performing the second quantization on the data to be
processed based on the second quantization scale
generated by the quantization scale generation circuit.
EFFECT OF THE INVENTION
[0012] According to the present invention, when
performing second quantization on data to be processed
and obtained by performing inverse quantization after
performing the first quantization, it is possible to
provide a data processing apparatus, the method and a
coding apparatus for suitably performing the second
quantization in terms of image" quality and a coding
efficiency.
BRIEF DESCRIPTION OF DRAWINGS
[0013] FIG. 1 is a view of the configuration of a
communication system according to a first embodiment of
the present invention;
FIG. 2 is a functional block diagram of a coding
apparatus shown in FIG. 1;
FIG. 3A and FIG. 3B are views for explaining frame
coding and field coding used in the MPEG2 method;
FIG. 4A and FIG. 4B are views for explaining frame
coding and field coding in unit of picture used in the
JVT method;
FIG. 5 is a view for explaining frame coding and
field coding in unit of macro block used in the JVT
method;
FIG. 6 is a view for explaining processing of
performing field coding on MPEG image data in unit of
picture in the JVT method;
FIG. 7 is a view for explaining processing for
performing field coding on MPEG image data in unit of
macro block in the JVT method;
FIG. 8 is a view for explaining processing in an
activity calculation circuit shown in FIG. 2 when
performing field coding in unit of picture as shown in
FIG. 6 in the coding apparatus 2 shown in FIG. 2;
FIG. 9 is a view for explaining an operation
example on determination of a quantization scale and
quantization in the coding apparatus shown in FIG. 2; and
FIG. 10 is a flowchart for explaining processing of
the activity calculation circuit shown in FIG. 2 when
performing field coding in unit of macro block pair as
shown in FIG. 7 in the coding apparatus 2 shown in FIG. 2
EXPLANATION OF REFERENCE
[0014] 1... COMMUNICATION SYSTEM, 2... CODING
APPARATUS, 3... DECODING APPARATUS, 22... A/D CONVERSION
CIRCUIT, 23... PICTURE RELOCATING CIRCUIT, 24...
CALCULATION CIRCUIT, 25... ORTHOGONAL TRANSFORMATION
CIRCUIT, 26... QUANTIZATION CIRCUIT, 27...
REVERSIBLE(LOSSLESS) CODING CIRCUIT, 28... BUFFER, 29...
INVERSE QUANTIZATION CIRCUIT, 30... INVERSE ORTHOGONAL
TRANSFORMATION CIRCUIT, 31... RESTRUCTURING CIRCUIT, 32..
DEBLOCK FILTER, 33... MEMORY, 41... INTRA PREDICTION
CIRCUIT, 42... MOTION PREDICTION/COMPENSATION CIRCUIT,
51... MPEG2 DECODING CIRCUIT, 52... PICTURE TYPE BUFFER,
53... ACTIVITY CALCULATION CIRCUIT, 54... RATE CONTROL
CIRCUIT
BEST MODE FOR CARRYING OUT THE INVENTION
[0015] Below, a JVT method coding apparatus
according to embodiments of the present invention will be
explained.
First Embodiment
In the present embodiment, a JVT method coding
apparatus will be explained with reference to FIG. 1 to
FIG. 9.
First, corresponding relationship of components of
the present invention and components of the present
embodiment will be explained.
In the present embodiment, a function of generating
a quantization scale based on a quantization parameter QP
among functions of an activity calculation circuit 53, a
rate control circuit 54 and a quantization circuit 26
corresponds to a quantization scale generation means of
the first and third inventions.
Also, a function of performing quantization based
on the quantization scale in functions of the
quantization circuit 26 in the present embodiment
corresponds to a quantization means of the first and
third inventions.
Also, an MPEG2 decoding circuit 51 in the present
embodiment corresponds to a decoding means of the third
invention.
[0016] FIG. 1 is a conceptual view of a
communication system 1 of the present embodiment.
As shown in FIG. 1, the communication system 1 has
a coding apparatus 2 provided on the transmission side
and a decoding apparatus 3 provided on the receiving side.
