Moving-picture compression encoding method, apparatus and program

An apparatus for performing compression encoding of a moving-picture, comprising:

• means for deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator, in a case where a subblock smaller than a block size exists in the motion-compensated predicting unit;
• means for generating prediction signals by applying filter processing to the input signals;
• means for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to the moving-picture compression encoding apparatus;
• means for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• means for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• means for selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values;
• means for deciding input signals of the motion-compensated predicting unit based upon the motion vector that has been generated by the motion estimator, in a case where a predetermined block has been selected by the block selection;
• means for generating prediction signals by applying filter processing to the input signals;
• means for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and the moving-picture signal that is input to the moving-picture compression encoding apparatus;
• means for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• means for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• means for selecting one block size from the preliminarily selected block sizes based upon the second evaluation values.

An apparatus for performing compression encoding of a moving-picture, comprising:

• means for deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator, in a case where a subblock smaller than a block size exists in the motion-compensated predicting unit;
• means for generating prediction signals by applying filter processing to the input signals;
• means for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to the moving-picture compression encoding apparatus;
• means for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• means for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• means for selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values;
• means for deciding input signals of a motion-compensated predicting unit based upon the motion vector that has been generated by the motion estimator, in a case where a predetermined block has been selected by said means for selecting the block size;
• means for generating prediction signals by applying filter processing to the input signals;
• means for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and the moving-picture signal that is input to the moving-picture compression encoding apparatus;
• means for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• means for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• means for selecting one block size from the preliminarily selected block sizes based upon the second evaluation values;
• wherein number of block sizes preliminarily selected in case of a block differs from that in case of a subblock.

An apparatus for performing compression encoding of a moving-picture, comprising a motion estimator, a motion-compensated predicting unit and a frame memory, said motion-compensated predicting unit comprising:

• a prediction signal generator that includes:
• a plurality of input signal call units, each reading in a reconstructed signal of a block from said frame memory based upon a motion vector that has been generated by said motion estimator, and delivering an input signal to an associated one of a plurality of filters;
• said plurality of filters, which are for generating prediction signals by applying filter processing to the input signals from corresponding ones of said input signal call units;
• and
• a plurality of prediction signal memories for storing and outputting respective ones of the plurality of prediction signals that have been generated by respective ones of said plurality of filters;
• a block size decision unit that includes:
• a first cost calculator for calculating difference information indicative of differences between the prediction values and an input moving-picture signal and outputting first evaluation values;
• a preliminary selector for selecting at least two block sizes from block sizes of small first evaluation values and outputting the block sizes selected;
• a second cost calculator for calculating and outputting second evaluation values obtained by applying a frequency conversion to difference information corresponding to the block sizes that have been output by said preliminary selector;
• and
• a block size selector for selecting a block size based upon the second evaluation values that have been output by said second cost calculator and outputting the selected block size and the evaluation value thereof as results of motion-compensated prediction;
• and
• a prediction signal call circuit for reading in a prediction signal, which corresponds to the block size output from said block size decision unit, from said prediction signal memory and outputting the prediction signal together with second evaluation value and block size as results of motion-compensated prediction.

An apparatus for performing compression encoding of a moving-picture, comprising a motion estimator, a motion-compensated predicting unit and a frame memory, said motion-compensated predicting unit comprising:

• a prediction signal generator that includes:
• a plurality of input signal call units, each reading in a reconstructed signal of a block from said frame memory based upon a motion vector that has been generated by said motion estimator, and delivering an input signal to an associated one of a plurality of filters;
• said plurality of filters, which are for generating prediction signals by applying filter processing to the input signals from corresponding ones of said input signal call units;
• and
• a plurality of prediction signal memories for storing and outputting respective ones of the plurality of prediction signals that have been generated by respective ones of said plurality of filters;
• a block size decision unit that includes:
• a first cost calculator for calculating first evaluation values from difference information indicative of differences between an input moving-picture signal and the prediction values output from said prediction signal generator, and outputting the first evaluation values;
• a block size selector for selecting a block size based upon the first evaluation values and outputting the block size and the difference information thereof;
• and
• a second cost calculator for applying a frequency conversion to the difference information of this block size, calculating a second evaluation value and outputting the block size and the second evaluation value as results of motion-compensated prediction;
• and
• a prediction signal call circuit for reading in a prediction signal, which corresponds to the block size output from said block size decision unit, from said prediction signal memory and outputting the prediction signal together with the second evaluation value and block size as results of motion-compensated prediction.

