- From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
- Date: Tue, 12 Nov 2013 06:41:50 +0800
- To: Brendan Long <B.Long@cablelabs.com>
- Cc: Eric Winkelman <E.Winkelman@cablelabs.com>, Bob Lund <B.Lund@cablelabs.com>, "public-inbandtracks@w3.org" <public-inbandtracks@w3.org>
On Tue, Nov 12, 2013 at 5:31 AM, Brendan Long <B.Long@cablelabs.com> wrote: >> I agree on using a string, and personally lean toward using a hex >> representation. > > I think it would be best if the translation between binary-in-the-file to > track id was format-specific, so we can use the most appropriate > representation for each format. We can define how to convert an inband ID representation into the value for the @id attribute of tracks in HTML, but we have to stick with the resulting @id value being conformant with the HTML spec. In the HTML spec, @id is always DOMString, for all types of tracks: http://www.w3.org/html/wg/drafts/html/master/embedded-content-0.html#texttrack http://www.w3.org/html/wg/drafts/html/master/embedded-content-0.html#audiotrack http://www.w3.org/html/wg/drafts/html/master/embedded-content-0.html#videotrack So, just convert whichever unique track identifier a format has into a DOMString and expose that in the @id attribute. It seems like that is: * the PID for MPEG-TS * the TrackUID for matroksa * the name in Ogg Skeleton Regards, Silvia. > For example, it seems reasonable to > represent Matroska TrackUID's as decimal numbers, but for Ogg Skeleton files > it could be the name (which can be anything, as long as it's per-file > unique). > > Or maybe it would be useful if the ID was always a number? For Ogg we can > use the stream serial number if we wanted to, but the current HTML spec says > we should use the name header. The stream serial number has the advantage of > always existing.
Received on Monday, 11 November 2013 22:42:39 UTC