In the communication system 1, in the coding
apparatus 2 on the transmission side, after generating
frame image data (bit stream) compressed by orthogonal
transformation, such as discrete cosine transformation
and Karhunen-Loeve transformation, and motion
compensation and modulating the frame image data, it is
transmitted via transmission media, such as a satellite
broadcast wave, a cable TV network, a telephone line
network and a cellular phone network.
On the receiving side, after decoding a received
image signal, frame image data decompressed by inverse
transformation of the orthogonal transformation at the
time of the above modulation and motion compensation is
generated and used.
Note that the transmission media may be recording
media, such as an optical disk, a magnetic disk and a
semiconductor memory.
A decoding apparatus 3 shown in FIG. 1 performs
decoding in accordance with coding by the coding
apparatus.
[0017] Below, the coding apparatus 2 shown in FIG. 1
will be explained.
FIG. 2 is a view of the overall configuration of
the coding apparatus 2 shown in FIG. 1.
As shown in FIG. 2, the coding apparatus 2 includes,
for example, an A/D conversion circuit 22, a picture
relocating circuit 23, a calculation circuit 24, an
orthogonal transformation circuit 25, a quantization
circuit 26, a reversible coding circuit 27, a buffer 28,
an inverse quantization circuit 29, an inverse orthogonal
transformation circuit 30, a restructuring circuit 31, a
deblock filter 32, a memory 33, an intra prediction
circuit 41, a motion prediction/compensation circuit 42,
a selection circuit 44, an MPEG2 decoding circuit 51, a
picture type buffer memory 52, an activity calculation
circuit 53 and a rate control circuit 54.
[0018] Below, an outline of the coding apparatus 2
will be explained.
In the coding apparatus 2, an MPEG image data Sll
coded by the MPEG2 in the MPEG2 decoding circuit 51 is
decoded to generate image data S51, and the image data
S51 is coded by the JVT method.
The MPEG2 decoding circuit 51 extracts a
quantization scale Qm (first quantization scale of the
present invention) of each macro block MB used in
quantization in the MPEG2 coding step (first quantization
of the present invention) from the MPEG image data Sll
and outputs to the activity calculation circuit 53.
The activity calculation circuit 53 calculates an
activity "Nact" based on the quantization scale Qm and
outputs the same to the rate control circuit 54 as will
be explained later on.
The rate control circuit 54 calculates a
quantization parameter QP of each macro block MB based on
the activity "Nact" input from the activity calculation
circuit 53 and outputs the same to the quantization
circuit 26.
The quantization circuit 26 performs quantization
(the second quantization of the present invention) on the
image data S25 by using a quantization scale (the second
quantization scale of the present invention) determined
based on the quantization parameter QP input from the
rate control circuit 54.
[0019] Next, MPEG2 and JVT coding methods will be
explained.
In either of the MPEG2 and JVT, there are noninterlace
scanning image data and interlace scanning
image data in image data input to the coding apparatus,
and it is possible to select coding in unit of field data
(field coding) and coding in unit of frame data (frame
coding).
In the MPEG2, frame coding may be performed on a
macro block MB composed of data of 16 pixels by 16 pixels,
for example, as shown in FIG. 3A, or field coding may be
performed on each top field data and bottom field data by
dividing to data of 16 pixels by 8 pixels as shown in FIG.
3B.
[0020] Also, in the JVT, it is possible to select
coding in unit of picture as shown in FIG. 4A and FIG. 4B
and coding in unit of macro block as shown in FIG. 5.
As coding in unit of picture, it is possible to
select frame coding shown in FIG. 4A and field coding
shown in FIG. 4B.
Also, as coding in unit of macro block, it is
possible to select the case of performing frame coding or
field coding in unit of single macro block and the case
of performing frame coding or field coding in unit of two
macro blocks MB (MB pair), that is data of 16 pixels by
32 pixels.
[0021] Also, in the present embodiment, as shown in
FIG. 6, respective macro blocks MB(i) and MB(i+l)
adjacent in the vertical direction in frame data FR_m
composing the image data S51 obtained by decoding in the
MPEG2 decoding circuit 51 are quantized based on
quantization scales Qm(i) and Qm(i+l), respectively, in
MPEG coding performed in the past.
The MPEG2 decoding circuit 51 extracts the
quantization scales Qm(i) and Qm(i+l) in the step of
decoding the MPEG image data Sll and outputs to the
activity calculation circuit 53.
