Dependent Scalar Quantizer Decisions for High-Throughput Entropy Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding technologies face challenges in achieving high throughput and coding efficiency when implementing dependent scalar quantization, particularly due to increased regular coded bins and interleaved bypass coded bins, which degrade performance compared to HEVC and VTM-1 designs.
Innovation Solution
A decision scheme for scalar quantizer selection in dependent scalar quantization that uses SIG-based state transition and context modeling, determining the quantizer based on the significance flag (SIG) of preceding coefficients, to maintain high throughput and coding efficiency similar to HEVC and VTM-1.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If dependent scalar quantization is implemented with traditional coefficient coding, then coding efficiency is improved, but throughput deteriorates due to increased regular coded bins and interleaved bypass coded bins
Solution Approach 1:
The patent segments the transform coefficients into different groups based on their positions and significance. By dividing coefficients into first transform coefficients (with non-zero SIG flags) and second transform coefficients (with zero SIG flags), the patent applies different coding strategies to each segment. This segmentation allows the system to maintain coding efficiency for significant coefficients while using simpler coding for less significant ones, thereby resolving the contradiction between coding efficiency and throughput.
Solution Approach 2:
The patent applies local quality by differentiating the coding method based on the local characteristics of each transform coefficient. Significant coefficients (with SIG=1) receive detailed coding with multiple flags (gt1, gt2), while non-significant coefficients (with SIG=0) receive simplified coding. This local differentiation ensures that coding resources are allocated efficiently, maintaining high coding efficiency where needed while improving throughput where simple coding suffices.
2Manufacturing precision
If context modeling depends on preceding transform coefficients, then coding efficiency is improved, but decoding complexity increases due to sequential dependency requirements
Solution Approach 1:
The patent applies preliminary action by pre-determining the coding strategy for each transform coefficient based on its significance flag before actually performing the coding. The SIG flag is set in advance based on the coefficient value, and this pre-computed information guides the subsequent coding process. This preliminary action allows the context model to be configured efficiently without requiring complex real-time decisions, thus maintaining coding efficiency while managing decoding complexity.
Solution Approach 2:
The patent implements feedback by using the decoded significance flag (SIG) of each transform coefficient to determine the coding parameters for subsequent coefficients. The context model for coding transform coefficients at position (x0, y0) depends on the decoded SIG flag of the previous coefficient, creating a feedback loop that adapts the coding strategy based on actual coefficient characteristics. This feedback mechanism enables efficient context modeling while keeping the complexity manageable through systematic dependency management.
3Manufacturing precision
If multiple scan passes are used for entropy decoding, then coding efficiency is improved, but processing time increases
Solution Approach 1:
The patent segments the entropy decoding process into multiple passes, where the first pass handles significant transform coefficients and subsequent passes handle non-significant coefficients. By organizing the decoding into segmented passes rather than a single monolithic process, the patent achieves better coding efficiency for different coefficient types while managing processing time through structured execution. The segmentation allows parallel processing opportunities and reduces the time required for each individual pass.
Solution Approach 2:
The patent applies partial action by performing detailed entropy decoding only for transform coefficients that are significant (SIG=1), while using simpler decoding methods for non-significant coefficients (SIG=0). This partial application of complex decoding only where necessary maintains high coding efficiency for important coefficients while reducing overall processing time by avoiding unnecessary complex processing for less important coefficients.
Data Source
AI summary
When dependent scalar quantization is used, the choice of the quantizer depends on the decoding of the preceding transform coefficient, and the entropy decoding of a transform coefficient depends on quantizer choice. To maintain high throughput in hardware implementations for transform coefficient entropy coding, several decision schemes of the scaler quantizer are proposed. In one implementation, the state transition and the context model selection are based on only regular coded bins. For example, the state transition can be based on the sum of the SIG, gt1 and gt2 flags, the exclusive-or function of the SIG, gt1 and gt2 flags, or based on only the gt1 or gt2 flag. When a block of transform coefficients is coded, the regular mode bins can be coded first in one or more scan passes, and the remaining bypass coded bins are grouped together in another one or more scan passes.