Undefined HRD syntax elements and variables

[vdrugeon]

In Annex C, it seems that some syntax elements and variables are checked without being always defined.

In C.1 General step 9: "When the PT SEI message associated with AU 0 has pt_cpb_alt_timing_info_present_flag equal to 1, …". But pt_cpb_alt_timing_info_present_flag is not always signalled. So adding an inference for this syntax element may be necessary.

In C.2.2 and C.2.3, the variable DefaultInitCpbParamsFlag is checked without necessarily being defined, since it is defined in C.1 only when pt_cpb_alt_timing_info_present_flag is equal to 1. Either the value of pt_cpb_alt_timing_info_present_flag should be added as condition before checking the value of DefaultInitCpbParamsFlag or there should be a default value of DefaultInitCpbParamsFlag specified when pt_cpb_alt_timing_info_present_flag is equal to 0.

Is the second condition in the definition of the numbers n2 and n3 correct: "The associated BP SEI message has bp_alt_cpb_params_present_flag equal to 1"? Why is bp_alt_cpb_params_present_flag rather than pt_cpb_alt_timing_info_present_flag considered given that pt_cpb_alt_timing_info_present_flag seems to be the syntax element used everywhere else when using alternate timings?

In C.2.3, the PT SEI syntax elements cpb_alt_initial_removal_delay_delta, cpb_alt_initial_removal_offset_delta, cpb_delay_offset and dpb_delay_offset are mentioned. These syntax elements do not exist. The prefix "pt_nal" or "pt_vcl" is missing.

[ysanchez]

Good comments. Please see my answers inline.

Replying to vdrugeon:

In Annex C, it seems that some syntax elements and variables are checked without being always defined.

In C.1 General step 9: "When the PT SEI message associated with AU 0 has pt_cpb_alt_timing_info_present_flag equal to 1, …". But pt_cpb_alt_timing_info_present_flag is not always signalled. So adding an inference for this syntax element may be necessary.

[YS] This could be an option but as pointed out below, it would be preferable to use bp_alt_cpb_params_present_flag in the BP SEI message instead, as this is the indication whether there are alternative timings or not.

In C.2.2 and C.2.3, the variable DefaultInitCpbParamsFlag is checked without necessarily being defined, since it is defined in C.1 only when pt_cpb_alt_timing_info_present_flag is equal to 1. Either the value of pt_cpb_alt_timing_info_present_flag should be added as condition before checking the value of DefaultInitCpbParamsFlag or there should be a default value of DefaultInitCpbParamsFlag specified when pt_cpb_alt_timing_info_present_flag is equal to 0.

[YS] Good catch. DefaultInitCpbParamsFlag needs to be 1 when the alt_timing is not used.

Is the second condition in the definition of the numbers n2 and n3 correct: "The associated BP SEI message has bp_alt_cpb_params_present_flag equal to 1"? Why is bp_alt_cpb_params_present_flag rather than pt_cpb_alt_timing_info_present_flag considered given that pt_cpb_alt_timing_info_present_flag seems to be the syntax element used everywhere else when using alternate timings?

[YS] I think that using bp_alt_cpb_params_present_flag is the right thing to do here. As pointed out above, I think the that the condition n C.1 General step 9 should be modified to use bp_alt_cpb_params_present_flag as well instead of using pt_cpb_alt_timing_info_present_flag. If bp_alt_cpb_params_present_flag is equal to 0 DefaultInitCpbParamsFlag is set to 0 and if bp_alt_cpb_params_present_flag is one the steps under 9 are taken. Would this fix the issues above?

In C.2.3, the PT SEI syntax elements cpb_alt_initial_removal_delay_delta, cpb_alt_initial_removal_offset_delta, cpb_delay_offset and dpb_delay_offset are mentioned. These syntax elements do not exist. The prefix "pt_nal" or "pt_vcl" is missing.

[YS] Good catch.

[vdrugeon]

Thank you for your comments. I mostly agree, but I have a question about the following comment:

[YS] I think that using bp_alt_cpb_params_present_flag is the right thing to do here. As pointed out above, I think the that the condition n C.1 General step 9 should be modified to use bp_alt_cpb_params_present_flag as well instead of using pt_cpb_alt_timing_info_present_flag. If bp_alt_cpb_params_present_flag is equal to 0 DefaultInitCpbParamsFlag is set to 0 and if bp_alt_cpb_params_present_flag is one the steps under 9 are taken. Would this fix the issues above?

Is it a typo that if bp_alt_cpb_params_present_flag is equal to 0 DefaultInitCpbParamsFlag is set to 0? My understanding is that if alternative timing is not used, then DefaultInitCpbParamsFlag should be set to 1 (as you mentioned yourself above). Could you please explain in case it is not a typo?

[ysanchez]

Yes. Definitely a typo. When bp_alt_cpb_params_present_flag is equal to 0 DefaultInitCpbParamsFlag should be set to 1.

Replying to vdrugeon:

Thank you for your comments. I mostly agree, but I have a question about the following comment:

[YS] I think that using bp_alt_cpb_params_present_flag is the right thing to do here. As pointed out above, I think the that the condition n C.1 General step 9 should be modified to use bp_alt_cpb_params_present_flag as well instead of using pt_cpb_alt_timing_info_present_flag. If bp_alt_cpb_params_present_flag is equal to 0 DefaultInitCpbParamsFlag is set to 0 and if bp_alt_cpb_params_present_flag is one the steps under 9 are taken. Would this fix the issues above?

Is it a typo that if bp_alt_cpb_params_present_flag is equal to 0 DefaultInitCpbParamsFlag is set to 0? My understanding is that if alternative timing is not used, then DefaultInitCpbParamsFlag should be set to 1 (as you mentioned yourself above). Could you please explain in case it is not a typo?

[yk]

Thanks! These have been fixed in S2001-v6.