Unified DCT Matrix Architecture for Large-Block Video Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards, such as HEVC, face challenges with high implementation complexity due to large block size transforms and directional transforms, which are not efficiently implementable in hardware and software environments, particularly on SIMD machines.
Innovation Solution
A transform and quantization architecture that uses a unified DCT matrix for various block sizes, allowing for configurable DCT and IDCT operations with minimal dependency and control logic, suitable for parallel processing, and reduces the need for separate matrices and norm correction matrices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If large block size transforms (16×16, 32×32, 64×64 DCT) and directional transforms are used, then coding efficiency is improved, but implementation complexity increases significantly
Solution Approach 1:
The patent segments the transform process by separating the transform coefficient generation from the transform matrix multiplication. It uses a simplified transform matrix with pre-calculated coefficients that can be applied through simple element-wise multiplication rather than full matrix multiplication, thereby reducing computational complexity while maintaining coding efficiency.
Solution Approach 2:
The patent changes the parameter representation by using pre-computed transform coefficients stored in lookup tables instead of performing real-time matrix multiplications. This parameter transformation allows the system to achieve the same transform effect with significantly reduced computational operations, making large block transforms feasible in hardware and software implementations.
2Measurement precision
If separate transform matrices and norm correction matrices are used for different block sizes, then transform accuracy is maintained, but memory requirements increase
Solution Approach 1:
The patent creates a universal transform coefficient table that can be used across multiple transform block sizes (4×4, 8×8, 16×16, 32×32, 64×64). Instead of storing separate transform matrices for each block size, the system uses a single set of pre-computed coefficients that can be applied to any block size through appropriate scaling and selection, thereby maintaining transform accuracy while significantly reducing memory requirements.
Solution Approach 2:
The patent uses copying by storing pre-computed transform coefficients in lookup tables that can be reused across different transform operations. Instead of recalculating transform matrices for each block size, the system copies and applies appropriate coefficients from the universal table, maintaining precision while reducing the memory footprint compared to storing multiple complete transform matrices.
3Measurement precision
If full matrix multiplication is used for directional transforms, then transform accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent extracts the computationally intensive matrix multiplication operation and replaces it with pre-computed coefficient application. By taking out the complex calculation and pre-solving it offline, the system achieves the same transform accuracy through simple element-wise multiplication with stored coefficients, thereby dramatically reducing runtime computational complexity for directional transforms.
Solution Approach 2:
The patent performs preliminary action by pre-computing all transform coefficients offline and storing them in lookup tables before runtime. This preliminary computation allows the actual transform operation to proceed with minimal computational overhead, maintaining accuracy while reducing the computational complexity during video encoding and decoding operations.
Data Source
AI summary
A method of encoding a video stream in a video encoder is provided that includes computing an offset into a transform matrix based on a transform block size, wherein a size of the transform matrix is larger than the transform block size, and wherein the transform matrix is one selected from a group consisting of a DCT transform matrix and an IDCT transform matrix, and transforming a residual block to generate a DCT coefficient block, wherein the offset is used to select elements of rows and columns of a DCT submatrix of the transform block size from the transform matrix.