Probability Table Compression for Bandwidth-Limited Entropy Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The transmission of probability tables in entropy coding systems consumes significant bandwidth, creating bottlenecks in resource-constrained environments and limiting data communication throughput, as these tables often include unused contexts and require full representation of frequency values for all symbols.
Innovation Solution
The proposed solution involves encoding probability tables such that unused contexts are not transmitted, and the representation of frequency values is dynamically reduced based on the maximum possible value of each entry, allowing for fewer bits to represent the same information, thereby reducing the overall size of the encoded table.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If full probability tables with all M entries are transmitted, then complete information is provided for decoding, but bandwidth consumption increases significantly
Solution Approach 1:
The patent extracts only the necessary probability entries from the full probability table for transmission. By identifying and removing unused or redundant context entries, the system transmits minimal required data while maintaining complete decoding capability, directly resolving the contradiction between information completeness and bandwidth consumption
Solution Approach 2:
The probability table is segmented into used and unused contexts. The encoder transmits only the used context segments, allowing the decoder to reconstruct the full probability table by combining received segments with predetermined default values for unused segments, thereby reducing transmission data while preserving information completeness
2Device complexity
If fixed K bits are used for each of M entries, then encoding is simple, but the representation size is M*K bits which is excessive
Solution Approach 1:
The patent applies dynamic bit allocation where the number of bits used to represent probability entries varies based on the actual data characteristics. Instead of fixed K bits per entry, the system dynamically adjusts the number of bits required for each context, reducing the total representation size from M*K bits to fewer bits while maintaining encoding simplicity through adaptive schemes
Solution Approach 2:
The system changes the parameter of bit representation from a fixed value K to a variable number of bits based on the maximum possible value of each entry. By analyzing the actual probability distribution and adjusting the bit representation parameter dynamically, the system achieves compact representation without excessive complexity
3Reliability
If all context entries are transmitted, then decoding accuracy is maintained, but transmission efficiency decreases
Solution Approach 1:
The decoder is designed to self-reconstruct the complete probability table by combining received used context entries with predetermined default values for unused contexts. This self-service capability allows the system to maintain full decoding accuracy without transmitting all M entries, thereby improving transmission efficiency while preserving reliability
Data Source
AI summary
This disclosure provides methods, devices, and systems for data compression. The present implementations more specifically relate to encoding techniques for compressing probability tables used for entropy coding. In some aspects, an entropy encoder may encode a probability table so that one or more contexts are represented by fewer bits than would otherwise be needed to represent the frequency of each symbol as a proportion of the total frequency of all symbols associated with such contexts. For example, if a given row of the probability table (prior to encoding) includes a number (M) of entries each having a binary value represented by a number (K) of bits, the same row of entries may be represented by fewer than M*K bits in the encoded probability table.