Look-up table for enhanced multiple transform

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

for a current coefficient block of a video block encoded according to one of a plurality of prediction modes, selecting a set of horizontal and vertical transform pair combinations from all available sets of horizontal and vertical transform pair combinations for a selected set of transforms, wherein the selected set of horizontal and vertical transform pair combinations comprises one or more horizontal and vertical transform pair combinations, wherein selecting the set of horizontal and vertical transform pair combinations comprises applying a mapping based on a height of the current coefficient block, a width of the current coefficient block, or both the height and the width of the current coefficient block;

selecting a horizontal and vertical transform pair combination from the set of horizontal and vertical transform pair combinations;

selecting a horizontal transform and a vertical transform from the selected horizontal and vertical transform pair combination;

applying an inverse transform using the selected horizontal transform and the selected vertical transform to the current coefficient block to determine a current transform block; and

reconstructing the video block based on the current transform block and a predictive block.

2. The method of claim 1 , wherein the selecting the set of transforms comprises selecting N transforms, wherein all the available horizontal and vertical transform pair combinations for the selected set of transforms comprises N 2 horizontal and vertical transform pair combinations, and wherein the all available horizontal and vertical transform pair combinations comprises fewer than N 2 horizontal and vertical transform pair combinations.

3. The method of claim 1 , wherein all the available sets of horizontal and vertical transform pair combinations consist of: {DST-7, DST-7}, {DST-4, DST-4}, {DST-4, DST-7}, {DST-7, DST-4}, {DCT-8, DST-7}, {DCT-8, DST-4}, {DST-7, DCT-5}, {DCT-5, DST-7}, {DST-7, DCT-8}, {DST-4, DCT-5}, {DST-1, DST-7}, {DST-1, DST-4}, {DST-1, DCT-5}, and {DCT-5, DCT-5} horizontal and vertical transform pair combinations,

wherein DST refers to a type of discrete sine transform, and DCT refers to a type of discrete cosine transform.

4. The method of claim 1 , wherein all the available sets of horizontal and vertical transform pair combinations consist of: {DST-7, DST-7}, {DST-4, DST-4}, {DST-4, DST-7}, {DST-7, DST-4}, {DCT-8, DST-7}, {DCT-8, DST-4}, {DST-7, DCT-5}, {DCT-5, DST-7}, {DST-7, DCT-8}, {DST-4, DCT-5}, {DST-1, DST-7}, {DST-1, DST-4}, {DST-1, DCT-5}, {DCT-5, DCT-5}, {DST-7, ID}, {DST-4, ID}, {DCT-8, ID}, (DCT-5, ID}, {DST-1, ID} and {ID, ID} horizontal and vertical transform pair combinations,

wherein DST refers to a type of discrete sine transform, DCT refers to a type of discrete cosine transform, and ID refers to an identity transform.

5. The method of claim 1 , wherein applying the mapping based on the height of the current coefficient block, the width of the current coefficient block, or both the height and the width of the current coefficient block comprises searching a two-stage mapping to the set of horizontal and vertical transform pair combinations.

6. The method of claim 5 , wherein the two-stage mapping comprises: a first mapping that maps block height, width and intra mode to a transform pair set index and a second mapping that maps the transform pair set index to the set of horizontal and vertical transform pair combinations.

7. The method of claim 1 , wherein selecting the horizontal and vertical transform pair combination from the selected set of horizontal and vertical transform pair combinations comprises:

determining an Enhanced Multiple Transform (EMT) index value; and

selecting the horizontal and vertical transform pair combination from the selected set of horizontal and vertical transform pair combinations based on the determined EMT index value.

8. The method of claim 1 , wherein selecting the set of horizontal and vertical transform pair combinations comprises selecting a set that includes an identity (ID) transform for a horizontal or vertical transform based on the one of the plurality of prediction modes comprising a horizontal or vertical intra prediction mode.

9. The method of claim 1 , wherein selecting the set of horizontal and vertical transform pair combinations comprises selecting a set that includes an identity (ID) transform for a horizontal or vertical transform based on the one of the plurality of prediction modes comprising an intra prediction mode within a threshold associated with a horizontal or vertical intra prediction mode.

10. The method of claim 9 , wherein the threshold is based on a size of the current coefficient block.

11. The method of claim 9 , wherein the threshold is based on a distance between an angle associated with the intra prediction mode and the horizontal or vertical intra prediction mode.

