Video encoder and video encoding method as well as video decoder and video decoding method

A video encoding method for encoding a coding-target block among a plurality of blocks obtained by dividing a coding-target picture included in video data, in a bidirectional prediction mode, the video encoding method comprising:

• calculating, for the coding-target block, a first motion vector representing an amount of spatial movement between a reference region in a first reference picture positioned in a first direction from the coding-target picture and the coding-target block, and calculating, for the coding-target block, a second motion vector representing an amount of spatial movement between a reference region in a second reference picture positioned in a second direction from the coding-target picture and the coding-target block;
• selecting, from among motion vectors calculated for respective blocks encoded before the coding-target block, a plurality of candidates for a first predictive vector for predicting the first motion vector and a plurality of candidates for a second predictive vector for predicting the second motion vector;
• calculating first difference vectors representing errors between the first motion vector and respective candidates for the first predictive vector, and calculating second difference vectors representing errors between the second motion vector and respective candidates for the second predictive vector;
• setting first selection information for identifying, as the first predictive vector, a candidate corresponding to a first difference vector whose amount of information is the smallest in the first difference vectors, from among the plurality of candidates for the first predictive vector, and setting second selection information for identifying, as the second predictive vector, a candidate corresponding to a second difference vector whose amount of information is the smallest in the second difference vectors, from among the plurality of candidates for the second predictive vector;
• performing variable length coding on the first selection information and the second selection information;
• and
• applying, on the basis of the first selection information, predetermined code to the second selection information representing that the second predictive vector is a candidate determined on the basis of a motion vector in the second direction for the encoded block corresponding to the first predictive vector, from among the plurality of candidates for the second predictive vector.

A video decoder that decodes a decoding-target bidirectional-predictive-coded block among a plurality of blocks obtained by dividing a decoding-target picture included in encoded video data, the video decoder comprising:

• a variable length decoding unit that regenerates first selection information of a first predictive vector for a first motion vector, and a first difference vector between the first motion vector and the first predictive vector, first motion vector representing an amount of spatial movement between the decoding-target block and a reference region in a first reference picture positioned in a first direction from the decoding-target picture, and that regenerates second selection information of a second predictive vector for a second motion vector, and a second difference vector between the second motion vector and the second predictive vector, the second motion vector representing an amount of spatial movement between the decoding-target block and a reference region in a second reference picture positioned in a second direction from the decoding-target picture;
• a predictive vector selecting unit that
• selects a plurality of candidates for the first predictive vector from among motion vectors of respective blocks decoded before the decoding-target block,
• identifies the first predictive vector from among the plurality of candidates for the first predictive vector, on the basis of the first selection information,
• selects a plurality of candidates for the second predictive vector from among the blocks decoded before the decoding-target block, and
• identifies the second predictive vector from among the plurality of candidates for the second predictive vector by judging, on the basis of the first selection information, whether or not the second selection information is a predetermined code indicating that a candidate determined on the basis of a motion vector in the second direction, for the decoded block corresponding to the first predictive vector from among the plurality of candidates for the second predictive vector, is the second predictive vector;
• and
• a motion vector decoding unit that decodes the first motion vector by adding the first difference vector to the first predictive vector, and the second motion vector by adding the second difference vector to the second predictive vector.

A video decoding method for decoding a decoding-target bidirectional-predictive-coded block among a plurality of blocks obtained by dividing a decoding-target picture included in encoded video data, the video decoding method comprising:

• regenerating first selection information of a first predictive vector for a first motion vector, and a first difference vector between the first motion vector and the first predictive vector, the first motion vector representing an amount of spatial movement between the decoding-target block and a reference region in a first reference picture positioned in a first direction from the decoding-target picture, and regenerating second selection information of a second predictive vector for a second motion vector, and a second difference vector between the second motion vector and the second predictive vector, the second motion vector representing an amount of spatial movement between the decoding-target block and a reference region in a second reference picture positioned in a second direction from the decoding-target picture;
• selecting a plurality of candidates for the first predictive vector from among motion vectors of respective blocks decoded before the decoding-target block;
• identifying the first predictive vector from among the plurality of candidates for the first predictive vector, on the basis of the first selection information;
• selecting a plurality of candidates for the second predictive vector from among the blocks decoded before the decoding-target block;
• identifying the second predictive vector from among the plurality of candidates for the second predictive vector by judging, on the basis of the first selection information, whether or not the second selection information is a predetermined code indicating that a candidate determined on the basis of a motion vector in the second direction, for the decoded block corresponding to the first predictive vector from among the plurality of candidates for the second predictive vector is the second predictive vector;
• and
• decoding the first motion vector by adding the first difference vector to the first predictive vector, and the second motion vector by adding the second difference vector to the second predictive vector.