Variable Length Coding of Coefficient Clusters for Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing joint amplitude and position coding methods for variable length coding of quantized transform coefficients in image and video compression require large code table sizes, which can be prohibitive in low-cost and low-power implementations, despite their coding efficiency.
Innovation Solution
The proposed method uses a three-dimensional or two-dimensional joint encoding approach to encode clusters of non-zero-valued coefficients, limiting the number of trailing coefficients to reduce code table size while maintaining coding efficiency, by encoding events with parameters such as the number of preceding zero-valued coefficients, the number of non-zero-valued coefficients, and the number of trailing amplitude-1 coefficients.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If joint amplitude and position coding is used to encode clusters of non-zero coefficients, then coding efficiency is improved, but code table size grows exponentially
Solution Approach 1:
The patent segments the coefficient encoding process into multiple independent stages: run-length encoding for zero coefficients, cluster length encoding for non-zero coefficient groups, and amplitude encoding for significant coefficients. This segmentation allows each stage to use a small, fixed-size code table rather than one large exponential table, resolving the contradiction between coding efficiency and code table size.
Solution Approach 2:
The patent transforms the traditional 2D joint amplitude-position coding into a multi-dimensional sequential encoding process involving run length, cluster length, and amplitude dimensions. By adding the run-length dimension and separating cluster length from amplitude, the patent reduces the complexity of any single code table while maintaining overall coding efficiency through the combined multi-dimensional approach.
2Loss of information
If large code tables are used for joint encoding, then bit-rate performance is improved, but memory usage increases
Solution Approach 1:
The patent divides the encoding task into separate segments (run-length encoding, cluster-length encoding, amplitude encoding), each with its own small code table. This segmentation maintains bit-rate performance by preserving the joint encoding capability while reducing memory usage by using multiple small tables instead of one large table.
Solution Approach 2:
The patent combines multiple simple encoding stages (run-length, cluster-length, amplitude) to achieve the same compression效果 as complex joint encoding, but with reduced memory requirements. Each stage uses a small code table, and their combination provides the overall compression performance.
3Manufacturing precision
If comprehensive joint encoding of run-length, cluster-length, and amplitude is performed, then coding precision is improved, but device complexity increases
Solution Approach 1:
The patent segments the comprehensive encoding into three simple sequential stages with fixed code tables, reducing device complexity while maintaining coding precision. Each stage handles one aspect (run-length, cluster-length, or amplitude) independently, avoiding the need for complex multi-dimensional table lookups.
Solution Approach 2:
The patent uses dynamic adaptive code tables that can be adjusted based on the specific characteristics of the coefficient data (e.g., different tables for different cluster lengths or amplitude ranges). This allows the system to maintain high coding precision while keeping the base device complexity low through the use of simple, adaptable encoding stages.
Data Source
AI summary
A coding method, apparatus, and medium with software encoded thereon to implement a coding method. The coding method includes encoding cluster of consecutive non-zero-valued coefficients, the encoding of a cluster including jointly encoding joint events that each are defined by at least two parameters: the number of zero-valued coefficients preceding the cluster, and the number of non-zero-valued coefficients in the cluster. The encoding of the cluster also includes encoding a parameter indicative of the number of amplitude-1 trailing non-zero-valued coefficients in the cluster, in one version with the parameter indicative of the number of trailing amplitude-1 coefficients part of the joint events such that the coding is according to a 3-dimensional joint variable length coding table. The method further includes encoding the amplitudes of the non-zero-valued coefficients that are not encoded by the joint encoding, e.g., encoding the amplitudes of the other than the trailing amplitude-1 coefficients.