Multi-format Video Decoder Using Vector Processing Units
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding and decoding methods, such as H.264 and VP8, face limitations in performance and speed, particularly in handling stereoscopic video signals and scalable video coding, which affects the efficiency and flexibility of video processing in devices with varying capabilities.
Innovation Solution
A multi-format video decoder system that utilizes vector processing units and software libraries to configure and decode various video coding formats, including H.264, SVC, and VP8, by generating entropy decoded data and performing parallel matrix and filter operations to produce a decoded video signal suitable for diverse video display devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video decoding methods (H.264, VP8) are used, then decoding functionality is provided, but performance and speed are limited
Solution Approach 1:
The patent replaces conventional scalar processing with vector processing instructions. The processor executes vector instructions that operate on multiple data elements simultaneously, substituting sequential mechanical operations with parallel vectorized operations to achieve both high speed and accuracy in video decoding
Solution Approach 2:
The patent changes the processing parameter from scalar to vector level. By transforming individual pixel operations into vector operations that process multiple pixels simultaneously, the system achieves improved decoding performance while maintaining accuracy through preserved processing precision
2Adaptability or versatility
If multiple video coding formats are supported, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent implements a universal video decoder architecture that can handle multiple video coding formats (H.264, VP8, and scalable video coding) through a common set of vector processing instructions. The same vector processing unit performs different decoding functions by executing format-specific vector instructions, eliminating the need for separate dedicated hardware for each format
Solution Approach 2:
The patent uses software libraries that contain decoded reference pictures and configuration data for different video formats. These libraries are copied and loaded into memory, allowing the vector processor to access pre-prepared decoding parameters for various formats without requiring complex hardware configuration for each format
3Adaptability or versatility
If scalable video coding is implemented, then flexibility for different device capabilities is improved, but processing requirements increase
Solution Approach 1:
The patent segments the video decoding process into distinct stages: entropy decoding, inverse quantization, inverse transform, and motion compensation. Each stage can be independently optimized with vector instructions, allowing efficient processing of scalable video streams where different quality levels are decoded separately
Solution Approach 2:
The patent implements pipelined processing where multiple decoding operations proceed simultaneously at different stages. The vector processor continuously executes decoding operations on different picture blocks and quality levels without idle time, maintaining continuous useful action to process scalable video efficiently
Data Source
AI summary
A video decoder includes an entropy decoding device that generates entropy decoded (EDC) data from an encoded video signal. A multi-format video decoding device includes a plurality of vector processor units that generate a decoded video signal from the EDC data. The plurality of vector processing units are programmed via VPU instructions formatted to include a vector instruction portion, a scalar instruction portion, and a branching instruction portion.