Adaptive Binarization in Arithmetic Coding for Faster Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing bin-to-bit ratio in modern codecs like H.264 and HEVC leads to a higher number of computations, hampering decoding/coding speeds.
Innovation Solution
The proposed solution involves selecting a binarization for currently decoded/encoded information values based on previously decoded/encoded information values, using binary arithmetic decoding/encoding to improve coding/decoding efficiency and speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If context-adaptive arithmetic coding (CABAC) is used to improve coding efficiency, then the bin-to-bit ratio increases, but the decoding/coding speed decreases due to increased computations
Solution Approach 1:
The patent applies dynamics by making the binarization method adaptive and variable rather than fixed. The decoder dynamically selects between different binarization approaches (uniform binarization for values ≤ threshold, exponential binarization for values > threshold) based on the current information value being decoded. This dynamic adaptation allows the system to optimize the bin-to-bit ratio for each specific value while maintaining manageable computational complexity through the threshold-based decision rule.
2Loss of information
If the bin-to-bit ratio is increased to improve coding efficiency, then more binary decisions are required per bit, but the number of computations increases, hampering processing speed
Solution Approach 1:
The patent changes parameters by introducing a threshold value that determines which binarization method to apply. By adjusting this threshold parameter, the system can control the balance between bin-to-bit ratio and computational complexity. The threshold acts as a configurable parameter that allows optimization for different scenarios, enabling the system to achieve high coding efficiency when needed while limiting computational overhead by switching to simpler binarization for smaller values.
Data Source
AI summary
A more effective coding/decoding of a sequence of information values and, for instance, one which enables implementations of increased processing speed, is achieved by rendering the binarization of the information values dependent on previously coded/decoded information values.