JVET-T0070 AHG9/AHG8: Scalability dimension SEI message [Y.-K. Wang, L. Zhang, K. Zhang, Z. Deng (Bytedance)]
See section 6.2.
JVET-T0070 AHG9/AHG8: Scalability dimension SEI message [Y.-K. Wang, L. Zhang, K. Zhang, Z. Deng (Bytedance)]
This was discussed in session 21 at 2310 on Thursday 15 October 2020 (chaired by GJS & JRO).
scalability_dimension( payloadSize ) { | Descriptor |
sd_max_layers_minus1 | u(6) |
sd_multiview_info_flag | u(1) |
sd_auxilary_info_flag | u(1) |
if( sd_multiview_info_flag | | sd_auxilary_info_flag ) { | |
if( sd_multiview_info_flag ) | |
sd_view_id_len | u(4) |
for( i = 0; i <= vps_max_layers_minus1; i++ ) { | |
if( sd_multiview_info_flag ) | |
sd_view_id_val[ i ] | u(v) |
if( sd_auxilary_info_flag ) | |
sd_aux_id[ i ] | u(8) |
} | |
} | |
} |
This is proposed to be at the CVS level, having information for all the layers in a CVS.
Additional information would be needed for complete interpretation such as camera parameters, depth channel interpretation, alpha channel interpretation specification.
It was suggested to consider the potential use of a very large number of layers, saying that we should make sure to produce something relatively “future proof”.
It was commented that “vps_max_layers_minus1” should be “sd_max_layers_minus1”, although perhaps those should be required to be equal to each other.
It was commented that the SEI message approach seems like a good one for handling this sort of information.
Further study was encouraged toward development of a more complete approach.