The constraint on RPL for dynamic subpicture IDs should be for pictures within a layer

[yk]

The constraint in JVET-R2001-v9 is as follows:

For each value of i in the range of 0 to sps_num_subpics_minus1, inclusive, when the value of SubpicIdVal[ i ] of a current picture is not equal to the value of SubpicIdVal[ i ] of a reference picture, the active entries of the RPLs of the coded slices in the i-th subpicture of the current picture shall not include that reference picture.

The adoption was from JVET-R0276, for relaxing the existing constraint requiring that a subpicture with a different subpicture ID compared to the collocated subpicture in the previous picture in the CLVS needing to have IRAP NUTs. That "existing" constraint was as follows in JVET-Q2001:

When the current picture is not the first picture of the CLVS, for each value of i in the range of 0 to sps_num_subpics_minus1, inclusive, if the value of SubpicIdVal[ i ] is not equal to the value of SubpicIdVal[ i ] of the previous picture in decoding order in the same layer, the nal_unit_type for all coded slice NAL units of the subpicture in the current picture with subpicture index i shall be equal to a particular value in the range of IDR_W_RADL to CRA_NUT, inclusive.

As can be seen, the intent of the adoption is still a constraint (but relaxed) for pictures within the same CLVS/layer. Therefore, the text should be changed to be as follows:

For each value of i in the range of 0 to sps_num_subpics_minus1, inclusive, when the value of SubpicIdVal[ i ] of a current picture is not equal to the value of SubpicIdVal[ i ] of a reference picture that is in the same layer as the current picture, the active entries of the RPLs of the coded slices in the i-th subpicture of the current picture shall not include that reference picture.

[Vadim]

In JVET-Q2001, same layer condition is redundant since the sub-picture layout is the same across layers. This behavior was changed with the adoption of JVET-R0058, where reference layer may have just one sub-picture.

I agree if both together then the same layer condition should be added.

[yk]

Thanks Vadim for confirming! The change will be made to JVET-R2001-vA/v10.