The concept of Quality Layers that has been introduced in Scalable Video Coding (SVC) amendment of MPEG4-AVC is presented. By using the Quality Layers post-processing to evaluate and signal the impact on rate and distortion of the various enhancement information pieces, a significant gain is achieved: Quality Layers significantly outperform the basic standard extractor that was initially proposed in SVC. For the standard set of test sequences, in a range of acceptable video quality, an average quality gain of up to 0.5 dB is achieved. Furthermore, the technique can be used for combined (spatial, temporal and quality) scalability. Thanks to the signaling of this information in the header of the Network Abstraction Layer units or in a Supplemental Enhancement Information message, the adaptation can be performed with a simple parser, e.g., at decoder side or in an intelligent network node designed for rate adaptation.
