Signaling picture size in video coding

A method of decoding video data comprising:

• obtaining, from a video bitstream for a coded video sequence, the coded video sequence including a plurality of coded pictures, each of the plurality of coded pictures including a respective plurality of largest coding units (LCUs) including a respective plurality of coding units, each of the coding units corresponding to a respective leaf node of a respective quadtree data structure, and each of the coding units including syntax elements defining prediction data and residual data for a plurality of samples of the corresponding coding unit, each of the coding units having one of a plurality of different coding unit sizes between 4×4 pixels and 64×64 pixels;
• decoding, from the video bitstream, a sequence parameter set (SPS), comprising:
• decoding a logarithmically encoded value for a first syntax element of the SPS indicating a smallest coding unit size of the plurality of different coding unit sizes used to code the plurality of coded pictures of the coded video sequence, the smallest coding unit size being equal to or larger than 4×4 pixels and smaller than a largest coding unit (LCU) size;
• decoding a logarithmically and differentially encoded value for a second syntax element of the SPS representing the LCU size used to code the plurality of coded pictures of the coded video sequence;
• decoding a value of a third syntax element of the SPS representing a first integer multiple of the smallest coding unit size represented by the value for the first syntax element of the SPS that specifies a picture width, in luma samples, of the plurality of pictures of the video sequence;
• and
• decoding a value of a fourth syntax element of the SPS representing a second integer multiple of the smallest coding unit size represented by the value for the first syntax element of the SPS that indicates a picture height of the plurality of pictures of the video sequence;
• decoding the respective quadtree data structures, decoding the respective quadtree data structures comprising decoding split flags of the quadtree data structures at one or more of a plurality of nodes of the quadtree data structures, the split flags representing whether the coding units corresponding to the nodes of the quadtree data structures are partitioned into smaller coding units and determining how the quadtree data structures are partitioned based on both the split flags and the decoded value of the first syntax element of the SPS indicating the smallest coding unit size;
• decoding the pictures based on the decoded value of the first syntax element representing the smallest coding unit size, the largest coding unit size determined based on the decoded value of the second syntax element, the decoded value of the third syntax element representing the picture width, and the decoded value of the fourth syntax element representing the picture height, decoding the pictures comprising decoding the coding units of the plurality of pictures, decoding the coding units comprising decoding, for each of the coding units, the syntax elements defining the prediction data and the residual data for the plurality of samples of the corresponding coding units;
• decoding cropping window data indicating an amount by which to crop the plurality of pictures of the video sequence to achieve a display size for the plurality of pictures, the display size comprising a display width and a display height, the display width being smaller than the picture width, and the display height being smaller than the picture height;
• cropping the decoded pictures according to the cropping window data to form cropped decoded pictures having the display size;
• and
• outputting the cropped decoded pictures to a display.

A device for decoding video data, the device comprising:

• a decoded picture buffer configured to store video data;
• and
• one or more processors implemented in circuitry and configured to:
• obtain, from a video bitstream for a coded video sequence of the video data, the coded video sequence including a plurality of coded pictures, each of the plurality of coded pictures including a respective plurality of largest coding units (LCUs) including a respective plurality of coding units, each of the coding units corresponding to a respective leaf node of a respective quadtree data structure, and each of the coding units including syntax elements defining prediction data and residual data for a plurality of samples of the corresponding coding unit, each of the coding units having one of a plurality of different coding unit sizes between 4×4 pixels and 64×64 pixels;
• decode, from the video bitstream, a sequence parameter set (SPS), the one or more processors being configured to:
• decode a logarithmically encoded value for a first syntax element of the SPS indicating a smallest coding unit size of the plurality of different coding unit sizes used to code the plurality of coded pictures of the coded video sequence, the smallest coding unit size being equal to or larger than 4×4 pixels and smaller than a largest coding unit (LCU) size;
• decode a logarithmically and differentially encoded value for a second syntax element of the SPS representing the LCU size used to code the plurality of coded pictures of the coded video sequence;
• decode a value of a third syntax element of the SPS representing a first integer multiple of the smallest coding unit size represented by the value for the first syntax element of the SPS that specifies a picture width, in luma samples, of the plurality of pictures of the video sequence;
• and
• decode a value of a fourth syntax element of the SPS representing a second integer multiple of the smallest coding unit size represented by the value for the first syntax element of the SPS that indicates a picture height of the plurality of pictures of the video sequence;
• decode the respective quadtree data structures, decoding the respective quadtree data structures comprising decoding split flags of the quadtree data structures at one or more of a plurality of nodes of the quadtree data structures, the split flags representing whether the coding units corresponding to the nodes of the quadtree data structures are partitioned into smaller coding units and determining how the quadtree data structures are partitioned based on both the split flags and the decoded value of the first syntax element of the SPS indicating the smallest coding unit size;
• decode the pictures based on the decoded value of the first syntax element representing the smallest coding unit size, the largest coding unit size determined based on the decoded value of the second syntax element, the decoded value of the third syntax element representing the picture width, and the decoded value of the fourth syntax element representing the picture height, decoding the pictures comprising decoding the coding units of the plurality of pictures, decoding the coding units comprising decoding, for each of the coding units, the syntax elements defining the prediction data and the residual data for the plurality of samples of the corresponding coding units;
• decode cropping window data indicating an amount by which to crop the plurality of pictures of the video sequence to achieve a display size for the plurality of pictures, the display size comprising a display width and a display height, the display width being smaller than the picture width, and the display height being smaller than the picture height;
• crop the decoded pictures according to the cropping window data to form cropped decoded pictures having the display size;
• and
• output the cropped decoded pictures to a display.