Video Decoder Parallel Processing JPEG2000 Bottleneck
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current JPEG2000 decoding methods are computationally intensive, making real-time processing challenging, especially for software decoders, and often require costly hardware or result in reduced picture quality by skipping data stream decoding.
Innovation Solution
A video decoder that decodes transformation coefficient blocks of different pictures on different computing kernels of a SIMD group simultaneously, exploiting parallelism to reduce decoding time and increase efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If JPEG2000 decoding is performed using conventional sequential methods, then decoding accuracy is maintained, but decoding speed is too slow to achieve real-time processing
Solution Approach 1:
The patent divides the picture sequence into multiple independent picture units that can be decoded separately. Each picture is further divided into transformation coefficient blocks that can be processed independently on different computing kernels, enabling parallel decoding without compromising the integrity of the decoding process
Solution Approach 2:
The patent introduces temporal parallelism by decoding multiple pictures from different time points simultaneously across different computing kernels. This adds a temporal dimension to the traditionally sequential decoding process, allowing multiple decoding operations to occur concurrently while maintaining causal dependencies within each individual picture
2Productivity
If costly hardware is used to achieve real-time JPEG2000 decoding, then decoding speed is improved, but device complexity and cost increase
Solution Approach 1:
The patent designs a decoding system where a single computing device can handle multiple pictures simultaneously through parallel processing on different computing kernels. This multi-functional approach allows one device to perform what would traditionally require multiple dedicated hardware decoders, reducing overall system complexity
Solution Approach 2:
The patent uses computing kernels that can be instantiated multiple times across different processing units. These kernels represent copied functional units that execute the same decoding algorithm in parallel, enabling real-time performance through software-based parallelization rather than complex custom hardware
3Speed
If picture quality is reduced by skipping data stream decoding to increase speed, then decoding speed is improved, but picture quality deteriorates
Solution Approach 1:
The patent performs complete decoding of all transformation coefficient blocks for each picture, rather than skipping parts of the data stream. By fully decoding multiple pictures in parallel, the system achieves both high speed and high quality, avoiding the trade-off present in selective decoding approaches
Data Source
AI summary
A video decoder for decoding a sequence of pictures, each of which is coded into a plurality of transformation coefficient blocks, is configured to decode transformation coefficient blocks of different pictures on different computing kernels of a first SIMD group at the same time.