Characteristics of traffic and quality traces of wavelet encoded video

Wavelet-based encoding is now emerging as an efficient way to encode video for stream-ing over the Internet and for wireless applications. Wavelet-based video coding has been recently added to the JPEG-2000 video standards. However, due to the lack of long wavelet encoded video streams, most research has so far been based on short video traces. This thesis paper presents a public library of traces of long, as well as short, wavelet encoded videos. Currently, all traces in this library are for intra-coded video sequences. The library includes traces for over 10 one-hour movies, encoded with a wavelet-based codec, as well as several test sequences both in the Quarter Common Intermediate Format (QCIF) and the Common Intermediate Format (CIF) format. Both the frame sizes as well as the frame qualities (Peak Signal to Noise Ratio (PSNR) values) in the traces are included in this library. The statistical characteristics of the traces, including their long range dependence have also been studied. A hump like behavior of the coefficient of variation (CoVX) of frame sizes is observed for increasing video bit rates. A similar hump behavior is observed for the peak to mean ratio of the frame sizes as well. From the video trace analysis, it is clear that the video frame sizes can not be scaled for simulations, as the scaling of frame sizes would not change the variations observed at different bit rates