Indexing of resSamples in Section 8.5.8

[bheng]

In Section 8.5.8, the resSamples output is supposed to be a (nTbW)x(nTbH) array.

However, the SBT derivation produces samples ranging over x = xTb0Sub..xTb0Sub + nTbWSub − 1, y = yTb0Sub..yTb0Sub + nTbHSub − 1 in Equation (1069).

Here, the location ( xTb0Sub, yTb0Sub) is the top-left sample of the current SBT subblock relative to the top left luma sample of the current picture.

Since they are relative to the picture, not the coding block, in general xTb0Sub and yTb0Sub will be completely outside the resSamples array.

[jlchen]

Thanks for reporting the bug.

I will fix the issue by removing the bold elements in the following equations.

xTb0Sub = xTb0 + ( nTbW * xPartIdx / 4 ) (1065)
yTb0Sub = yTb0 + ( nTbH * yPartIdx / 4 ) (1066)

[bbross]

This will be fixed in JVET-S2001-vH as suggested by Jianle.