Table 125 - Context initialization table for abs_level_gtx_flag in wrong order

[bheng]

I believe the context initialization table for abs_level_gtx_flag has the gt1 and gtx contexts reversed.

The table in the text follows VTM which adds the gtx contexts first, and the gt1 contexts second. However, the context derivation in Section 9.3.4.2.9 assumes the gt1 contexts are first and the gtx contexts are second.

Specifically, for I-slices, table entries [0-31] and [32-63] should be swapped. And, for transform skip, table entries [64-67] and [68-71] should be swapped.

Similarly, the same changes for P-slices and B-slices would be required as well.

[fbossen]

The suggested fix looks correct to me.

[bbross]

Good catch, thanks!
This will be fixed in JVET-S2001-vH by swapping:
[0-31] and [32-63]
[64-67] and [68-71]

[72-103] and [104-135]
[136-139] and [140-143]

[144-175] and [176-207]
[208-211] and [212-215]

Also, I fixed the issue that indices 128..143 appared twice in the index row while 144..159 are missing.