In some examples, a method of decoding video data may comprise reconstructing a block of video data comprising chroma samples, applying an adaptive loop filter to the chroma samples, and applying a cross-component adaptive loop filter to the chroma samples. Applying the cross-component adaptive loop filter may include determining an offset, and applying the offset to a particular chroma sample being filtered, wherein the offset is a function of a difference between a collocated luma sample that is collocated with the particular chroma sample being filtered and a plurality of neighboring luma samples that are spatial neighbors to the collocated luma sample.