Packet loss protection of scalable video bitstreams using forward error correction and feedback

We propose a system for packet loss protection of quality scalable video coders. The compressed bitstream is composed of a non-scalable base layer and an embedded enhancement layer as in MPEG4-fine granularity scalability (FGS) and the emerging H26L-FGS. But the system can also be used with embedded video bitstreams like the 3D-SPIHT bitstream. Our system allocates a given transmission bitrate between the base layer bitstream and the enhancement layer bitstream in an efficient way using feedback. The base layer is protected with equal packet loss protection. For the enhancement layer, unequal packet loss protection is used to provide graceful degradation of video quality in the presence of packet erasure. Preliminary results for a hypothetical operational distortion-rate function and the 3D-SP1HT video coder show promising performance of the proposed system.