Simplified dual prime video motion estimation

A method for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the method comprising the steps of:

• generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, wherein the current picture is a field of the video sequence, the first reference picture is a previous opposite-parity field and the second reference picture is a previous same-parity field;
• and
• generating from the first and second match vectors a base motion vector and a delta motion vector such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement or the current block from the respective first and second reference blocks,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of:
• setting the base motion vector equal to the second match vector generated from the previous same-parity field;
• and
• selecting the delta motion vector such that an opposite parity motion vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector and that vertical and horizontal components of the delta motion vector arc selected from the set {-1, 0, 1}.

A method for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the method comprising the steps of:

• generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, wherein the current picture is a field of the video sequence, the first reference picture is a previous opposite-parity field and the second reference picture is a previous same-parity field;
• and
• generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of:
• setting an opposite parity motion vector to be generated by a decoder equal to the first match vector generated from the previous opposite-parity field;
• and
• selecting the base motion vector and the delta motion vector such that the base motion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector.

A method for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the method comprising the steps of;

• generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, wherein the current picture is a field of the video sequence, the first reference picture is a previous opposite-parity field and the second reference is a previous same-parity field;
• and
• generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from respective first and second reference blocks,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of:
• setting the base motion vector equal to the second match vector generated from the previous same-parity field;
• setting a horizontal component of the delta motion vector equal to positive one if a horizontal component of the first match vector is greater than the closest integer to a horizontal component of the base motion vector divided by two that is no larger than the horizontal component of the base motion vector divided by two;
• setting the horizontal component of the delta motion vector equal to negative one if the horizontal component of the first match vector is less than the closest integer to the horizontal component of the base motion vector divided by two that is not larger than the horizontal component of the base motion vector divided by two;
• and
• setting the horizontal component of the delta motion vector equal to zero if the horizontal component of the first match vector is equal to the closest integer to the horizontal component of the base motion vector divided by two that is not larger than the horizontal component of the base motion vector divided by two.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• a first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vapors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures:
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generating from the first and second match vectors a base motion vector and a delta motion vector such that the delta motion vector is determined as a function of at least one of the first and second match vectors the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• and
• a mode decider having inputs connected to receive the first and second match vectors from the first motion vector generator and the base and delta motion vectors from the second motion vector generator, and outputs corresponding to selected vectors corresponding to a particular estimation mode.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• a first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• and
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generation from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks, wherein the second motion vector generator generates the base motion vector and delta motion vector by setting the base motion vector equal to the second match vector generated from the previous same-parity field, and selecting the delta motion vector such that an opposite parity motion vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• a first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• and
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks, wherein the second motion vector generator generates the base motion vector and delta motion vector by setting the base motion vector equal to the second match vector generated from the previous same-parity field, and selecting the delta motion vector such that an opposite parity motion vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector and that vertical and horizontal components of the delta motion vector are selected from the set {-1, 0, 1}.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• a first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• and
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks, wherein the second motion vector generator generates the base motion vector and delta motion vector by setting an opposite parity motion vector to be generated by a decoder equal to the first match vector generated from the previous opposite-parity field, and selecting the base motion vector and the delta motion vector such that the base motion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• a first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• and
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks, wherein the second motion vector generator generates the base motion vector and delta motion vector by setting an opposite parity motion vector to be generated by a decoder equal to the first match vector generated from the previous opposite-parity field, and selecting the base motion vector and the delta motion vector such that the base notion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector and that vertical and horizontal components of the delta motion vector are selected from the set {-1, 0, 1}.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• a first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• and
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks wherein the second motion vector generator is further operative to determine whether to set the base motion vector equal to the second match vector generated from the previous same-parity field or to set an opposite parity motion vector equal to the first match vector, based on the result of an error measure between the current block and at least one of the first and second reference blocks.

An apparatus for performing motion estimation on a video picture of a video sequence, the picture including a plurality of blocks of pixel data, the apparatus comprising:

• first motion vector generator for generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• and
• a second motion vector generator for receiving the first and second match vectors from the first motion vector generator, the second motion vector generator generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks, wherein the second motion vector generator generates the base motion vector and delta motion vector by setting the base motion vector equal to the second match vector generated from the previous same-parity field, setting a horizontal component of the delta motion vector equal to positive one if a horizontal component of the First match vector is greater than the closest integer to a horizontal component of the base motion vector divided by two that is no larger than the horizontal component of the base motion vector divided by two, setting the horizontal component of the delta motion vector equal to negative one if the horizontal component of the first match vector is less than the closest integer to the horizontal component of the base motion vector divided by two that is not larger than the horizontal component of the base motion vector divided by two, and setting the horizontal component of the delta notion vector equal to zero if the horizontal component of the first match vector is equal to the closest integer to the horizontal component of the base motion vector divided by two that is not larger than the horizontal component of the base motion vector divided by two.

A method of encoding a video picture sequence to generate a compressed video bit stream, the method comprising the steps of:

• receiving the video picture sequence from a video source, each of the video pictures including a plurality of blocks of pixel data;
• generating first and second match vectors for a current block of a current picture of the video sequence, the first aid second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to generate spatially-compressed blocks;
• and
• utilizing the base and delta motion vectors and the spatially-compressed blocks to generate the compressed video bit stream,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of:
• setting the base motion vector equal to the second match vector generated from the previous same-parity field;
• and
• selecting the delta motion vector such that an opposite parity motion Vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector.

