Re: Proposal for initial CG focus

On 11/8/13 1:38 PM, "Brendan Long" <B.Long@cablelabs.com> wrote:

>> >- To identify the type of the track data, I don't think we should rely
>> >on the label.
>>
>> [2] uses the label for two different purposes. For 'track description'
>> metadata text tracks, the label identifies the MIME of the media
>>resource
>> and that this metadata track holds the other track metadata. The MIME is
>> needed for JS to know how to parse the metadata Cues. For other tracks,
>> the label contains the stream ID of the the in-band stream that the
>>track
>> represents. Someone has suggested that track.id could be used for this
>> purpose.
>
>I don't think we should use the label this way, since it's meant to be
>displayed to users, and showing users technical information like codecs
>and PIDs isn't user-friendly.

Agree, but we need alternate solutions.

> It seems like "id" is already fine for exposing the PID / TrackUID /
>whatever,

Track.id would seem to work for containing that track's in-band ID

>but we may need a new attribute to expose the MIME type.

Yes, perhaps an attribute on the htmlMediaElement

> Is this not what the metadata dispatch type attribute is for?

No. The inBandTrackMetadataDispatchType contains an encoding of the
metadata describing what's in a metadata text track. It contains no MIME
info and it only applies to in-band metadata text tracks.

As noted in an earlier response to Cyril, replacing the
inBandŠDispatchType attribute with a 'metadata' attribute, that exists on
all video, audio and text tracks, removes the need for a text track
containing the media resource metadata.

Received on Friday, 8 November 2013 21:13:06 UTC