Television – Bandwidth reduction system – Data rate reduction
Patent
1997-06-05
1999-11-23
Lee, Richard
Television
Bandwidth reduction system
Data rate reduction
348420, H04N 732
Patent
active
059909617
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
The present invention relates to a method for encoding digital picture video signals in order to reduce the amount of data possibly to be transmitted. More particularly, the invention concerns how to encode more than one picture at the same time, or more precisely, how to encode a number of blocks from different pictures simultaneously.
Several standard methods exist for compressing video signals. In a number of these methods (ITU recommendation H.261 and ISO/IEC 11172) the following techniques are substantially utilized in order to achieve compression of the data to be transmitted:
1. The pictures, which consist of a number of pixels, are divided into a number of blocks consisting of, for example, 8.times.8 or 16.times.16 pixels. The size utilized depends on what is to be performed.
2. Since pictures following each other in direct temporal succession often are relatively similar, a picture block can often be approximately described by using parts of previously encoded pictures. If one presupposes that a block of 8.times.8 pixels is to be encoded, wherein this block is referred to as ORIG(i,j) wherein i,j=1 . . . 8, one will search for a suitable block of 8.times.8 pixels in a previously encoded picture. This block in the previously encoded picture will typically have a shifted position in the picture as compared to ORIG, which fact reflects movements in the picture content. The magnitude of the shift is indicated by means of a so-called vector stating the number of pixels along which the block has been shifted in a horizontal and a vertical direction. The block which has been found to be rather similar to the ORIG block is referred to as a prediction for the ORIG block, and it is defined as PRED(i,j), wherein i,j=1 . . . 8. Algorithms exist for finding a best possible PRED block in a previously encoded picture, however such algorithms are not herein discussed in detail, since detailed formulations of such algorithms do not constitute any part of the present invention.
3. Since a receiver itself is capable of generating PRED (the receiver has stored the necessary previously reconstructed encoded/decoded pictures, and the receiver itself is also able to compute PRED in accordance with the same algorithm as the transmitter), there is no need to transmit the full ORIG. It is sufficient to transmit a so-called difference signal defined as follows:
4. In order to transmit DIFF in a form as highly compressed as possible, one usually converts the DIFF block (which in the present example is 8.times.8 pixels) into a matrix which contains 8.times.8 transform coefficients which thus represent DIFF. To a large degree, the so-called two-dimensional Cosine transform is utilized which has favorable characteristics. The transform coefficients obtained by transforming DIFF are referred to as TRANS, and are defined as follows:
5. Subsequent to transforming DIFF into TRANS, the transform coefficients are quantized, and the transformed and quantized coefficients, which through quantizing now constitute an approximate representation of DIFF, can then be transmitted to a receiver in accordance with a certain matrix reading strategy (for example see the reading strategy and formatting into transmittable data described in Norwegian patent no. 175,080 belonging to the same applicant as in the present invention). In the receiver, inverse quantizing and inverse transforming are performed in order to reconstruct DIFF. However, in actuality the reconstructed values DIFF' deviate somewhat from DIFF. The reason for the deviation is the quantization. It is to be noted that the two last mentioned operations must be made in an equal manner at both the transmitting and the receiving sides. The obtained DIFF' values are utilized for reconstructing a representation ORIG' of the picture block which was to be encoded:
When picture compression is implemented, it is necessary both to compress in order to represent the pictures by as few bits as possible, and de-compress to be able to display the pictures again. An exp
REFERENCES:
patent: 5173773 (1992-12-01), Ueda et al.
patent: 5198901 (1993-03-01), Lynch
patent: 5212549 (1993-05-01), Ng et al.
patent: 5293229 (1994-03-01), Iu
patent: 5453801 (1995-09-01), Kim
patent: 5510840 (1996-04-01), Yonemitsu et al.
Signal Processing: Image Communication, vol. 2, No. 2, Aug. 1990 (The Netherlands), A. Puri et al., "Video Coding With Motion-Compensated Interpoliation for CD-ROM Applications", pp. 127 --pp. 144, see especially pp. 131-132.
Signal Processing: Image Communication, vol. 2, No. 2, Aug. 1990 (The Netherlands), Atsushi Nagata et al., "Moving Picture Coding System for Digital Storage Media Using Hybrid Coding", pp.109 --pp. 116, see especially pp.112-113.
Lee Richard
Telenor AS
LandOfFree
Method for compression encoding of digital picture video signals 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 for compression encoding of digital picture video signals, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for compression encoding of digital picture video signals will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1227646