Note that each of the macro blocks MB in the MPEG
image data Sll corresponding to the macro blocks MB(i)
and MB(i-l-l) includes both of the quantization scales
Qm(i) and Qm(i + l) .
[0022] Also, when field coding in unit of picture by
the JVT method is performed, JVT image data S2 includes
in a macro block MBjt(i) in a top field TF_j
corresponding to a macro block MBm(i) a quantization
scale Qjt(i) used in the quantization as shown in FIG. 6.
Also, in a macro block MBjb(i) in a bottom field BF_j
corresponding to a macro block MBm(i+l), a quantization
scale Qjb(i) used in the quantization is included.
[0023] On the other hand, furthermore, when field
coding in unit of macro block pair by the JVT method is
performed, in JVT image data S2, as shown in FIG. 7, a
macro block MBj(i) corresponding to the macro block
MBm(i) and a macro block MBj(i+1) corresponding to the
macro block MBm(i+l) are arranged in the same field FI_j.
The macro block MBj(i) includes a quantization
scale Qj(i) used in the quantization, and the macro block
MBj(i+1) includes a quantization scale Qj(i+1) used in
the quantization.
[0024] Below, components of the coding apparatus 2
will be explained.
The A/D conversion circuit 22 converts image data
S10 to be coded and composed of input analog luminance
signal Y and color-difference signals Pb and Pr to image
data S22 in digital, and outputs the same to the picture
relocating circuit 23.
The screen relocating circuit 23 outputs image data
S23 obtained by relocating image data S22 input from the
A/D conversion circuit 22 or image data S51 input from
the MPEG2 decoding circuit 51 in an order of coding in
accordance with the GOP (group of pictures) structure
composed of the picture types I, P and B to the
calculation circuit 24, the intra prediction circuit 41
and the motion prediction/compensation circuit 42.
Below, in the present embodiment, the case where
the screen relocating circuit 23 performs processing on
image data S51 input from the MPEG2 decoding circuit 51.
[0025] The calculation circuit 24 generates image
data S24 indicating a difference between the image data
S23 and prediction image data PI input from a selection
circuit 44 and outputs the same to the orthogonal
transformation circuit 25.
The orthogonal transformation circuit 25 performs
orthogonal transformation, such as discrete cosine
transformation and Karhunen-Loeve transformation, on the
image data S24 to generate image data (for example, a DCT
coefficients) S25 and outputs the same to the
quantization circuit 26.
The quantization circuit 26 performs quantization
on the image data S25 based on the quantization scale
regulated based on the quantization parameter QP input
from the rate control circuit 32 and in accordance with
the quantization parameter QP to generate image data S26
and outputs the same to the reversible coding circuit 27
and inverse quantization circuit 29.
[0026] The reversible coding circuit 27 stores in
the buffer 28 image data obtained by performing variable
length coding or calculation coding on the image data S26,
At this time, when the selection data S44 indicates
that inter prediction coding is selected, the reversible
coding circuit 27 performs coding on a motion vector MV
input from the motion prediction/compensation circuit 58
and stores the same in the header data.
Alternately, when selection data S44 indicates that
intra prediction coding is selected, the reversible
coding circuit 27 stores an intra prediction mode IPM
input from the intra prediction circuit 41 in the header
data, etc.
Also, the reversible coding circuit 27 makes the
quantization scale used in quantization in the
quantization circuit 26 included in respective macro
blocks MB.
[0027] Image data stored in the buffer 28 is
transmitted after being modulated, etc.
The inverse quantization circuit 29 performs
inverse quantization on the image data S26 based on the
quantization scale used in the quantization circuit 26
and outputs the same to the inverse orthogonal
transformation circuit 30.
The inverse orthogonal transformation circuit 30
performs inverse orthogonal transformation corresponding
to the orthogonal transformation used in the orthogonal
transformation circuit 25 on inversely quantized image
data input from the inverse quantization circuit 29 and
outputs the same to the restructuring circuit 31.
The restructuring circuit 31 adds prediction image
data PI input from the selection circuit 44 and image
data input from the inverse orthogonal transformation
circuit 30 to generate restructuring image data and
outputs the same to the deblock filter 32.
After eliminating block strain of image data input
from the restructuring circuit 31, the deblock filter 32
writes the same as reference image data in the memory 33.
