Modelling power consumption of a H.263 video encoder

As video transmission is integrated into wireless communication systems, the theory of power control should also be expanded to consider both signal processing power and transmission power when designing new algorithms, since video coding consumes a significant portion of power. To better understand the interaction between signal processing and transmission, it helps to develop power consumption models for video coding. The goal of this work is to model the power consumption of an H.263 video encoder, in which motion estimation is the most computation intensive component. Different models for algorithms using 1) full search motion estimation; and 2) a fast algorithm using spiral order motion estimation are presented for a software H.263 encoder. We observe that one set of model parameters fits all test sequences for full search, whereas the model parameters are sequence specific for the fast algorithm.