VLC Table Array Selection Memory Reduction via Difference Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies require significant memory to store VLC table array selection tables, which can be inefficient in terms of memory usage, especially when dealing with different block characteristics such as prediction types and sizes.
Innovation Solution
The approach involves storing only one VLC table array selection table and using an indication of the difference between this table and another to reconstruct the necessary entries, thereby reducing memory usage by eliminating the need to store multiple tables.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple VLC table array selection tables are stored to support different block characteristics, then coding performance is maintained, but memory usage increases
Solution Approach 1:
The VLC table array selection tables are segmented into a base table and difference tables. The base table contains common entries, while difference tables contain only the specific modifications needed for different block characteristics. This segmentation allows the system to maintain multiple table configurations without storing complete redundant tables, thus reducing memory usage while preserving coding performance.
Solution Approach 2:
The difference tables are nested within the base table structure. The decoder reconstructs full VLC tables by combining the base table with the appropriate difference tables, similar to nested dolls where smaller elements are contained within larger structures. This nesting allows compact storage of multiple table configurations in a hierarchical manner.
2Adaptability or versatility
If multiple complete VLC tables are stored for different prediction types and block sizes, then all coding scenarios are supported, but memory requirements increase significantly
Solution Approach 1:
The base VLC table array selection table serves as a universal foundation that can be adapted to support multiple coding scenarios through the application of different difference tables. This universal base table contains common entries applicable to all block types, while the difference tables provide scenario-specific modifications, allowing one base table to fulfill multiple functions.
Solution Approach 2:
Instead of storing completely different tables for different scenarios, the patent applies local quality changes through difference tables that contain only the specific entries that differ for each scenario. This allows the system to maintain adaptability for different prediction types and block sizes by modifying only the necessary local entries rather than storing entirely separate tables.
Data Source
AI summary
This disclosure describes techniques for coding transform coefficients for a block of video data. According to these techniques, a video coder (a video encoder or video decoder) stores a first VLC table array selection table in memory, and an indication of at least one difference between the first VLC table array selection table and a second VLC table array selection table. The video coder reconstructs at least one entry of the second VLC table array selection table based on the first VLC table array selection table using the stored indication of the difference between the first VLC table array selection table and a second VLC table array selection table. The video coder uses the reconstructed at least one entry of the second VLC table array selection table to code at least one block of video data.