Open-Loop Prediction in H.264/AVC for High Definition Sequences

Tiago da Fonseca, Ricardo de Queiroz, Yuxin Liu

DOI: 10.14209/sbrt.2007.31174

Evento: XXV Simpósio Brasileiro de Telecomunicações (SBrT2007)

Keywords: H.264/AVC mode decision high-definition parallelization

Abstract

H.264/AVC is the newest, state-of-the-art, video compression standard. It leads to substantial performance improvement compared to other existing standards. However, like other video standards, it is a hybrid predictive-transform coder. In this paper we propose a more efficient approach to implement the prediction stage. The idea is to employ original data rather than reconstructed ones to perform prediction tests before choosing the best mode. The residue, however, is evaluated using previously decoded data to avoid drifting. Results show that, for high definition sequences, the quality loss is negligible. We are, thus, able to parallelize the intra and inter-prediction parts, which are the most time consuming operations in H.264/AVC. “Inter” prediction generates a prediction macroblock from one or more previously encoded video frames using blockbased motion estimation and compensation. This model is responsible for almost 90% of the complexity of an H.264/AVC baseline encoder [7]. Important advances from earlier video standards include the support for a range of block sizes (16x16 and down, in Fig. 2) and refined motion vectors (quartersample resolution for the luminance component).

Download