[0028] The intra prediction circuit 41 performs
intra prediction coding on the respective macro blocks MB
composing image data read from the memory 33 to generate
prediction image data, for example, based on respective
intra prediction modes regulated in advance by the JVT
and detects a difference DIF between the prediction image
data and the image data S23.
Then, the intra prediction circuit 41 specifies an
intra prediction mode corresponding to the minimum
difference in the above difference generated respectively
for the above plurality of intra prediction modes and
outputs the specified intra prediction mode IPM to the
reversible coding circuit 27.
Also, the intra prediction circuit 41 outputs the
prediction image data PI by the specified intra
prediction mode and the difference DIF to the selection
circuit 44.
[0029] The motion prediction/compensation circuit 42
performs motion prediction processing in unit of frame
data and field data on the image data S23 as explained
with reference to FIG. 4 and FIG. 5 and determines a
motion vector MV based on the reference image data KEF
read from the memory 33.
Namely, the motion/compensation circuit 42
determines a motion vector MV to make the difference DIF
between prediction image data PI regulated by the motion
vector MV and the reference image data KEF and the image
data S23 minimum.
The motion prediction/compensation circuit 42
outputs the prediction image data PI and the difference
DIF to the selection circuit 44 and outputs the motion
vector MV to the reversible coding circuit 27.
Note that the motion prediction/compensation
circuit 42 performs motion prediction/compensation
processing on the respective frame data and field data
based on picture type data PIC_T read from the picture
type buffer memory 52 by applying the same picture type
used in the MPEG coding.
The selection circuit 44 compares the difference
DIF input from the intra prediction circuit 41 and the
difference DIF input from the motion
prediction/compensation circuit 42.
When the selection circuit 44 determines that the
difference DIF input from the intra prediction circuit 41
is smaller from the above comparison, it selects the
prediction image data PI input from the intra prediction
circuit 41 and outputs to the calculation circuit 24.
[0030] When the selection circuit 44 determines that
the difference DIF input from the motion
prediction/compensation circuit 42 is smaller from the
above comparison, it selects the prediction image data PI
input from the motion prediction/compensation circuit 58
and outputs to the calculation circuit 24.
Also, the selection circuit 44 outputs to the
reversible coding circuit 27 selection data S44
indicating that inter prediction coding is selected when
the prediction image data PI input from the intra
prediction circuit 41 is selected, while outputs to the
reversible coding circuit 27 selection data S44
indicating that intra prediction coding is selected when
the prediction data PI input from the motion
prediction/compensation circuit 58 is selected.
[0031] The MPEG2 decoding circuit 51 receives as an
input, for example, MPEG image data Sll, and decodes the
MPEG image data Sll by the MPEG2 to generate image data
S51 and outputs the same to the screen relocating circuit
23.
Also, the MPEG2 decoding circuit 51 writes in the
picture type buffer memory 52 the picture type data PIC_T
included in a header of the image data Sll and indicating
a picture kind of each macro block.
The MPEG2 decoding circuit 51 extracts a
quantization scale Qm of the each macro block used in the
quantization in the MPEG2 coding step from the MPEG image
data Sll in the above decoding and outputs to the
activity calculation circuit 53.
[0032] The picture type data PIC_T stored in the
picture type buffer memory 52 is read by the selection
circuit 44 and the motion prediction/compensation circuit
58.
[0033] The activity calculation circuit 53
calculates an activity Nact based on the quantization
scale Qm input from the MPEG2 decoding circuit 51 and
outputs the same to the rate control circuit 54. FIG. 8
is a view for explaining processing in the activity
calculation circuit 53 shown in FIG. 2 when performing
field coding in unit of picture as shown in FIG. 6 in JVT
coding.
Below, an explanation will be made by taking as an
example calculation of the activity Nact used for
generating macro blocks MBjt(i) and MBjb(t) in the JVT
image data S2 shown in FIG. 6.
Step ST11:
The activity calculation circuit 53 receives as an
input a quantization scale Qm(i) of the macro block
MBm(i) and a quantization scale Qm(i+l) of a macro block
MBm(i+l) shown in FIG. 6 from the MPEG2 decoding circuit
51.
