JVET-AF0096 [AHG9] On miscellaneous aspects of NNPFGC and NNPFGA [Hendry, J. Nam, S. Kim, J. Lim (LGE)]

This contribution proposes several modifications that are asserted to provide improvement to some aspects of the NNPFGC and NNPFGA SEI messages design.

Item 1 relates to complexity: Complexity for alternative to be signalled should be maximum complexity.

It was commented that the complexity information can also be determined from the individual complexities of the filters in the group, and a decoder that is not capable of processing the maximum complexity might still be capable to process an alternative configuration. From that point of view, signalling minimum complexity (or both min and max of the group) might be better. Further study appears necessary which complexity is useful to be signalled for a group for best benefit of a decoder. Such a study on the needs of signalling group complexity would also be useful for SPO, where it is not signalled. There, the consideration of minimum complexity would apply to cascaded configuration where it would only include the essential filters.

Item 2 relates to input pictures information signalling for first NNPF in the group. It was commented that this signalling was introduced by purpose in cases where different input pictures are candidates that could be selected. No action was considered necessary on this.

AHG9: SEI processing order SEI message aspects other than PPF grouping (5)

Contributions in this area were discussed at 0835–1235 on Sunday 15 Oct. 2023 (chaired by JRO).

Remaining aspects of JVET-AF0174 were also presented on Sunday Oct. 15 (see in section 6.2).