Multi-Sink Wi-Fi Display Encoding with Shared Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Wi-Fi display (WFD) multi-sink topology, concurrent encoding sessions for multiple devices increase power consumption and latency due to redundant encoding processes with different resolutions and bitrates.
Innovation Solution
A multi-sink multi-session encoding system that uses hints from a compositor to optimize encoding processes, including still scene, layer geometry, dirty rectangle, and transform matrix hints, to skip unnecessary encoding stages and share motion estimation across sessions, reducing redundant processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple concurrent encoding sessions are performed for each WFD sink device, then each session can be optimized for specific resolution and bitrate requirements, but power consumption and latency are increased due to redundant encoding processes
Solution Approach 1:
The patent merges multiple encoding sessions into a single shared encoding session by having the compositor generate one encoded bitstream that is then scaled and decoded separately for each sink device. This eliminates redundant encoding operations while maintaining the ability to serve multiple sinks with different resolution and bitrate requirements.
Solution Approach 2:
The single encoding session produces a universal encoded bitstream that serves multiple sink devices with different requirements. The encoded data is then adapted through scaling and decoding operations to meet specific resolution and bitrate needs of each sink, making the encoding process multi-functional.
2Adaptability or versatility
If multiple concurrent encoding sessions are performed for each WFD sink device, then each session can be optimized for specific resolution and bitrate requirements, but latency is increased due to redundant encoding processes
Solution Approach 1:
The patent merges multiple encoding sessions into a single shared encoding session by having the compositor generate one encoded bitstream that is then scaled and decoded separately for each sink device. This eliminates redundant encoding operations while maintaining the ability to serve multiple sinks with different resolution and bitrate requirements.
Solution Approach 2:
The compositor performs preliminary encoding of the video content once, creating a master encoded bitstream before distribution to multiple sinks. This preliminary action eliminates the need for repeated encoding operations and reduces overall latency in the multi-sink scenario.
3Use of energy by moving object
If a single encoded bitstream is shared across multiple encoding sessions, then power consumption and latency are reduced, but the system must handle complex scaling and decoding operations for each sink
Solution Approach 1:
The patent extracts the scaling and decoding operations from the encoding process itself, separating these functions into distinct stages. The encoding session focuses solely on compression, while subsequent scaling and decoding operations handle adaptation to different sink requirements, simplifying the overall system architecture.
Solution Approach 2:
The patent segments the video processing pipeline into distinct functional stages: encoding, scaling, and decoding. Each stage handles specific tasks independently, allowing the system to share the encoding workload while distributing scaling and decoding operations to individual sink devices or dedicated processing units.
4Manufacturing precision
If concurrent encoding sessions perform full hierarchical motion estimation for each session, then each session achieves optimal encoding quality, but redundant processing increases power consumption and latency
Solution Approach 1:
The patent merges the motion estimation process into a single shared operation performed during the common encoding session. The motion vectors and estimation results are then reused across multiple sink adaptations through scaling operations, eliminating redundant motion estimation computations while maintaining encoding quality.
Solution Approach 2:
The patent discards redundant motion estimation operations for each individual sink and recovers the necessary motion information through scaling transformations applied to the master encoded bitstream. This approach maintains encoding quality by preserving motion data while avoiding repetitive computational overhead.
Data Source
AI summary
Systems, apparatuses, and methods may provide for multi-session encoding to optimize multiple encoding sessions on Wi-Fi display (WFD) source devices when the WFD source devices are connected to multiple sink devices. The multiple encoding sessions may be optimized with encoding hints that are generated by a compositor and transmitted to a pre-encoding checking device. The encoding session that has the highest encoding resolution is subjected to hierarchical motion estimation (HME) processing, and the encoding sessions that have lower resolutions are optimized based on a motion vector prediction hint generated by the encoding session that has the highest encoding resolution and a scaling factor.