receiving prediction information of a current block in a current picture from a coded video bitstream, the prediction information indicating that a filter with a central symmetric shape is applied to the current block, the filter including a plurality of coefficients, each of the plurality of coefficients being associated with a respective first coordinate value along a first axis and a respective second coordinate value along a second axis;

grouping the plurality of coefficients of the filter into a plurality of groups based on a transposition operation to be performed on the filter, the coefficients in each of the plurality of groups being a same distance from a center position of the filter and stored in a respective same chunk of contiguous memory;

reordering the coefficients of a first group of the plurality of groups based on the transposition operation; and

reconstructing filtered samples of the current block by applying the filter with the central symmetric shape to the current block based on the reordered coefficients of the filter.