Joint sub-pixel interpolation filter for temporal prediction

An apparatus configured to encode or decode video content, the apparatus comprising:

• one or more computer processors;
• and
• a computer-readable storage medium comprising instructions, that when executed, control the one or more computer processors to be configured for:
• determining a plurality of interpolation filters for use in interpolating sub-pel values for a temporal prediction process of video content, wherein each interpolation filter of at least some of the plurality of interpolation filters is designed for a respective combination of sub-pel offsets for two reference blocks, the sub-pel offsets representing respective sub-pel motion for each of the two reference blocks;
• determining a first sub-pel offset for a first reference block for a first block of video content;
• determining a second sub-pel offset for a second reference block for the first block of video content, the second reference block different from the first reference block;
• determining a set of interpolation filters to interpolate a set of sub-pel pixel values for use in the temporal prediction process for the first block of video content by comparing the first sub-pel offset and the second sub-pel offset to combinations of the sub-pel offsets within the plurality of interpolation filters;
• wherein determining the set of interpolation filters comprises:
• determining a first interpolation filter for the first reference block, the first interpolation filter designed with a first delay to cover a first portion of pixel values, the first delay indicating how many taps of the first interpolation filter are to one side of a sub-pel position to be interpolated using the first reference block and how many taps of the interpolation filter are to the other side of the sub-pel position to be interpolated using the first reference block;
• and
• determining a second interpolation filter for the second reference block, the second interpolation filter designed with a second delay to cover a second portion of pixel values, the second delay indicating how many taps of the second interpolation filter are to one side of the sub-pel position to be interpolated using the second reference block and how many taps of the interpolation filter are to the other side of the sub-pel position to be interpolated using the second reference block;
• wherein:
• the first portion of pixel values is different from the second portion of pixel values;
• interpolating a first sub-pel pixel value for the first reference block using the first interpolation filter;
• interpolating a second sub-pel value for the second reference block using the second interpolation filter;
• and
• averaging the first sub-pel pixel value and the second sub-pel pixel value for the temporal prediction process for the first block.

A non-transitory computer-readable storage medium comprising instructions for encoding or decoding video content, the instructions, when executed, control a computer system to be operable for:

• determining a plurality of interpolation filters for use in interpolating sub-pel values for a temporal prediction process of video content, wherein each interpolation filter of at least some of the plurality of interpolation filters is designed for a respective combination of sub-pel offsets for two reference blocks, the sub-pel offsets representing respective sub-pel motion for each of the two reference blocks;
• determining a first sub-pel offset for a first reference block for a first block of video content;
• determining a second sub-pel offset for a second reference block for the first block of video content, the second reference block different from the first reference block;
• and
• determining a set of interpolation filters to interpolate a set of sub-pel pixel values for use in the temporal prediction process for the first block of video content by comparing the first sub-pel offset and the second sub-pel offset to combinations of the sub-pel offsets within the plurality of interpolation filters;
• interpolating a first sub-pel pixel value for the first reference block;
• and
• interpolating a second sub-pel value for the second reference block, wherein the plurality of interpolation filters is designed for all combinations of available sub-pel offsets for each of the first reference block and the second reference block.