Entropy Coding Partition Scheme for Adaptive Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing entropy coding schemes, such as Huffman codes and arithmetic coding, face challenges in adapting to time-varying source statistics and higher-order probability modeling, leading to inefficiencies in compression efficiency and complexity.
Innovation Solution
The proposed solution involves decomposing syntax elements into multiple source symbols and categorizing them into substreams for VLC and PIPE coding, allowing for better adaptation to statistics and improved compression efficiency through separate encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If Huffman codes with VLC tables are used, then implementation is simple and easy to realize, but adaptation to time-varying source statistics is demanding and complex
Solution Approach 1:
The source symbols are divided into multiple substreams based on their statistical characteristics. Each substream is encoded separately using appropriate coding methods (VLC or PIPE), allowing different parts of the data to be handled with different complexity levels matched to their specific needs.
Solution Approach 2:
The coding system dynamically selects between VLC and PIPE coding methods for different substreams based on the actual statistical properties of the source data. This dynamic adaptation allows the system to optimize for both simplicity and efficiency depending on the input characteristics.
2Adaptability or versatility
If arithmetic coding is used, then handling of adaptive and higher-order probability modeling is improved, but implementation complexity is substantially increased
Solution Approach 1:
The data stream is segmented into substreams with similar statistical properties. PIPE coding is applied to specific substreams where it provides advantages in handling adaptive probability modeling, while other substreams use simpler VLC coding, thus distributing the complexity only where necessary.
Solution Approach 2:
Different coding methods are applied to different substreams based on their local statistical characteristics. PIPE coding with its superior adaptive modeling capabilities is applied locally to substreams that benefit from it, rather than applying complex arithmetic coding to the entire data stream.
3Device complexity
If PIPE coding is used, then complexity is reduced compared to arithmetic coding, but it still has higher complexity than VLC coding
Solution Approach 1:
The source sequence is divided into multiple substreams, allowing the system to apply VLC coding (lower complexity) to some substreams and PIPE coding (higher complexity but better compression) to others. This segmentation enables the system to achieve good overall compression efficiency without applying high complexity to all data.
Solution Approach 2:
Instead of applying PIPE coding to all source symbols, the system applies it only to specific substreams where the additional complexity yields meaningful compression improvements. This partial application optimizes the trade-off between complexity and productivity.
Data Source
AI summary
Decomposing a value range of the respective syntax elements into a sequence of n partitions with coding the components of z laying within the respective partitions separately with at least one by VLC coding and with at least one by PIPE or entropy coding is used to greatly increase the compression efficiency at a moderate coding overhead since the coding scheme used may be better adapted to the syntax element statistics. Accordingly, syntax elements are decomposed into a respective number n of source symbols si with i=1 . . . n, the respective number n of source symbols depending on as to which of a sequence of n partitions into which a value range of the respective syntax elements is sub-divided, a value z of the respective syntax elements falls into, so that a sum of values of the respective number of source symbols si yields z, and, if n>1, for all i=1 . . . n−1, the value of si corresponds to a range of the ith partition.