Content-Sensitive FEC for Low-Latency Media Transport
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing forward error correction (FEC) techniques struggle with variable packet loss rates and bit rates in packet-switched networks, leading to inefficiencies in bandwidth overhead and latency, particularly in variable bit rate (VBR) video streams, where the fixed percentage of checksum packets results in excessive latency and jitter due to unpredictable data block processing times.
Innovation Solution
Implementing a content-sensitive FEC mechanism that bounds the size of FEC blocks to a fixed number of audio and video frames, allowing for variable allocation of checksum packets based on content type, and using Galois Field FEC to enable recovery of any subset of data packets within a block, thereby eliminating the need for interleaving and reducing latency and jitter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed percentage of checksum packets is added to FEC blocks, then error correction capability is provided, but latency and jitter increase due to unpredictable data block processing times in VBR streams
Solution Approach 1:
The patent implements dynamic FEC block sizing that adapts to variable bit rate streams by bounding block size to a fixed number of audio/video frames rather than a fixed number of packets. This dynamic approach allows the system to maintain low latency by limiting the maximum time span of each FEC block while preserving error correction capability through adaptive checksum packet allocation based on frame type and content sensitivity.
Solution Approach 2:
The system changes the parameter of FEC block definition from packet-count-based to frame-count-based, and implements variable checksum packet allocation based on content sensitivity. Critical frames like I-frames receive more checksum packets while less critical frames receive fewer, optimizing both error correction reliability and latency performance across VBR streams.
2Reliability
If more checksum packets are added to handle high packet loss rates, then packet recovery capability improves, but bandwidth overhead increases
Solution Approach 1:
The patent applies local quality by allocating checksum packets differentially based on content sensitivity. Different frame types (I-frames, P-frames, B-frames) and different network conditions receive different levels of FEC protection. This allows the system to provide strong error correction for critical content while reducing overhead for less critical content, optimizing the trade-off between packet recovery capability and bandwidth overhead.
3Reliability
If FEC blocks are sized to accommodate VBR bursts, then error correction coverage is improved, but processing latency increases
Solution Approach 1:
The system dynamically bounds FEC block size by a fixed number of frames (e.g., 30 video frames or 1 second of audio) rather than allowing blocks to grow with VBR bursts. This dynamic constraint ensures that even during high-bitrate bursts, the FEC processing latency remains bounded and predictable, while still providing adequate error correction coverage for the framed content.
4Reliability
If interleaving is used to handle burst packet loss, then error recovery robustness improves, but system complexity and latency increase
Solution Approach 1:
The patent extracts and removes the interleaving step from the FEC processing pipeline. Instead of interleaving packets before FEC encoding, the system applies FEC directly to frame-based blocks and uses Galois Field arithmetic to provide robust error recovery without the complexity and latency of interleaving operations.
Data Source
AI summary
A forward error correction (FEC) encoding system and method optimized for protecting real-time audio-video streams for transmission over packet-switched networks with minimal latency. Embodiments of this invention provide bandwidth-efficient and low-latency FEC for both variable and constant bit-rate MPEG-encoded audio and video streams. To maximize bandwidth-efficiency and playable frame rate for recovered media streams, embodiments of the invention may sort packets by content type and aggregate them into FEC blocks weighted by sensitivity in the recovered stream to packet loss of a particular content type. Embodiments of this invention may use temporal constraints to limit FEC block size and thereby facilitate their use in the transport of VBR streams.