W3C home > Mailing lists > Public > public-media-annotation@w3.org > May 2009

Re: Properties for media fragment?

From: David Singer <singer@apple.com>
Date: Tue, 5 May 2009 11:28:01 -0700
Message-Id: <p062408adc626350d7c2d@[17.202.35.52]>
To: Pierre-Antoine Champin <pierre-antoine.champin@liris.cnrs.fr>, Felix Sasaki <felix.sasaki@fh-potsdam.de>
Cc: "public-media-annotation@w3.org" <public-media-annotation@w3.org>
At 22:02  +0100 4/05/09, Pierre-Antoine Champin wrote:
>Hi,
>
>I agree with David that "my favourite scenes" is neither 'intrinsic' not
>'published'. I also agree to call that 'user' metadata, and would point
>out that our use case 5.6 [1] is exactly about that.
>
>But the properties proposed by Felix are not, IMO, limited to user
>generated metadata. Some videos are complex, and different fragments may
>have different intrinsic or published metadata. A canonical example is a
>news report, which is a composite media object comprising several parts.
>Both the whole report and each part deserve theit own metadata
>(intrinsic and published), but it should be possible to express the
>relation between the whole and the parts.

right, time-variant metadata is a big question. 
Most annotation/metadata systems think of it as 
time-invariuant ('applies to the whole 
resource').  However, some things are better 
time-variant (e.g. the copyright owner of a movie 
assembled from pieces might vary by piece).

It's easy to design an interface "at  time T in 
this movie, what is the answer to X" and have the 
reply be "at any time, the answer is Y" 
(time-invariant label).  It's harder to answer 
"what is the answer to X" with "well, that 
depends on what time in the movie you ask about".

worth pondering...

>
>   pa
>
>[1]
>http://www.w3.org/TR/2009/WD-media-annot-reqs-20090119/#uc-user-generated-metadata
>
>Felix Sasaki a écrit :
>>
>>
>>  2009/5/5 David Singer <singer@apple.com <mailto:singer@apple.com>>
>>
>>      At 7:14  +0200 2/05/09, Felix Sasaki wrote:
>>
>>          List of my favorite scences of a video, as part of the video
>>          metadata? Does that not make sense?
>>
>>
>>      Oh wow, this is a new category of metadata.
>>
>>
>>
>>  Is it so new? At least the mechanisms listed in the media fragments
>>  draft seem to exist for some time. So my assumption was that the
>>  mechanisms are actually used, and that mapping betweeen them is useful.
>>  If they exist only as a new category, to be filled by the new fragment
>>  identifier syntax defined by the media fragments WG, than it does not
>>  make sense to invest time in their mapping.
>>
>>  Felix
>>
>> 
>>
>>       So far I have been seeing
>>
>>      * 'intrinsic' properties of the media itself (duration, whether it
>>      has video, audio etc.)
>>      * published annotations for the media (copyright, title, etc.)
>>
>>      both of these are 'source supplied'.
>>
>>      'My favorite scenes' is user-supplied.  Hm.
>>
>> 
>>
>>
>>      --
>>      David Singer
>>      Multimedia Standards, Apple Inc.
>>
>>
>
>
>
>Content-Type: application/pgp-signature; name="signature.asc"
>Content-Description: OpenPGP digital signature
>Content-Disposition: attachment; filename="signature.asc"
>
>Attachment converted: DaveG49:signature 311.asc (    /    ) (00219A47)


-- 
David Singer
Multimedia Standards, Apple Inc.
Received on Tuesday, 5 May 2009 18:30:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 5 May 2009 18:30:35 GMT