Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2000-10-02
2003-06-10
Philippe, Gims S. (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240280, C348S387100, C348S423100
Reexamination Certificate
active
06577682
ABSTRACT:
This invention concerns the handling of compressed or bit-rate reduced signals.
In the most important example, this invention relates to the MPEG2 video compression system and the following description will take MPEG2 as an illustrative compression (or bit-rate reduction) system.
As compressed video increasingly find its way into the programme chain, it becomes more likely that bitstreams will be decoded and re-encoded. Examples of where this happens include switching between different bitstreams and changing between bit-rates and compression formats. It is important that unnecessary loss of quality is avoided in these cases.
There is proposed in EP-A-O 765 576 a solution in which an information bus accompanies a decoded signal and is available in a subsequent encoding process. In apparatus and channels designed for this purpose, it is possible for the decoded, picture signal to be accompanied by an information bus which contains all the side information which is contained in the MPEG signal, alongside the DCT coefficients. The information bus may also contain additional information which is related to the control of the coding or decoding process but which does not form part of the coded bitstream. It can include, for example, statistical information, information concerning errors together with instructions for use in error concealment, candidate motion vectors or recommended quantizer step sizes.
There are disclosed in WO-A-9803017, techniques by which an information bus can pass along a video pathway. One method uses an information bus embedded in the lowest bits of the decoded video signal (in ITU-R Rec. 656 format), such that the picture quality is not visibly affected. In one version of this, the video bits used to convey a particular coding decision are spatially and temporally locked to the area of the picture to which it corresponds. Alternatively, the information may take a different form, for instance using non-active parts of the picture, such as the vertical blanking interval (VBI), or by using a separate channel.
A problem arises where it is not possible to embed the information bus in the video signal without suffering excessive data loss and where the equipment concerned has not been designed with an appropriate channel to carry the information bus separately. An example of this is when the video signal is recorded by a digital VTR that employs DCT-based compression, which tends to change the lowest bits of the signal.
Accordingly, the present invention consists in one aspect in a video compression system in which an encoder operates on a video signal and takes coding decisions in a compression encoding process to generate an encoded signal and in which a decoder operates on the encoded signal in a compression decoding process to generate an decoded signal, wherein a representation of the coding decisions or some of them is provided in bit-rate reduced form and accompanies the decoded signal through a bit-rate reduced channel.
The bit-rate reduced channel may take a variety of forms including one or more audio channels (such as VT audio tracks or a separate audio recorder); unused VBI lines, a video channel with significant error correction and appropriate data channels (including DVCpro data channel).
For the case of an MPEG-2 bitstream, the coding decisions typically include:
Picture dimensions
Frame rate
Picture structure (frame-coded or field-coded)
Picture type (I, P or B)
Whether macroblocks intra-coded or use prediction
Whether forward, backward or bi-directional prediction is used
Whether frame-based or field-based prediction is used
Motion vectors
Quantiser visibility weighting matrices
Quantiser step
Buffer state of a decoder that is decoding the bitstream
Information on bit-rate usage within the bitstream
When it is possible to keep all (or a sufficient sub-set of) the decisions the same on recoding, it can be shown that the switch is transparent. In other cases, such as when a new bitrate is used on recoding, it is still advantageous to reuse many decisions (e.g. picture type and motion vectors), and take account of others (e.g. quantiser scale).
In one form of the invention, the bit-rate reduced coding decisions pass through a convenient separate channel such as one or more audio channels. An alternative is to use the bit-rate reduction of the coding decisions to enable them to be embedded in the video signal in a more robust manner, such that they can be recovered without errors.
Where the chosen channel has insufficient capacity to support sending the coding decisions with fixed length codewords, variable length coding can be used. However, use of variable length coding means that the short-term bit-rate can exceed that which fixed-length coding would give, and some method of reducing this is required.
This invention, in one aspect, provides for decisions to be modified to reduce their bit-rate whilst signaling to the recoder that the decisions have been changed and so no longer correspond to the original bitstream.
Another aspect of this invention addresses the problem that where an information bus is embedded in a video signal, it is not possible to detect where parts of the picture have been changed by caption insertion, cross-fades, and the like.
Another embodiment of this invention can be used for converting between an information bus embedded in the video signal and a side-channel based information bus. This would be typically used for integrating a DCT-based VTR into an ITU-R Rec. 656 infrastructure.
On embodiment of this invention consists in a system for producing a bit-rate reduced representation of coding decisions suitable for passing through a channel (comprised of e.g. one or more audio channels) in which the available bit-rate may be insufficient to carry the coding decisions themselves. Preferably, variable length coding of the information is employed.
The following are examples of this aspect of the invention:
A. An MPEG-compliant video bitstream, with “dummy” coefficients and coded block patterns inserted to keep the bitstream legal. These would not represent the coefficients and coded block patterns used for the original bitstream, but would give the shortest legal codewords.
B. A bitstream that uses a modified version of the MPEG format in which the macroblock syntactical element is altered so that it no longer includes coefficients or a coded block pattern. This would mean a small change in the way in which the bitstream is decoded.
C. As form B. but in addition, some of the coding decisions, in particular the motion vectors, are coded in a non-compliant way in order to reduce their bit-rate. Depending on how this is done (methods for this are described below), there may be a small or a large change to the way in which the bitstream is decoded.
REFERENCES:
patent: 6088393 (2000-07-01), Knee et al.
patent: 6285716 (2001-09-01), Knee et al.
Brightwell Peter John
Tudor Phillip Nicholas
Wells Nicholas Dominic
British Broadcasting Corporation
Michael & Best & Friedrich LLP
Philippe Gims S.
LandOfFree
Video processing system also compressing coding decision 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 Video processing system also compressing coding decision data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video processing system also compressing coding decision data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3131768