CH-1015 Lausanne

Current video coding standards, and their reference im-plementations, are architected as large monolithic and se-quential algorithms, in spite of the considerable overlap of functionality between standards, and the fact that they are frequently implemented on highly parallel computing plat-forms. The former leads to unnecessary complexity in the standardization process, while the latter implies that imple-mentations have to be rebuilt from the ground up to reflect the parallel nature of the target. The upcoming Reconfigurable Video Coding (RVC) stan-dard currently developed at MPEG attempts to address these issues by building a framework that supports the con-struction of video standards as libraries of coding tools. These libraries can be incrementally updated and extended, and the tools in them can be aggregated to form complete codecs using a streaming (or dataflow) programming model, which preserves the inherent parallelism of the coding algo-rithm. This paper presents the RVC framework and its un-derlying dataflow programming model, along with the tool support and initial results