W3C home > Mailing lists > Public > public-tt@w3.org > January 2013

TTWG Meeting Minutes Jan 17, 2013

From: Monica Martin (MS OPEN TECH) <momartin@microsoft.com>
Date: Thu, 17 Jan 2013 16:48:25 +0000
To: "'public-tt@w3.org'" <public-tt@w3.org>
Message-ID: <b2158fc4947b49779d27b4a2e491191c@BY2PR03MB059.namprd03.prod.outlook.com>


      [1] http://www.w3.org/

                               - DRAFT -

                Timed Text Working Group Teleconference

17 Jan 2013

   See also: [2]IRC log

      [2] http://www.w3.org/2013/01/17-tt-irc


          glenn, mijordan, mmartin3, mike, +1.408.771.aaaa,
          Andreas, Thomas_Bause_Mason, +44.118.900.aabb, sean,
          courtney, +1.310.210.aacc, pal




     * [3]Topics
         1. [4]IETF Registration
     * [5]Summary of Action Items

   <trackbot> Date: 17 January 2013

   thanks glenn


   Sean Hayes

   <scribe> scribenick: mmartin3

IETF Registration


   <trackbot> ISSUE-175 -- media type and file extension not
   registered -- open


      [6] http://www.w3.org/AudioVideo/TT/tracker/issues/175

   mike: Issue-179 encoding is separate from registration.


   <trackbot> ISSUE-179 -- Interpreting the pixel measure -- open


      [7] http://www.w3.org/AudioVideo/TT/tracker/issues/179

   scribe: RFC 3023 in TTML 1.0
   ... provided redlines to glenn
   ... Encoding requirements Section 4.3.3 XML 1.0 - gives rules
   and provides an Annex that describes an algorithm.
   ... Issues with transferring the charset for encoding.
   ... Suggest we constrain to UTF-8 and -16, and that would be
   supported by an XML processor.

   glenn: What about US ASCII.

   mike: Wouldn't matter. Use UTF-8.

   glenn: Constraint on all interchange of documents?

   mike: All identified as ttml/xml.

   glenn: Constrain to XML and the use of the MIME type charset.
   ... Write registration to constrain to UTF-8 and -16.

   mike: Don't specify what happens if optional charset is not

   sean: Doesn't constrain use of other encodings.

   glenn: Charset parameter is ignored and what is specified in
   the document is used.

   sean: Japanese may have questions if their encodings are not
   considered. Don't know if an issue for charset.

   mike: Only the encoding is used not the charset.

   sean: Japanese may still have questions.

   mike: Think this is orthogonal.

   glenn: Specify encoding only without a character set.
   ... Suggest charset be ignored.

   sean: Contradicts XML specification.
   ... Just don't use.

   mike: RFC 3023 has "if present xxxxx" language. Better not to

   glenn: HTML5 changes that.

   mike: Could specify should not use, but if you do xxxx.

   <glenn> should not use charset param, but if used, must be
   identical to encoding parameter in document instance


   <trackbot> ISSUE-175 -- media type and file extension not
   registered -- open


      [8] http://www.w3.org/AudioVideo/TT/tracker/issues/175


   <trackbot> ISSUE-179 -- Interpreting the pixel measure -- open


      [9] http://www.w3.org/AudioVideo/TT/tracker/issues/179

   <glenn> RESOLVED: accept the above language re:

   RESOLVED On text re: encoding.


   <trackbot> ISSUE-175 -- media type and file extension not
   registered -- open


     [10] http://www.w3.org/AudioVideo/TT/tracker/issues/175

   mike: Provided text.
   ... No comments from Philippe.
   ... Appendix C.
   ... Close Issue-175 and adopt the redlines proposed by Mike.
   ... Action to Glenn.

   RESOLVED to use edits from Mike on Issue-175 and action to
   Glenn to process. Leave both open.

   <glenn> ACTION-136?

   <trackbot> ACTION-136 -- Glenn Adams to integrate changes from
   Action-133 into the TTML 1.0 Second Edition -- due 2013-01-14
   -- OPEN


     [11] http://www.w3.org/AudioVideo/TT/tracker/actions/136

   <glenn> ISSUE-199?

   <trackbot> ISSUE-199 -- timeExpression to video frame
   calculation is unclear in some cases -- raised


     [12] http://www.w3.org/AudioVideo/TT/tracker/issues/199

   mike: Differences of view between Glenn and I.
   ... timeExpression
   ... Conversion to frames of video
   ... Time values have different frame representation. If 2 are
   different they resolve to the same frame in the video.
   ... Do math.
   ... Non-integer frame rates only.
   ... One algorithm is to take the seconds and use real frame
   rate to get frame number. Add frame field to that with final
   frame number. Sometimes that two different ones resolve to the
   same frame. TTML is silent on the math.

   glenn: Agree current specification is silent on this topic.
   ... Question is what algorithm do we use?
   ... Intent of timeExpression was that it was different that
   implementers have interpreted.
   ....hh: mm:ss didn't represent media type but frame type.
   ...hh: mm:ss converted to seconds value.
   ... Seconds multiplied frame rate to get frame count.
   ... Fraction could be added to frame count. Could be
   ... Convert to real media type - divide by actual frame rate
   (frame rate x frame rate multiplier).
   ... Explains mechanics.
   ... Alleviates issue.
   ... Easier to understand in implement in my view.
   ... Haven't accounted for drop frame rate.
   ... Need to account for any algorithm.

   mike: Time values diverge from media time.

   sean: My previous work was similar to Glenn's. That number
   should uniquely identify what is in the media and it is of
   concern if they diverge.
   ... No position at this point.

   mike: Significant divergence in a feature film.

   glenn: Frames are well understood. How do systems handle
   non-integer values?

   mike: SMPTE repeating combinations are prohibited.
   ... Solved by not allowing.

   glenn: What is the SMPTE interpretation of seconds and frames?
   ... Discuss illegal combos.

   sean: Explains scenario with timeExpression and use of

   mike: We have 2 different views of what timeExpression means.

   glenn: There may be another option.
   ... Constrain the use of frames and sub-frames to apply in the
   SMPTE not the media mode.

   sean: Allow frame rate to be present but not apply in a
   specific case.

   mike: Forbid frame expression for media time.
   ... Your option may also be fatal.

   Discuss scenario for breaking up frames with a specific rate.

   glenn: Can't convert to an integer value if you don't have an
   integer frame rate.

   mike: Still have the issue that if you have a non-integer rate
   that isn't the real media time.

   sean: Effectively use frameRateMultiplier to get to real time.

   glenn: Restrict frameRateMultiplier to 1 over 1.

   mike: Can't do that.
   ... Need to signal conversions when you have a non-integer
   frame rate.

   sean: Table the discussion and bring back next week.

   glenn: Try to get more information how this is done in the

   mijordan: HBO using SMPTE.

   mike: May be able to determine what algorithm was used. Need
   audio and video.
   ... DECE implemented as frames.

   sean: Discuss next week.

   trackbot, end meeting

Summary of Action Items

   [End of minutes]

Monica J Martin
Senior Program Manager
Microsoft Open Technologies, Inc.
A subsidiary of Microsoft Corporation
Received on Thursday, 17 January 2013 16:52:15 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 5 October 2017 18:24:07 UTC