The activity calculation circuit 53 receives as an
input the quantization scales Qm(i) and Qm(i+l) as
arguments of a function ft() shown in the formula (1)
below regulated in advance for a top field TF_j and
specifies a quantization scale Qa_t(i).
[0034] [Formula 1]
(Formula Removed)
[0035] The activity calculation circuit 53 receives
as an input the quantization scales Qm(i) and Qm(i+l) as
arguments of a function fb() shown in the formula (2)
regulated in advance for a bottom field BF_j and
specifies a quantization scale Qa_b(i).
[0036] [Formula 2]
(Formula Removed)
[0037] As the ft() and fb(), for example as shown in
the formula (3) , the smaller of the quantization scales
Qm(i) and Qm(i+l) is selected and used for a function for
specifying the quantization scales Qa_t(i) and Qa_b(i).
[0038] [Formula 3]
(Formula Removed)
[0039] Note that as the functions ft() and fb() , for
example, a function for calculating the quantization
scales Qa_t(i) and Qa_b(i) by calculation shown in the
formula (4) below.
[0040] [Formula 4]
(Formula Removed)
[0041] Step ST12:
The activity calculation circuit 53 calculates an
average value aveQa_t of quantization scales Qa_t(i) of
all block data in the top field TF_j , to which the macro
block MBjt(i) belongs, based on the formula (5) below.
[0042] [Formula 5]
(Formula Removed)
[0043] Also, the activity calculation circuit 53
calculates an average value aveQa_b of quantization
scales Qa_b(i) of all block data in the bottom field BF_j,
to which the macro block MBjb(i) belongs, based on the
formula (6) below.
[0044] [Formula 6]
(Formula Removed)
[0045] Step ST13:
The activity calculation circuit 53 calculates
activity Nact_t(i) by dividing the quantization scale
Qa_t(i) calculated in the step ST11 by the average value
aveQa_t calculated in the step ST12 for each of the macro
blocks MB belonging to the top field TF_j as shown in the
formula (7) below.
[0046] [Formula 7]
(Formula Removed)
[0047] Also, the activity calculation circuit 53
calculates activity Nact_b(i) by dividing the
quantization scale Qa_b(i) calculated in the step ST11 by
the average value aveQa_b calculated in the step ST12 for
each of the macro blocks MB belonging to the bottom field
BF_j as shown in the formula (8) below.
[0048] [Formula 8]
(Formula Removed)
[0049] Step ST14:
The activity calculation circuit 53 outputs the
activities Nact_t(i) and Nact_b(i) calculated in the step
ST13 to the rate control circuit 54.
[0050] The rate control circuit 54 calculates a
quantization parameter QP for each macro block MB based
on the activities Nact t(i) and Nact b(i) input from the
activity calculation circuit 53 and outputs the same to
the quantization circuit 26.
Here, when expressing the activities Nact_t(i) and
Nact_b(i) by the activity Nact(i), a quantization
parameter QP(i) of each macro block MB is expressed by
the formulas (9) and (10) below. Note that "round" in the
formula (9) indicates integer processing by rounding, and
"QPr" in the formula (10) is a reference quantization
parameter regulated by the JVT method, which is regulated
for field data or frame data.
(Formula Removed)
[0053] The rate control circuit 54 outputs a
quantization parameter QP(i) generated as explained above
to the quantization circuit 26.
The quantization circuit 26 performs quantization
on the image data S25 by a quantization scale regulated
in accordance with the quantization parameter QP(i) input
from the rate control circuit 54 to generate image data
S26.
Note that, in the present embodiment, the
quantization scale is regulated to be doubled when the
quantization parameter QP(i) increases by "6".
[0054] FIG. 9 is a view for explaining an operation
example of the coding apparatus 2 regarding determination
of a quantization scale and quantization.
Step ST21:
The MPEG2 decoding circuit 51 extracts a
quantization scale Qm of each macro block used in
quantization in an MPEG2 coding step from the MPEG image
data Sll in the decoding above and outputs to the
activity calculation circuit 53.
Step ST22:
The activity calculation circuit 53 calculates
activities Nact(Nact_t(i) and Nact_b(i)) based on the
quantization scale Qm input from the MPEG2 decoding
circuit 51 in the step ST21 and outputs the same to the
rate control circuit 54.
