Dual Memory Module Segmentation for Video Decoder Bandwidth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video communication systems face inefficiencies in memory access bandwidth, particularly in motion compensation processing, where large memory access is required for sub-pixel displacements, leading to wasteful utilization of bandwidth and resources.
Innovation Solution
A dual-memory module system is implemented, where a first memory module with wider access handles non-reference video information and a second memory module with narrower access efficiently stores and retrieves reference video information, optimizing memory access by segregating data types and access widths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If larger memory modules and multiple memory accesses per read cycle are used to increase memory access bandwidth, then video decoder performance is improved, but the cost increases significantly
Solution Approach 1:
The patent divides the memory system into two separate memory modules: a first memory module for storing current frame data and a second memory module for storing reference frame data. This segmentation allows each memory module to be optimized independently, with the second memory module having narrower access width suited for reference frame retrieval patterns, thereby reducing overall system cost while maintaining decode performance
Solution Approach 2:
The patent applies different memory access widths to different memory modules based on their specific usage requirements. The first memory module uses wider access for current frame data, while the second memory module uses narrower access for reference frame data. This local optimization matches memory resources to actual access patterns, reducing unnecessary bandwidth consumption and cost
2Speed
If wider memory access is used for motion compensation processing, then data access speed is improved, but bandwidth is wasted when only smaller portions of data are needed
Solution Approach 1:
The patent optimizes memory access width locally for reference frame data by using a second memory module with narrower access width that matches the actual data retrieval requirements. This prevents wasting wide memory bus bandwidth when only small portions of reference frame data are needed, improving bandwidth utilization efficiency while maintaining adequate access speed
3Device complexity
If a single memory module is used for both current and reference video information, then device complexity is reduced, but memory access bandwidth efficiency decreases
Solution Approach 1:
The patent segments memory storage into two distinct memory modules: one for current frame data and another for reference frame data. This segmentation enables independent optimization of each memory module's access width according to its specific access patterns, improving overall memory access bandwidth efficiency despite the increased device complexity
Data Source
AI summary
A system and method for processing video information. Various aspects of the present invention may provide a decoder module that decodes block encoded video information. The system may, for example, include a first memory module, communicatively coupled to the decoder module, that stores video processing information utilized by the decoder module for decoding a current video block from a current video frame. The system may also, for example, include a second memory module, communicatively coupled to the decoder module, that stores reference video information from a previous video frame utilized by the decoder module for decoding the current video block. In a non-limiting exemplary scenario, the first memory module and the second memory module may be communicatively coupled to the decoder module with independent respective data and/or address buses.