VR Live Streaming With Split Rendering for Precise Virtual Edges
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Virtual reality (VR) all-in-one machines suffer from limited rendering performance, resulting in biased-pixelized characters and background edges, leading to suboptimal video streaming experiences in virtual reality live streaming.
Innovation Solution
Utilize a high-performance workstation to render virtual picture data, encode it into a video stream, and seamlessly fuse it with local rendering at a VR device to enhance the visual experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If VR all-in-one machine is used to fuse and render virtual character picture and audience picture, then interaction between performer and audience is increased, but rendering performance is limited resulting in pixelized characters and sawtooth edges
Solution Approach 1:
The system divides rendering tasks into two segments: complex performer rendering is performed locally on the VR all-in-one machine, while audience rendering is performed remotely on a server. This segmentation allows each component to be optimized for its specific function, resolving the contradiction between interaction capability and rendering precision.
Solution Approach 2:
A server acts as an intermediary between the VR all-in-one machine and the audience. The server receives rendering tasks for audience members and returns rendered images, enabling the VR device to maintain high interaction capability while offloading complex rendering to a more powerful system.
2Adaptability or versatility
If VR all-in-one machine performs local rendering of audience video stream, then interaction field is achieved, but picture quality deteriorates due to insufficient rendering performance
Solution Approach 1:
The rendering workload is segmented between local and remote processing. The VR all-in-one machine handles performer rendering locally to maintain interaction field capability, while audience rendering is segmented and performed on a server to ensure high picture quality.
Solution Approach 2:
The system moves the audience rendering task from the local spatial dimension (VR all-in-one machine) to a remote dimension (server), effectively adding a network dimension to the system architecture. This allows the VR device to maintain interaction field capability while achieving high picture quality through remote rendering.
3Manufacturing precision
If rendering distribution is increased to improve picture quality, then device complexity increases beyond VR all-in-one machine capabilities
Solution Approach 1:
A server is introduced as an intermediary to handle complex rendering tasks. This externalizes the complexity from the VR all-in-one machine to a dedicated rendering system, allowing the VR device to maintain simplicity while achieving high picture quality through the intermediary server.
Solution Approach 2:
The server is designed to handle multiple functions: rendering audience members, managing video streams, and coordinating with the VR all-in-one machine. This multi-functionality consolidates complexity into a single universal system rather than requiring the VR device itself to handle all complex tasks.
Data Source
AI summary
Embodiments of the disclosure provide a method, an apparatus, an electronic device, and a storage medium for video live streaming. The video live streaming method comprises: determining first virtual picture data rendered by a first device end; obtaining a corresponding first device end video stream by encoding the first virtual picture data; pushing the first device end video stream to a second device end, to cause the first device end video stream, during the second device end performing video playback, being fused and displayed with second virtual picture data rendered by the second device end.