Step ST23:
The rate control circuit 54 calculates a
quantization parameter QP of each macro block MB based on
the activity Nact input from the activity calculation
circuit 53 in the step ST22 and outputs the same to the
quantization circuit 26.
Step ST24:
The quantization circuit 26 performs quantization
on the image data S25 by a quantization scale regulated
in accordance with the quantization parameter QP(i) input
from the rate control circuit 54 in the step ST23 to
generate image data S26.
[0055] Below, an example of an overall operation of
the coding apparatus 2 when coding by the JVT method the
image data S51 obtained by decoding the MPEG image data
Sll will be explained.
First, the MPEG image data Sll coded by the MPEG2
is input to the MPEG2 decoding circuit 51.
Next, the MPEG2 decoding circuit 51 decodes the
MPEG image data Sll to generate the image data S51 and
outputs the same to the screen relocating circuit 23.
At this time, the MPEG2 decoding circuit 51
extracts a quantization scale Qm of each macro block used
in quantization in the MPEG2 coding step from the MPEG
image data Sll in the above decoding and outputs the same
to the activity calculation circuit 53.
Then, the activity calculation circuit 53
calculates an activity Nact based on the quantization
scale Qm and outputs the same to the rate control circuit
54.
Then, the rate control circuit 54 calculates a
quantization parameter QP of each macro block MB based on
the activity Nact and outputs the same to the
quantization circuit 26.
[0056] Also intra prediction is performed in the
intra prediction circuit 41 and a difference DIF with the
prediction image data PI is output to the selection
circuit 44.
Also, in the motion prediction/compensation circuit
42, motion prediction/compensation processing is
performed and a motion vector MV is specified, and the
prediction image data PI and the difference DIF are
output to the selection circuit 44.
Then, the selection circuit 44 outputs prediction
image data PI corresponding to the smaller difference DIF
of the difference DIF input from the intra prediction
circuit 41 and the difference DIF input from the motion
prediction/compensation circuit 58 to the calculation
circuit 24.
[0057] Next, the calculation circuit 24 generates
image data S24 indicating a difference between the image
data S23 and the prediction image data PI input from the
selection circuit 44 and outputs the same to the
orthogonal transformation circuit 25.
Next, the orthogonal transformation circuit 25
performs orthogonal transformation, such as discrete
cosine transformation and Karhunen-Loeve transformation,
on the image data S24 to generate image data (for example
a DCT coefficients) S25 and outputs the same to the
quantization circuit 26.
Next, the quantization circuit 26 performs
quantization on the image data S25 based on the
quantization scale regulated in accordance with the
quantization parameter QP based on the quantization
parameter QP input from the rate control circuit 32, and
outputs the same to the reversible coding circuit 27 and
the inverse quantization circuit 29.
Next, the reversible coding circuit 27 stores in
the buffer 28 the image data obtained by performing
variable length coding or calculation coding on the image
data S26.
[0058] As explained above, in the coding apparatus 2,
when performing JVT coding on the image data S51 decoded
in the MPEG2 decoding circuit 51, a quantization
parameter QP (quantization scale) of each macro block
used for quantization of the quantization circuit 26
based on the quantization scale Qm used for generating
each macro block MBm of MPEG image data is determined.
Therefore, according to the coding apparatus 2, it
is possible to perform high quality quantization with
less waste in the JVT coding by considering
characteristics of quantization in the MPEG coding
comparing with the case of determining a quantization
parameter QP to be used for quantization by the
quantization circuit 26 without using a quantization
scale Qm.
[0059] Also, according to the coding apparatus 2, as
explained above, quantization scales Qa_t(i) and Qa_b(i)
are generated based on the quantization scales Qm(i) and
Qm(i + l) as shown in the above formulas (3) and (4) in the
activity calculation circuit 53, and by determining a
quantization scale to be used in the quantization circuit
26 based thereon, it is possible to prevent selecting an
extremely larger or smaller quantization scale than the
quantization scale used in the MPEG method coding in
quantization in the JVT method coding.
Therefore, according to the coding apparatus 2, it
is possible to perform suitable quantization in the
quantization circuit 26 in terms of image quality and a
coding efficiency. Namely, it is possible to prevent in
the JVT coding a wasteful loss of information held in
MPEG coding or assignment of an unnecessarily large
amount of bits to information already lost in the MPEG
coding.
