Arithmetic Coding of Spectral Coefficients With Reduced Context Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing arithmetic coding methods face challenges with high encoding/decoding latency and memory capacity requirements due to the large number of contexts and corresponding probability density functions, leading to complexity and resource inefficiencies.
Innovation Solution
A method and device for arithmetic encoding and decoding that utilize non-uniform quantization and context classification to reduce the number of contexts and probability density functions, using neighborhoods of previously encoded/decoded spectral bins to determine context classes and probability density functions, with adaptive variance estimation and mapping to reduce memory and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a large neighbourhood of previously decoded coefficients is used for context-based arithmetic coding, then compression efficiency is improved, but the number of contexts and probability density functions increases tremendously, leading to increased encoding/decoding latency and memory capacity requirements
Solution Approach 1:
The patent segments the continuous neighbourhood of previously decoded coefficients into discrete context groups by comparing coefficients against threshold values. This segmentation transforms the overwhelming number of possible contexts into a manageable set of context groups, each associated with a probability density function, thereby reducing encoding/decoding latency and memory requirements while preserving compression efficiency
Solution Approach 2:
The patent changes the parameter representation by introducing threshold-based context grouping. Instead of using the raw values of neighbourhood coefficients directly as context identifiers, the system transforms them into discrete context groups through threshold comparisons, reducing the dimensionality and complexity of the context space
2Productivity
If the neighbourhood size is increased to improve compression, then better compression is achieved, but encoding/decoding latency increases due to the need to process and store more probability density functions
Solution Approach 1:
The patent segments the neighbourhood coefficients into discrete context groups using threshold comparisons, which reduces the computational burden of processing large neighbourhoods. This segmentation allows the system to efficiently handle larger neighbourhood sizes without proportionally increasing encoding/decoding latency, as the threshold-based grouping is computationally simpler than processing all possible context combinations
Solution Approach 2:
The patent applies partial action by using only the most significant threshold comparisons needed to distinguish between different context groups. Instead of processing all possible neighbourhood configurations, the system uses a subset of threshold-based distinctions that capture the essential contextual information needed for effective compression, thereby reducing latency
3Productivity
If more probability density functions are stored to handle all possible contexts, then compression efficiency improves, but memory capacity requirements increase
Solution Approach 1:
The patent merges multiple similar contexts into unified context groups by comparing neighbourhood coefficients against threshold values. This merging reduces the total number of probability density functions that need to be stored in memory, as multiple distinct contexts that share similar statistical characteristics are grouped together and handled by a single probability density function
Solution Approach 2:
The patent creates universal context groups that can represent multiple specific contexts through threshold-based comparison. Each context group serves as a universal representation for multiple neighbourhood configurations, allowing a single probability density function to handle multiple contexts, thereby reducing memory capacity requirements
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention proposes a method and a device for arithmetic encoding of a current spectral coefficient using preceding spectral coefficients. Said preceding spectral coefficients are already encoded and both, said preceding and current spectral coefficients, are comprised in one or more quantized spectra resulting from quantizing time-frequency-transform of video, audio or speech signal sample values. Said method comprises processing the preceding spectral coefficients, using the processed preceding spectral coefficients for determining a context class being one of at least two different context classes, using the determined context class and a mapping from the at least two different context classes to at least two different probability density functions for determining the probability density function, and arithmetic encoding the current spectral coefficient based on the determined probability density function wherein processing the preceding spectral coefficients comprises non-uniformly quantizing absolutes of the preceding spectral coefficients for use in determining of the context class.