Typo in section 6.4.1

[vdrugeon]

One of the conditions to derive allowBtSplit in 6.4.1 is:
– Otherwise if all of the following conditions are true, allowBtSplit is set equal to FALSE
– btSplit is equal to SPLIT_BT_VER
– x0 + cbWidth is greater than pic_width_in_luma_samples
– y0 + cbHeight is greater than pic_width_in_luma_samples

In the last comparison, y0 + cbHeight is compared to pic_width_in_luma_samples instead of pic_height_in_luma_samples. Is that a typo?

[bbross]

Yes, that's a typo and it will be fixed in v4.

[vdrugeon]

Here is a list of other small typos I found:

1. In the syntax of multi_type_tree, when doing the splitting for SPLIT_BT_HOR, y1 is calculated as y1 = y0 + ( cbHeight / 4 ). For binary splitting, it should be y1 = y0 + ( cbHeight / 2 ).

2. In coding_unit syntax:

if( intra_luma_mpm_flag[ x0 ][ y0 ] )

intra_luma_mpm_idx[ x0 ][ y0 ]

Else

intra_luma_mpm_remainder[ x0 ][ y0 ]

The "Else" should probably be written "else" without the capital letter because that is the way it is written in the rest of the document.

3. Section 7.4.3.1 in the semantics of sps_cclm_enabled_flag, there are two typos where "chroma componenent" should be "component".

4. Section 7.4.5.5 in the semantics of pred_mode_flag, there are two "is" in "the variable CuPredMode[ x ][ y ] is is inferred".

5. Section 7.4.5.5 in the semantics of intra_luma_mpm_flag, there is a "," in "neighbouring intr,a-predicted prediction unit according to clause 8.2.2"

[vdrugeon]

Another small issue in the intra prediction specification:
In section 8.2.2 in the process to derive the luma intra prediction mode, step 3 should actually be part of step 2, because it is still deriving candIntraPredModeX with X being replaced by A or B as stated at the beginning of step 2.

[vdrugeon]

Here is a list of other small typos or potential issues that I found in section 8.2.4.1:

In section 8.2.4.1, "– If nTbW is greater than maxTbSize or nTbH is greater tha maxTbSize" the second "than" is missing an "n".

In section 8.2.4.1 after "– Otherwise, the following ordered steps apply", in step 2 the luma location ( xTbY, yTbY ) is used as input. How are xTbY and yTbY defined? Should it be (xTb0, yTb0) instead?

Furthermore, in the following step 3, there is no output for the process. Should it not be the final output for the whole 8.2.4.1 process? In that case, I think that it would be helpful to add to step 3 "and the output is a modified reconstructed picture before in-loop filtering" as done when calling many of the processes.

[bbross]

Fixed in VVC D2 v4