No Title




Thus, the code length is elevated in accordance with the displacement or norm of motion vector. Information signifying the optimum more… prediction perform decided as described above is equipped, on send message the identical time, to the code compressor 7 via line 108b. The code compressor 7 performs compressing transformation so that data is expressed by a minimized variety of codes by use of knowledge signifying the optimum prediction perform, the prediction error sign which is output from the quantizer circuit three when that optimum prediction perform is used, and the synchronization sign which becomes needed on the time of decoding. The reworked sign is then sent out through a buffer memory 300 on a line 107 to a transmission line 109 or a storage medium.

However, as will be seen from the foregoing description, because the transmission pace decreases, in order that motion vector information occupies a percentage that cannot be uncared for. In an inter-frame coding system, the prediction error such as prediction error energy or brightness having a big magnitude doesn't occur frequently for immobile pictures and quasi-motionless pictures and less data will therefore be passable. On the other hand, the amount of information to be generated increases for footage subject to much motion or displacement. Accordingly, in inter-frame coding, coding efficiency is higher for pictures subject to much less motion than for footage subject to much motion. Means for coding a prediction error sign right into a variable word size code having the size which will increase because the amplitude of said error sign increases. In the coding control mode, the sub-sample and sub-line designation from line 111 is selectively supplied to the respective arithmetic items 8-1, 8-2, 8-3 and 8-4 via an arithmetic unit selector 8-15 and lines 111-1, 111-2, 111-3, and 111-4; and the quantizing characteristic designation from the road 112 is immediately fed to the respective arithmetic items. This invention relates to predictive coding system for television signals.

The vector code size might preferably be of variable code length. Usually, a motion vector corresponding to an inter-frame prediction occurs most frequently, in order that this motion vector is represented by the shortest code . However, when a variable code size method, for instance, modified Huffman coding method is applied to a collection of the shortest codes, the common code size per one movement vector turns into a lot shorter than 1.

The sum of those two code lengths is taken into account to be an amount of knowledge generated when the actual prediction perform is used. With regard to the other prediction capabilities, the sum of two code lengths is equally decided. A prediction perform which is obtained when the outcomes of comparison of the sum of the two code lengths reveals a minimal is set as an optimum prediction function, that's, a movement vector. It is feasible to make smaller the quantity of the transmission information when the motion vector is decided in this method than when the motion vector is set by using only the prediction error information as within the prior art. The share of the codes essential to symbolize this movement vector with respect to the entire transmission info varies greatly relying upon the transmission speed at the transmission path used.

5, a sign for scanning line S1 together with image parts a11 to a14 is saved in a line memory 1-1 and a sign for scanning line S2 including a21 to a24 is saved in a line reminiscence 1-2 sequentially. Subsequently, when scanning lines S3 and S4, a31 to a34 and a41 to a44 are saved in line reminiscences 1-3 and 1-4, respectively, line recollections 1-1 and 1-2 are learn out within the order of blocks, that's, a11, a12, a21, a22, a13, a14, a23, a24, and so forth. Since studying of scanning strains S1 and S2 in such converted form takes the identical time as that for studying the scanning lines without scan conversion, S1 and S2 may have been learn when the S4 line sign has just been stored within the line reminiscence 1-4. Then, the road recollections 1-3 and 1-4 are read out in the above-mentioned transformed form, while succeeding indicators for scanning traces S5 and S6 are stored in the line reminiscences 1-1 and 1-2 sequentially.

This worth was obtained when the sampling frequency was eight get redirected here MHz, in order that info of about zero.8M bits/sec. Was devoted to the motion vector information alone. Accordingly, the place a television sign is transmitted at a transmission velocity of 1.5M bits/sec., the motion vector information would occupy a little more than 50% of the entirety of knowledge transmitted. A predictive coding methods operates based on a principle that a distinction between a sign to be transmitted and a prediction sign is transmitted so as to decrease the amount of transmission info. This invention contemplates an adaptive predictive coding system wherein a plurality of prediction functions that give the prediction signals are used and an optimum one of many prediction capabilities is used for performing a prediction coding operation. Thus, the controller 301 receives a sign representative of the occupancy of the buffer memory 300 by way of a line a hundred and ten, and designates the quantizer 3, optimum prediction detector 8 and code compressor 7 by way of line 111, and designates the quantizer three, the optimum prediction detector eight, the code compressor 7 and the interpolator 302 through line 112.

One of four parallel picture component outputs is equipped through line 108-6a to a subtractor 8-12a, shown in FIG. The place of learn more an image component processed by the arithmetic unit 8-1 shall be called hereinafter picture component point D1 as proven in FIG. 3A. Similarly, different arithmetic models 8-2 via 8-4 are used for computing image component points D2 by way of D4, and 4 arithmetic items in whole are utilized in parallel for concurrent computations for one block. Line reminiscence pair 1-1 and 1-2 and one other line reminiscence pair 1-3 and 1-4 are saved and read out alternately, so that the video signal which is rendered scan conversion i.e. transformed into two-dimensional blocks, is output in a string of picture parts. One of two outputs from a line reminiscence pair in reading operation is selectively output through the multiplexor 1-5. The block-formation video sign on a line one hundred and one from the multiplexor 1-5 branches to the lines 101a and 101b, then is equipped to the delay circuit 2a and optimum prediction detector 8, respectively. The above-mentioned scan conversion is advantageous since the predictive coding circuit using the two-dimensional blocks is simple.