In one embodiment, a method receives a plurality of sample values. The method then determines a filter to determine an up-sampled value for a first layer for a video, wherein the filter has a set of coefficient values of [−1, 3, −9, 47, 31, −10, 4, −1] or [−1, 4, −10, 31, 47, −9, 3, −1] assigned to the filter. The up-sampled value is determined by applying the set of coefficient values of [−1, 3, −9, 47, 31, −10, 4, −1] or [−1, 4, −10, 31, 47, −9, 3, −1] to the plurality of sample values. The method then outputs the up-sampled value for use in coding a second layer of a higher resolution than the first layer. The up-sampled values may be for the 6/16 and −6/16 phase offsets.