Scalable live video in Max/MSP/Jitter

This thesis describes the mcl.jit software library we developed to support scalable live video coding and transmission in Max/MSP/Jitter. Video codecs from this library have been successfully used in several telematic dance performances created by dancers and media artists from the School for the Contemporary Arts at Simon Fraser University. The mcl.jit library includes Region-Of-Interest (ROI) coding and motion detection objects, which can be used in a variety of interactive multimedia applications besides distributed dance performance. We also developed a combined bit rate and frame rate control method for live video for the mcl.jit library. This method differs from previously developed frame rate control approaches in that it does not assume that video is pre-recorded before frame rate adjustment. The proposed method was compared to another state-of-the-art method through an extensive subjective evaluation study, the results of which indicate the superiority of the proposed approach.