Method and apparatus for sub-pixel motion compensation
View Patent ↗A method of motion compensation for an input block is disclosed. The method generally includes the steps of (A) generating a plurality of tap values in response to a motion vector for the input block, (B) generating an interpolated block by programmable filtering a reference block using the tap values, the interpolated block being spatially offset from the reference block by less than a half-pel separation and (C) generating an output block in response to the input block and the interpolated block.
1. A method of motion compensation for an input block performed by an interpolator, comprising the step of:
(A) using said interpolator, generating a plurality of tap values in response to a motion vector for said input block;
(B) generating an interpolated block by programmable filtering a reference block using said tap values, said interpolated block being spatially offset from said reference block by less than a half-pel separation; and
(C) generating a motion compensated output block in response to said input block and said interpolated block, wherein generating said tap values is in further response to a plurality of parameters common to a plurality of interpolation levels possible for said interpolated block.
2. The method according to claim 1 , wherein said parameters comprise a plurality of half-pel tal values.
3. The method according to claim 1 , further comprising the step of transmitting said parameters in a bitstream.
4. The method according to claim 1 , further comprising the step of extracting said parameters from a bitstream.
5. The method according to claim 1 , further comprising the step of:
fixed filtering said interpolated block at least one interpolation level higher after said programmable filtering.
6. The method according to claim 5 , wherein said tap values for a quarter-pel interpolation match said tap value for a half-pel interpolation.
7. The method according to claim 1 , further comprising the step of:
altering said tap value to adjust a sharpness of said output block.
8. The method according to claim 1 , wherein said plurality of tap values comprise at least eight tap values.
9. A system comprising:
a programmable filter (i) having a plurality of programmable taps and (ii) configured to generate an interpolated block directly from a reference block having an integer-pel resolution, said interpolated block being at least two interpolation levels above said reference block;
an adder configured to generate a reconstructed block by adding said interpolated block to a residue block; and
a tap module configured to generate a plurality of tap values for said programmable taps, wherein said tap module is further configured to generate said tap values in response to at least one parameter common to said interpolation levels.
10. The system according to claim 9 , further comprising a fixed filter configured to interpolate said interpolated block at least one finer interpolation level.
11. The system according to claim 10 , further comprising a tap module configured to generate a plurality of tap values for said programmable taps.
12. A method of motion compensation for an input block performed by an interpolator, comprising the step of:
(A) using said interpolator, generating a first block greater than one interpolation level above a reference block by first filtering said reference block;
(B) generating a second block at least one interpolation level above said first block by second filtering said first block;
(C) generating a motion compensated output block by adding said second block to said input block;
(D) programming a plurality of tap values for said first filtering; and
(E) generating said tap values from a plurality of parameters common to said interpolation levels.
13. The method according to claim 12 , wherein said first block consist of a quarter-pel interpolated block and said second block consist of an eighth-pel interpolatated block.
14. The method according to claim 12 , futher comprising the step of:
generating a third block by third filtering said reference block to a half-pel interpolation level.
15. The method according to claim 14 , further comprising the step of:
generating a plurality of tap values said third filtering based on a plurality of parameters common with said first filtering.