An apparatus for performing compression encoding of a moving-picture, comprising a motion estimator, a motion-compensated predicting unit and a frame memory, said motion-compensated predicting unit comprising:

• a prediction signal generator that includes:
• a plurality of input signal call units, each reading in a reconstructed signal of a block from said frame memory based upon a motion vector that has been generated by said motion estimator, and delivering an input signal to an associated one of a plurality of filters;
• said plurality of filters, which are for generating prediction signals by applying filter processing to the input signals from corresponding ones of said input signal call units;
• and
• a plurality of prediction signal memories for storing and outputting respective ones of the plurality of prediction signals that have been generated by respective ones of said plurality of filters;
• a block size decision unit that includes:
• a difference calculator for calculating and outputting difference information between an input moving-picture signal and prediction signals from said prediction signal generator;
• a first cost calculator for applying a frequency conversion to difference information of block sizes, calculating evaluation values and outputting the evaluation values;
• and
• a block size selector for selecting a block size based upon the evaluation values that have been output from said first cost calculator and outputting the selected block size and the evaluation value thereof as results of motion-compensated prediction;
• and
• a prediction signal call circuit for reading in a prediction signal, which corresponds to the block size output from said block size decision unit, from said prediction signal memory and outputting the prediction signal together with the second evaluation value and block size as results of motion-compensated prediction.

A method for performing compression encoding of a moving-picture, said method comprising the steps of:

• deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator;
• generating prediction signals by applying filter processing to the input signals;
• calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to a moving-picture compression encoding apparatus;
• preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values.

A method for performing compression encoding of a moving-picture, said method comprising the steps of:

• deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator, in a case where a subblock smaller than a block size exist in the motion-compensated predicting unit;
• generating prediction signals by applying filter processing to the input signals;
• calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to a moving-picture compression encoding apparatus;
• preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values;
• deciding input signals of a motion-compensated predicting unit based upon the motion vector that has been generated by the motion estimator, in a case where a predetermined block has been selected by the block selection;
• generating prediction signals by applying filter processing to the input signals;
• calculating first evaluation values based upon difference information indicative of differences between the prediction signals and the moving-picture signal that is input to the moving-picture compression encoding apparatus;
• preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• selecting one block size from the preliminarily selected block sizes based upon the second evaluation values.

A method for performing compression encoding of a moving-picture, said method comprising the steps of:

• deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator, in a case where a subblock smaller than a block size exists in the motion-compensated predicting unit;
• generating prediction signals by applying filter processing to the input signals;
• calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to a moving-picture compression encoding apparatus;
• preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values;
• deciding input signals of a motion-compensated predicting unit based upon the motion vector that has been generated by the motion estimator, in a case where a predetermined block has been selected by said step for selecting the block size;
• generating prediction signals by applying filter processing to the input signals;
• calculating first evaluation values based upon difference information indicative of differences between the prediction signals and the moving-picture signal that is input to the moving-picture compression encoding apparatus;
• preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• selecting one block size from the preliminarily selected block sizes based upon the second evaluation values;
• wherein number of block sizes preliminarily selected in case of a block differs from that in case of a subblock.

A computer program for causing a computer constituting a moving-picture compression encoding apparatus to execute the following processing:

• processing for deciding input signals of a motion-compensated prediction unit based upon a motion vector generated by a motion estimator;
• processing for generating prediction signals by applying filter processing to the input signals;
• processing for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to the moving-picture compression encoding apparatus;
• processing for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• processing for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• processing for selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values.

A computer program for causing a computer constituting a moving-picture compression encoding apparatus to execute the following processing:

• processing for deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator, in a case where a subblock smaller than a block size exists in a motion-compensated predicting unit;
• processing for generating prediction signals by applying filter processing to the input signals;
• processing for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to the moving-picture compression encoding apparatus;
• processing for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• processing for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• processing for selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values;
• processing for deciding input signals of a motion-compensated predicting unit based upon the motion vector that has been generated by the motion estimator, in a case where a predetermined block has been selected by the block selection;
• processing for generating prediction signals by applying filter processing to the input signals;
• processing for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and the moving-picture signal that is input to the moving-picture compression encoding apparatus;
• processing for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• processing for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• processing for selecting one block size from the preliminarily selected block sizes based upon the second evaluation values.

A computer program for causing a computer constituting a moving-picture compression encoding apparatus to execute the following processing:

• processing for deciding input signals of a motion-compensated predicting unit based upon a motion vector generated by a motion estimator, in a case where a subblock smaller than a block size exists in a motion-compensated predicting unit;
• processing for generating prediction signals by applying filter processing to the input signals;
• processing for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and a moving-picture signal that is input to the moving-picture compression encoding apparatus;
• processing for preliminarily selecting a plurality of block sizes based upon the first evaluation values, wherein number of block sizes preliminarily selected in case of a block differs from that in case of a subblock;
• processing for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• processing for selecting one block size from the plurality of preliminarily selected block sizes based upon the second evaluation values;
• processing for deciding input signals of a motion-compensated predicting unit based upon the motion vector that has been generated by the motion estimator, in a case where a predetermined block has been selected by said processing of selecting the block size;
• processing for generating prediction signals by applying filter processing to the input signals;
• processing for calculating first evaluation values based upon difference information indicative of differences between the prediction signals and the moving-picture signal that is input to the moving-picture compression encoding apparatus;
• processing for preliminarily selecting a plurality of block sizes based upon the first evaluation values;
• processing for calculating second evaluation values based upon difference information of the block sizes preliminarily selected;
• and
• processing for selecting one block size from the preliminarily selected block sizes based upon the second evaluation values.