Rate-distortion optimized hybrid error control for real-time packetized video transmission

Area of Science:

• Computer Science
• Electrical Engineering
• Information Theory

Background:

• Real-time video transmission over packet-lossy networks faces challenges in maintaining quality.
• Traditional joint source-channel coding (JSCC) approaches often treat source and channel coding sequentially.
• Application-layer error control is crucial for mitigating packet loss effects.

Purpose of the Study:

• To propose and evaluate an integrated joint source-channel coding (IJSCC) framework for enhanced video transmission quality.
• To compare the performance of the IJSCC framework against sequential JSCC approaches.
• To analyze different hybrid error control strategies, including forward error correction (FEC) and retransmissions.

Main Methods:

• Developed an integrated joint source-channel coding (IJSCC) framework.
• Jointly optimized error resilient source coding, channel coding, and error concealment.
• Evaluated performance under various packet loss rates and round-trip times.
• Compared pure FEC, pure retransmission, and hybrid FEC-retransmission schemes.

Main Results:

• The proposed IJSCC framework demonstrates advantages over sequential JSCC by achieving better integration of coding components.
• Pure FEC and pure retransmissions show optimal performance under specific network conditions (packet loss rates, round-trip time).
• A hybrid approach combining FEC and retransmissions consistently outperforms individual methods, offering greater flexibility and improved video quality.

Conclusions:

• The IJSCC framework provides a more effective approach to application-layer error control for real-time video.
• Hybrid error correction strategies, leveraging both FEC and retransmissions, are essential for robust video delivery over lossy networks.
• The flexibility of hybrid schemes allows for adaptation to diverse network conditions, maximizing video quality.