Multi-type-tree framework for transform in video coding

1. A method of decoding video data, the method comprising:

receiving an encoded video bitstream that forms a representation of a coded picture of the video data;

determining a partitioning of the coded picture of the video data into a plurality of coded units, the partitioning being according to a first tree structure and the plurality of coded units including a leaf node of the first tree structure;

determining that a residual block of the leaf node is recursively split into a plurality of transform units according to a second tree structure comprising at least one of a binary tree, an asymmetric binary tree, or a triple tree;

decoding, from the encoded video bitstream, a first syntax element for the leaf node indicating that only a single transform unit partition structure is allowed in the second tree structure for the residual block of the leaf node, the single transform structure being one of the binary tree, the asymmetric binary tree, or the triple tree;

decoding, from the encoded video bitstream, a second syntax element for a subset of the plurality of transform units indicative of the structure into which the corresponding group of the transform unit is split;

determining the second tree structure for the residual block based on the first and second syntax elements; and

reconstructing the coded picture based on the determined first and second tree structures.

2. The method of claim 1 , further comprising decoding, from the encoded video bitstream, a syntax element for each of the plurality of transform units indicative of a portion of the second tree structure into which the corresponding transform unit is split.

3. The method of claim 1 , further comprising decoding, from the encoded video bitstream, a syntax element associated with a root of a transform unit tree indicating that the plurality of transform units in the tree are transformed according to a specified transform.

4. The method of claim 3 , wherein the syntax element indicates that the plurality of TUs share parameters associated with one of an enhanced multiple transform (EMT), non-separable secondary transform (NSST) or a transform skip (TS).

5. The method of claim 1 , further comprising decoding, from the encoded video bitstream, a syntax element associated with a root of a transform unit tree indicating at least one of a coded block flag or last position of a non-zero coefficient for the transform unit tree.

6. The method of claim 1 , wherein the first tree structure is different than the second tree structure.

7. A method of encoding video data, the method comprising:

obtaining a picture for encoding into an encoded video bitstream;

determining a partitioning of the picture of the video data into a plurality of coded units, the partitioning being according to a first tree structure and the plurality of coded units including a leaf node of the first tree structure;

determining that a residual block of the leaf node is recursively split into a plurality of transform units according to a second tree structure comprising at least one of a binary tree, an asymmetric binary tree, or a triple tree;

encoding a second syntax element into the encoded video bitstream indicating for the leaf node that only a single transform unit partition structure is allowed in the second tree for the residual block of the leaf node, the single transform structure being one of the binary tree, the asymmetric binary tree, or the triple tree;

encoding a second syntax element into the encoded video bitstream for a subset of the plurality of transform units indicative of the structure into which the corresponding group of the transform unit is split; and

encoding, into the encoded video bitstream, the picture based on the determined first and second tree structures.

8. The method of claim 7 , further comprising encoding a syntax element into the encoded video bitstream for each of the plurality of transform units indicative of a portion of the second tree structure into which the corresponding transform unit is split.

9. The method of claim 7 , further comprising encoding a syntax element into the encoded video bitstream associated with a root of a transform unit tree indicating that the plurality of transform units in the tree are transformed according to a specified transform.

10. The method of claim 9 , wherein the syntax element indicates that the plurality of TUs share parameters associated with one of an enhanced multiple transform (EMT), non-separable secondary transform (NSST) or a transform skip (TS).

11. The method of claim 7 , further comprising encoding a syntax element into the encoded video bitstream associated with a root of a transform unit tree indicating at least one of a coded block flag or last position of a non-zero coefficient for the transform unit tree.

12. An apparatus for decoding video data, the apparatus comprising:

a memory configured to store a coded picture of the video data; and

a video processor configured to:

receive an encoded video bitstream that forms a representation of the coded picture of the video data;

determine a partitioning of the coded picture of the video data into a plurality of coded units, the partitioning being according to a first tree structure and the plurality of coded units including a leaf node of the first tree structure;

determine that a residual block of the leaf node is recursively split into a plurality of transform units according to a second tree structure that comprises at least one of a binary tree, an asymmetric binary tree, or a triple tree;

decode, from the encoded video bitstream, a first syntax element indicating for the leaf node that only a single transform unit partition structure is allowed in the tree for the residual block of the leaf node, the single transform structure being one of the binary tree, the asymmetric binary tree, or the triple tree;

decode, from the encoded video bitstream, a second syntax element for a subset of the plurality of transform units indicative of the structure into which the corresponding group of the transform unit is split;

determine the second tree structure for the residual block based on the first and second syntax elements; and

reconstruct the coded picture based on the determined first and second tree structures.

13. The apparatus of claim 12 , further comprising a display configured to display the coded picture.