[0060] Second Embodiment
In the above first embodiment, an explanation was
made on the processing of the activity calculation
circuit 53 shown in FIG. 2 in the case of performing
field coding in unit of picture as shown in FIG. 6.
In the present embodiment, an explanation will be
made on processing of the activity calculation circuit 53
shown in FIG. 2 when performing field coding in unit of
micro block as shown in FIG. 7.
FIG. 10 is a flowchart for explaining processing of
the activity calculation circuit 53 show in FIG. 2 when
performing field coding in unit of macro block as shown
in FIG. 7.
Below, calculation of activity Nact of macro blocks
MBj(i) and MBj(i+1) in JVT image data S2 shown in FIG. 7
will be explained as an example.
Step ST31:
The activity calculation circuit 53 receives as an
input a quantization scale Qm(i) of the macro block
MBm(i) and a quantization scale Qm(i+l) of the macro
block MBm(i+l) shown in FIG. 6 from the MPEG2 decoding
circuit 51.
The activity calculation circuit 53 receives as an
input the quantization scales Qm(i) and Qm(i+l) as
arguments of the function fl() shown in the formula (11)
below and specifies a quantization scale Qa(i).
[0061] [Formula 11]
(Formula Removed)
[0062] Also, the activity calculation circuit 53
receives as an input the quantization scales Qm(i) and
Qm(i+l) as arguments of the function f2() shown in the
formula (12) below and specifies a quantization scale
Qa(i-fl) .
[0063] [Formula 12]
(Formula Removed)
[0064] As the functions fl() and f2(), for example,
those as same as the above formulas (3) and (4) are used.
[0065] Step ST32:
The activity calculation circuit 53 calculates an
average value aveQa of quantization scales Qa(i) and
Qa(i+l) of all block data in the field FI_j, to which the
macro blocks MBj(i) and MBj(i+1) belong, based on the
formula (13) below.
[0066] [Formula 13]
(Formula Removed)
[0067] Step ST33:
The activity calculation circuit 53 calculates an
activity Nact(i) by dividing the quantization scale Qa(i)
calculated in the step ST31 by the average value aveQa
calculated in the step ST32 as shown in the formula (14)
below.
[0068] [Formula 14]
(Formula Removed)
[0069] Also, the activity calculation circuit 53
calculates an activity Nact(i+l) by dividing the
quantization scale Qa(i+l) calculated in the step ST41 by
the average value aveQa calculated in the step ST32 as
shown in the formula (15) below.
[0070] [Formula 15]
(Formula Removed)
[0071] Step ST34:
The activity calculation circuit 53 outputs the
activities Nact(i) and Nact(i+l) calculated in the step
ST33 to the rate control circuit 54.
[0072] The same effects as those in the first
embodiment can be obtained also in the present embodiment,
[0073] The present invention is not limited to the
above embodiments.
For example, in the above embodiments, in the
coding apparatus 2, the case of performing field coding
by the JVT method was explained as an example, but frame
coding may be performed.
In this case, for example, in the step ST12 shown
in FIG. 8, the activity calculation circuit 53 calculates
an average value aveQa of quantization scales Qa of all
block data in the frame data, to which the macro block
belongs, and based thereon, an activity Nact is generated.
Also, in the above embodiments, motion image data
was explained as an example of data to be processed in
the present invention, but data to be processed in the
present invention may be still image data or audio data.
INDUSTRIAL APPLICABILITY
[0074] The present invention can be applied to
coding systems for coding image data.

What is claimed is:
1. A data processing apparatus for performing a second
quantization on data to be processed and obtained by
performing inverse quantization after performing a first
quantization by a first quantization scale, comprising:
a quantization scale generation means for
generating a second quantization scale based on the first
quantization scale; and
a quantization means for performing the second
quantization on the data to be processed based on the
second quantization scale generated by the quantization
scale generation means.
2. A data processing apparatus as set forth in claim 1,
wherein:
the quantization scale generation means generates
the second quantization scale based on the first
quantization scale for each of a plurality of block data
composing image data as the data to be processed; and
the quantization means performs the second
quantization on the block data based on the second
quantization scale generated by the second quantization
scale generation means in accordance with the block data.
