Back to Search Document details
17th Meeting: Brussels, January 2020 2020-01-02 08:45
AHG9: On temporal sub-layer switching
Abstract
To guarantee the presence of the referenced PPS/APS NAL unit with TemporalId greater than 0 in the bitstream after temporal up-switching, the following constraint is proposed:
JVET-Q0282 AHG9: On temporal sublayer switching [B. Choi, S. Wenger, S. Liu (Tencent)] [late]

Discussed Track A Saturday 11 January at 0930 (chaired by GJS).

To guarantee the presence of the referenced PPS/APS NAL unit with TemporalId greater than 0 in the bitstream after temporal up-switching, the following constraint is proposed:

When a non-VCL NAL unit with nal_unit_type equal to PPS_NUT, PREFIX_APS_NUT, or SUFFIX_APS_NUT and TemporalId equal to k (k is greater than 0) referred to by a VCL NAL unit with TemporalId equal to k is included in an AU, the non-VCL NAL unit shall be included in an AU with TemporalId equal to k between the nearest STSA prior to the VCL NAL unit in decoding order and the VCL NAL unit referring to the non-VCL NAL unit inclusive, or included in an AU with TemporalId less than k prior to the VCL NAL unit in decoding order.

It was commented that JVET-P0359 of the previous meeting was similar and that there is an SEI message (the referenced parameter sets SEI message) that is sufficient to indicate that such constraints are applied.

It was further commented that there are quite a few other aspects of random access that are not guaranteed – e.g., PSs are not even required at IDR pictures of the base layer (unless indicated by an SEI message or required by some systems specification); such constraints may be imposed elsewhere. No action was thus taken on this.

Decisions
It was further commented that there are quite a few other aspects of random access that are not guaranteed – e.g., PSs are not even required at IDR pictures of the base layer (unless indicated by an SEI message or required by some systems specification); such constraints may be imposed elsewhere. No action was thus taken on this.
Citation