Video Decoder Regular Coded Bin Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding and decoding techniques introduce lossy compression artifacts, such as blockiness, due to transform and quantization processes, which affect the quality of decoded video data.
Innovation Solution
The proposed solution involves determining a threshold number of regular coded bins for decoding passes, context decoding syntax elements for the first set of coefficients, and bypass decoding additional syntax elements for the second set of coefficients, using a Rice parameter derived from neighboring coefficients to optimize transform coefficient coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If transform and quantization processes are applied to compress video data, then compression efficiency is improved, but lossy compression artifacts such as blockiness are introduced
Solution Approach 1:
The patent segments the coefficient coding process into two distinct passes: a first pass that codes significance flags and parity level flags using context modeling, and a second pass that codes remaining coefficients using bypass decoding. This segmentation allows different coding strategies to be applied to different parts of the data, optimizing both compression efficiency and quality by handling significant coefficients differently from less significant ones.
Solution Approach 2:
The patent applies different coding methods to different coefficients based on their importance. Significant coefficients (those with non-zero values) are coded using context modeling in the first pass, while less significant coefficients are handled in the second pass using bypass decoding. This local differentiation optimizes the balance between compression efficiency and quality preservation for each coefficient based on its individual characteristics.
2Measurement precision
If context decoding is used for all coefficients, then decoding accuracy is maintained, but decoding complexity and processing time increase
Solution Approach 1:
The patent divides coefficient decoding into two passes where the first pass uses context decoding for significance and parity level flags, while the second pass uses bypass decoding for remaining coefficients. This segmentation reduces overall decoding complexity by applying the more complex context decoding only where necessary (for significant coefficients) while using simpler bypass decoding for less significant coefficients.
Solution Approach 2:
The patent applies context decoding partially - only to the extent necessary for coding significant coefficients in the first pass. The second pass uses bypass decoding which requires less computational complexity. This partial application of the more complex decoding method achieves sufficient accuracy for significant coefficients while avoiding the unnecessary complexity of applying context decoding to all coefficients.
3Measurement precision
If more syntax elements are decoded for each coefficient, then coefficient coding accuracy is improved, but the number of coded bins increases
Solution Approach 1:
The patent segments the syntax element decoding into two passes: the first pass decodes significance flags and parity level flags (fewer bins), while the second pass decodes remaining coefficients using bypass decoding (also fewer bins). This segmentation reduces the total number of coded bins compared to decoding all possible syntax elements for every coefficient, while maintaining accuracy for significant coefficients.
Solution Approach 2:
The patent applies partial decoding - only decoding the necessary syntax elements for each coefficient based on its significance. Significant coefficients receive context decoding for their flags in the first pass, while the second pass uses bypass decoding that requires fewer bins. This partial approach maintains sufficient coding accuracy without the excessive bin count that would result from decoding all possible syntax elements for every coefficient.
Data Source
AI summary
A video decoder can be configured to determine a threshold number of regular coded bins for a first decoding pass; for a first set of coefficients, context decode syntax elements of a coefficient group until the threshold number of regular coded bins is reached; and in response to reaching the threshold number of regular coded bins, for a second set of coefficients, bypass decode additional syntax elements, wherein to bypass decode the additional syntax elements the video decoder, for a coefficient of the second set of coefficients, derives a value for a Rice parameter based on a sum of absolute values of neighboring coefficients to the coefficient; determine values for the first set of coefficients of the transform unit based on the context decoded syntax elements; and determine values for the second set of coefficients of the transform unit based on the additional syntax elements.