Motion video signal processing for recording or reproducing – Local trick play processing – With randomly accessible medium
Utility Patent
1999-03-01
2001-01-02
Vincent, David R. (Department: 2732)
Motion video signal processing for recording or reproducing
Local trick play processing
With randomly accessible medium
C386S349000
Utility Patent
active
06169846
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method and apparatus for recording compressed digital video data, and in particular to a method and apparatus whereby variable-speed playback of the recorded data can be performed and which are applicable to recording of digital video data that have been compressed using an coding algorithm which utilizes motion compensation predictive coding to encode at least some of a stream of pictures, such as the MPEG algorithm.
2. Description of the Prior Art
The MPEG (Moving Pictures Experts Group) coding algorithm is now widely used for digital video signal compression. In the following, the MPEG-1 format will be assumed, although as described hereinafter the invention is equally applicable to the MPEG-2 format. The MPEG algorithm is basically a hybrid coding method, which employs motion compensation predictive coding and DCT (Discrete Cosine Transform) processing, together with quantization and variable-length coding, to attain a high level of data compression. Predictive coding is effective in achieving compression of a stream of data which expresses a moving picture as successive video pictures expressed by respective digital video signal frames, and utilizes the generally strong degree of correlation between corresponding regions of pictures which are closely time-adjacent. However in order to transmit a picture by predictive coding, it is always necessary to utilize the data of a specific preceding (or succeeding) picture as a reference, or to use both a specific preceding and succeeding picture. With the MPEG system, pictures are periodically inserted into the compressed video data stream as coded data which are derived independently from all other pictures, i.e. using intra-coding. When pictures are received which have been data-compressed by predictive coding it is not possible to randomly access these pictures, i.e. random access is only possible for the intra-coded data. In the case of video recording, the capability for random access may be important, in particular with regard to providing a variable-speed playback capability (i.e. for providing a “fast-forward” playback mode) for a digital video tape recorder. That is due to the fact that in a video tape recorder (VTR) which utilizes a a helical scan rotary recording head to scan successive diagonal tracks of a magnetic tape during a normal recording or playback mode, some tracks will be periodically skipped over by the recording head when the VTR is operated in a variable-speed playback mode. With recording of compressed video data, however, problems arise with respect to variable-speed playback. This is basically because:
(a) The data of one independently-coded picture may be recorded on a plurality of successive recording tracks; and
(b) The independently-coded pictures may not occur at regular fixed periods within the stream of compressed video data.
The problems will be described in more detail, for the case of MPEG compressed video data. With the MPEG standards, the video data are arranged in a high-level data layer of units which are called GOP (Group of Pictures) units, each consisting of a plurality of pictures (expressed as respective compressed digital video signal frames) which are of three different types. Firstly, there are the aforementioned pictures which have been encoded independently, by intra-coding, and which will be referred to as I-pictures. Secondly are pictures referred to in the following as P-pictures, which have been encoded by predictive coding, using one other I-picture or P-picture as a reference. Thirdly, there are B-pictures, which have been coded by bi-directional predictive coding using a preceding I-picture or P-picture and a preceding P-picture or I-picture.
In the case of an I-picture, only the contents of an original picture are utilized. Data compression is executed by applying orthogonal transform processing, specifically, DCT (Discrete Cosine Transform) processing, to each of respective (8×8) picture element blocks of the original picture. These blocks convey either luminance or color information, and are arranged in sets of six blocks to form respective macroblocks of a digital color signal frame, as is well known, so that detailed description will be omitted. The DCT processing produces a set of DCT coefficients for each of these blocks, including a DC coefficient and a plurality of AC coefficients which are of successively higher order, extending from a lowest-order AC coefficient, within a 2-dimensional transform space. The DCT coefficient values thus obtained are then quantized, and to achieve further data compression, variable-length coding is applied to the quantized DCT AC coefficients, i.e. run-length coding which is applied (after applying quantization) to successive positions within the transform space along a zig-zag path.
With the GOP data structure, the I-pictures can theoretically be utilized for random access or variable-speed playback operation of a recording apparatus, if the GOP units occur at regular intervals within the compressed video data stream (although in general that is not the case).
FIG. 1
illustrates an example of the GOP structure. In this example (assuming that GOP the units occur at fixed intervals) the I-pictures occur in the data flow at a rate of one in every 15 pictures, the P-pictures at a rate of one in every 3 pictures, and the B-pictures at a rate of 2 in every 3 pictures. The respective average proportions of data conveyed by the I, P and B pictures respectively, however, are 1:0.4:0.15.
Since as described above the P and B pictures alone cannot be used to recover the original video data, e.g. from a playback signal obtained from a video data recording apparatus, if it is required that the recording apparatus provide a variable-speed playback capability, it becomes necessary to ensure that the recorded video data includes a high proportion of intra-coded data, i.e. a high proportion of I-pictures, and to ensure that the recorded intra-coded data can be efficiently captured during variable-speed playback operation. However, as noted above, even if the I-pictures constitute only a small proportion of the total number of pictures of compressed video data, they account for a very high proportion of the compressed video data. This is due to the fact that the I-pictures do not utilize time axis correlation for data compression, and so contain much larger amounts of data than the P or B-pictures. Hence, if the proportion of I-pictures were to be increased so as to obtain more effective variable-speed playback operation, then the data compression efficiency would be lowered substantially. Thus, although it would be theoretically possible to achieve an improved variable-speed playback capability for the compressed video data if the proportion of intra-coded data within that data were to be increased, in practice that proportion must be held near the minimum amount that is necessary to ensure satisfactory picture quality in the finally obtained display picture during normal-speed playback.
In particular, in the case of a video signal which has been compressed by MPEG data compression, the compressed data flow must further include synchronizing signals, ID (identification) information, error correction information etc., which have all been subjected to a high degree of data compression before being inserted into the compressed video data stream. It is thus difficult to increase the proportion of I-pictures in the compressed video data flow, without reducing the data compression efficiency.
As a result, it has been found in the prior art that the display picture which is obtained by variable-speed playback of such recorded compressed video data does not have a natural appearance. That is due to the fact that the frequency of updating the display picture is excessively low. The reasons for that will be described in more detail referring to
FIGS. 1 and 2
. In
FIG. 2
, a GOP set of 15 pictures (shown in
FIG. 1
) are recorded by a VTR (video tape recorder)
Lowe Hauptman Gopstein Gilman & Berner LLP
Victor Company of Japan Ltd.
Vincent David R.
LandOfFree
Compressed video data recording method and recording... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Compressed video data recording method and recording..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Compressed video data recording method and recording... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2471601