Hybrid Video Coder Selective Adaptive Context Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for efficient video encoding and decoding techniques is hindered by the power and size constraints of memory required for adaptive context models, which limits the storage and updating of context models needed for high throughput in video coding systems.
Innovation Solution
The implementation of a hybrid coding system that selectively uses adaptive or non-adaptive context models, where adaptive models are stored in fast memory for updating and non-adaptive models are stored in read-only memory to reduce power consumption and increase storage capacity, allowing for efficient entropy coding without modifying existing arithmetic coding engines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If adaptive context models are used for entropy coding, then compression efficiency is improved, but memory size and power consumption increase
Solution Approach 1:
The patent segments the memory system into two distinct parts: a first memory for storing adaptive context models and a second memory for storing non-adaptive context models. This segmentation allows the system to maintain both adaptive and non-adaptive coding capabilities while managing memory resources efficiently, as each memory type has different access and update requirements.
Solution Approach 2:
The patent applies local quality by assigning different memory characteristics to different context model types. The first memory is optimized for adaptive models requiring frequent updates, while the second memory is optimized for non-adaptive models requiring only read access. This localized optimization resolves the contradiction by matching memory properties to specific usage patterns rather than using a uniform memory solution.
2Loss of information
If adaptive context models are updated during bitstream processing, then coding gain is improved, but power consumption and memory access complexity increase
Solution Approach 1:
The patent segments the coding process into two pathways: one using adaptive context models from the first memory for syntax elements requiring coding gain, and another using non-adaptive context models from the second memory for syntax elements where coding gain is less critical. This segmentation allows selective application of adaptive coding, reducing overall power consumption while maintaining necessary coding gain.
Solution Approach 2:
The patent implements partial adaptive action by applying adaptive context model updates only to specific syntax elements where they provide significant benefit, rather than updating all context models adaptively. This partial application reduces the power consumption and computational overhead associated with continuous adaptive updates while maintaining adequate compression performance.
3Productivity
If all context models are stored in fast memory for adaptive updating, then coding performance is improved, but memory cost and device complexity increase
Solution Approach 1:
The patent segments the memory architecture into two distinct memory systems with different characteristics and access patterns. The first memory handles adaptive context models with update operations, while the second memory handles non-adaptive context models with read-only access. This segmentation simplifies the overall device complexity by creating specialized memory pathways rather than requiring a single complex unified memory system.
Solution Approach 2:
The patent creates a universal memory architecture that can handle both adaptive and non-adaptive context modeling requirements through the two-memory system. This multi-functional design allows the same entropy coding engine to operate with different context model types without requiring separate dedicated hardware for each mode, thereby reducing overall device complexity while maintaining coding performance.
Data Source
AI summary
Techniques are described of hybrid coders that are configured to selectively use adaptive or non-adaptive coding techniques. A video coder (e.g., video encoder or video decoder) may code (e.g., encode or decode) first video data (e.g., a syntax element or value), for coding a first block, based on an adaptive context model (e.g., one or more adaptive context models) and code second video data, for coding a second block, based on a non-adaptive context model (e.g., one or more non-adaptive context models).