Unified Video Transform Scaling for Inverse Quantization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The H.264 video coding standard requires complex and inefficient inverse quantization and transformation processes, particularly for 4×8, 8×4, and 8×8 block sizes, which limit processing efficiency and quality due to the need for multiple tables and multiplication operations, and restrict the highest quality achievable to 4×4 block sizes.
Innovation Solution
A unified transform method that scales coefficients using a single look-up table based on a quantization parameter, block size, and coefficient position, eliminating the need for matrix multiplication and intermediate scaling, and allowing efficient inverse quantization and transformation for 4×4, 4×8, 8×4, and 8×8 block sizes using only addition, subtraction, and shift operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple different arrays V are used for different block sizes (4×4, 4×8, 8×4, 8×8), then inverse quantization accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies universality by using a single array V that serves all block sizes (4×4, 4×8, 8×4, 8×8) through a unified indexing scheme. The index into array V is calculated as index = (QP % 6) * 16 + (QP / 6) * 4 + offset, where QP is the quantization parameter and offset depends on block size and coefficient position. This single array replaces the previous need for multiple separate arrays, reducing device complexity while maintaining inverse quantization accuracy across all block sizes.
2Measurement precision
If matrix multiplication is used for inverse transformation, then transformation accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent replaces the mechanical matrix multiplication system with a simplified computational approach using only addition, subtraction, and bit-shift operations. The inverse transformation is performed by scaling the quantized coefficients using pre-computed values from array V, then applying a simplified transformation that avoids full matrix multiplication. This substitution reduces computational complexity while maintaining transformation accuracy through carefully designed scaling factors and shift operations.
Solution Approach 2:
The patent changes the parameters of the transformation by using a unified quantization parameter QP that applies to all block sizes, and by using a single array V with indices derived from QP and block size. This parameter unification simplifies the transformation process across different block sizes, eliminating the need for separate transformation matrices for each block size while maintaining accuracy through adaptive scaling based on the unified parameters.
3Measurement precision
If intermediate scaling is performed between horizontal and vertical inverse transforms, then processing accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts the intermediate scaling operation from the transformation process and integrates it into the coefficient scaling step using array V. Instead of performing separate scaling operations between horizontal and vertical transforms, the scaling is incorporated into the initial coefficient processing using the unified array V and indexing scheme. This extraction and integration eliminates redundant scaling operations, reducing processing time while maintaining accuracy through the unified approach.
4Measurement precision
If 16-bit registers are used for inverse transformation, then processing precision is improved, but device complexity increases
Solution Approach 1:
The patent changes the parameter of numerical precision by using scaled integer arithmetic with carefully chosen shift amounts instead of requiring 16-bit registers. The transformation results are scaled using bit-shift operations that maintain sufficient precision for video coding applications without requiring wider registers. This parameter change allows the use of standard integer registers while achieving the necessary processing precision through adaptive scaling based on block size and position.
Data Source
AI summary
A method and apparatus is disclosed herein for decoding data (e.g., video data) using transforms. In one embodiment, the decoding process comprises scaling a block of coefficients using a scaling factor determined for each coefficient by computing an index for said each coefficient and indexing a look-up table (LUT) using the index. The index is based on a quantization parameter, a size of the block of coefficients, and a position of said each coefficient within the block. The method also comprises applying a transform to the block of scaled coefficients.