Abstract: Data-driven character animation is a hot research topic in computer graphics. It relies on data reusing of pre-captured motion data. Therefore, it is very important to compress the motion capture（mocap） data effectively and in real time for easy storing and transmitting. This paper introduces the latest research results of mocap compression from the view of eliminating data redundancy in temporal, spatial and compound segmental level. First, it gives an overview of the fundamental compression pipeline, and presents a guideline for classifying mocap data compression methods. Then, revisits and discusses the recently developed mocap data compression schemes from the view point of data redundancy reduction. Finally, gives some discussions on possible directions for future works in mocap data compression.