||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 3DSPIHT
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-SPIHT video coder show promising performance
of the proposed system.