Dynamic slice resizing while encoding video
AssigneeIntel Corporation
InventorDariusz Dzik
Life of this patent
Estimating the term…
Abstract
An apparatus having a plurality of processors is disclosed. The processors may be configured to (i) gather a plurality of statistics by encoding a current picture in a video signal, (ii) calculate a plurality of complexity values in response to both the statistics and a plurality of coefficients and (iii) partition a next picture in the video signal into a plurality of slices in response to the complexity values such that each of the slices has a similar coding complexity. The statistics generally characterize how the current picture was encoded. The coefficients may correspond to a plurality of coding modes.
Claims16
An apparatus comprising:
- a plurality of processors configured to (i) gather a plurality of statistics by encoding a current picture in a video signal, (ii) calculate a plurality of complexity values in response to both said statistics and a plurality of coefficients and (iii) partition a next picture in said video signal into a plurality of slices in response to said complexity values such that each of said slices has a similar coding complexity, wherein (a) said statistics characterize how said current picture was encoded and (b) said coefficients correspond to a plurality of coding modes, wherein:
- said complexity values correspond to a plurality of respective macroblock rows in said current picture;
- and
- said statistics per each of said macroblock rows comprise (i) a number of bits created by said encoding, (ii) a number of inter-coded macroblocks, (iii) a number of intra-coded macroblocks, (iv) a number of skip-coded macroblocks, (v) a number of sub-pixel motion estimations, (vi) a number of intra-block motion estimations and (vii) a number of transformations from a spatial domain to a frequency domain.
A method for dynamic slice resizing while encoding video, comprising the steps of:
- (A) gathering a plurality of statistics by encoding a current picture in a video signal, wherein said statistics characterize how said current picture was encoded by a plurality of processors;
- (B) calculating a plurality of complexity values in response to both said statistics and a plurality of coefficients, wherein said coefficients correspond to a plurality of coding modes;
- and (C) partitioning a next picture in said video signal into a plurality of slices in response to said complexity values such that each of said slices has a similar coding complexity, wherein:
- said complexity values correspond to a plurality of respective macroblock rows in said current picture;
- and
- said statistics per each of said macroblock rows comprise (i) a number of bits created by said encoding, (ii) a number of inter-coded macroblocks, (iii) a number of intra-coded macroblocks, (iv) a number of skip-coded macroblocks, (v) a number of sub-pixel motion estimations, (vi) a number of intra-block motion estimations and (vii) a number of transformations from a spatial domain to a frequency domain.
An apparatus comprising:
- means for gathering a plurality of statistics by encoding a current picture in a video signal, wherein said statistics characterize how said current picture was encoded;
- means for calculating a plurality of complexity values in response to both said statistics and a plurality of coefficients, wherein said coefficients correspond to a plurality of coding modes;
- and
- means for partitioning a next picture in said video signal into a plurality of slices in response to said complexity values such that each of said slices has a similar coding complexity, wherein:
- said complexity values correspond to a plurality of respective macroblock rows in said current picture;
- and
- said statistics per each of said macroblock rows comprise (i) a number of bits created by said encoding, (ii) a number of inter-coded macroblocks, (iii) a number of intra-coded macroblocks, (iv) a number of skip-coded macroblocks, (v) a number of sub-pixel motion estimations, (vi) a number of intra-block motion estimations and (vii) a number of transformations from a spatial domain to a frequency domain.
Dependencies are inferred from the claim text.