System and method for optimizing context-adaptive binary arithmetic coding
A system and method is provided for ordering intervals rLPS and rMPS of a range to increase speed of binary symbol decoding in a binary arithmetic decoder. The method comprises the steps of: placing rLPS at a bottom of the range; enabling subtraction for rMPS to occur in parallel with comparison of rLPS and offset; and, reducing time that it takes to decode a bin. A method is also provided for performing context selection for a given syntax element, comprising the steps of: first, comparing information regarding properties of neighboring pixels with a threshold; second, adding results of threshold comparison of neighboring pixels, to provide a secondary result; and using the secondary result to select a context.
1. A method of ordering intervals rLPS and rMPS of a range to increase speed of decoding a binary symbol (bin) in a binary arithmetic decoder, comprising the steps of:
placing rLPS at a bottom of the range;
subtracting the rLPS from a provided offset; and
comparing the provided offset to the rLPS to determine if the offset is greater than or equal to the rLPS,
wherein the steps of subtracting and comparing are performed substantially in parallel, reducing time it takes to decode the bin.
2. The method of claim further comprising the step of:
if the offset less than the rLPS, updating the range to the value of the rLPS, otherwise updating the range to the value of the rMPS.
3. The method of claim 1 , wherein the range comprises a current interval divided into subintervals based on the probability of the bins.
4. The method of claim 3 , further comprising the step of updating the range after every bin is decoded.
5. The method of claim 4 , further comprising the step of renormalizing the updated range and the offset.
6. The method of claim 1 , wherein the binary arithmetic decoder is configured to perform context-adaptive binary arithmetic coding (CABAC).
7. The method of claim 1 , further comprising the step of obtaining the rLPS from a look up table.
8. A method of ordering intervals rLPS and rMPS of a range to increase speed of encoding a binary symbol (bin) in a binary arithmetic encoder, comprising the steps of:
placing rLPS at a bottom of the range;
subtracting the rLPS from a provided offset; and
comparing the bin to a provided MPS to determine if the bin equals the MPS,
wherein the steps of subtracting and comparing are performed substantially in parallel, reducing time it takes to encode the bin.
9. The method of claim 8 , further comprising the step of:
if the bin is not equal to the rLPS, updating the range to the value of the rLPS, otherwise updating the range to the value of the rMPS.
10. The method of claim 8 , wherein the range comprises a current interval divided into subintervals based on the probability of the bins.
11. The method of claim 9 , further comprising the step of updating the range after every bin is encoded.
12. The method of claim 11 , further comprising the step of renormalizing the updated range and the offset.
13. The method of claim 8 , wherein the binary arithmetic encoder is configured to perform context-adaptive binary arithmetic coding (CABAC).
14. The method of claim 8 , further comprising the step of obtaining the rLPS from a look up table.
15. A method of performing context selection for a given syntax element, comprising the steps of:
first, comparing information regarding properties of neighboring, pixels with a threshold, resulting in a threshold comparison;
second, adding results of the threshold comparison of neighboring pixels, to provide a secondary result; and
using the secondary result to select a context.
16. The method of claim 15 , wherein the context selection is for a motion vector difference (mvd).
17. The method of claim 15 , wherein the threshold comparison results are stored in a single bit memory.