Cross-component sample offset (CCSO) with adaptive multi-tap-filter classifiers

A method of encoding video data, comprising:

• receiving video data comprising a current block of a current image frame;
• determining whether to apply a cross-component sample offset (CCSO) mode for the current block, the CCSO mode indicating whether a first sample offset of a first color sample of the current block is based on one or more luma samples;
• determining a filter type index and a filter shape index for the current block, the filter type index indicating one of a plurality of filter types of a loop filter selected for the current block and the filter shape index identifying a set of one or more neighboring locations used by the loop filter;
• when the filter type index selects a four-tap filter, concurrently selecting two of a plurality of distinct combinations of two neighboring locations as one or more selected neighboring luma samples based on the filter shape index;
• determining one or more difference values between the one or more selected neighboring luma samples and a first luma sample;
• quantizing the one or more difference values to generate one or more quantized difference values;
• classifying the first color sample based on the one or more quantized difference values to determine the first sample offset of the first color sample;
• encoding the current block at least by adjusting the first color sample based on the first sample offset of the first color sample;
• and
• signaling, in a video bitstream, a first syntax element indicating the CCSO mode, and a second syntax element indicating the filter type index and the filter shape index.

A non-transitory computer-readable storage medium storing a video bitstream that is generated by a video encoding method, the video bitstream comprising:

• coded information for a plurality of blocks, including a current block in a current frame;
• a first syntax element for a cross-component sample offset (CCSO) mode indicating whether a first sample offset of a first color sample of the current block is determined based on one or more luma samples;
• and
• a second syntax element for a filter type index and a filter shape index, wherein:
• the filter type index indicates a four-tap filter is selected for the current block, the four-tap filter being selected from a plurality of filter types for a loop filter;
• the filter shape index identifies a set of one or more neighboring locations used by the loop filter;
• and
• the four-tap filter indicates concurrent selection of two of a plurality of distinct combinations of two neighboring locations as one or more selected neighboring luma samples based on the filter shape index;
• wherein the coded information for the current block includes one or more quantized difference values for the current block, the one or more quantized difference values indicating a difference between the one or more selected neighboring luma samples and a first luma sample of the current block;
• and
• wherein the current block is encoded at least by adjusting the first color sample using the first sample offset for the first color sample, the first sample offset being based on a classification of the one or more quantized difference values.