Image analysis – Image compression or coding – Lossless compression
Reexamination Certificate
1999-03-17
2002-11-19
Couso, Jose L. (Department: 2621)
Image analysis
Image compression or coding
Lossless compression
C375S240230
Reexamination Certificate
active
06483947
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to video signal processing apparatuses and, more particularly, to a video signal processing apparatus that reduces the quantity of data which have been digitized and compressively coded.
BACKGROUND OF THE INVENTION
When a video which is an analog signal is digitized to be handled as a digital video signal, it can be handled integrally with digitized audio signal, character data, program and the like and, moreover, data processing such as data compression or encryption is easily carried out.
Image data of a digital video signal or the like is an array of pixels which are discrete digital data having pixel values indicating luminance and chrominance, and it is processed for every frame corresponding to one image. When image data of each frame is compressively coded for recording or transmission, intra-frame coding based on spatial correlation within one frame is performed, and the compression ratio can be increased by performing, as well as intra-frame coding, inter-frame coding based on temporal correlation between frames which are close each other in time series. According to intra-frame coding, I pictures, which can be decoded independently, are obtained. According to inter-frame coding, P pictures are obtained by performing prediction in the forward direction while B pictures are obtained by performing prediction in both of the forward and backward directions. Although the compression ratio of the P or B pictures is increased by processing the difference data, these pictures cannot be decoded independently.
In a general compressive coding technique such as MPEG, discrete cosine transform (DCT) is carried out. DCT is performed on each pixel block comprising a predetermined number of pixels. In the state before DCT, the pixel values are randomly distributed over the pixel block. However, in DCT coefficients obtained as the result of DCT, large pixel values are unevenly distributed in a low frequency band. So, the DCT coefficients are subjected to quantization (i.e., division) to make the values of high-band components
0
, and further subjected to variable-length coding based on a series of the same value, thereby obtaining compressed variable-length code ward data. Usually, the variable-length code word data is recorded or transmitted.
When recording such digitized and compressively coded data in a recording medium, a transparent recording method represented by MPEG2-TS is used as well as a base band recording method which has conventionally been used. According to the transparent recording method, data (e.g., variable-length code word data) transmitted as a bit stream is recorded as it is and, therefore, this method has the advantage of needing no encoder or decoder for a digital video signal, and keeping the image quality of the transmitted data as it is.
As an example of a recording medium used for recording and retaining a digital video signal, there is a digital video cassette tape. In “DVB (Digital Video Broadcasting) Signal Recording DVCR (Digital Video Cassette Recorder) Standard” by HD-DVCR conference held at December 1995, a recording format for a digital TV broadcast signal processed in accordance with MPEG is standardized, and this standard includes, as an option, the standard of recording of data for trick play (TP). Hereinafter, a description is given of recording of data for rapid play in this standard.
It is assumed that two kinds of data for rapid play having different search speeds are to be recorded. When recording data on a tape record medium, a signal is recorded on part called tracks. The data for rapid play is repeatedly recorded on plural tracks so that it can be searched at a high speed. Therefore, the maximum search speed is decided according to the number of repetitions of recording. Here, it is assumed that TPH data which can be read at a search speed 18 times as high as the normal speed at the maximum, and TPL data which can be read at a search speed 4 times as high as the normal speed at the maximum, are recorded.
FIG. 14
is a diagram for explaining recording of such data on a tape record medium, illustrating a track pattern of a DVCR in which MPEG2-TS data is allocated on a tape. As shown in
FIG. 14
, in this track format, a unit comprising four tracks (F
0
, F
1
, F
0
, F
2
) is repeated. When data for trick play (rapid play) is recorded in this format (since it is an option as mentioned above), the data should be allocated in a specific position in the tape.
Data to be recorded in the tape is based on a sync block of a fixed data quantity, and the data for rapid play is recorded for every five sync blocks in a specific position in a video data recording area shown in FIG.
14
. To be specific, TPH data is recorded in positions on the F
0
track, which positions are indicated by sync block numbers
40
-
44
,
62
-
66
,
84
-
88
,
106
-
110
,
128
-
132
, and
150
-
154
, that is, it is recorded in 30 sync blocks in total (5 sync blocks×6 positions). Further, TPL data is recorded in a position in the F
1
track, which position is indicated by sync block numbers
131
-
155
, that is, it is recorded in 25 sync blocks.
As described above, in this standard, recording of data for trick play is an option, so that sync block extra headers given to the respective sync blocks are used for identification of data recorded in the above-mentioned positions.
FIG. 15
shows the state of a sync block extra header given to data for rapid play. The sync block extra header is a 1-byte header included in each sync block, and one bit shown by b
7
in the figure is used to distinguish between data for normal-speed play and data for rapid play. In
FIG. 15
, the value of b
7
is 1, and this indicates that the sync block is data for trick play (rapid play).
Amongst the other bits composing the sync block extra header, one bit shown by b
6
is a bit used for deciding whether the sync block is new frame data or not, and this is a bit of a frame toggle used as a toggle. It is defined that frame toggling should not be performed more than two times within one track. Bits shown by b
2
-b
5
are used as a counter for distinguishing areas of data for trick play. Further, b
0
and b
1
are reserved.
As described above, in the DVCR, data for rapid play is recorded in a specific position in accordance with the optional standard, and the effect that the data recorded in that position is data for trick play (rapid play) is identified by the header.
As described above, when handling digital video data, the data is compressively coded by DCT, quantization, and variable-length coding, to be recorded or transmitted. Since variable-length code ward data obtained through these processes includes high-band (high-frequency) components of DCT coefficients, it is possible to reduce the data quantity for recording or transmission by removing the high-band components which have less influence on the quality of reproduced image. However, such reduction of data quantity has hardly been performed.
A thinkable method for reducing the data quantity is as follows. That is, variable-length code word data (digital data) is once analogized, and high-band components are removed from the analog signal and, thereafter, the analog signal is again digitized to obtain variable-length code word data. According to this method, high-band components are appropriately removed from the variable-length code word data to reduce the data quantity. In this method, however, since the reverse processes of digitization and compressive coding to generate the variable-length code word data are performed and then further digitization and compressive coding are performed, the circuit scale and the processing time are increased. Especially the circuit scale and the number of processes for the digital-to-analog conversion are considerable, it is not practical to reduce the data quantity according to this method.
Further, in the case where the digital data so obtained is recorded on a tape record medium or the like by the above-mentioned transparent recording metho
Kashiro Takao
Murata Hisaji
Yano Katsuhisa
Yasue Ken
Couso Jose L.
Matsushita Electric - Industrial Co., Ltd.
Patel Kanji
Wenderoth , Lind & Ponack, L.L.P.
LandOfFree
Video signal processing apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Video signal processing apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video signal processing apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2939233