Filtering wavelet based video streams for wireless interworking

The dynamic variation in quality of service (QoS) is a crucial difference between wireless systems and communication in fixed high-speed network technology. Mobile multimedia systems must provide some components to cope with the variability inherent to wireless systems. Video filtering can be used to dynamically adjust video streams to match receivers´ processing and hardware capabilities and the QoS characteristics of wired and wireless links. The paper focuses on filter algorithms developed for a wavelet based video coder. We have implemented a number of different filter operations and evaluated the algorithms with respect to bandwidth scalability and processing time requirements. Processing time for filtering the streams are almost negligible so that filters are suitable for real time connections. Wireless inter-working can be achieved by cascading several filters, reducing bandwidth step-by-step. Additionally, a wireless proxy server houses filter services for wireless sub-networks. Finally, we developed a traffic estimation algorithm for WaveVideo streams which trades off link utilization and average re-negotiation interval. The model can be used for deriving traffic descriptors for admission control tests