International organisation for standardisation organisation internationale de normalisation

Action Points

6. Timed Text & Other Visual Overlays (14496-30)

   1. Topics

2. Contributions

.

Number	Session	Title	Source	Dispositions
m27992	File Format	Summary of Voting on ISO/IEC CD 14496-30	SC 29 Secretariat	Accepted N13266
m28166	File Format	Suggested edits to the Timed-Text in MP4 specification (part 30)	David Singer	Accepted N13267
m27728	Liaison	Liaison comments on 14496-12AMD2 & -30	Michael Dolan	Noted

3. Summary of discussions

1. m28166 Suggested edits to the Timed-Text in MP4 specification (part 30)

Question: can we reference VTT given it’s (a) not final and (b) a CG product?

We need a note to say that a derived specification may require that current-times match decode/presentation times.

On the MIME sample entry, we don’t actually define the ‘mime’ 4CC; this is missing from the amendment. The sentence saying what the behavior is should only apply to this TYPE, not the derived classes.

On the sub-sample, alternative designs include at least (a) putting all images, fonts, etc. in the meta box and making the sample pure XML (b) using sample auxiliary information for the additional images, fonts, etc. (c) changing the design to box up the sample contents. Options (a) and (b) leave the actual data exactly where it is.

When we allow meta-boxes in movie fragments, we need to document that construction_method==0 documents offsets the same as default-base-is-moof==1, i.e. they are all fragment-relative. This needs to be fixed in the amendment.

We change the text of the TTML not to mention fonts explicitly in the samples, but simply ‘resources’ with a specific e.g. of images only.

On the ‘current time’ (currently implicitly the same as decode time) – we could use a box if option (c) above is used; we could use sample auxiliary information (this is not so different from initialization vectors, after all).

1. m27728 Liaison comments on 14496-12AMD2 & -30

1. 
   width and height: we say “The default coordinate system used by a subtitle or timed text track should correspond be appropriate forto the width and height of the video track (as declared in the track header) it is intended to overlay.” We also clarify that the text format might also define its own layout, and that the file format provides the default root container (as defined by TTML).
   
2. 
   z-index: thank you, we use the word ‘stacking’
   
3. 
   empty time: clarified, thank you.
   
4. 
   multiple languages: we need to improve part 12 and delete this section
   
5. 
   width and height of root container: clarified, thank you
   
6. 
   shared images and fonts: thank you, clarified
   
7. 
   URN or URI: agreed, thanks
   
8. 
   namespaces: agreed, most of this should not be a note
   

We realize that when e.g. DASH supplies one text stream to be used with multiple video streams that vary in size, something in DASH needs to say how the text stream is appropriately scaled as DASH adaptively switches video streams.

4. Action Points

7. Contract Expression Language (21000-20)

   1. Topics

      1. ISO/IEC 21000-20 Contract Expression Language

2. Contributions

Number	Session	Title	Source		Dispositions
m28134	MPEG-M	DMAG-UPC contribution to ISO/IEC 21000-20 (MPEG-21 Contract Expression Language) Reference Software	Jonathan Florido, Eva Rodríguez, Jaime Delgado	Accepted N13269
m28177	MPEG-M	Proposal for Amendment to MPEG-21 CEL	Xin Wang	Noted
m28126	MPEG-M	Contribution to Reference Software for MPEG-21 Media Contract Ontology	L.Boch	Accepted N13269

3.