ISSUE-270: Appendix N assumption that root temporal extent corresponds with the beginning of a related media object

nigel

Appendix N assumption that root temporal extent corresponds with the beginning of a related media object

State:
OPEN
Product:
TTML2
Raised by:
Nigel Megitt
Opened on:
2013-08-12
Description:
From Appendix N.2:

"The above formalisms assumes that the Root Temporal Extent corresponds with the beginning of a related media object. If this assumption doesn't hold, then an additional offset that accounts for the difference may be introduced when computing media time M."

In a streaming environment it is very likely, especially for live-created subtitles, that a set of TTML samples will be produced each with a different Root Temporal Extent that starts at a different point relative to a single Related Media Object (which may itself be packaged up into an independent set of samples). However I would not in general account for this with an offset and assume (or hope) that every sample's begin and end times are zero-based; rather I'd expect the begin of the TT to be coincident with the point within the Related Media Object at which that sample begins. As a side-effect this should reduce the complexity of the transformations needed to accumulate the [short] samples together to create a longer one.

For this reason I propose that we reverse this assumption.
Related Actions Items:
No related actions
Related emails:
  1. April F2F Meeting minutes (from nigel.megitt@bbc.co.uk on 2015-04-22)
  2. {agenda} TTWG Meeting 2015-01-08 (from nigel.megitt@bbc.co.uk on 2015-01-07)
  3. {agenda} TTWG Meeting 2/10/2014 (from nigel.megitt@bbc.co.uk on 2014-10-01)
  4. Re: [TTML2] Document relationship to media timings - re Issue-346, Issue-270, Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-26)
  5. Re: [TTML2] Document relationship to media timings - re Issue-346, Issue-270, Issue-335 (from glenn@skynav.com on 2014-09-26)
  6. Re: [TTML2] Document relationship to media timings - re Issue-346, Issue-270, Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-26)
  7. Re: [TTML2] Document relationship to media timings - re Issue-346, Issue-270, Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-26)
  8. Re: [TTML2] Document relationship to media timings - re Issue-346, Issue-270, Issue-335 (from glenn@skynav.com on 2014-09-26)
  9. [TTML2] Document relationship to media timings - re Issue-346, Issue-270, Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-26)
  10. RE: ISSUE-346: Need ttp:mediaDuration parameter [TTML2] (from John.Birch@screensystems.tv on 2014-09-25)
  11. Re: ISSUE-346: Need ttp:mediaDuration parameter [TTML2] (from glenn@skynav.com on 2014-09-25)
  12. Re: ISSUE-346: Need ttp:mediaDuration parameter [TTML2] (from nigel.megitt@bbc.co.uk on 2014-09-25)
  13. RE: ISSUE-346: Need ttp:mediaDuration parameter [TTML2] (from John.Birch@screensystems.tv on 2014-09-25)
  14. Re: ISSUE-346: Need ttp:mediaDuration parameter [TTML2] (from nigel.megitt@bbc.co.uk on 2014-09-25)
  15. {agenda} TTWG Meeting 25/9/2014 (from nigel.megitt@bbc.co.uk on 2014-09-24)
  16. Re: Issue-270 and Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-24)
  17. Re: Issue-270 and Issue-335 (from cyril.concolato@telecom-paristech.fr on 2014-09-24)
  18. Re: Issue-270 and Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-24)
  19. Re: Issue-270 and Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-24)
  20. Re: Issue-270 and Issue-335 (from glenn@skynav.com on 2014-09-23)
  21. Re: Issue-270 and Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-23)
  22. Re: Issue-270 and Issue-335 (from glenn@skynav.com on 2014-09-22)
  23. Issue-270 and Issue-335 (from nigel.megitt@bbc.co.uk on 2014-09-22)
  24. TTML Minutes for 11/11/13 (from nigel.megitt@bbc.co.uk on 2013-11-21)
  25. {minutes} TTML Meeting of 24/10/13 (from glenn@skynav.com on 2013-10-24)
  26. RE: TTML Agenda for 24/10/13 (from mdolan@newtbt.com on 2013-10-23)
  27. TTML Agenda for 24/10/13 (from nigel.megitt@bbc.co.uk on 2013-10-23)
  28. TTML Agenda for 15/08/13 (from Sean.Hayes@microsoft.com on 2013-08-15)
  29. ISSUE-270 (nigel): Appendix N assumption that root temporal extent corresponds with the beginning of a related media object [TTML 1.0] (from sysbot+tracker@w3.org on 2013-08-12)

Related notes:

At present, there is only one root temporal extent for a given document and it corresponds to the related media object's temporal extent when operating in media or smpte time bases. If we want to introduce a mechanism to specify an offset between a related media object's temporal extent and the root temporal extent, then we would need to define a new parameter, e.g., ttp:mediaOffset, which would need to be done in TTML2, not TTML1.

The current assumption documented in the note is accurate for TTML1. I am consequently taking no action in TTML1 but moving this issue to TTML2.

Glenn Adams, 24 Aug 2013, 00:03:52

Nigel: P1

Philippe Le Hégaret, 24 Oct 2013, 15:46:38

Addressed in https://dvcs.w3.org/hg/ttml/rev/3cbc109b90bd

Glenn Adams, 21 Sep 2014, 12:39:15

The proposed solution is the opposite of that proposed in the issue, and I believe introduces more problems than it resolves.

There are a number of problems with the current draft of TTML2 that may or may not also be present in TTML1, when taken together:

1. Definition of root temporal extent "The temporal extent (interval) defined by the temporal beginning and ending of a document instance in relationship with some external application or presentation context."
2. N.2 Media Time Base. R = playRate * M + epochOffset; M="0 corresponding to the beginning of the root temporal extent"; epochOffset is the difference between the local real time at the beginning of the root temporal extent and the local real time at the the beginning of epoch E, expressed in seconds.

Together these mean that the media time of the earliest computed begin value in the document (or 00:00:00 if unspecified) is somehow related to local real time. The problem here is that the earliest begin time can override the mapping, and specify a new datum that presumably is equivalent to the start of media playback. At least this is one possible reading of the definition of Root Temporal Extent in this context.

In reality I don't think this is what anyone wants. More realistically media time 00:00:00 should always relate the start of the media, and anyone wanting to apply an offset should do so in an external wrapper. If media time (e.g. through some kind of timecode, though really this should be SMPTE time base) doesn't start at zero, say it starts at 10:00:00 in timecode, then it is certainly worth including a 'media begins at 10:00:00' attribute in the document. However I would expect the media times in the document also to begin at 10:00:00 in that case, and no additions or subtractions would occur to process media times within the document.

Nigel Megitt, 22 Sep 2014, 13:52:05

This is related to Issue-361.

Nigel Megitt, 5 Jan 2015, 11:51:45

Display change log ATOM feed


David Singer <singer@apple.com>, Nigel Megitt <nigel.megitt@bbc.co.uk>, Chairs, Thierry Michel <tmichel@w3.org>, Philippe Le Hégaret <plh@w3.org>, Staff Contacts
Tracker: documentation, (configuration for this group), originally developed by Dean Jackson, is developed and maintained by the Systems Team <w3t-sys@w3.org>.
$Id: index.php,v 1.325 2014-09-10 21:42:02 ted Exp $