Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-04-15
2003-09-30
Philippe, Gims S. (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C382S236000, C348S699000
Reexamination Certificate
active
06628713
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a device and a method for data encoding and a method for data transmission, and more particularly, is suitably applied to a device and a method for data encoding and a method for data transmission for compressively coding picture data, for example.
2. Description of Related Art
Various compressive coding methods have been provided as methods for recording and transmitting picture data in a reduced amount of information when picture data is recorded in the record media such as a magneto-optical disk and a magnetic tape as digital data or when picture data is transmitted via a specified transmission media. As typical of them, there is a Moving Picture Experts Group Phase 2 (MPGE2) method. Nowadays a digital broadcasting system for compressively coding picture data using that MPEG2 method and broadcasted it via ground waves or satellite waves is begun.
In this compressive coding method by the MPEG2 method, picture data is encoded with defining picture data composed of fifteen frames as one processing unit called group of pictures (GOP). For instance, as shown in
FIGS. 1A and 1B
, in one GOP (frames F
0
-F
14
), there are an intra-picture: intraframe-coded picture (I-picture), a predictive-picture: interframe forward predictively-coded picture (P-picture) and a bidirectionally predictive-picture: bidirectionally predictively-coded picture (B picture).
The I-picture is used to keep the independence of GOP, and the entire picture is coded (intra-coding). The P-picture is predictively coded in the forward direction using an I-picture and a P-picture existing in the temporal past (forward predictive coding). And the B-picture is predictively coded from both directions using an I-picture or a P-picture existing in the temporal past and future (bidirectional predictive coding).
By the way, as shown in
FIG. 2
, if a scene changes in a series of pictures to be transmitted, pictures are largely different between a scene A and a scene B. So that, even if a frame picture after the scene change is predicted according to a reference picture before the scene change, an extremely large error is included in difference data between the predicted frame picture and the reference picture, and thus, image quality deteriorates. Moreover, if another frame picture is compressively coded using that frame picture deteriorated in quality as a reference picture, image quality further deteriorates.
To prevent such things, as shown in
FIG. 3
, the GOP
2
of the scene A is ended at the time point of a scene change, and from a scene B after the scene change, an I-picture is inserted and a new GOP
3
is begun. Thereby, predictive coding can be performed without using the frame pictures of the GOP
2
as reference pictures but the I-picture in the GOP
3
after the scene change as a reference picture. Thus, the propagation of the deterioration of image quality can be prevented.
Here, the configuration of a data encoding device for performing scene change detection is shown in FIG.
4
. In
FIG. 4
, the data encoding device
100
supplies picture data D
100
to a scene change detection circuit
101
and a delay circuit
103
individually.
The scene change detection circuit
101
calculates the absolute difference sum of brightness, e.g., between two successive frames. For instance, if absolute difference sum of brightness between the P-picture of a frame F
5
and the B-picture of the frame F
6
in the GOP
2
exceeds a prescribed threshold value, the scene change detection circuit
101
judges that a scene is changed from the scene A to the scene B, and transmits scene change information data D
101
to a picture type decision circuit
102
at this time.
To prevent the propagation of the deterioration of image quality to be caused by predictively coding a frame picture after the scene change using the frame picture before the scene change as a reference frame, the picture type decision circuit
102
specifies its picture type (for instance, the I-picture in the GOP
3
in
FIG. 5
is to be a reference frame) when a frame picture after the scene change is encoded, and transmits specifying data D
102
to a data encoding device
50
.
On the other hand, the picture data D
100
inputted to the delay circuit
103
is delayed for a time when the scene change detection circuit
101
and the picture type decision circuit
102
are performing the processing. The picture data D
100
is supplied to the data encoding device
50
synchronizing with the timing that the specifying data D
102
is supplied.
Here, the data encoding device
50
inputs the picture data D
100
to a pre-filter
51
as shown in FIG.
6
. The pre-filter
51
applies band limitation on the picture data D
100
according to a frequency characteristic control signal S
55
supplied from a quantization rate control part
55
, and reducing the high frequency component of the above picture data D
100
, and transmits it to pixel conversion part
52
as band-limited picture data D
51
.
The pixel conversion part
52
applies pixel conversion processing on band-limited picture data D
51
. That is, assuming a number of pixels in the horizontal direction of the picture data D
100
as horizontal pixel M and a number of pixels in the vertical direction of that as vertical pixel N, the pixel conversion part
52
reduces the horizontal pixel M and the vertical pixel N of the band-limited picture data D
51
that has been obtained by band-limiting the above picture data D
100
into horizontal pixel m and vertical pixel n that satisfy the relation of m<M and n<N, and transmits them to an encoding part
53
as pixel-converted picture data D
52
.
Here, the horizontal pixel m and the vertical pixel n are set according to the image contents of the picture data D
100
, in large values in image contents requiring being high image quality, and in small values in image .contents not requiring.
The encoding part
53
performs compressive coding to the pixel-converted picture data D
52
by applying motion compensation processing, discrete cosine transform (DCT) processing, quantization processing and variable length coding (VLC) processing, and then transmits it to a transmission buffer
54
as variable-length coded data D
53
. At this time, the encoding part
53
modulates a quantization step size in quantization processing based on a quantization control signal S
56
supplied from the quantization rate control part
55
.
Moreover, an encoding part control part
57
supplies picture types based on the specifying data D
102
supplied from the picture type decision circuit
102
, coding timings and a search range of motion vectors in motion compensation processing, to the encoding part
53
as encoding part control information D
57
. Thereby, the above encoding part
53
sets picture types, encoding timings and a motion vector search range in motion compensation processing according to the above encoding part control information D
57
.
A code amount to be generated in each frame of the variable-length coded data D
53
varies according to the image content of the picture data D
100
. Therefore, the encoding part
53
controls a quantization step size in the quantization processing and band limitation in the pre-filter
51
and accumulates the variable-length coded data D
53
for several frames in the transmission buffer
54
, and then outputs as fixed-length coded data D
50
in which a code amount generated per GOP has been controlled in fixed amount for a fixed period.
That is, the quantization rate control part
55
always monitors the accumulation of the variable-length coded data D
53
in the transmission buffer
54
and .obtaining this accumulation as occupancy rate information S
54
, and generate a quantization control signal S
56
and a frequency characteristic control signal S
55
based on the above occupancy rate information S
54
, and supplies this to the encoding part
53
and the pre-filter
51
. In this manner, the encoding part
53
generates the fixed-length coded data D
50
in
Funado Shigeto
Kojima Takashi
Koyanagi Hideki
Murakami Yoshihiro
Philippe Gims S.
Sonnenschein Nath & Rosenthal
LandOfFree
Method and device for data encoding and method for data... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for data encoding and method for data..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for data encoding and method for data... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3056840