GPU Texture Storage for Video Frame Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video playing methods degrade GPU rendering speed due to the need for the Graphic Processing Unit (GPU) to communicate with Random-Access Memory (RAM) for decoded video frames, leading to inefficient video playing performance.
Innovation Solution
Storing decoded video frames as texture data in a texture storage area of the GPU, allowing the GPU to directly access and render frames without needing to communicate with RAM, and implementing jump instructions to manage video playback efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If decoded video frames are stored in RAM, then the video playing system can maintain a simple memory architecture, but the GPU needs to communicate with RAM which consumes GPU resources and degrades rendering speed
Solution Approach 1:
The patent merges the texture storage function into the GPU's internal memory system by storing decoded video frames as texture data in the GPU's texture storage area. This integration eliminates the need for separate RAM storage and communication between GPU and RAM, thereby improving rendering speed while maintaining architectural simplicity.
2Ease of operation
If decoded video frames are stored in RAM, then video frames can be easily accessed, but the communication process consumes GPU resources and influences rendering speed
Solution Approach 1:
The patent introduces texture data as an intermediary form between decoded video frames and rendered video frames. By converting decoded frames into texture data stored in the GPU's texture storage area, the system enables direct GPU access without RAM communication, thus maintaining ease of access while eliminating resource consumption.
3Quantity of substance
If video frames are stored in RAM, then the system can handle video playback, but electronic devices with low memory experience degraded performance
Solution Approach 1:
The patent transitions the storage dimension from system RAM to GPU internal memory. By storing decoded video frames as texture data within the GPU's own memory architecture, the system bypasses the limitations of system RAM capacity, enabling devices with low memory to achieve improved video playing performance.
Data Source
AI summary
A video playing method, apparatus and electronic device. The method comprises: obtaining to-be-played video frames; inputting the obtained to-be-played video frames into a preset decoder, and decoding the to-be-played video frames by the preset decoder to obtain decoded to-be-played video frames; storing the decoded to-be-played video frames as texture data in a texture storage area of a graphic processing unit; determining, from the decoded to-be-played video frames stored in the texture storage area, a decoded to-be-played video frame that satisfies a rendering condition as a to-be-rendered video frame; rendering and the determined to-be-rendered video frames by the graphic processing unit and playing the rendered video frame. With the embodiments of the present application, rendering speed of a graphic processing unit (GPU) can be increased to a certain extent. Therefore, video playing performance of the electronic device can be improved.