Reference-Order AL-FEC Encoding for Video Packet Loss Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional AL-FEC methods, such as Block AL-FEC and time-order AL-FEC, face limitations in error correction performance, particularly in Scalable Video Coding (SVC) scenarios due to inadequate consideration of video encoding dependencies between frames, leading to inferior performance when Packet Loss Rate is high.
Innovation Solution
The reference-order AL-FEC system addresses these limitations by using a packetizer, encoder, and decoder that construct and recover source symbols based on a reference order, forming encoding windows that include reference frames, allowing for effective FEC encoding and decoding without additional delay, even in SVC scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Block AL-FEC is used for error correction, then error correction capability is improved, but decoding delay increases when multi-frame FEC coding is applied
Solution Approach 1:
The patent implements a dynamic AL-FEC encoding window that adapts its size and composition based on real-time network conditions and video frame types. The encoding window dynamically adjusts to include only necessary reference frames, avoiding unnecessary processing delay while maintaining error correction capability. This dynamic adaptation resolves the contradiction between robust error correction and minimal decoding delay.
Solution Approach 2:
The patent segments the video stream into independent AL-FEC encoding windows, each processing a specific set of frames with their reference dependencies. By segmenting the processing into manageable units with clear boundaries, the system avoids global multi-frame processing delays while maintaining error correction within each segment. This segmentation allows parallel processing and reduces overall decoding delay.
2Loss of time
If time-order AL-FEC is used, then decoding delay is reduced, but error correction performance deteriorates in SVC scenarios due to inadequate consideration of video encoding dependencies
Solution Approach 1:
The patent applies local quality by treating different video frames and reference relationships differently within the AL-FEC encoding window. Instead of uniform time-order processing, the system identifies and prioritizes frames based on their specific reference dependencies and importance to the current frame. This localized adaptation to frame-specific requirements improves error correction performance in SVC scenarios without adding significant delay.
Solution Approach 2:
The patent performs preliminary analysis of frame reference dependencies before constructing the AL-FEC encoding window. By pre-identifying which frames are critical references and establishing their priority order, the system prepares the optimal encoding window in advance. This preliminary action ensures that error correction prioritizes the most important frames while maintaining efficient processing with minimal delay.
3Reliability
If retransmission is used to improve transmission robustness, then reliability is improved, but it is not suitable for single-direction channels and channels with large round-trip-time
Solution Approach 1:
The patent implements beforehand cushioning by proactively adding AL-FEC redundant packets to the video stream before transmission. These redundant packets are embedded in advance to compensate for potential packet loss, eliminating the need for reactive retransmission. This approach provides transmission robustness suitable for single-direction channels and channels with large round-trip-times, as the error correction capability is prepared beforehand without requiring feedback channels.
Data Source
AI summary
A reference-order AL-FEC system for recovering network video data packet loss during real-time video communication includes a packetizer, a reference-order AL-FEC encoder, a reference-order AL-FEC decoder and a depacketizer. The packetizer constructs source symbols from source packets of a current frame. The encoder generates a repair symbol from the source symbols of the current frame and other reference frames based on the reference-order, not time-order, between the frames within an encoding window. The encoder also generates a repair packet based on the repair symbol. The decoder recovers a lost source symbol based on the source symbols of the frames of the encoding window and the repair symbol by decoding the repair packet. The decoding is achieved by solving a linear system of the repair symbol.