TTWG minutes 10/31/08

TTWG minutes, 10/31/2008

Attending:
Sean Hayes (SH; chair)
Geoff Freed (GF; scribe)
Andrew Kirkpatrick (AK; scribe for first few minutes)
Glenn Adams (GA)
Philippe Le Hegaret (PH)

Regrets:
David Kirby (DK)
Mike Dolan (MD)


SH (reviewing outstanding action items):  PH did action #9. (Sean  
can't edit database - needs to rejoin group officially.) Sean sent  
action #11 for one implementation, Expression Media Encoder.  Other  
item won't likely be done. Geoff did #13 - turning in implementation  
report for NCAM implementations.

SH:  SMPTE has new WG for delivery of broadband.  Subgroups are  
packaging activity and one to look at timed-text and timed-essence  
data.  Latter is now available for participation; suggests that  
everyone in TTWG sign up.  DFXP may be a starting point for  
discussion of timed text.

PH:  What is timeline?

SH:  Not sure; SMPTE can be slow.  Probably one year-18 months.   
Scope is not just to validate DFXP; is to create end-to-end system,  
working with delivery and creation.  DFXP might be part of the solution.

PH: A profile?

SH:  Not necessarily.  More like a guidebook.  More like a now to use/ 
how to agree on features.

PH:  Adding a streaming capability required?

SH:  Yes, probably not using DFXP.  Will want something that can be  
embedded in a MPEG4-type container for broadcast.  Also want it to be  
interoperable with existing broadcast solutions.  Interoperability  
with CEA-708 very important.

GF:  Pointer for sign-up?

SH:  As soon as something is available will send a notice to member  
list.

SH:  Will send the work statements to the list for study.

ACTION:  Sean to send work statements to TTWG.

SH:  Mozilla group is doing work on Ogg video codecs and what they  
want to do with accessibility.  Sylvia Pfeiffer is in touch with Sean  
regarding the a11y requirements for accessibility. Link to Ogg group  
is in 10/31/08 agenda (https://wiki.mozilla.org/Accessibility/ 
Video_a11y_requirements).  They are evaluating DFXP now to see if it  
meets their needs.

PH:  Looking at what captioning format is to be recommended in  
HTML5.  Would be good for having a profile for on-line captioning so  
it can be used in HTML5.

SH:  Sylvia is in Australia, so time zone is a problem for attending  
TTWG calls.  May want to formally join the TTWG.  A11y group is  
looking at many formats, is working with HTML5 WG, too.  Would be  
good for them to participate in TTWG.

GF:  Would be good for DFXP members to join a11y group as well.

PH:  Pointer to join group?

SH:  Look at link in agenda (https://wiki.mozilla.org/Accessibility/ 
Video_a11y_requirements) and poke around the Wiki for info on  
joining.  (Sends link to IRC.)

SH:  Will leave CEA-708 discussion until Mike Dolan is on the call.

SH:  MSFT Expression encoder doesn't support styling.  Basically  
takes data and allows you to do captioning.  Is a partial  
implementation, but definitely uses timed-text files.

GF:  Similar to NCAM's implementation.

SH:  New implementations to report?

GF:  May be new implementation at WGBH.  Is not yet complete, but may  
be relatively soon.  Will submit when ready.

SH:  Geoff was going to try the test suite.

GF:  Made attempts.  Can't actually test the XML files because  
ccPlayer can't play many of the attributes used in tests.  Regions  
are a problem.

PH:  Could clean up XML files so they reflect what NCAM can support?   
Make the tests finer-grained; that is, test only one thing at a time.

ACTION:  GF to investigate editing the tests and getting rid of non- 
NCAM-supported features and test again; also look at creating editing  
to files test one feature at a time.

PH:  Edit features to be more usable.

GA:  May be some issues with the spec... If one has a timed-text file  
w/o region specified, what is the default region and behavior?

SH:  Good point.  MSFT creates a default region; position where you  
like.  Can style, etc.  Can apply region styling to that.  Applies no  
default semantics to it.  That's up to the programmer.  What should  
the default region be?

GA:  The way things work now, if you don't define a region then the  
current spec would have nothing appear.

SH:  True.  We could take the expected behavior and create a default  
style sheet and use that if nothing is applied by author.  Might be  
problematic given differences in US and European captioning styles.   
Maybe have a two-line region in lower third?  Might be good to invent  
a default style sheet.

PH:  Would be good to know what default region is in current  
implementations.

SH:  Silverlight leaves options up to author.

AK:  We do something similar (Flash); our default playback component  
lacks full region support.  Supports primary region only.

GA:  No notion of primary region, so nothing should appear if region  
isn't specified.

ACTION:  SH will define what is a "default" region based on US and UK  
television behavior.

SH:  Submitted some code to the list that was part of timing tests.   
Not sure if useful for test suite.

GA:  Will incorporate this code into new tests to add more timing  
uses.  Also working at filling out the ttp: parameter attributes so  
we will have tests for those.

SH:  Have been working on creating demos.  Used the recent of MSFT  
advertisements to create DFXP-captioned videos with audio  
descriptions as well.

PH:  Send to me for incorporation to W3C server.

SH:  Will do.  Will be presenting these in Barcelona; will send demos  
after November 11.

AK:  Demo of Adobe DFXP captions in use at http:// 
my.adobe.acrobat.com/accessiblevideo.

SH:  Report on JW player?

AK:  Received data from three different implementers.  Will compile  
and discuss next week.  Regarding test suite... many tests require  
multiple regions, is problematic for verification.

PH:  GF will see if files can be edited so they can be used by  
players that don't have full support.  (See earlier discussion.)   
Tests need to concentrate on one specific item.

AK: Adobe would fail tests now primarily due to requirements of  
region support.  (Ditto NCAM.)

GA:  Current tests were created to test many things all at once,  
using different regions simultaneously.  Some attributes *do* need  
multiple regions in order to be tested, however.

AK:  Logic makes sense but doesn't work for Adobe player.

SH:  Complicated tests need to be simplified to reflect one attribute  
at a time.

SH:  Other items?

Nothing else; call ends.

Received on Friday, 31 October 2008 15:09:05 UTC