These functions use two reference blocks for prediction. At first, predictions are calculated for each reference block, and then prediction is calculated as average of two predictions.
Copyright © 2000 - 2011, Intel Corporation. All rights reserved.