12. The method of claim 1 , wherein selecting the set of horizontal and vertical transform pair combinations comprises selecting the set of horizontal and vertical transform pair combinations of four horizontal and vertical transform pairs.

13. The method of claim 12 , wherein the one of the plurality of prediction modes comprises an angular intra prediction mode.

14. The method of claim 1 , wherein applying the mapping comprises determining values for a subset of prediction modes of the plurality of prediction modes and determining values for another subset of prediction modes of the plurality of prediction modes derived from the values for the subset of prediction modes.

15. A device for decoding video data, the device comprising:

a memory configured to store the video data; and

one or more processors configured to:

for a current coefficient block of a video block of the video data encoded according to one of a plurality of prediction modes, select a set of horizontal and vertical transform pair combinations from all available sets of horizontal and vertical transform pair combinations for a selected set of transforms, wherein the selected set of horizontal and vertical transform pair combinations comprises one or more horizontal and vertical transform pair combinations, wherein to select the set of horizontal and vertical transform pair combinations, the one or more processors are configured to apply a mapping based on a height of the current coefficient block, a width of the current coefficient block, or both the height and the width of the current coefficient block;

select a horizontal and vertical transform pair combination from the selected set of horizontal and vertical transform pair combinations;

select a horizontal transform and a vertical transform from the selected horizontal and vertical transform pair combination;

apply an inverse transform using the selected horizontal transform and the selected vertical transform to the current coefficient block to determine a current transform block; and

reconstruct the video block based on the current transform block and a predictive block.

16. The device of claim 15 , wherein the selecting the set of transforms comprises selecting N transforms, wherein all the available horizontal and vertical transform pair combinations for the selected set of transforms comprises N 2 horizontal and vertical transform pair combinations, and wherein the all available horizontal and vertical transform pair combinations comprises fewer than N 2 horizontal and vertical transform pair combinations.

17. The device of claim 15 , wherein all the available sets of horizontal and vertical transform pair combinations consist of: {DST-7, DST-7}, {DST-4, DST-4}, {DST-4, DST-7}, {DST-7, DST-4}, {DCT-8, DST-7}, {DCT-8, DST-4}, {DST-7, DCT-5}, {DCT-5, DST-7}, {DST-7, DCT-8}, {DST-4, DCT-5}, {DST-1, DST-7}, {DST-1, DST-4}, {DST-1, DCT-5}, and {DCT-5, DCT-5} horizontal and vertical transform pair combinations,

wherein DST refers to a type of discrete sine transform, and DCT refers to a type of discrete cosine transform.

18. The device of claim 15 , wherein all the available sets of horizontal and vertical transform pair combinations consist of: {DST-7, DST-7}, {DST-4, DST-4}, {DST-4, DST-7}, {DST-7, DST-4}, {DCT-8, DST-7}, {DCT-8, DST-4}, {DST-7, DCT-5}, {DCT-5, DST-7}, {DST-7, DCT-8}, {DST-4, DCT-5}, {DST-1, DST-7}, {DST-1, DST-4}, {DST-1, DCT-5}, {DCT-5, DCT-5}, {DST-7, ID}, {DST-4, ID}, {DCT-8, ID}, (DCT-5, ID}, {DST-1, ID} and {ID, ID} horizontal and vertical transform pair combinations,

wherein DST refers to a type of discrete sine transform, DCT refers to a type of discrete cosine transform, and ID refers to an identity transform.

19. The device of claim 15 , wherein selection of the set of horizontal and vertical transform pair combinations from all the available sets of horizontal and vertical transform pair combinations comprises search of a two-stage mapping to the set of horizontal and vertical transform pair combinations.

20. The device of claim 15 , wherein to select the horizontal and vertical transform pair combination from the selected set of horizontal and vertical transform pair combinations, the one or more processors are configured to:

determine an Enhanced Multiple Transform (EMT) index value; and

select the horizontal and vertical transform pair combination from the selected set of horizontal and vertical transform pair combinations based on the determined EMT index value.

21. The device of claim 15 , wherein to apply the mapping, the one or more processors are configured to determine values for a subset of prediction modes of the plurality of prediction modes and determine values for another subset of prediction modes of the plurality of prediction modes derived from the values for the subset of prediction modes.