Quantization Coefficient Region Coding for Zero-Run Bitrate Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video compression, particularly in AOMedia Video Model (AVM), the scanning of quantization coefficients results in numerous consecutive zero coefficients, leading to significant coding overload and poor performance due to excessive bitrate.
Innovation Solution
The method involves partitioning quantization coefficients into regions based on scanning order and using region-level syntax elements to determine whether coding is required, employing continuous scanning order indexes for efficient encoding and decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If quantization coefficients are scanned in a specific scanning order, then the coefficients can be processed systematically, but numerous consecutive zero coefficients occur leading to significant coding overhead and large bitrate
Solution Approach 1:
The patent divides the transform block into multiple coefficient groups (CGs) based on the scanning order. Each coefficient group contains a specific range of coefficients (e.g., 4x4 blocks in an 8x8 transform). By segmenting the coefficients into groups, the patent can independently encode each group and skip entire groups that contain only zero coefficients, thereby reducing coding overhead and bitrate while maintaining systematic processing.
2Productivity
If each transform block is divided into different coefficient groups for processing, then coding efficiency can be improved, but the complexity of determining and processing multiple groups increases
Solution Approach 1:
The patent performs preliminary division of the transform block into coefficient groups based on predetermined scanning orders (e.g., zigzag, diagonal, or raster scan) before the actual encoding process. The grouping structure is established in advance with clear boundaries and indexing rules, so that during encoding, the encoder can directly determine which groups to process without complex real-time calculations. This preliminary organization reduces the complexity of group determination while maintaining coding efficiency.
3Loss of information
If consecutive zero coefficients are encoded individually, then complete coefficient information is transmitted, but significant coding overload occurs resulting in too large bitrate
Solution Approach 1:
The patent merges multiple consecutive zero coefficients into a single coefficient group representation. When a coefficient group contains only zero coefficients, the entire group is encoded as a single run-length value or skipped entirely using syntax elements like `last_subset_non_zero_coeff` and `coeff_group_significance`. This merging approach transmits the same information (that these coefficients are zero) with significantly fewer bits compared to encoding each zero coefficient individually, thereby reducing bitrate while preserving complete coefficient information.
Data Source
AI summary
The embodiments of the present disclosure disclose a decoding method, which includes: parsing a bitstream to determine first syntax element information of a current block; determining a coefficient region where to-be-decoded coefficients of the current block are located based on an end of block of the current block indicated by the first syntax element information; where the coefficient region includes one or more coefficient regions, and scanning orders of respective coefficient regions are continuous; determining second syntax element information of the coefficient region; where the second syntax element information indicates whether coefficient decoding is required to be performed on the coefficient region; and determining a quantization coefficient value of the current block based on the second syntax element information.