5 is a block diagram showing in detail the scan converter 1 of FIG. 4A. In this embodiment, 2×2 two-dimensional blocks every consisting of two image parts within the horizontal course and two strains within the vertical path, as shown in FIG. One block includes 4 picture elements, e.g. a11, a12, a21 and a22, enclosed by damaged line. In a basic tv system, the display is scanned from left to right in the order of, for example, a11, a12 , a13 and a14, and parts a21, a22, a23 and a24 are included on the succeeding scanning line.
In this case, picture parts in a two-dimensional block across a number of scanning traces are rearranged, and movie parts in every block are reworked so that picture parts in one block continue to these in the subsequent block and all image parts in a two-dimensional block are aligned serially. Accordingly, presence of scanning lines need not be of concern when finishing up prediction coding for image parts in a two-dimensional block, and a two-dimensional block could be treated identically to a one-dimensional block produced from M picture components on a single scanning line. As may be seen from the above, when scan conversion is carried out in advance, predictive coding by use of two-dimensional blocks could be not more than foxz168 predictive coding by use of one-dimensional blocks. According to the current invention, the above process is achieved for dedication of the optimum predictive perform by repetitive use of M arithmetic items for N instances, thus, hardware can be made small when the number of image parts M included in a block is small. Moreover, the prediction error energy for a block for each perform may be obtained by single computation, allowing the selection of an prediction function to be compared next while carrying out the computation. It implies that all of N-times computations are not at all times necessary for determining the optimum prediction perform.
The transmission path may include conductive wiring traces or a spatial transmission path. In another manner of utilization, a coded video signal is stored once in a storage or a recording medium and thereafter read out right into a decoder as shown in dotted line in FIG. The current system is particularly directed to the previous scheme.
The delay time depends on the variety of image elements included inside a block. The video signal is supplied from the delay circuit 2a to a subtractor 2b. Here the difference between the image sign, and the prediction signal supplied through a line 105a branching from an output line 105 of a prediction signal generator 5, is produced, and the distinction, i.e. the prediction error signal, is then provided to a quantizer three which capabilities to restrict the variety of potential ranges.
This system is the best one for assuring a high coding effectivity even for a moving image. With this motion compensation inter-frame prediction system, for the purpose of sustaining a high effectivity for various actions, many prediction capabilities are used. More particularly, a plurality of prediction functions comparable to movements at varied speeds and in varied instructions are prepared beforehand, and a prediction function most adaptable for an really occurring motion is used as an optimum function.
4A is a block diagram showing an instance of a predictive coding system, which is situated on the facet of a transmitter and operates to transmit a sign to a predictive decoding system on the facet of a receiver. An enter video signal is equipped via line 100 to a scan converter 1, in which a plurality of blocks, each consisting of a quantity of picture parts, are fashioned. The block-formation video signal is then equipped to a delay circuit 2a and an optimum prediction detector eight via strains 101a and 101b, respectively, branched from line 101. The delay circuit 2a delays the propagation of the block-formation video signal for a size of time required for determining the optimum prediction function.
If any computation in two-dimensional block units have been meant with out use of the scan conversion, the foxz168 สมัคร video sign can be scanned in the order of a11, a12, a13, a14, and so on, in FIG. 6, and computation for a two-dimensional block together with a11, a12, a21 and a22 must await image components a21, a22, a23 and a24 on the subsequent scanning line, resulting in many blocks whose computations are left suspended till the subsequent line is scanned. More notably, in a conference TV, the frequency of prevalence of movement vector is the utmost for a motion vector indicative of a standstill picture, that is, corresponding to a inter-frame prediction, and tends to lower as the norm of motion vector will increase. For this reason, it's possible to efficiently convert an input vector right into a variable size code in accordance with the sample of variable code length as proven in FIG.