International organisation for standardisation organisation internationale de normalisation

Summary of discussions

1. m27963 Defect report and its correction about Audio Sample Entry on ISO/IEC 14496-12

Editors to integrate 28160/27963 handling into the amendment in process. Can we amend the amendment name?

2. m28043 DECE Liaison 14496-12 request for clarification

We agree that it should be documented, as an offset into the DATA of the idat box. We may need a note, as this is different from mdat, sidx, base_offset_is_moof.

3. m28160 Improved audio support in the ISO base media file format

Use this text, or something like it, in the amendment (see above) for high-sample-rate audio. The rest was reviewed in audio (see audio minutes for details). Into a TuC for Part 12, for study.

4. Other

Comment from one vendor that the duration of the initial movie of a fragmented file is not clearly documented; they wrote the total length there, but our consensus has been it should be of the initial ‘moov’ box only. Clarify?

• 
  into the fix (COR/AMD), as an explicit statement.
  

We need to clarify what the language in the media header means (the track is suitable for …). We also need to allow full BCP47 strings, and multiple ones.

1. Action Points

1. MPEG-4 AVC File Format (14496-15)

1. Topics

   1. ISO/IEC 14496-15:2010 AMD 2 Carriage of HEVC

2. Contributions

Number	Session	Title	Source	Dispositions
m27593	File Format	HEVC file format update	Y.-K. Wang (Qualcomm)	Accepted N13264
m27961	File Format	Comments on ISO/IEC 14496-15:2010/PDAM2	Takehiko Nakano, Mitsuhhiro Hirabayashi	Accepted N13264
m28165	File Format	Proposal to store startcode in mdat in ISO/IEC 14496-15	T. Suzuki, T. Nakano, M. Hirabayashi	Accepted N13264
m27591	File Format	A proposal of MVC+D file format	Y.-K. Wang, Y. Chen (Qualcomm)	Accepted N13295
m28322	File Format	Comments on HEVC file format	Min Woo Park, Yong Jin Cho, Byeongdoo Choi, Chanyul Kim, Jaeyeon Song (Samsung), Ye-Kui Wang (Qualcomm)	Accepted N13264

3. Summary of discussions

1. m27593 HEVC file format update

Editors to integrate into the study. Editors to also integrate an annex for the codecs sub-parameters for part 15 streams (lifted from RFC 6381, with new declarations for HEVC).

2. m27961 Comments on ISO/IEC 14496-15:2010/PDAM2

See m28165.

3. m28165 Proposal to store startcode in mdat in ISO/IEC 14496-15

We agree to relax the prohibition on filler_data removal to a recommendation or permission, to allow it particularly if HRD Type I is desired.

Some opposition to always having start-codes; one solution might be to have them always, and provide the offsets in a table – sub-sample information, or in sample auxiliary information.

There are two issues here (a) transcoding between environments and (b) maintaining HRD conformance.

Is it really true that Type II HRD parameters when found in an environment without start-codes, mean to apply to that stream without start-codes? Yes. We need a flag trust_type_II_HRD? Yes.

Editors to draft option 4 (samples contain only start-codes, no lengths/offsets) with NAL boundaries [optionally, to study] indicated by sub-sample information.

4. m27591 A proposal of MVC+D file format

Note that this contribution does not support unpaired output of video+depth. The attached document also seems to have quite a few changes of an editorial or quality nature not directly related to MVC+D; editors to check that they are corrected in the new base specification.

Make sure that a regular MVC player, playing an integrated stream, will see the information it needs (e.g. in its config record).

What is the right acronym? 3VC seems rather general.

7.6.1 has an issue with the 4^th paragraph; the view order is different for MVC and MVC+D readers. Delete the parenthesis?

We should use ‘auxiliary video’ tracks for separate depth tracks; the track dependency is right (it says ‘depth’).

Do we need/want the texture_in_stream or depth_in_stream flags, especially if their values are fragile with respect to extraction etc.?

Would the viewidentifierbox be better as two boxes (one for textures, one for depths)?

To WD of AMD1 of 14496-15:2013.

5. m28322 Comments on HEVC file format

Editors to align the study with the HEVC final text.

4.