Encoding video data according to target decoding device decoding complexity

1. A method of encoding video data for decoding on a target decoding device having parallel processing, comprising:

(a) receiving uncoded video data;

(b) computing a relationship between each of a plurality of encoding parameters and target decoding device decoding costs;

(c) selecting HEVC encoding parameters to minimize encoding costs, wherein the encoding costs include the target decoding device decoding costs; and

(d) encoding the uncoded video data according to the selected encoding parameters,

(e) wherein when the target decoding device performs synchronized parallel processing by means of multiple cores, shared cache wavefront-parallel-processing (WPP) is used for encoding as indicated by the selected encoding parameters;

(f) wherein when the target decoding device performs free running parallel processing using multiple cores or shared cache, tiling is used for encoding as indicated by the selected encoding parameters,

(g) wherein free running comprises freely defining a size and shape of the tiles by controlling a number of row and column partitions wherein the selected encoding parameters set the tile size and shape as the wavefront to determine which of the parallel cores are used for processing each of the tiles.

2. The method of claim 1 , wherein the encoding parameters are selected subject to a target decoding cost determined on a unit basis, where the unit is a tile or wavefront.

3. The method of claim 1 , wherein the encoding parameters are selected subject to an upper bound on moving average decoding cost.

4. The method of claim 1 , wherein computing a relationship between each of a plurality of encoding parameters and target decoding device decoding costs comprises:

associating each encoding parameter with a cost factor indicative of target decoding complexity;

computing cost factor values for each encoding parameter;

computing target decoding cost from the cost factor values.

5. The method of claim 1 , wherein the encoded video data comprises a plurality of variants of an adaptive bitstream, and the method further comprises:

generating the plurality of variants, by performing steps (a)-(d) for each variant, wherein each of the variants has a different upper bound on moving average decoding cost;

signaling at least one of the plurality of variants in the encoded video data.

6. The method of claim 1 , wherein the encoded video data comprises a complexity-scalable video stream comprising a plurality of complexity layers, and the method further comprises:

generating the plurality of complexity layers, by performing steps (a)-(d) for each complexity layer;

signaling at least one of the plurality of complexity layers in the encoded video data.

7. The method of claim 1 , wherein the encoded video data comprises a plurality of variants, each variant generated with a different parallelization coding parameter associated with a class of target decoding devices.

8. The method of claim 1 , wherein the encoding parameters include at least one of:

a bit rate first coding parameter;

a prediction block size first coding parameter;

a transform unit size first coding parameter; and

a quantization factor (Q p ) of a transform block first coding parameter.

9. The method of claim 8 , wherein:

the prediction block first coding parameter is selected based on one or more of motion compensation, motion vector prediction and intra prediction processes by the target decoding device;

the transform unit size first coding parameter is selected based on a complexity of one or more of an inverse discrete cosine transform and reconstruction processes at the target decoding device; and

the quantization factor is selected according to a distortion and a number of bits to encode the transform block.

10. The method of claim 8 , wherein the target encoding costs (C T ) include at least one of:

an entropy decoding cost (C ENTROPY ) associated with the bit rate encoding parameter;

a motion vector and compensation complexity cost (C MC, MVP ), associated with the prediction block size;

a reconstruction time cost (C RECON ), associated with the transform unit size; and

an inverse discrete cosine transfer cost (C IDCT ), associated with the quantization factor of the transform block.

11. The method of claim 10 , wherein the encoding costs are represented by J=D+λR+μC t , and wherein:

J is a measure of the encoding costs;

D is a measure of a distortion of the encoded video data;

R is a measure of a number of bits needed to encode a block of the video data;

λ is a measure of a slope of a ratio of R and D;

μ is a decoding complexity weighting factor; and

C T is a total cost to decode the block of the video data.

12. The method of claim 11 , wherein C T =C Entropy +C MC,MVP +C IDCT +C Recon .

13. The method of claim 1 , wherein the plurality of encoding parameters includes enabling or disabling of decoder in-loop filtering of the target decoding device.