A method of encoding a video picture sequence to generate a compressed video bit stream, the method comprising the steps of:

• receiving the video picture sequence from a video source, each of the video pictures including a plurality of blocks of pixel data;
• generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence generate spatially-compressed blocks;
• and
• utilizing the base and delta motion vectors and the spatially-compressed blocks to generate the compressed video bit stream,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of.
• setting an opposite parity motion vector to be generated by a decoder equal the first match vector generated from the previous opposite-parity field;
• and
• selecting the base motion vector and the delta motion vector such that the base motion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector.

A video encoder for encoding a video picture sequence to generate a compressed video bit stream, the encoder comprising:

• a memory for storing video pictures of the video picture sequence, each of the video pictures including a plurality of blocks of pixel data;
• a motion estimator for performing motion estimation operations on the video pictures of the video sequence, the motion estimator generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, and generating from the first and second match vectors abase motion vector and delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, wherein the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• and
• a processor for performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to thereby generate spatially-compressed blocks, wherein the processor utilizes the base and delta motion vectors and the spatially-compressed blocks to generate the compressed video bit stream,
• wherein the motion estimator generates the base motion vector and delta motion vector by setting the base motion vector equal to the second match vector generated from the previous same-parity field, and selecting the delta motion vector such that an opposite parity motion vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector.

A video encoder encoding a video picture sequence to generate a compressed video bit stream, the encoder comprising:

• a memory for storing video pictures of the video picture sequence, each of the video pictures including a plurality of blocks of pixel data;
• a motion estimator for performing motion estimation operations on the video pictures of the video sequence, the motion estimator generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, and generating from the first and second match vectors a base motion vector and a delta motion vectors such that the delta motion vector is determined as a function of at least one of the first and second match vectors wherein the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• and
• a processor for performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to thereby generate spatially-compressed blocks wherein the processor utilizes the base and delta motion vectors and the spatially-compressed blocks to generate the compressed video bit stream, wherein the motion estimator generates the base motion vector and delta motion vector by setting an opposite parity motion vector to be generated by a decoder equal to the first match vector generated from the previous opposite-parity field, and selecting the base motion vector and the delta motion vector such that the base motion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector.

A method of processing a video picture sequence, the method comprising the steps of:

• receiving the video picture sequence from a video source, each of the video pictures including a plurality of blocks of pixel data;
• generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to generate spatially-compressed blocks;
• and
• utilizing the base and delta motion vectors and the spatially-compressed blocks to generate a compressed video bit stream,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of;
• setting the base motion vector equal to the second match vector generated from the previous same-parity field;
• and
• selecting the delta motion vector such that an opposite parity motion vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector.

An apparatus for processing a video picture sequence, the apparatus comprising:

• a memory for storing video pictures of the video picture sequence, each of the video pictures including a plurality of blocks of pixel data;
• a motion estimator for performing motion estimation operations on the video pictures of the video sequence, the motion estimator generating first and second match vectors for a current block of a current picture of the video sequence the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, and generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, wherein the base motion vector and delta motion vector indicating displacement of the current block from the respective first and second reference blocks;
• a
• a processor for performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to Thereby generate spatially-compressed blocks, wherein the processor utilizes the base and delta motion vectors and the spatially-compressed blocks to generate a compressed video bit stream,
• wherein the motion estimator generates the base motion vector and delta motion vector by setting the base motion vector equal to the second match vector generated from the previous same-parity field, and selecting the delta motion vector such that an opposite parity motion vector to be generated by a decoder is as close as possible to the first match vector generated from the previous opposite-parity field given that the base motion vector is equal to the second match vector.

A method of processing a video picture sequence, the method comprising the steps of:

• receiving the video picture sequence from a video source, each of the video pictures including a plurality of blocks of pixel data;
• generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures;
• generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, the base motion vector and delta motion vector indicating resultant displacement of the current block from the respective first and second reference blocks;
• performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to generate spatially-compressed blocks;
• utilizing the base and delta motion vectors and the spatially-compressed blocks to generate a compressed video bit stream;
• and
• transmitting the compressed video bit stream,
• wherein the step of generating from the first and second match vectors a base motion vector and a delta motion vector further includes the steps of:
• setting an opposite parity motion vector to be generated by a decoder equal to the first match vector generated from the previous opposite-parity field;
• and
• selecting the base motion vector and the delta motion vector such that the base motion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector.

An apparatus for processing a video picture sequence, the apparatus comprising:

• a memory for storing video pictures of the video picture sequence, each of the video pictures including a plurality of blocks of pixel data;
• a motion estimator for performing motion estimation operations on the video pictures of the video sequence, the motion estimator generating first and second match vectors for a current block of a current picture of the video sequence, the first and second match vectors indicating displacement of the current block from respective first and second reference blocks of respective first and second reference pictures, and generating from the first and second match vectors a base motion vector and a delta motion vector, such that the delta motion vector is determined as a function of at least one of the first and second match vectors, wherein the base motion vector and delta motion vector indicating displacement of the current block from the respective first and second reference blocks;
• and
• a processor for performing spatial compression operations on the blocks of pixel data of the video pictures of the video picture sequence to thereby generate spatially-compressed blocks, wherein the processor utilizes the base and delta motion vectors and the spatially-compressed blocks to generate a compressed video bit stream,
• wherein the motion estimator generates the base motion vector and delta motion vector by setting an opposite parity motion vector to be generated by a decoder equal to the first match vector generated from the previous opposite-parity field, and selecting the base motion vector and the delta motion vector such that the base motion vector is as close as possible to the second match vector generated from the previous same-parity field given that the opposite parity motion vector is set equal to the first match vector.