Video Encoder FIFO Buffering for Bidirectional Motion Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding systems experience performance degradation due to processing stalls caused by interrupts for generating intraframe and interframe cost data, and inaccurate motion vector predictions due to unidirectional data computation, leading to reduced video quality and efficiency.
Innovation Solution
Implement a first-in-first-out (FIFO) memory to store intraframe and interframe data for multiple blocks, allowing simultaneous retrieval and processing without stalling, and generate both forward and backward motion vector data by accessing previous and next media frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional video encoders process interrupts to generate intraframe and interframe cost data, then compression efficiency is improved, but processing stalls occur and performance decreases
Solution Approach 1:
The patent pre-generates and stores motion vector data for all blocks in reference frames before the actual encoding process begins. This preliminary action allows the encoder to retrieve pre-computed motion vector data during encoding without stalling, thus maintaining both compression efficiency and encoder performance. The motion vector data is stored in a buffer accessible by the encoder, enabling continuous processing without interrupts.
2Device complexity
If motion vector data is computed in only one direction based on previous frames, then processing is simplified, but prediction accuracy decreases and video quality is reduced
Solution Approach 1:
The patent segments the motion vector computation into two independent directional components: forward prediction from previous frames and backward prediction from next frames. Each direction is processed separately and their results are combined. This segmentation allows the system to maintain simplicity in individual processing streams while achieving high prediction accuracy through the combination of bidirectional data, thus resolving the contradiction between processing complexity and prediction accuracy.
Data Source
AI summary
Various embodiments include techniques for generating coding unit tree statistical data for video blocks included in a media frame. The disclosed video encoder includes a first-in-first-out (FIFO) memory for storing encoding data for multiple video blocks. A first set of units within the video encoder generates encoding data for each video block. After storing the encoding data in FIFO memory, the first set of units can proceed with encoding additional blocks of the media frame without having to wait for a second set of units within the video encoder to retrieve and process the encoding data. Subsequently, the second set of units can efficiently retrieve and process data for multiple blocks of the media frame at a time. Further, the video encoder can generate both forward looking prediction data and backward looking prediction data to encode a current block of the media frame, resulting in improved video quality.