This paper discusses a methodology for developing a new video coding standard through collaboration; use of the reference model. It is a coding model which may start from a simple framework, is used as a yardstick to evaluate new coding elements and/or improvements of existing elements and allows evolution by incorporating these new or improved elements whose effectiveness has been demonstrated into the next generation model. Then this new generation model serves as an updated yardstick for the next round of study. This process is iterated until a sufficient level of performance is obtained. An outstanding feature of this method is to enable decision making based on purely technical characteristics The work to develop ITU-T Recommendation H.261 is detailed in this paper as evidence of its productiveness by describing techniques employed as well as some potential techniques which were not accepted in the end. This method has also been verified by the subsequent successful development of MPEG-1 and MPEG-2 standards in ISO/IEC JTC1. Furthermore, it is pointed out that the reference model method also contributes to the efficient dissemination of the state of the art of the video coding technology. As a conclusion, some guidelines to apply this methodology are listed.
展开▼