3. A data processing apparatus as set forth in claim 2,
wherein the quantization scale generation means generates
index data indicating a complexity degree of the block
data to be processed based on the first quantization
scale used in the first quantization for obtaining the
block data to be processed or the block data around the
block data, and generates the second quantization scale
of the block data to be processed based on the index data,
4. A data processing apparatus as set forth in claim 3,
wherein:
when two block data MBm(i) and MBm(i+l)
respectively corresponding to adjacent two block image
positions in the vertical direction in the image data are
subjected to the first quantization based on the first
quantization scales Qm(i) and Qm(i+l), respectively,
the quantization scale generation means calculates
the second quantization scale Q(i) to be used when
performing the second quantization on the block data
MBm(i) based on both of the first quantization scales
Qm(i) and Qm(i+l) and the second quantization scale
Q(i+l) to be used when performing the second quantization
on the block data MBm(i+l); and
the quantization means performs the second
quantization on the block data MBm(i) based on the second
quantization scale Q(i) calculated by the quantization
scale generation means and performs the second
quantization on the block data MBm(i+l) based on the
second quantization scale Q(i + l).
5. A data processing apparatus as set forth in claim 4,
furthermore comprising:
a control means for generating first field data
configured based on block data MBjt(i) obtained by
performing the second quantization on the block data
MBm(i); and second field data to form a pair with the
first field data, configured based on block data MBjb(i)
obtained by performing the second quantization on the
block data MBm(i+l), in the case where field coding at a
picture level is performed on the image data.
6. A data processing apparatus as set forth in claim 4,
further comprising:
a control means for generating field data
configured based on block data MBj(i) and MBj(i+1)
obtained respectively by performing the second
quantization on the block data MBm(i) and MBm(i-t-l) in the
case where field coding in unit of the two block data
MBm(i) and MBm(i+l) is performed on the image data.
7. A data processing apparatus as set forth in claim 4,
wherein:
the quantization scale generation means specifies a
quantization scale Qa based on a predetermined function
using the first quantization scales Qm(i) and Qm(i+l) as
arguments, and calculates the second quantization scales
Q(i) and Q(i+l) based on the specified quantization scale
Qa.
8. A data processing apparatus as set forth in claim 1,
wherein:
the quantization scale generation means specifies
the quantization scale Qa based on the function using the
smaller of the first quantization scales Qm and Qm(i+l)
as a quantization scale Qa.
9. A data processing apparatus as set forth in claim 7,
wherein:
the quantization scale generation means specifies
the quantization scale Qa based on the function for
calculating a quantization scale Qa by calculating
(Qm(i)+Qm(i + l)+l)/2.
10. A data processing apparatus as set forth in claim 1,
wherein:
the quantization scale generation means
calculates an average value ave of the
quantization scales Qa of all of the block data in field
data or frame data, to which the block data to be
processed belongs,
calculates an activity Nact by dividing the
quantization scale Qa of the block data to be processed
by the average value ave, and
calculates the second quantization scale of the
block data to be processed based on the activity Nact.
11. A data processing method for performing second
quantization on data to be processed and obtained by
performing inverse quantization after performing first
quantization by a first quantization scale, including:
a first step of generating a second quantization
scale based on the first quantization scale; and
a. second step of performing the second quantization
on the data to be processed based on the second
quantization scale generated in the first step.
12. A coding apparatus, comprising:
a decoding means for generating decoding data by
decoding coding data generated by performing coding on
motion image data by a first coding method and obtained
by performing first quantization based on a first
quantization scale in the coding step;
a quantization scale generation means for
generating a second quantization scale based on the first
quantization scale; and
a quantization means for performing second
quantization on the decoding data based on the second
quantization scale generated by the quantization scale
generation means in a step of performing coding in a
second coding method which is different from the first
coding method on the decoding data generated by the
decoding means.
13. A data processing apparatus for performing second
quantization on data to be processed and obtained by
performing inverse quantization after performing first
quantization by a first quantization scale, comprising:
a quantization scale generation circuit for
generating a second quantization scale based on the first
quantization scale; and
a quantization circuit for performing the second
quantization on the data to be processed based on the
second quantization scale generated by the quantization
scale generation circuit.