Video Coder Memory Reduction via Single Probability State
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video coders configured for context initialization in entropy coding face increased memory usage, which can hinder processing performance and power efficiency due to the need to store multiple probability states for quantization parameters.
Innovation Solution
Implementing a method where a video coder determines and stores a single probability state for multiple quantization parameters using temporal prediction, reducing memory usage while maintaining improved coding efficiency through context initialization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a video coder stores a probability state for each quantization parameter using context initialization, then coding efficiency is improved, but memory usage increases
Solution Approach 1:
The patent merges multiple probability states for different quantization parameters into a single shared probability state. Instead of maintaining separate probability states for each quantization parameter (QP), the invention uses one common probability state that is updated based on the current QP value, thereby reducing memory usage while preserving coding efficiency through context initialization.
Solution Approach 2:
The single probability state serves multiple quantization parameters simultaneously, making it a universal resource. This probability state is adaptively updated based on the current QP value, allowing it to function effectively for different QP values without requiring separate dedicated states for each parameter.
2Measurement precision
If a video coder stores multiple probability states for quantization parameters, then context initialization accuracy is improved, but processing performance deteriorates due to increased memory access
Solution Approach 1:
By combining multiple probability states into a single shared state, the patent reduces the number of memory accesses required during entropy coding. The single probability state can be updated and accessed more quickly than multiple separate states, thereby improving processing performance while maintaining context initialization accuracy through adaptive updates based on current QP values.
Solution Approach 2:
The invention changes the parameter management approach by using a single probability state whose behavior is dynamically adjusted based on the current quantization parameter value. This parameter-driven adaptation allows the system to maintain accuracy without the overhead of storing multiple full probability states.
3Measurement precision
If a video coder implements context initialization with multiple probability states, then entropy coding precision is improved, but power efficiency decreases
Solution Approach 1:
The patent merges multiple probability states into a single shared state, which reduces the computational overhead and power consumption associated with maintaining and updating multiple separate states. The single probability state is efficiently updated based on current QP values, preserving entropy coding precision while improving power efficiency by eliminating redundant state management operations.
Solution Approach 2:
The invention extracts only the essential probability information needed for entropy coding into a single shared state, rather than maintaining complete separate states for each quantization parameter. This extraction approach retains the necessary precision for accurate entropy coding while removing the unnecessary overhead of multiple full probability states, thereby reducing power consumption.
Data Source
AI summary
A method of decoding video data includes determining, by a video decoder implemented in processing circuitry, a context initialization using a plurality of quantization parameters for a portion of a picture of video data decoded using temporal prediction. The method includes storing, by the video decoder, the context initialization with a single probability state for the plurality of quantization parameters. The method includes entropy decoding, by the video decoder, a set of syntax elements in a bitstream using the single probability state.