Re: MPEG-2 TS Closed Caption mapping

On Wed, May 14, 2014 at 9:32 PM, Aaron Colwell <acolwell@google.com> wrote:
> On Wed, May 14, 2014 at 11:20 AM, Silvia Pfeiffer
> <silviapfeiffer1@gmail.com> wrote:
>>
>> On Tue, May 13, 2014 at 7:34 AM, Clift, Graham <Graham.Clift@am.sony.com>
>> wrote:
>> > The problem I identified related to passing a caption service of type
>> > line
>> > 21 as an audio or video track
>>
>> If line21 captions are burnt into a video track, then exposing them as
>> a video track makes sense. I don't understand how they would be
>> exposed as an audio track.
>
>
> Why does it make sense to expose captions data as a video track?

Only where captions are actually burnt into video.

> ISTM that
> these should be exposed as TextTracks since this is actually text/metadata
> right?

Yes, where it's text and can be extracted, it should go into a TextTrack.

> I would expect a video with line 21 data to create a VideoTrack for
> the video and a TextTrack for the line21 caption data. That seems like the
> most natural mapping to me.

I don't know how the line21 data is encoded into MPEG - is it text or
bitmaps or burnt-in?

I agree, you should choose the most natural mapping.

HTH,
Silvia.

Received on Wednesday, 14 May 2014 12:44:07 UTC